aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorArthur She <arthur.she@linaro.org>2014-09-18 14:17:16 -0700
committerArthur She <arthur.she@linaro.org>2014-09-18 14:17:16 -0700
commita261017e3da7cac1b887764dced3c1c0964c2ba2 (patch)
tree117c21aa0b42865a321b078ddb71c410506c51c9
parentdbb5bc6a61eff1f9cd72c597b4d3a4b0d4411f04 (diff)
Update to 72efa0ffec6a4361ef33d44ae0094714d37f4d6b from repo http://git.linaro.org/kernel/linux-linaro-tracking.git branch linux-linaro72efa0ffec6a4361ef33d44ae0094714d37f4d6blinux-linaro
-rw-r--r--Makefile49
-rw-r--r--README.txt61
-rw-r--r--breakpoints/Makefile23
-rw-r--r--breakpoints/breakpoint_test.c394
-rw-r--r--cpu-hotplug/Makefile9
-rw-r--r--cpu-hotplug/on-off-test.sh269
-rw-r--r--efivarfs/Makefile12
-rw-r--r--efivarfs/create-read.c38
-rw-r--r--efivarfs/efivarfs.sh198
-rw-r--r--efivarfs/open-unlink.c63
-rw-r--r--firmware/Makefile27
-rw-r--r--firmware/fw_filesystem.sh62
-rw-r--r--firmware/fw_userhelper.sh89
-rw-r--r--ipc/Makefile25
-rw-r--r--ipc/msgque.c252
-rw-r--r--kcmp/.gitignore2
-rw-r--r--kcmp/Makefile28
-rw-r--r--kcmp/kcmp_test.c96
-rw-r--r--memfd/.gitignore4
-rw-r--r--memfd/Makefile41
-rw-r--r--memfd/fuse_mnt.c110
-rw-r--r--memfd/fuse_test.c311
-rw-r--r--memfd/memfd_test.c913
-rw-r--r--memfd/run_fuse_test.sh14
-rw-r--r--memory-hotplug/Makefile9
-rw-r--r--memory-hotplug/on-off-test.sh238
-rw-r--r--mount/Makefile17
-rw-r--r--mount/unprivileged-remount-test.c242
-rw-r--r--mqueue/.gitignore2
-rw-r--r--mqueue/Makefile10
-rw-r--r--mqueue/mq_open_tests.c500
-rw-r--r--mqueue/mq_perf_tests.c743
-rw-r--r--net/.gitignore3
-rw-r--r--net/Makefile25
-rw-r--r--net/psock_fanout.c312
-rw-r--r--net/psock_lib.h127
-rw-r--r--net/psock_tpacket.c805
-rw-r--r--net/run_afpackettests26
-rw-r--r--net/run_netsocktests12
-rw-r--r--net/socket.c92
-rw-r--r--powerpc/Makefile39
-rw-r--r--powerpc/copyloops/Makefile29
-rw-r--r--powerpc/copyloops/asm/ppc_asm.h89
-rw-r--r--powerpc/copyloops/asm/processor.h0
l---------powerpc/copyloops/copyuser_64.S1
l---------powerpc/copyloops/copyuser_power7.S1
l---------powerpc/copyloops/memcpy_64.S1
l---------powerpc/copyloops/memcpy_power7.S1
-rw-r--r--powerpc/copyloops/validate.c99
-rw-r--r--powerpc/harness.c114
-rw-r--r--powerpc/mm/Makefile18
-rw-r--r--powerpc/mm/hugetlb_vs_thp_test.c72
-rw-r--r--powerpc/pmu/Makefile38
-rw-r--r--powerpc/pmu/count_instructions.c147
-rw-r--r--powerpc/pmu/ebb/Makefile33
-rw-r--r--powerpc/pmu/ebb/back_to_back_ebbs_test.c106
-rw-r--r--powerpc/pmu/ebb/busy_loop.S271
-rw-r--r--powerpc/pmu/ebb/close_clears_pmcc_test.c59
-rw-r--r--powerpc/pmu/ebb/cpu_event_pinned_vs_ebb_test.c93
-rw-r--r--powerpc/pmu/ebb/cpu_event_vs_ebb_test.c89
-rw-r--r--powerpc/pmu/ebb/cycles_test.c58
-rw-r--r--powerpc/pmu/ebb/cycles_with_freeze_test.c117
-rw-r--r--powerpc/pmu/ebb/cycles_with_mmcr2_test.c91
-rw-r--r--powerpc/pmu/ebb/ebb.c478
-rw-r--r--powerpc/pmu/ebb/ebb.h77
-rw-r--r--powerpc/pmu/ebb/ebb_handler.S365
-rw-r--r--powerpc/pmu/ebb/ebb_on_child_test.c86
-rw-r--r--powerpc/pmu/ebb/ebb_on_willing_child_test.c92
-rw-r--r--powerpc/pmu/ebb/ebb_vs_cpu_event_test.c86
-rw-r--r--powerpc/pmu/ebb/event_attributes_test.c131
-rw-r--r--powerpc/pmu/ebb/fixed_instruction_loop.S43
-rw-r--r--powerpc/pmu/ebb/fork_cleanup_test.c79
-rw-r--r--powerpc/pmu/ebb/instruction_count_test.c164
-rw-r--r--powerpc/pmu/ebb/lost_exception_test.c100
-rw-r--r--powerpc/pmu/ebb/multi_counter_test.c91
-rw-r--r--powerpc/pmu/ebb/multi_ebb_procs_test.c109
-rw-r--r--powerpc/pmu/ebb/no_handler_test.c61
-rw-r--r--powerpc/pmu/ebb/pmae_handling_test.c106
-rw-r--r--powerpc/pmu/ebb/pmc56_overflow_test.c93
-rw-r--r--powerpc/pmu/ebb/reg.h49
-rw-r--r--powerpc/pmu/ebb/reg_access_test.c39
-rw-r--r--powerpc/pmu/ebb/task_event_pinned_vs_ebb_test.c91
-rw-r--r--powerpc/pmu/ebb/task_event_vs_ebb_test.c83
-rw-r--r--powerpc/pmu/ebb/trace.c300
-rw-r--r--powerpc/pmu/ebb/trace.h41
-rw-r--r--powerpc/pmu/event.c131
-rw-r--r--powerpc/pmu/event.h43
-rw-r--r--powerpc/pmu/l3_bank_test.c48
-rw-r--r--powerpc/pmu/lib.c300
-rw-r--r--powerpc/pmu/lib.h42
-rw-r--r--powerpc/pmu/loop.S43
-rw-r--r--powerpc/pmu/per_event_excludes.c114
-rw-r--r--powerpc/subunit.h52
-rw-r--r--powerpc/tm/Makefile15
-rw-r--r--powerpc/tm/tm-resched-dscr.c98
-rw-r--r--powerpc/utils.h49
-rw-r--r--ptrace/Makefile10
-rw-r--r--ptrace/peeksiginfo.c218
-rw-r--r--rcutorture/.gitignore6
-rw-r--r--rcutorture/bin/config2frag.sh25
-rwxr-xr-xrcutorture/bin/configNR_CPUS.sh45
-rwxr-xr-xrcutorture/bin/configcheck.sh54
-rwxr-xr-xrcutorture/bin/configinit.sh74
-rwxr-xr-xrcutorture/bin/cpus2use.sh41
-rw-r--r--rcutorture/bin/functions.sh223
-rwxr-xr-xrcutorture/bin/kvm-build.sh71
-rwxr-xr-xrcutorture/bin/kvm-recheck-lock.sh51
-rwxr-xr-xrcutorture/bin/kvm-recheck-rcu.sh51
-rwxr-xr-xrcutorture/bin/kvm-recheck.sh69
-rwxr-xr-xrcutorture/bin/kvm-test-1-run.sh227
-rw-r--r--rcutorture/bin/kvm.sh413
-rwxr-xr-xrcutorture/bin/parse-build.sh57
-rwxr-xr-xrcutorture/bin/parse-console.sh41
-rwxr-xr-xrcutorture/bin/parse-torture.sh106
-rw-r--r--rcutorture/configs/lock/BUSTED6
-rw-r--r--rcutorture/configs/lock/BUSTED.boot1
-rw-r--r--rcutorture/configs/lock/CFLIST1
-rw-r--r--rcutorture/configs/lock/CFcommon2
-rw-r--r--rcutorture/configs/lock/LOCK016
-rw-r--r--rcutorture/configs/lock/ver_functions.sh43
-rw-r--r--rcutorture/configs/rcu/BUSTED7
-rw-r--r--rcutorture/configs/rcu/BUSTED.boot1
-rw-r--r--rcutorture/configs/rcu/CFLIST13
-rw-r--r--rcutorture/configs/rcu/CFcommon2
-rw-r--r--rcutorture/configs/rcu/SRCU-N7
-rw-r--r--rcutorture/configs/rcu/SRCU-N.boot1
-rw-r--r--rcutorture/configs/rcu/SRCU-P7
-rw-r--r--rcutorture/configs/rcu/SRCU-P.boot1
-rw-r--r--rcutorture/configs/rcu/TINY0112
-rw-r--r--rcutorture/configs/rcu/TINY0212
-rw-r--r--rcutorture/configs/rcu/TREE0121
-rw-r--r--rcutorture/configs/rcu/TREE01.boot1
-rw-r--r--rcutorture/configs/rcu/TREE0224
-rw-r--r--rcutorture/configs/rcu/TREE02-T24
-rw-r--r--rcutorture/configs/rcu/TREE0321
-rw-r--r--rcutorture/configs/rcu/TREE0423
-rw-r--r--rcutorture/configs/rcu/TREE04.boot1
-rw-r--r--rcutorture/configs/rcu/TREE0523
-rw-r--r--rcutorture/configs/rcu/TREE05.boot1
-rw-r--r--rcutorture/configs/rcu/TREE0624
-rw-r--r--rcutorture/configs/rcu/TREE0722
-rw-r--r--rcutorture/configs/rcu/TREE0824
-rw-r--r--rcutorture/configs/rcu/TREE08-T24
-rw-r--r--rcutorture/configs/rcu/TREE08.boot1
-rw-r--r--rcutorture/configs/rcu/TREE0919
-rw-r--r--rcutorture/configs/rcu/v0.0/CFLIST14
-rw-r--r--rcutorture/configs/rcu/v0.0/N1-S-T-NH-SD-SMP-HP18
-rw-r--r--rcutorture/configs/rcu/v0.0/N2-2-t-nh-sd-SMP-hp20
-rw-r--r--rcutorture/configs/rcu/v0.0/N3-3-T-nh-SD-SMP-hp22
-rw-r--r--rcutorture/configs/rcu/v0.0/N4-A-t-NH-sd-SMP-HP18
-rw-r--r--rcutorture/configs/rcu/v0.0/N5-U-T-NH-sd-SMP-hp22
-rw-r--r--rcutorture/configs/rcu/v0.0/NT1-nh23
-rw-r--r--rcutorture/configs/rcu/v0.0/NT3-NH20
-rw-r--r--rcutorture/configs/rcu/v0.0/P1-S-T-NH-SD-SMP-HP19
-rw-r--r--rcutorture/configs/rcu/v0.0/P2-2-t-nh-sd-SMP-hp20
-rw-r--r--rcutorture/configs/rcu/v0.0/P3-3-T-nh-SD-SMP-hp20
-rw-r--r--rcutorture/configs/rcu/v0.0/P4-A-t-NH-sd-SMP-HP22
-rw-r--r--rcutorture/configs/rcu/v0.0/P5-U-T-NH-sd-SMP-hp27
-rw-r--r--rcutorture/configs/rcu/v0.0/PT1-nh23
-rw-r--r--rcutorture/configs/rcu/v0.0/PT2-NH22
-rw-r--r--rcutorture/configs/rcu/v0.0/ver_functions.sh33
-rw-r--r--rcutorture/configs/rcu/v3.12/CFLIST17
-rw-r--r--rcutorture/configs/rcu/v3.12/N1-S-T-NH-SD-SMP-HP19
-rw-r--r--rcutorture/configs/rcu/v3.12/N2-2-t-nh-sd-SMP-hp20
-rw-r--r--rcutorture/configs/rcu/v3.12/N3-3-T-nh-SD-SMP-hp22
-rw-r--r--rcutorture/configs/rcu/v3.12/N4-A-t-NH-sd-SMP-HP18
-rw-r--r--rcutorture/configs/rcu/v3.12/N5-U-T-NH-sd-SMP-hp22
-rw-r--r--rcutorture/configs/rcu/v3.12/N6---t-nh-SD-smp-hp19
-rw-r--r--rcutorture/configs/rcu/v3.12/N7-4-T-NH-SD-SMP-HP26
-rw-r--r--rcutorture/configs/rcu/v3.12/N8-2-T-NH-SD-SMP-HP22
-rw-r--r--rcutorture/configs/rcu/v3.12/NT1-nh23
-rw-r--r--rcutorture/configs/rcu/v3.12/NT3-NH20
-rw-r--r--rcutorture/configs/rcu/v3.12/P1-S-T-NH-SD-SMP-HP20
-rw-r--r--rcutorture/configs/rcu/v3.12/P2-2-t-nh-sd-SMP-hp20
-rw-r--r--rcutorture/configs/rcu/v3.12/P3-3-T-nh-SD-SMP-hp20
-rw-r--r--rcutorture/configs/rcu/v3.12/P4-A-t-NH-sd-SMP-HP22
-rw-r--r--rcutorture/configs/rcu/v3.12/P5-U-T-NH-sd-SMP-hp27
-rw-r--r--rcutorture/configs/rcu/v3.12/P6---t-nh-SD-smp-hp18
-rw-r--r--rcutorture/configs/rcu/v3.12/P7-4-T-NH-SD-SMP-HP30
-rw-r--r--rcutorture/configs/rcu/v3.12/P7-4-T-NH-SD-SMP-HP-all30
-rw-r--r--rcutorture/configs/rcu/v3.12/P7-4-T-NH-SD-SMP-HP-none30
-rw-r--r--rcutorture/configs/rcu/v3.12/P7-4-T-NH-SD-SMP-hp30
-rw-r--r--rcutorture/configs/rcu/v3.12/PT1-nh23
-rw-r--r--rcutorture/configs/rcu/v3.12/PT2-NH22
-rw-r--r--rcutorture/configs/rcu/v3.3/CFLIST14
-rw-r--r--rcutorture/configs/rcu/v3.3/N1-S-T-NH-SD-SMP-HP19
-rw-r--r--rcutorture/configs/rcu/v3.3/N2-2-t-nh-sd-SMP-hp20
-rw-r--r--rcutorture/configs/rcu/v3.3/N3-3-T-nh-SD-SMP-hp22
-rw-r--r--rcutorture/configs/rcu/v3.3/N4-A-t-NH-sd-SMP-HP18
-rw-r--r--rcutorture/configs/rcu/v3.3/N5-U-T-NH-sd-SMP-hp22
-rw-r--r--rcutorture/configs/rcu/v3.3/NT1-nh23
-rw-r--r--rcutorture/configs/rcu/v3.3/NT3-NH20
-rw-r--r--rcutorture/configs/rcu/v3.3/P1-S-T-NH-SD-SMP-HP20
-rw-r--r--rcutorture/configs/rcu/v3.3/P2-2-t-nh-sd-SMP-hp20
-rw-r--r--rcutorture/configs/rcu/v3.3/P3-3-T-nh-SD-SMP-hp20
-rw-r--r--rcutorture/configs/rcu/v3.3/P4-A-t-NH-sd-SMP-HP22
-rw-r--r--rcutorture/configs/rcu/v3.3/P5-U-T-NH-sd-SMP-hp27
-rw-r--r--rcutorture/configs/rcu/v3.3/PT1-nh23
-rw-r--r--rcutorture/configs/rcu/v3.3/PT2-NH22
-rw-r--r--rcutorture/configs/rcu/v3.3/ver_functions.sh44
-rw-r--r--rcutorture/configs/rcu/v3.5/CFLIST14
-rw-r--r--rcutorture/configs/rcu/v3.5/N1-S-T-NH-SD-SMP-HP19
-rw-r--r--rcutorture/configs/rcu/v3.5/N2-2-t-nh-sd-SMP-hp20
-rw-r--r--rcutorture/configs/rcu/v3.5/N3-3-T-nh-SD-SMP-hp22
-rw-r--r--rcutorture/configs/rcu/v3.5/N4-A-t-NH-sd-SMP-HP18
-rw-r--r--rcutorture/configs/rcu/v3.5/N5-U-T-NH-sd-SMP-hp22
-rw-r--r--rcutorture/configs/rcu/v3.5/NT1-nh23
-rw-r--r--rcutorture/configs/rcu/v3.5/NT3-NH20
-rw-r--r--rcutorture/configs/rcu/v3.5/P1-S-T-NH-SD-SMP-HP20
-rw-r--r--rcutorture/configs/rcu/v3.5/P2-2-t-nh-sd-SMP-hp20
-rw-r--r--rcutorture/configs/rcu/v3.5/P3-3-T-nh-SD-SMP-hp20
-rw-r--r--rcutorture/configs/rcu/v3.5/P4-A-t-NH-sd-SMP-HP22
-rw-r--r--rcutorture/configs/rcu/v3.5/P5-U-T-NH-sd-SMP-hp27
-rw-r--r--rcutorture/configs/rcu/v3.5/PT1-nh23
-rw-r--r--rcutorture/configs/rcu/v3.5/PT2-NH22
-rw-r--r--rcutorture/configs/rcu/v3.5/ver_functions.sh57
-rw-r--r--rcutorture/configs/rcu/ver_functions.sh57
-rw-r--r--rcutorture/doc/TINY_RCU.txt40
-rw-r--r--rcutorture/doc/TREE_RCU-kconfig.txt94
-rw-r--r--rcutorture/doc/initrd.txt90
-rw-r--r--rcutorture/doc/rcu-test-image.txt42
-rw-r--r--sysctl/Makefile19
-rw-r--r--sysctl/common_tests109
-rw-r--r--sysctl/run_numerictests10
-rw-r--r--sysctl/run_stringtests77
-rw-r--r--timers/Makefile8
-rw-r--r--timers/posix_timers.c221
-rw-r--r--user/Makefile13
-rw-r--r--vm/.gitignore4
-rw-r--r--vm/Makefile15
-rw-r--r--vm/hugepage-mmap.c92
-rw-r--r--vm/hugepage-shm.c100
-rw-r--r--vm/hugetlbfstest.c84
-rw-r--r--vm/map_hugetlb.c79
-rw-r--r--vm/run_vmtests93
-rw-r--r--vm/thuge-gen.c254
236 files changed, 16764 insertions, 0 deletions
diff --git a/Makefile b/Makefile
new file mode 100644
index 0000000..36ff2e4
--- /dev/null
+++ b/Makefile
@@ -0,0 +1,49 @@
+TARGETS = breakpoints
+TARGETS += cpu-hotplug
+TARGETS += efivarfs
+TARGETS += kcmp
+TARGETS += memfd
+TARGETS += memory-hotplug
+TARGETS += mqueue
+TARGETS += mount
+TARGETS += net
+TARGETS += ptrace
+TARGETS += timers
+TARGETS += vm
+TARGETS += powerpc
+TARGETS += user
+TARGETS += sysctl
+TARGETS += firmware
+
+TARGETS_HOTPLUG = cpu-hotplug
+TARGETS_HOTPLUG += memory-hotplug
+
+all:
+ for TARGET in $(TARGETS); do \
+ make -C $$TARGET; \
+ done;
+
+run_tests: all
+ for TARGET in $(TARGETS); do \
+ make -C $$TARGET run_tests; \
+ done;
+
+hotplug:
+ for TARGET in $(TARGETS_HOTPLUG); do \
+ make -C $$TARGET; \
+ done;
+
+run_hotplug: hotplug
+ for TARGET in $(TARGETS_HOTPLUG); do \
+ make -C $$TARGET run_full_test; \
+ done;
+
+clean_hotplug:
+ for TARGET in $(TARGETS_HOTPLUG); do \
+ make -C $$TARGET clean; \
+ done;
+
+clean:
+ for TARGET in $(TARGETS); do \
+ make -C $$TARGET clean; \
+ done;
diff --git a/README.txt b/README.txt
new file mode 100644
index 0000000..2660d5f
--- /dev/null
+++ b/README.txt
@@ -0,0 +1,61 @@
+Linux Kernel Selftests
+
+The kernel contains a set of "self tests" under the tools/testing/selftests/
+directory. These are intended to be small unit tests to exercise individual
+code paths in the kernel.
+
+On some systems, hot-plug tests could hang forever waiting for cpu and
+memory to be ready to be offlined. A special hot-plug target is created
+to run full range of hot-plug tests. In default mode, hot-plug tests run
+in safe mode with a limited scope. In limited mode, cpu-hotplug test is
+run on a single cpu as opposed to all hotplug capable cpus, and memory
+hotplug test is run on 2% of hotplug capable memory instead of 10%.
+
+Running the selftests (hotplug tests are run in limited mode)
+=============================================================
+
+To build the tests:
+
+ $ make -C tools/testing/selftests
+
+
+To run the tests:
+
+ $ make -C tools/testing/selftests run_tests
+
+- note that some tests will require root privileges.
+
+To run only tests targeted for a single subsystem: (including
+hotplug targets in limited mode)
+
+ $ make -C tools/testing/selftests TARGETS=cpu-hotplug run_tests
+
+See the top-level tools/testing/selftests/Makefile for the list of all possible
+targets.
+
+Running the full range hotplug selftests
+========================================
+
+To build the tests:
+
+ $ make -C tools/testing/selftests hotplug
+
+To run the tests:
+
+ $ make -C tools/testing/selftests run_hotplug
+
+- note that some tests will require root privileges.
+
+Contributing new tests
+======================
+
+In general, the rules for for selftests are
+
+ * Do as much as you can if you're not root;
+
+ * Don't take too long;
+
+ * Don't break the build on any architecture, and
+
+ * Don't cause the top-level "make run_tests" to fail if your feature is
+ unconfigured.
diff --git a/breakpoints/Makefile b/breakpoints/Makefile
new file mode 100644
index 0000000..e18b42b
--- /dev/null
+++ b/breakpoints/Makefile
@@ -0,0 +1,23 @@
+# Taken from perf makefile
+uname_M := $(shell uname -m 2>/dev/null || echo not)
+ARCH ?= $(shell echo $(uname_M) | sed -e s/i.86/i386/)
+ifeq ($(ARCH),i386)
+ ARCH := x86
+endif
+ifeq ($(ARCH),x86_64)
+ ARCH := x86
+endif
+
+
+all:
+ifeq ($(ARCH),x86)
+ gcc breakpoint_test.c -o breakpoint_test
+else
+ echo "Not an x86 target, can't build breakpoints selftests"
+endif
+
+run_tests:
+ @./breakpoint_test || echo "breakpoints selftests: [FAIL]"
+
+clean:
+ rm -fr breakpoint_test
diff --git a/breakpoints/breakpoint_test.c b/breakpoints/breakpoint_test.c
new file mode 100644
index 0000000..a0743f3
--- /dev/null
+++ b/breakpoints/breakpoint_test.c
@@ -0,0 +1,394 @@
+/*
+ * Copyright (C) 2011 Red Hat, Inc., Frederic Weisbecker <fweisbec@redhat.com>
+ *
+ * Licensed under the terms of the GNU GPL License version 2
+ *
+ * Selftests for breakpoints (and more generally the do_debug() path) in x86.
+ */
+
+
+#include <sys/ptrace.h>
+#include <unistd.h>
+#include <stddef.h>
+#include <sys/user.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <signal.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+
+
+/* Breakpoint access modes */
+enum {
+ BP_X = 1,
+ BP_RW = 2,
+ BP_W = 4,
+};
+
+static pid_t child_pid;
+
+/*
+ * Ensures the child and parent are always "talking" about
+ * the same test sequence. (ie: that we haven't forgotten
+ * to call check_trapped() somewhere).
+ */
+static int nr_tests;
+
+static void set_breakpoint_addr(void *addr, int n)
+{
+ int ret;
+
+ ret = ptrace(PTRACE_POKEUSER, child_pid,
+ offsetof(struct user, u_debugreg[n]), addr);
+ if (ret) {
+ perror("Can't set breakpoint addr\n");
+ exit(-1);
+ }
+}
+
+static void toggle_breakpoint(int n, int type, int len,
+ int local, int global, int set)
+{
+ int ret;
+
+ int xtype, xlen;
+ unsigned long vdr7, dr7;
+
+ switch (type) {
+ case BP_X:
+ xtype = 0;
+ break;
+ case BP_W:
+ xtype = 1;
+ break;
+ case BP_RW:
+ xtype = 3;
+ break;
+ }
+
+ switch (len) {
+ case 1:
+ xlen = 0;
+ break;
+ case 2:
+ xlen = 4;
+ break;
+ case 4:
+ xlen = 0xc;
+ break;
+ case 8:
+ xlen = 8;
+ break;
+ }
+
+ dr7 = ptrace(PTRACE_PEEKUSER, child_pid,
+ offsetof(struct user, u_debugreg[7]), 0);
+
+ vdr7 = (xlen | xtype) << 16;
+ vdr7 <<= 4 * n;
+
+ if (local) {
+ vdr7 |= 1 << (2 * n);
+ vdr7 |= 1 << 8;
+ }
+ if (global) {
+ vdr7 |= 2 << (2 * n);
+ vdr7 |= 1 << 9;
+ }
+
+ if (set)
+ dr7 |= vdr7;
+ else
+ dr7 &= ~vdr7;
+
+ ret = ptrace(PTRACE_POKEUSER, child_pid,
+ offsetof(struct user, u_debugreg[7]), dr7);
+ if (ret) {
+ perror("Can't set dr7");
+ exit(-1);
+ }
+}
+
+/* Dummy variables to test read/write accesses */
+static unsigned long long dummy_var[4];
+
+/* Dummy functions to test execution accesses */
+static void dummy_func(void) { }
+static void dummy_func1(void) { }
+static void dummy_func2(void) { }
+static void dummy_func3(void) { }
+
+static void (*dummy_funcs[])(void) = {
+ dummy_func,
+ dummy_func1,
+ dummy_func2,
+ dummy_func3,
+};
+
+static int trapped;
+
+static void check_trapped(void)
+{
+ /*
+ * If we haven't trapped, wake up the parent
+ * so that it notices the failure.
+ */
+ if (!trapped)
+ kill(getpid(), SIGUSR1);
+ trapped = 0;
+
+ nr_tests++;
+}
+
+static void write_var(int len)
+{
+ char *pcval; short *psval; int *pival; long long *plval;
+ int i;
+
+ for (i = 0; i < 4; i++) {
+ switch (len) {
+ case 1:
+ pcval = (char *)&dummy_var[i];
+ *pcval = 0xff;
+ break;
+ case 2:
+ psval = (short *)&dummy_var[i];
+ *psval = 0xffff;
+ break;
+ case 4:
+ pival = (int *)&dummy_var[i];
+ *pival = 0xffffffff;
+ break;
+ case 8:
+ plval = (long long *)&dummy_var[i];
+ *plval = 0xffffffffffffffffLL;
+ break;
+ }
+ check_trapped();
+ }
+}
+
+static void read_var(int len)
+{
+ char cval; short sval; int ival; long long lval;
+ int i;
+
+ for (i = 0; i < 4; i++) {
+ switch (len) {
+ case 1:
+ cval = *(char *)&dummy_var[i];
+ break;
+ case 2:
+ sval = *(short *)&dummy_var[i];
+ break;
+ case 4:
+ ival = *(int *)&dummy_var[i];
+ break;
+ case 8:
+ lval = *(long long *)&dummy_var[i];
+ break;
+ }
+ check_trapped();
+ }
+}
+
+/*
+ * Do the r/w/x accesses to trigger the breakpoints. And run
+ * the usual traps.
+ */
+static void trigger_tests(void)
+{
+ int len, local, global, i;
+ char val;
+ int ret;
+
+ ret = ptrace(PTRACE_TRACEME, 0, NULL, 0);
+ if (ret) {
+ perror("Can't be traced?\n");
+ return;
+ }
+
+ /* Wake up father so that it sets up the first test */
+ kill(getpid(), SIGUSR1);
+
+ /* Test instruction breakpoints */
+ for (local = 0; local < 2; local++) {
+ for (global = 0; global < 2; global++) {
+ if (!local && !global)
+ continue;
+
+ for (i = 0; i < 4; i++) {
+ dummy_funcs[i]();
+ check_trapped();
+ }
+ }
+ }
+
+ /* Test write watchpoints */
+ for (len = 1; len <= sizeof(long); len <<= 1) {
+ for (local = 0; local < 2; local++) {
+ for (global = 0; global < 2; global++) {
+ if (!local && !global)
+ continue;
+ write_var(len);
+ }
+ }
+ }
+
+ /* Test read/write watchpoints (on read accesses) */
+ for (len = 1; len <= sizeof(long); len <<= 1) {
+ for (local = 0; local < 2; local++) {
+ for (global = 0; global < 2; global++) {
+ if (!local && !global)
+ continue;
+ read_var(len);
+ }
+ }
+ }
+
+ /* Icebp trap */
+ asm(".byte 0xf1\n");
+ check_trapped();
+
+ /* Int 3 trap */
+ asm("int $3\n");
+ check_trapped();
+
+ kill(getpid(), SIGUSR1);
+}
+
+static void check_success(const char *msg)
+{
+ const char *msg2;
+ int child_nr_tests;
+ int status;
+
+ /* Wait for the child to SIGTRAP */
+ wait(&status);
+
+ msg2 = "Failed";
+
+ if (WSTOPSIG(status) == SIGTRAP) {
+ child_nr_tests = ptrace(PTRACE_PEEKDATA, child_pid,
+ &nr_tests, 0);
+ if (child_nr_tests == nr_tests)
+ msg2 = "Ok";
+ if (ptrace(PTRACE_POKEDATA, child_pid, &trapped, 1)) {
+ perror("Can't poke\n");
+ exit(-1);
+ }
+ }
+
+ nr_tests++;
+
+ printf("%s [%s]\n", msg, msg2);
+}
+
+static void launch_instruction_breakpoints(char *buf, int local, int global)
+{
+ int i;
+
+ for (i = 0; i < 4; i++) {
+ set_breakpoint_addr(dummy_funcs[i], i);
+ toggle_breakpoint(i, BP_X, 1, local, global, 1);
+ ptrace(PTRACE_CONT, child_pid, NULL, 0);
+ sprintf(buf, "Test breakpoint %d with local: %d global: %d",
+ i, local, global);
+ check_success(buf);
+ toggle_breakpoint(i, BP_X, 1, local, global, 0);
+ }
+}
+
+static void launch_watchpoints(char *buf, int mode, int len,
+ int local, int global)
+{
+ const char *mode_str;
+ int i;
+
+ if (mode == BP_W)
+ mode_str = "write";
+ else
+ mode_str = "read";
+
+ for (i = 0; i < 4; i++) {
+ set_breakpoint_addr(&dummy_var[i], i);
+ toggle_breakpoint(i, mode, len, local, global, 1);
+ ptrace(PTRACE_CONT, child_pid, NULL, 0);
+ sprintf(buf, "Test %s watchpoint %d with len: %d local: "
+ "%d global: %d", mode_str, i, len, local, global);
+ check_success(buf);
+ toggle_breakpoint(i, mode, len, local, global, 0);
+ }
+}
+
+/* Set the breakpoints and check the child successfully trigger them */
+static void launch_tests(void)
+{
+ char buf[1024];
+ int len, local, global, i;
+
+ /* Instruction breakpoints */
+ for (local = 0; local < 2; local++) {
+ for (global = 0; global < 2; global++) {
+ if (!local && !global)
+ continue;
+ launch_instruction_breakpoints(buf, local, global);
+ }
+ }
+
+ /* Write watchpoint */
+ for (len = 1; len <= sizeof(long); len <<= 1) {
+ for (local = 0; local < 2; local++) {
+ for (global = 0; global < 2; global++) {
+ if (!local && !global)
+ continue;
+ launch_watchpoints(buf, BP_W, len,
+ local, global);
+ }
+ }
+ }
+
+ /* Read-Write watchpoint */
+ for (len = 1; len <= sizeof(long); len <<= 1) {
+ for (local = 0; local < 2; local++) {
+ for (global = 0; global < 2; global++) {
+ if (!local && !global)
+ continue;
+ launch_watchpoints(buf, BP_RW, len,
+ local, global);
+ }
+ }
+ }
+
+ /* Icebp traps */
+ ptrace(PTRACE_CONT, child_pid, NULL, 0);
+ check_success("Test icebp");
+
+ /* Int 3 traps */
+ ptrace(PTRACE_CONT, child_pid, NULL, 0);
+ check_success("Test int 3 trap");
+
+ ptrace(PTRACE_CONT, child_pid, NULL, 0);
+}
+
+int main(int argc, char **argv)
+{
+ pid_t pid;
+ int ret;
+
+ pid = fork();
+ if (!pid) {
+ trigger_tests();
+ return 0;
+ }
+
+ child_pid = pid;
+
+ wait(NULL);
+
+ launch_tests();
+
+ wait(NULL);
+
+ return 0;
+}
diff --git a/cpu-hotplug/Makefile b/cpu-hotplug/Makefile
new file mode 100644
index 0000000..e9c28d8
--- /dev/null
+++ b/cpu-hotplug/Makefile
@@ -0,0 +1,9 @@
+all:
+
+run_tests:
+ @/bin/bash ./on-off-test.sh || echo "cpu-hotplug selftests: [FAIL]"
+
+run_full_test:
+ @/bin/bash ./on-off-test.sh -a || echo "cpu-hotplug selftests: [FAIL]"
+
+clean:
diff --git a/cpu-hotplug/on-off-test.sh b/cpu-hotplug/on-off-test.sh
new file mode 100644
index 0000000..98b1d65
--- /dev/null
+++ b/cpu-hotplug/on-off-test.sh
@@ -0,0 +1,269 @@
+#!/bin/bash
+
+SYSFS=
+
+prerequisite()
+{
+ msg="skip all tests:"
+
+ if [ $UID != 0 ]; then
+ echo $msg must be run as root >&2
+ exit 0
+ fi
+
+ taskset -p 01 $$
+
+ SYSFS=`mount -t sysfs | head -1 | awk '{ print $3 }'`
+
+ if [ ! -d "$SYSFS" ]; then
+ echo $msg sysfs is not mounted >&2
+ exit 0
+ fi
+
+ if ! ls $SYSFS/devices/system/cpu/cpu* > /dev/null 2>&1; then
+ echo $msg cpu hotplug is not supported >&2
+ exit 0
+ fi
+
+ echo "CPU online/offline summary:"
+ online_cpus=`cat $SYSFS/devices/system/cpu/online`
+ online_max=${online_cpus##*-}
+ echo -e "\t Cpus in online state: $online_cpus"
+
+ offline_cpus=`cat $SYSFS/devices/system/cpu/offline`
+ if [[ "a$offline_cpus" = "a" ]]; then
+ offline_cpus=0
+ else
+ offline_max=${offline_cpus##*-}
+ fi
+ echo -e "\t Cpus in offline state: $offline_cpus"
+}
+
+#
+# list all hot-pluggable CPUs
+#
+hotpluggable_cpus()
+{
+ local state=${1:-.\*}
+
+ for cpu in $SYSFS/devices/system/cpu/cpu*; do
+ if [ -f $cpu/online ] && grep -q $state $cpu/online; then
+ echo ${cpu##/*/cpu}
+ fi
+ done
+}
+
+hotplaggable_offline_cpus()
+{
+ hotpluggable_cpus 0
+}
+
+hotpluggable_online_cpus()
+{
+ hotpluggable_cpus 1
+}
+
+cpu_is_online()
+{
+ grep -q 1 $SYSFS/devices/system/cpu/cpu$1/online
+}
+
+cpu_is_offline()
+{
+ grep -q 0 $SYSFS/devices/system/cpu/cpu$1/online
+}
+
+online_cpu()
+{
+ echo 1 > $SYSFS/devices/system/cpu/cpu$1/online
+}
+
+offline_cpu()
+{
+ echo 0 > $SYSFS/devices/system/cpu/cpu$1/online
+}
+
+online_cpu_expect_success()
+{
+ local cpu=$1
+
+ if ! online_cpu $cpu; then
+ echo $FUNCNAME $cpu: unexpected fail >&2
+ elif ! cpu_is_online $cpu; then
+ echo $FUNCNAME $cpu: unexpected offline >&2
+ fi
+}
+
+online_cpu_expect_fail()
+{
+ local cpu=$1
+
+ if online_cpu $cpu 2> /dev/null; then
+ echo $FUNCNAME $cpu: unexpected success >&2
+ elif ! cpu_is_offline $cpu; then
+ echo $FUNCNAME $cpu: unexpected online >&2
+ fi
+}
+
+offline_cpu_expect_success()
+{
+ local cpu=$1
+
+ if ! offline_cpu $cpu; then
+ echo $FUNCNAME $cpu: unexpected fail >&2
+ elif ! cpu_is_offline $cpu; then
+ echo $FUNCNAME $cpu: unexpected offline >&2
+ fi
+}
+
+offline_cpu_expect_fail()
+{
+ local cpu=$1
+
+ if offline_cpu $cpu 2> /dev/null; then
+ echo $FUNCNAME $cpu: unexpected success >&2
+ elif ! cpu_is_online $cpu; then
+ echo $FUNCNAME $cpu: unexpected offline >&2
+ fi
+}
+
+error=-12
+allcpus=0
+priority=0
+online_cpus=0
+online_max=0
+offline_cpus=0
+offline_max=0
+
+while getopts e:ahp: opt; do
+ case $opt in
+ e)
+ error=$OPTARG
+ ;;
+ a)
+ allcpus=1
+ ;;
+ h)
+ echo "Usage $0 [ -a ] [ -e errno ] [ -p notifier-priority ]"
+ echo -e "\t default offline one cpu"
+ echo -e "\t run with -a option to offline all cpus"
+ exit
+ ;;
+ p)
+ priority=$OPTARG
+ ;;
+ esac
+done
+
+if ! [ "$error" -ge -4095 -a "$error" -lt 0 ]; then
+ echo "error code must be -4095 <= errno < 0" >&2
+ exit 1
+fi
+
+prerequisite
+
+#
+# Safe test (default) - offline and online one cpu
+#
+if [ $allcpus -eq 0 ]; then
+ echo "Limited scope test: one hotplug cpu"
+ echo -e "\t (leaves cpu in the original state):"
+ echo -e "\t online to offline to online: cpu $online_max"
+ offline_cpu_expect_success $online_max
+ online_cpu_expect_success $online_max
+
+ if [[ $offline_cpus -gt 0 ]]; then
+ echo -e "\t offline to online to offline: cpu $offline_max"
+ online_cpu_expect_success $offline_max
+ offline_cpu_expect_success $offline_max
+ fi
+ exit 0
+else
+ echo "Full scope test: all hotplug cpus"
+ echo -e "\t online all offline cpus"
+ echo -e "\t offline all online cpus"
+ echo -e "\t online all offline cpus"
+fi
+
+#
+# Online all hot-pluggable CPUs
+#
+for cpu in `hotplaggable_offline_cpus`; do
+ online_cpu_expect_success $cpu
+done
+
+#
+# Offline all hot-pluggable CPUs
+#
+for cpu in `hotpluggable_online_cpus`; do
+ offline_cpu_expect_success $cpu
+done
+
+#
+# Online all hot-pluggable CPUs again
+#
+for cpu in `hotplaggable_offline_cpus`; do
+ online_cpu_expect_success $cpu
+done
+
+#
+# Test with cpu notifier error injection
+#
+
+DEBUGFS=`mount -t debugfs | head -1 | awk '{ print $3 }'`
+NOTIFIER_ERR_INJECT_DIR=$DEBUGFS/notifier-error-inject/cpu
+
+prerequisite_extra()
+{
+ msg="skip extra tests:"
+
+ /sbin/modprobe -q -r cpu-notifier-error-inject
+ /sbin/modprobe -q cpu-notifier-error-inject priority=$priority
+
+ if [ ! -d "$DEBUGFS" ]; then
+ echo $msg debugfs is not mounted >&2
+ exit 0
+ fi
+
+ if [ ! -d $NOTIFIER_ERR_INJECT_DIR ]; then
+ echo $msg cpu-notifier-error-inject module is not available >&2
+ exit 0
+ fi
+}
+
+prerequisite_extra
+
+#
+# Offline all hot-pluggable CPUs
+#
+echo 0 > $NOTIFIER_ERR_INJECT_DIR/actions/CPU_DOWN_PREPARE/error
+for cpu in `hotpluggable_online_cpus`; do
+ offline_cpu_expect_success $cpu
+done
+
+#
+# Test CPU hot-add error handling (offline => online)
+#
+echo $error > $NOTIFIER_ERR_INJECT_DIR/actions/CPU_UP_PREPARE/error
+for cpu in `hotplaggable_offline_cpus`; do
+ online_cpu_expect_fail $cpu
+done
+
+#
+# Online all hot-pluggable CPUs
+#
+echo 0 > $NOTIFIER_ERR_INJECT_DIR/actions/CPU_UP_PREPARE/error
+for cpu in `hotplaggable_offline_cpus`; do
+ online_cpu_expect_success $cpu
+done
+
+#
+# Test CPU hot-remove error handling (online => offline)
+#
+echo $error > $NOTIFIER_ERR_INJECT_DIR/actions/CPU_DOWN_PREPARE/error
+for cpu in `hotpluggable_online_cpus`; do
+ offline_cpu_expect_fail $cpu
+done
+
+echo 0 > $NOTIFIER_ERR_INJECT_DIR/actions/CPU_DOWN_PREPARE/error
+/sbin/modprobe -q -r cpu-notifier-error-inject
diff --git a/efivarfs/Makefile b/efivarfs/Makefile
new file mode 100644
index 0000000..29e8c6b
--- /dev/null
+++ b/efivarfs/Makefile
@@ -0,0 +1,12 @@
+CC = $(CROSS_COMPILE)gcc
+CFLAGS = -Wall
+
+test_objs = open-unlink create-read
+
+all: $(test_objs)
+
+run_tests: all
+ @/bin/bash ./efivarfs.sh || echo "efivarfs selftests: [FAIL]"
+
+clean:
+ rm -f $(test_objs)
diff --git a/efivarfs/create-read.c b/efivarfs/create-read.c
new file mode 100644
index 0000000..7feef18
--- /dev/null
+++ b/efivarfs/create-read.c
@@ -0,0 +1,38 @@
+#include <stdio.h>
+#include <stdint.h>
+#include <stdlib.h>
+#include <unistd.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <fcntl.h>
+#include <errno.h>
+#include <string.h>
+
+int main(int argc, char **argv)
+{
+ const char *path;
+ char buf[4];
+ int fd, rc;
+
+ if (argc < 2) {
+ fprintf(stderr, "usage: %s <path>\n", argv[0]);
+ return EXIT_FAILURE;
+ }
+
+ path = argv[1];
+
+ /* create a test variable */
+ fd = open(path, O_RDWR | O_CREAT, 0600);
+ if (fd < 0) {
+ perror("open(O_WRONLY)");
+ return EXIT_FAILURE;
+ }
+
+ rc = read(fd, buf, sizeof(buf));
+ if (rc != 0) {
+ fprintf(stderr, "Reading a new var should return EOF\n");
+ return EXIT_FAILURE;
+ }
+
+ return EXIT_SUCCESS;
+}
diff --git a/efivarfs/efivarfs.sh b/efivarfs/efivarfs.sh
new file mode 100644
index 0000000..77edcdc
--- /dev/null
+++ b/efivarfs/efivarfs.sh
@@ -0,0 +1,198 @@
+#!/bin/bash
+
+efivarfs_mount=/sys/firmware/efi/efivars
+test_guid=210be57c-9849-4fc7-a635-e6382d1aec27
+
+check_prereqs()
+{
+ local msg="skip all tests:"
+
+ if [ $UID != 0 ]; then
+ echo $msg must be run as root >&2
+ exit 0
+ fi
+
+ if ! grep -q "^\S\+ $efivarfs_mount efivarfs" /proc/mounts; then
+ echo $msg efivarfs is not mounted on $efivarfs_mount >&2
+ exit 0
+ fi
+}
+
+run_test()
+{
+ local test="$1"
+
+ echo "--------------------"
+ echo "running $test"
+ echo "--------------------"
+
+ if [ "$(type -t $test)" = 'function' ]; then
+ ( $test )
+ else
+ ( ./$test )
+ fi
+
+ if [ $? -ne 0 ]; then
+ echo " [FAIL]"
+ rc=1
+ else
+ echo " [PASS]"
+ fi
+}
+
+test_create()
+{
+ local attrs='\x07\x00\x00\x00'
+ local file=$efivarfs_mount/$FUNCNAME-$test_guid
+
+ printf "$attrs\x00" > $file
+
+ if [ ! -e $file ]; then
+ echo "$file couldn't be created" >&2
+ exit 1
+ fi
+
+ if [ $(stat -c %s $file) -ne 5 ]; then
+ echo "$file has invalid size" >&2
+ exit 1
+ fi
+}
+
+test_create_empty()
+{
+ local file=$efivarfs_mount/$FUNCNAME-$test_guid
+
+ : > $file
+
+ if [ ! -e $file ]; then
+ echo "$file can not be created without writing" >&2
+ exit 1
+ fi
+}
+
+test_create_read()
+{
+ local file=$efivarfs_mount/$FUNCNAME-$test_guid
+ ./create-read $file
+}
+
+test_delete()
+{
+ local attrs='\x07\x00\x00\x00'
+ local file=$efivarfs_mount/$FUNCNAME-$test_guid
+
+ printf "$attrs\x00" > $file
+
+ if [ ! -e $file ]; then
+ echo "$file couldn't be created" >&2
+ exit 1
+ fi
+
+ rm $file
+
+ if [ -e $file ]; then
+ echo "$file couldn't be deleted" >&2
+ exit 1
+ fi
+
+}
+
+# test that we can remove a variable by issuing a write with only
+# attributes specified
+test_zero_size_delete()
+{
+ local attrs='\x07\x00\x00\x00'
+ local file=$efivarfs_mount/$FUNCNAME-$test_guid
+
+ printf "$attrs\x00" > $file
+
+ if [ ! -e $file ]; then
+ echo "$file does not exist" >&2
+ exit 1
+ fi
+
+ printf "$attrs" > $file
+
+ if [ -e $file ]; then
+ echo "$file should have been deleted" >&2
+ exit 1
+ fi
+}
+
+test_open_unlink()
+{
+ local file=$efivarfs_mount/$FUNCNAME-$test_guid
+ ./open-unlink $file
+}
+
+# test that we can create a range of filenames
+test_valid_filenames()
+{
+ local attrs='\x07\x00\x00\x00'
+ local ret=0
+
+ local file_list="abc dump-type0-11-1-1362436005 1234 -"
+ for f in $file_list; do
+ local file=$efivarfs_mount/$f-$test_guid
+
+ printf "$attrs\x00" > $file
+
+ if [ ! -e $file ]; then
+ echo "$file could not be created" >&2
+ ret=1
+ else
+ rm $file
+ fi
+ done
+
+ exit $ret
+}
+
+test_invalid_filenames()
+{
+ local attrs='\x07\x00\x00\x00'
+ local ret=0
+
+ local file_list="
+ -1234-1234-1234-123456789abc
+ foo
+ foo-bar
+ -foo-
+ foo-barbazba-foob-foob-foob-foobarbazfoo
+ foo-------------------------------------
+ -12345678-1234-1234-1234-123456789abc
+ a-12345678=1234-1234-1234-123456789abc
+ a-12345678-1234=1234-1234-123456789abc
+ a-12345678-1234-1234=1234-123456789abc
+ a-12345678-1234-1234-1234=123456789abc
+ 1112345678-1234-1234-1234-123456789abc"
+
+ for f in $file_list; do
+ local file=$efivarfs_mount/$f
+
+ printf "$attrs\x00" 2>/dev/null > $file
+
+ if [ -e $file ]; then
+ echo "Creating $file should have failed" >&2
+ rm $file
+ ret=1
+ fi
+ done
+
+ exit $ret
+}
+
+check_prereqs
+
+rc=0
+
+run_test test_create
+run_test test_create_empty
+run_test test_create_read
+run_test test_delete
+run_test test_zero_size_delete
+run_test test_open_unlink
+run_test test_valid_filenames
+run_test test_invalid_filenames
+
+exit $rc
diff --git a/efivarfs/open-unlink.c b/efivarfs/open-unlink.c
new file mode 100644
index 0000000..8c07644
--- /dev/null
+++ b/efivarfs/open-unlink.c
@@ -0,0 +1,63 @@
+#include <stdio.h>
+#include <stdint.h>
+#include <stdlib.h>
+#include <unistd.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <fcntl.h>
+
+int main(int argc, char **argv)
+{
+ const char *path;
+ char buf[5];
+ int fd, rc;
+
+ if (argc < 2) {
+ fprintf(stderr, "usage: %s <path>\n", argv[0]);
+ return EXIT_FAILURE;
+ }
+
+ path = argv[1];
+
+ /* attributes: EFI_VARIABLE_NON_VOLATILE |
+ * EFI_VARIABLE_BOOTSERVICE_ACCESS |
+ * EFI_VARIABLE_RUNTIME_ACCESS
+ */
+ *(uint32_t *)buf = 0x7;
+ buf[4] = 0;
+
+ /* create a test variable */
+ fd = open(path, O_WRONLY | O_CREAT);
+ if (fd < 0) {
+ perror("open(O_WRONLY)");
+ return EXIT_FAILURE;
+ }
+
+ rc = write(fd, buf, sizeof(buf));
+ if (rc != sizeof(buf)) {
+ perror("write");
+ return EXIT_FAILURE;
+ }
+
+ close(fd);
+
+ fd = open(path, O_RDONLY);
+ if (fd < 0) {
+ perror("open");
+ return EXIT_FAILURE;
+ }
+
+ if (unlink(path) < 0) {
+ perror("unlink");
+ return EXIT_FAILURE;
+ }
+
+ rc = read(fd, buf, sizeof(buf));
+ if (rc > 0) {
+ fprintf(stderr, "reading from an unlinked variable "
+ "shouldn't be possible\n");
+ return EXIT_FAILURE;
+ }
+
+ return EXIT_SUCCESS;
+}
diff --git a/firmware/Makefile b/firmware/Makefile
new file mode 100644
index 0000000..e23cce0
--- /dev/null
+++ b/firmware/Makefile
@@ -0,0 +1,27 @@
+# Makefile for firmware loading selftests
+
+# No binaries, but make sure arg-less "make" doesn't trigger "run_tests"
+all:
+
+fw_filesystem:
+ @if /bin/sh ./fw_filesystem.sh ; then \
+ echo "fw_filesystem: ok"; \
+ else \
+ echo "fw_filesystem: [FAIL]"; \
+ exit 1; \
+ fi
+
+fw_userhelper:
+ @if /bin/sh ./fw_userhelper.sh ; then \
+ echo "fw_userhelper: ok"; \
+ else \
+ echo "fw_userhelper: [FAIL]"; \
+ exit 1; \
+ fi
+
+run_tests: all fw_filesystem fw_userhelper
+
+# Nothing to clean up.
+clean:
+
+.PHONY: all clean run_tests fw_filesystem fw_userhelper
diff --git a/firmware/fw_filesystem.sh b/firmware/fw_filesystem.sh
new file mode 100644
index 0000000..3fc6c10
--- /dev/null
+++ b/firmware/fw_filesystem.sh
@@ -0,0 +1,62 @@
+#!/bin/sh
+# This validates that the kernel will load firmware out of its list of
+# firmware locations on disk. Since the user helper does similar work,
+# we reset the custom load directory to a location the user helper doesn't
+# know so we can be sure we're not accidentally testing the user helper.
+set -e
+
+modprobe test_firmware
+
+DIR=/sys/devices/virtual/misc/test_firmware
+
+OLD_TIMEOUT=$(cat /sys/class/firmware/timeout)
+OLD_FWPATH=$(cat /sys/module/firmware_class/parameters/path)
+
+FWPATH=$(mktemp -d)
+FW="$FWPATH/test-firmware.bin"
+
+test_finish()
+{
+ echo "$OLD_TIMEOUT" >/sys/class/firmware/timeout
+ echo -n "$OLD_PATH" >/sys/module/firmware_class/parameters/path
+ rm -f "$FW"
+ rmdir "$FWPATH"
+}
+
+trap "test_finish" EXIT
+
+# Turn down the timeout so failures don't take so long.
+echo 1 >/sys/class/firmware/timeout
+# Set the kernel search path.
+echo -n "$FWPATH" >/sys/module/firmware_class/parameters/path
+
+# This is an unlikely real-world firmware content. :)
+echo "ABCD0123" >"$FW"
+
+NAME=$(basename "$FW")
+
+# Request a firmware that doesn't exist, it should fail.
+echo -n "nope-$NAME" >"$DIR"/trigger_request
+if diff -q "$FW" /dev/test_firmware >/dev/null ; then
+ echo "$0: firmware was not expected to match" >&2
+ exit 1
+else
+ echo "$0: timeout works"
+fi
+
+# This should succeed via kernel load or will fail after 1 second after
+# being handed over to the user helper, which won't find the fw either.
+if ! echo -n "$NAME" >"$DIR"/trigger_request ; then
+ echo "$0: could not trigger request" >&2
+ exit 1
+fi
+
+# Verify the contents are what we expect.
+if ! diff -q "$FW" /dev/test_firmware >/dev/null ; then
+ echo "$0: firmware was not loaded" >&2
+ exit 1
+else
+ echo "$0: filesystem loading works"
+fi
+
+exit 0
diff --git a/firmware/fw_userhelper.sh b/firmware/fw_userhelper.sh
new file mode 100644
index 0000000..6efbade
--- /dev/null
+++ b/firmware/fw_userhelper.sh
@@ -0,0 +1,89 @@
+#!/bin/sh
+# This validates that the kernel will fall back to using the user helper
+# to load firmware it can't find on disk itself. We must request a firmware
+# that the kernel won't find, and any installed helper (e.g. udev) also
+# won't find so that we can do the load ourself manually.
+set -e
+
+modprobe test_firmware
+
+DIR=/sys/devices/virtual/misc/test_firmware
+
+OLD_TIMEOUT=$(cat /sys/class/firmware/timeout)
+
+FWPATH=$(mktemp -d)
+FW="$FWPATH/test-firmware.bin"
+
+test_finish()
+{
+ echo "$OLD_TIMEOUT" >/sys/class/firmware/timeout
+ rm -f "$FW"
+ rmdir "$FWPATH"
+}
+
+load_fw()
+{
+ local name="$1"
+ local file="$2"
+
+ # This will block until our load (below) has finished.
+ echo -n "$name" >"$DIR"/trigger_request &
+
+ # Give kernel a chance to react.
+ local timeout=10
+ while [ ! -e "$DIR"/"$name"/loading ]; do
+ sleep 0.1
+ timeout=$(( $timeout - 1 ))
+ if [ "$timeout" -eq 0 ]; then
+ echo "$0: firmware interface never appeared" >&2
+ exit 1
+ fi
+ done
+
+ echo 1 >"$DIR"/"$name"/loading
+ cat "$file" >"$DIR"/"$name"/data
+ echo 0 >"$DIR"/"$name"/loading
+
+ # Wait for request to finish.
+ wait
+}
+
+trap "test_finish" EXIT
+
+# This is an unlikely real-world firmware content. :)
+echo "ABCD0123" >"$FW"
+NAME=$(basename "$FW")
+
+# Test failure when doing nothing (timeout works).
+echo 1 >/sys/class/firmware/timeout
+echo -n "$NAME" >"$DIR"/trigger_request
+if diff -q "$FW" /dev/test_firmware >/dev/null ; then
+ echo "$0: firmware was not expected to match" >&2
+ exit 1
+else
+ echo "$0: timeout works"
+fi
+
+# Put timeout high enough for us to do work but not so long that failures
+# slow down this test too much.
+echo 4 >/sys/class/firmware/timeout
+
+# Load this script instead of the desired firmware.
+load_fw "$NAME" "$0"
+if diff -q "$FW" /dev/test_firmware >/dev/null ; then
+ echo "$0: firmware was not expected to match" >&2
+ exit 1
+else
+ echo "$0: firmware comparison works"
+fi
+
+# Do a proper load, which should work correctly.
+load_fw "$NAME" "$FW"
+if ! diff -q "$FW" /dev/test_firmware >/dev/null ; then
+ echo "$0: firmware was not loaded" >&2
+ exit 1
+else
+ echo "$0: user helper firmware loading works"
+fi
+
+exit 0
diff --git a/ipc/Makefile b/ipc/Makefile
new file mode 100644
index 0000000..74bbefd
--- /dev/null
+++ b/ipc/Makefile
@@ -0,0 +1,25 @@
+uname_M := $(shell uname -m 2>/dev/null || echo not)
+ARCH ?= $(shell echo $(uname_M) | sed -e s/i.86/i386/)
+ifeq ($(ARCH),i386)
+ ARCH := x86
+ CFLAGS := -DCONFIG_X86_32 -D__i386__
+endif
+ifeq ($(ARCH),x86_64)
+ ARCH := x86
+ CFLAGS := -DCONFIG_X86_64 -D__x86_64__
+endif
+
+CFLAGS += -I../../../../usr/include/
+
+all:
+ifeq ($(ARCH),x86)
+ gcc $(CFLAGS) msgque.c -o msgque_test
+else
+ echo "Not an x86 target, can't build msgque selftest"
+endif
+
+run_tests: all
+ ./msgque_test
+
+clean:
+ rm -fr ./msgque_test
diff --git a/ipc/msgque.c b/ipc/msgque.c
new file mode 100644
index 0000000..552f081
--- /dev/null
+++ b/ipc/msgque.c
@@ -0,0 +1,252 @@
+#include <stdlib.h>
+#include <stdio.h>
+#include <string.h>
+#include <errno.h>
+#include <linux/msg.h>
+#include <fcntl.h>
+
+#define MAX_MSG_SIZE 32
+
+struct msg1 {
+ int msize;
+ long mtype;
+ char mtext[MAX_MSG_SIZE];
+};
+
+#define TEST_STRING "Test sysv5 msg"
+#define MSG_TYPE 1
+
+#define ANOTHER_TEST_STRING "Yet another test sysv5 msg"
+#define ANOTHER_MSG_TYPE 26538
+
+struct msgque_data {
+ key_t key;
+ int msq_id;
+ int qbytes;
+ int qnum;
+ int mode;
+ struct msg1 *messages;
+};
+
+int restore_queue(struct msgque_data *msgque)
+{
+ int fd, ret, id, i;
+ char buf[32];
+
+ fd = open("/proc/sys/kernel/msg_next_id", O_WRONLY);
+ if (fd == -1) {
+ printf("Failed to open /proc/sys/kernel/msg_next_id\n");
+ return -errno;
+ }
+ sprintf(buf, "%d", msgque->msq_id);
+
+ ret = write(fd, buf, strlen(buf));
+ if (ret != strlen(buf)) {
+ printf("Failed to write to /proc/sys/kernel/msg_next_id\n");
+ return -errno;
+ }
+
+ id = msgget(msgque->key, msgque->mode | IPC_CREAT | IPC_EXCL);
+ if (id == -1) {
+ printf("Failed to create queue\n");
+ return -errno;
+ }
+
+ if (id != msgque->msq_id) {
+ printf("Restored queue has wrong id (%d instead of %d)\n",
+ id, msgque->msq_id);
+ ret = -EFAULT;
+ goto destroy;
+ }
+
+ for (i = 0; i < msgque->qnum; i++) {
+ if (msgsnd(msgque->msq_id, &msgque->messages[i].mtype,
+ msgque->messages[i].msize, IPC_NOWAIT) != 0) {
+ printf("msgsnd failed (%m)\n");
+ ret = -errno;
+ goto destroy;
+ };
+ }
+ return 0;
+
+destroy:
+ if (msgctl(id, IPC_RMID, 0))
+ printf("Failed to destroy queue: %d\n", -errno);
+ return ret;
+}
+
+int check_and_destroy_queue(struct msgque_data *msgque)
+{
+ struct msg1 message;
+ int cnt = 0, ret;
+
+ while (1) {
+ ret = msgrcv(msgque->msq_id, &message.mtype, MAX_MSG_SIZE,
+ 0, IPC_NOWAIT);
+ if (ret < 0) {
+ if (errno == ENOMSG)
+ break;
+ printf("Failed to read IPC message: %m\n");
+ ret = -errno;
+ goto err;
+ }
+ if (ret != msgque->messages[cnt].msize) {
+ printf("Wrong message size: %d (expected %d)\n", ret,
+ msgque->messages[cnt].msize);
+ ret = -EINVAL;
+ goto err;
+ }
+ if (message.mtype != msgque->messages[cnt].mtype) {
+ printf("Wrong message type\n");
+ ret = -EINVAL;
+ goto err;
+ }
+ if (memcmp(message.mtext, msgque->messages[cnt].mtext, ret)) {
+ printf("Wrong message content\n");
+ ret = -EINVAL;
+ goto err;
+ }
+ cnt++;
+ }
+
+ if (cnt != msgque->qnum) {
+ printf("Wrong message number\n");
+ ret = -EINVAL;
+ goto err;
+ }
+
+ ret = 0;
+err:
+ if (msgctl(msgque->msq_id, IPC_RMID, 0)) {
+ printf("Failed to destroy queue: %d\n", -errno);
+ return -errno;
+ }
+ return ret;
+}
+
+int dump_queue(struct msgque_data *msgque)
+{
+ struct msqid64_ds ds;
+ int kern_id;
+ int i, ret;
+
+ for (kern_id = 0; kern_id < 256; kern_id++) {
+ ret = msgctl(kern_id, MSG_STAT, &ds);
+ if (ret < 0) {
+ if (errno == -EINVAL)
+ continue;
+ printf("Failed to get stats for IPC queue with id %d\n",
+ kern_id);
+ return -errno;
+ }
+
+ if (ret == msgque->msq_id)
+ break;
+ }
+
+ msgque->messages = malloc(sizeof(struct msg1) * ds.msg_qnum);
+ if (msgque->messages == NULL) {
+ printf("Failed to get stats for IPC queue\n");
+ return -ENOMEM;
+ }
+
+ msgque->qnum = ds.msg_qnum;
+ msgque->mode = ds.msg_perm.mode;
+ msgque->qbytes = ds.msg_qbytes;
+
+ for (i = 0; i < msgque->qnum; i++) {
+ ret = msgrcv(msgque->msq_id, &msgque->messages[i].mtype,
+ MAX_MSG_SIZE, i, IPC_NOWAIT | MSG_COPY);
+ if (ret < 0) {
+ printf("Failed to copy IPC message: %m (%d)\n", errno);
+ return -errno;
+ }
+ msgque->messages[i].msize = ret;
+ }
+ return 0;
+}
+
+int fill_msgque(struct msgque_data *msgque)
+{
+ struct msg1 msgbuf;
+
+ msgbuf.mtype = MSG_TYPE;
+ memcpy(msgbuf.mtext, TEST_STRING, sizeof(TEST_STRING));
+ if (msgsnd(msgque->msq_id, &msgbuf.mtype, sizeof(TEST_STRING),
+ IPC_NOWAIT) != 0) {
+ printf("First message send failed (%m)\n");
+ return -errno;
+ };
+
+ msgbuf.mtype = ANOTHER_MSG_TYPE;
+ memcpy(msgbuf.mtext, ANOTHER_TEST_STRING, sizeof(ANOTHER_TEST_STRING));
+ if (msgsnd(msgque->msq_id, &msgbuf.mtype, sizeof(ANOTHER_TEST_STRING),
+ IPC_NOWAIT) != 0) {
+ printf("Second message send failed (%m)\n");
+ return -errno;
+ };
+ return 0;
+}
+
+int main(int argc, char **argv)
+{
+ int msg, pid, err;
+ struct msgque_data msgque;
+
+ if (getuid() != 0) {
+ printf("Please run the test as root - Exiting.\n");
+ exit(1);
+ }
+
+ msgque.key = ftok(argv[0], 822155650);
+ if (msgque.key == -1) {
+ printf("Can't make key\n");
+ return -errno;
+ }
+
+ msgque.msq_id = msgget(msgque.key, IPC_CREAT | IPC_EXCL | 0666);
+ if (msgque.msq_id == -1) {
+ err = -errno;
+ printf("Can't create queue\n");
+ goto err_out;
+ }
+
+ err = fill_msgque(&msgque);
+ if (err) {
+ printf("Failed to fill queue\n");
+ goto err_destroy;
+ }
+
+ err = dump_queue(&msgque);
+ if (err) {
+ printf("Failed to dump queue\n");
+ goto err_destroy;
+ }
+
+ err = check_and_destroy_queue(&msgque);
+ if (err) {
+ printf("Failed to check and destroy queue\n");
+ goto err_out;
+ }
+
+ err = restore_queue(&msgque);
+ if (err) {
+ printf("Failed to restore queue\n");
+ goto err_destroy;
+ }
+
+ err = check_and_destroy_queue(&msgque);
+ if (err) {
+ printf("Failed to test queue\n");
+ goto err_out;
+ }
+ return 0;
+
+err_destroy:
+ if (msgctl(msgque.msq_id, IPC_RMID, 0)) {
+ printf("Failed to destroy queue: %d\n", -errno);
+ return -errno;
+ }
+err_out:
+ return err;
+}
diff --git a/kcmp/.gitignore b/kcmp/.gitignore
new file mode 100644
index 0000000..5a9b373
--- /dev/null
+++ b/kcmp/.gitignore
@@ -0,0 +1,2 @@
+kcmp_test
+kcmp-test-file
diff --git a/kcmp/Makefile b/kcmp/Makefile
new file mode 100644
index 0000000..8aabd82
--- /dev/null
+++ b/kcmp/Makefile
@@ -0,0 +1,28 @@
+uname_M := $(shell uname -m 2>/dev/null || echo not)
+ARCH ?= $(shell echo $(uname_M) | sed -e s/i.86/i386/)
+ifeq ($(ARCH),i386)
+ ARCH := x86
+ CFLAGS := -DCONFIG_X86_32 -D__i386__
+endif
+ifeq ($(ARCH),x86_64)
+ ARCH := x86
+ CFLAGS := -DCONFIG_X86_64 -D__x86_64__
+endif
+
+CFLAGS += -I../../../../arch/x86/include/generated/
+CFLAGS += -I../../../../include/
+CFLAGS += -I../../../../usr/include/
+CFLAGS += -I../../../../arch/x86/include/
+
+all:
+ifeq ($(ARCH),x86)
+ gcc $(CFLAGS) kcmp_test.c -o kcmp_test
+else
+ echo "Not an x86 target, can't build kcmp selftest"
+endif
+
+run_tests: all
+ @./kcmp_test || echo "kcmp_test: [FAIL]"
+
+clean:
+ $(RM) kcmp_test kcmp-test-file
diff --git a/kcmp/kcmp_test.c b/kcmp/kcmp_test.c
new file mode 100644
index 0000000..dbba408
--- /dev/null
+++ b/kcmp/kcmp_test.c
@@ -0,0 +1,96 @@
+#define _GNU_SOURCE
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <signal.h>
+#include <limits.h>
+#include <unistd.h>
+#include <errno.h>
+#include <string.h>
+#include <fcntl.h>
+
+#include <linux/unistd.h>
+#include <linux/kcmp.h>
+
+#include <sys/syscall.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <sys/wait.h>
+
+static long sys_kcmp(int pid1, int pid2, int type, int fd1, int fd2)
+{
+ return syscall(__NR_kcmp, pid1, pid2, type, fd1, fd2);
+}
+
+int main(int argc, char **argv)
+{
+ const char kpath[] = "kcmp-test-file";
+ int pid1, pid2;
+ int fd1, fd2;
+ int status;
+
+ fd1 = open(kpath, O_RDWR | O_CREAT | O_TRUNC, 0644);
+ pid1 = getpid();
+
+ if (fd1 < 0) {
+ perror("Can't create file");
+ exit(1);
+ }
+
+ pid2 = fork();
+ if (pid2 < 0) {
+ perror("fork failed");
+ exit(1);
+ }
+
+ if (!pid2) {
+ int pid2 = getpid();
+ int ret;
+
+ fd2 = open(kpath, O_RDWR, 0644);
+ if (fd2 < 0) {
+ perror("Can't open file");
+ exit(1);
+ }
+
+ /* An example of output and arguments */
+ printf("pid1: %6d pid2: %6d FD: %2ld FILES: %2ld VM: %2ld "
+ "FS: %2ld SIGHAND: %2ld IO: %2ld SYSVSEM: %2ld "
+ "INV: %2ld\n",
+ pid1, pid2,
+ sys_kcmp(pid1, pid2, KCMP_FILE, fd1, fd2),
+ sys_kcmp(pid1, pid2, KCMP_FILES, 0, 0),
+ sys_kcmp(pid1, pid2, KCMP_VM, 0, 0),
+ sys_kcmp(pid1, pid2, KCMP_FS, 0, 0),
+ sys_kcmp(pid1, pid2, KCMP_SIGHAND, 0, 0),
+ sys_kcmp(pid1, pid2, KCMP_IO, 0, 0),
+ sys_kcmp(pid1, pid2, KCMP_SYSVSEM, 0, 0),
+
+ /* This one should fail */
+ sys_kcmp(pid1, pid2, KCMP_TYPES + 1, 0, 0));
+
+ /* This one should return same fd */
+ ret = sys_kcmp(pid1, pid2, KCMP_FILE, fd1, fd1);
+ if (ret) {
+ printf("FAIL: 0 expected but %d returned (%s)\n",
+ ret, strerror(errno));
+ ret = -1;
+ } else
+ printf("PASS: 0 returned as expected\n");
+
+ /* Compare with self */
+ ret = sys_kcmp(pid1, pid1, KCMP_VM, 0, 0);
+ if (ret) {
+ printf("FAIL: 0 expected but %d returned (%s)\n",
+ ret, strerror(errno));
+ ret = -1;
+ } else
+ printf("PASS: 0 returned as expected\n");
+
+ exit(ret);
+ }
+
+ waitpid(pid2, &status, P_ALL);
+
+ return 0;
+}
diff --git a/memfd/.gitignore b/memfd/.gitignore
new file mode 100644
index 0000000..afe87c4
--- /dev/null
+++ b/memfd/.gitignore
@@ -0,0 +1,4 @@
+fuse_mnt
+fuse_test
+memfd_test
+memfd-test-file
diff --git a/memfd/Makefile b/memfd/Makefile
new file mode 100644
index 0000000..ad4ab01
--- /dev/null
+++ b/memfd/Makefile
@@ -0,0 +1,41 @@
+uname_M := $(shell uname -m 2>/dev/null || echo not)
+ARCH ?= $(shell echo $(uname_M) | sed -e s/i.86/i386/)
+ifeq ($(ARCH),i386)
+ ARCH := x86
+endif
+ifeq ($(ARCH),x86_64)
+ ARCH := x86
+endif
+
+CFLAGS += -D_FILE_OFFSET_BITS=64
+CFLAGS += -I../../../../arch/x86/include/generated/uapi/
+CFLAGS += -I../../../../arch/x86/include/uapi/
+CFLAGS += -I../../../../include/uapi/
+CFLAGS += -I../../../../include/
+
+all:
+ifeq ($(ARCH),x86)
+ gcc $(CFLAGS) memfd_test.c -o memfd_test
+else
+ echo "Not an x86 target, can't build memfd selftest"
+endif
+
+run_tests: all
+ifeq ($(ARCH),x86)
+ gcc $(CFLAGS) memfd_test.c -o memfd_test
+endif
+ @./memfd_test || echo "memfd_test: [FAIL]"
+
+build_fuse:
+ifeq ($(ARCH),x86)
+ gcc $(CFLAGS) fuse_mnt.c `pkg-config fuse --cflags --libs` -o fuse_mnt
+ gcc $(CFLAGS) fuse_test.c -o fuse_test
+else
+ echo "Not an x86 target, can't build memfd selftest"
+endif
+
+run_fuse: build_fuse
+ @./run_fuse_test.sh || echo "fuse_test: [FAIL]"
+
+clean:
+ $(RM) memfd_test fuse_test
diff --git a/memfd/fuse_mnt.c b/memfd/fuse_mnt.c
new file mode 100644
index 0000000..feacf12
--- /dev/null
+++ b/memfd/fuse_mnt.c
@@ -0,0 +1,110 @@
+/*
+ * memfd test file-system
+ * This file uses FUSE to create a dummy file-system with only one file /memfd.
+ * This file is read-only and takes 1s per read.
+ *
+ * This file-system is used by the memfd test-cases to force the kernel to pin
+ * pages during reads(). Due to the 1s delay of this file-system, this is a
+ * nice way to test race-conditions against get_user_pages() in the kernel.
+ *
+ * We use direct_io==1 to force the kernel to use direct-IO for this
+ * file-system.
+ */
+
+#define FUSE_USE_VERSION 26
+
+#include <fuse.h>
+#include <stdio.h>
+#include <string.h>
+#include <errno.h>
+#include <fcntl.h>
+#include <unistd.h>
+
+static const char memfd_content[] = "memfd-example-content";
+static const char memfd_path[] = "/memfd";
+
+static int memfd_getattr(const char *path, struct stat *st)
+{
+ memset(st, 0, sizeof(*st));
+
+ if (!strcmp(path, "/")) {
+ st->st_mode = S_IFDIR | 0755;
+ st->st_nlink = 2;
+ } else if (!strcmp(path, memfd_path)) {
+ st->st_mode = S_IFREG | 0444;
+ st->st_nlink = 1;
+ st->st_size = strlen(memfd_content);
+ } else {
+ return -ENOENT;
+ }
+
+ return 0;
+}
+
+static int memfd_readdir(const char *path,
+ void *buf,
+ fuse_fill_dir_t filler,
+ off_t offset,
+ struct fuse_file_info *fi)
+{
+ if (strcmp(path, "/"))
+ return -ENOENT;
+
+ filler(buf, ".", NULL, 0);
+ filler(buf, "..", NULL, 0);
+ filler(buf, memfd_path + 1, NULL, 0);
+
+ return 0;
+}
+
+static int memfd_open(const char *path, struct fuse_file_info *fi)
+{
+ if (strcmp(path, memfd_path))
+ return -ENOENT;
+
+ if ((fi->flags & 3) != O_RDONLY)
+ return -EACCES;
+
+ /* force direct-IO */
+ fi->direct_io = 1;
+
+ return 0;
+}
+
+static int memfd_read(const char *path,
+ char *buf,
+ size_t size,
+ off_t offset,
+ struct fuse_file_info *fi)
+{
+ size_t len;
+
+ if (strcmp(path, memfd_path) != 0)
+ return -ENOENT;
+
+ sleep(1);
+
+ len = strlen(memfd_content);
+ if (offset < len) {
+ if (offset + size > len)
+ size = len - offset;
+
+ memcpy(buf, memfd_content + offset, size);
+ } else {
+ size = 0;
+ }
+
+ return size;
+}
+
+static struct fuse_operations memfd_ops = {
+ .getattr = memfd_getattr,
+ .readdir = memfd_readdir,
+ .open = memfd_open,
+ .read = memfd_read,
+};
+
+int main(int argc, char *argv[])
+{
+ return fuse_main(argc, argv, &memfd_ops, NULL);
+}
diff --git a/memfd/fuse_test.c b/memfd/fuse_test.c
new file mode 100644
index 0000000..67908b1
--- /dev/null
+++ b/memfd/fuse_test.c
@@ -0,0 +1,311 @@
+/*
+ * memfd GUP test-case
+ * This tests memfd interactions with get_user_pages(). We require the
+ * fuse_mnt.c program to provide a fake direct-IO FUSE mount-point for us. This
+ * file-system delays _all_ reads by 1s and forces direct-IO. This means, any
+ * read() on files in that file-system will pin the receive-buffer pages for at
+ * least 1s via get_user_pages().
+ *
+ * We use this trick to race ADD_SEALS against a write on a memfd object. The
+ * ADD_SEALS must fail if the memfd pages are still pinned. Note that we use
+ * the read() syscall with our memory-mapped memfd object as receive buffer to
+ * force the kernel to write into our memfd object.
+ */
+
+#define _GNU_SOURCE
+#define __EXPORTED_HEADERS__
+
+#include <errno.h>
+#include <inttypes.h>
+#include <limits.h>
+#include <linux/falloc.h>
+#include <linux/fcntl.h>
+#include <linux/memfd.h>
+#include <sched.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <signal.h>
+#include <string.h>
+#include <sys/mman.h>
+#include <sys/stat.h>
+#include <sys/syscall.h>
+#include <sys/wait.h>
+#include <unistd.h>
+
+#define MFD_DEF_SIZE 8192
+#define STACK_SIZE 65535
+
+static int sys_memfd_create(const char *name,
+ unsigned int flags)
+{
+ return syscall(__NR_memfd_create, name, flags);
+}
+
+static int mfd_assert_new(const char *name, loff_t sz, unsigned int flags)
+{
+ int r, fd;
+
+ fd = sys_memfd_create(name, flags);
+ if (fd < 0) {
+ printf("memfd_create(\"%s\", %u) failed: %m\n",
+ name, flags);
+ abort();
+ }
+
+ r = ftruncate(fd, sz);
+ if (r < 0) {
+ printf("ftruncate(%llu) failed: %m\n", (unsigned long long)sz);
+ abort();
+ }
+
+ return fd;
+}
+
+static __u64 mfd_assert_get_seals(int fd)
+{
+ long r;
+
+ r = fcntl(fd, F_GET_SEALS);
+ if (r < 0) {
+ printf("GET_SEALS(%d) failed: %m\n", fd);
+ abort();
+ }
+
+ return r;
+}
+
+static void mfd_assert_has_seals(int fd, __u64 seals)
+{
+ __u64 s;
+
+ s = mfd_assert_get_seals(fd);
+ if (s != seals) {
+ printf("%llu != %llu = GET_SEALS(%d)\n",
+ (unsigned long long)seals, (unsigned long long)s, fd);
+ abort();
+ }
+}
+
+static void mfd_assert_add_seals(int fd, __u64 seals)
+{
+ long r;
+ __u64 s;
+
+ s = mfd_assert_get_seals(fd);
+ r = fcntl(fd, F_ADD_SEALS, seals);
+ if (r < 0) {
+ printf("ADD_SEALS(%d, %llu -> %llu) failed: %m\n",
+ fd, (unsigned long long)s, (unsigned long long)seals);
+ abort();
+ }
+}
+
+static int mfd_busy_add_seals(int fd, __u64 seals)
+{
+ long r;
+ __u64 s;
+
+ r = fcntl(fd, F_GET_SEALS);
+ if (r < 0)
+ s = 0;
+ else
+ s = r;
+
+ r = fcntl(fd, F_ADD_SEALS, seals);
+ if (r < 0 && errno != EBUSY) {
+ printf("ADD_SEALS(%d, %llu -> %llu) didn't fail as expected with EBUSY: %m\n",
+ fd, (unsigned long long)s, (unsigned long long)seals);
+ abort();
+ }
+
+ return r;
+}
+
+static void *mfd_assert_mmap_shared(int fd)
+{
+ void *p;
+
+ p = mmap(NULL,
+ MFD_DEF_SIZE,
+ PROT_READ | PROT_WRITE,
+ MAP_SHARED,
+ fd,
+ 0);
+ if (p == MAP_FAILED) {
+ printf("mmap() failed: %m\n");
+ abort();
+ }
+
+ return p;
+}
+
+static void *mfd_assert_mmap_private(int fd)
+{
+ void *p;
+
+ p = mmap(NULL,
+ MFD_DEF_SIZE,
+ PROT_READ | PROT_WRITE,
+ MAP_PRIVATE,
+ fd,
+ 0);
+ if (p == MAP_FAILED) {
+ printf("mmap() failed: %m\n");
+ abort();
+ }
+
+ return p;
+}
+
+static int global_mfd = -1;
+static void *global_p = NULL;
+
+static int sealing_thread_fn(void *arg)
+{
+ int sig, r;
+
+ /*
+ * This thread first waits 200ms so any pending operation in the parent
+ * is correctly started. After that, it tries to seal @global_mfd as
+ * SEAL_WRITE. This _must_ fail as the parent thread has a read() into
+ * that memory mapped object still ongoing.
+ * We then wait one more second and try sealing again. This time it
+ * must succeed as there shouldn't be anyone else pinning the pages.
+ */
+
+ /* wait 200ms for FUSE-request to be active */
+ usleep(200000);
+
+ /* unmount mapping before sealing to avoid i_mmap_writable failures */
+ munmap(global_p, MFD_DEF_SIZE);
+
+ /* Try sealing the global file; expect EBUSY or success. Current
+ * kernels will never succeed, but in the future, kernels might
+ * implement page-replacements or other fancy ways to avoid racing
+ * writes. */
+ r = mfd_busy_add_seals(global_mfd, F_SEAL_WRITE);
+ if (r >= 0) {
+ printf("HURRAY! This kernel fixed GUP races!\n");
+ } else {
+ /* wait 1s more so the FUSE-request is done */
+ sleep(1);
+
+ /* try sealing the global file again */
+ mfd_assert_add_seals(global_mfd, F_SEAL_WRITE);
+ }
+
+ return 0;
+}
+
+static pid_t spawn_sealing_thread(void)
+{
+ uint8_t *stack;
+ pid_t pid;
+
+ stack = malloc(STACK_SIZE);
+ if (!stack) {
+ printf("malloc(STACK_SIZE) failed: %m\n");
+ abort();
+ }
+
+ pid = clone(sealing_thread_fn,
+ stack + STACK_SIZE,
+ SIGCHLD | CLONE_FILES | CLONE_FS | CLONE_VM,
+ NULL);
+ if (pid < 0) {
+ printf("clone() failed: %m\n");
+ abort();
+ }
+
+ return pid;
+}
+
+static void join_sealing_thread(pid_t pid)
+{
+ waitpid(pid, NULL, 0);
+}
+
+int main(int argc, char **argv)
+{
+ static const char zero[MFD_DEF_SIZE];
+ int fd, mfd, r;
+ void *p;
+ int was_sealed;
+ pid_t pid;
+
+ if (argc < 2) {
+ printf("error: please pass path to file in fuse_mnt mount-point\n");
+ abort();
+ }
+
+ /* open FUSE memfd file for GUP testing */
+ printf("opening: %s\n", argv[1]);
+ fd = open(argv[1], O_RDONLY | O_CLOEXEC);
+ if (fd < 0) {
+ printf("cannot open(\"%s\"): %m\n", argv[1]);
+ abort();
+ }
+
+ /* create new memfd-object */
+ mfd = mfd_assert_new("kern_memfd_fuse",
+ MFD_DEF_SIZE,
+ MFD_CLOEXEC | MFD_ALLOW_SEALING);
+
+ /* mmap memfd-object for writing */
+ p = mfd_assert_mmap_shared(mfd);
+
+ /* pass mfd+mapping to a separate sealing-thread which tries to seal
+ * the memfd objects with SEAL_WRITE while we write into it */
+ global_mfd = mfd;
+ global_p = p;
+ pid = spawn_sealing_thread();
+
+ /* Use read() on the FUSE file to read into our memory-mapped memfd
+ * object. This races the other thread which tries to seal the
+ * memfd-object.
+ * If @fd is on the memfd-fake-FUSE-FS, the read() is delayed by 1s.
+ * This guarantees that the receive-buffer is pinned for 1s until the
+ * data is written into it. The racing ADD_SEALS should thus fail as
+ * the pages are still pinned. */
+ r = read(fd, p, MFD_DEF_SIZE);
+ if (r < 0) {
+ printf("read() failed: %m\n");
+ abort();
+ } else if (!r) {
+ printf("unexpected EOF on read()\n");
+ abort();
+ }
+
+ was_sealed = mfd_assert_get_seals(mfd) & F_SEAL_WRITE;
+
+ /* Wait for sealing-thread to finish and verify that it
+ * successfully sealed the file after the second try. */
+ join_sealing_thread(pid);
+ mfd_assert_has_seals(mfd, F_SEAL_WRITE);
+
+ /* *IF* the memfd-object was sealed at the time our read() returned,
+ * then the kernel did a page-replacement or canceled the read() (or
+ * whatever magic it did..). In that case, the memfd object is still
+ * all zero.
+ * In case the memfd-object was *not* sealed, the read() was successfull
+ * and the memfd object must *not* be all zero.
+ * Note that in real scenarios, there might be a mixture of both, but
+ * in this test-cases, we have explicit 200ms delays which should be
+ * enough to avoid any in-flight writes. */
+
+ p = mfd_assert_mmap_private(mfd);
+ if (was_sealed && memcmp(p, zero, MFD_DEF_SIZE)) {
+ printf("memfd sealed during read() but data not discarded\n");
+ abort();
+ } else if (!was_sealed && !memcmp(p, zero, MFD_DEF_SIZE)) {
+ printf("memfd sealed after read() but data discarded\n");
+ abort();
+ }
+
+ close(mfd);
+ close(fd);
+
+ printf("fuse: DONE\n");
+
+ return 0;
+}
diff --git a/memfd/memfd_test.c b/memfd/memfd_test.c
new file mode 100644
index 0000000..3634c90
--- /dev/null
+++ b/memfd/memfd_test.c
@@ -0,0 +1,913 @@
+#define _GNU_SOURCE
+#define __EXPORTED_HEADERS__
+
+#include <errno.h>
+#include <inttypes.h>
+#include <limits.h>
+#include <linux/falloc.h>
+#include <linux/fcntl.h>
+#include <linux/memfd.h>
+#include <sched.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <signal.h>
+#include <string.h>
+#include <sys/mman.h>
+#include <sys/stat.h>
+#include <sys/syscall.h>
+#include <unistd.h>
+
+#define MFD_DEF_SIZE 8192
+#define STACK_SIZE 65535
+
+static int sys_memfd_create(const char *name,
+ unsigned int flags)
+{
+ return syscall(__NR_memfd_create, name, flags);
+}
+
+static int mfd_assert_new(const char *name, loff_t sz, unsigned int flags)
+{
+ int r, fd;
+
+ fd = sys_memfd_create(name, flags);
+ if (fd < 0) {
+ printf("memfd_create(\"%s\", %u) failed: %m\n",
+ name, flags);
+ abort();
+ }
+
+ r = ftruncate(fd, sz);
+ if (r < 0) {
+ printf("ftruncate(%llu) failed: %m\n", (unsigned long long)sz);
+ abort();
+ }
+
+ return fd;
+}
+
+static void mfd_fail_new(const char *name, unsigned int flags)
+{
+ int r;
+
+ r = sys_memfd_create(name, flags);
+ if (r >= 0) {
+ printf("memfd_create(\"%s\", %u) succeeded, but failure expected\n",
+ name, flags);
+ close(r);
+ abort();
+ }
+}
+
+static __u64 mfd_assert_get_seals(int fd)
+{
+ long r;
+
+ r = fcntl(fd, F_GET_SEALS);
+ if (r < 0) {
+ printf("GET_SEALS(%d) failed: %m\n", fd);
+ abort();
+ }
+
+ return r;
+}
+
+static void mfd_assert_has_seals(int fd, __u64 seals)
+{
+ __u64 s;
+
+ s = mfd_assert_get_seals(fd);
+ if (s != seals) {
+ printf("%llu != %llu = GET_SEALS(%d)\n",
+ (unsigned long long)seals, (unsigned long long)s, fd);
+ abort();
+ }
+}
+
+static void mfd_assert_add_seals(int fd, __u64 seals)
+{
+ long r;
+ __u64 s;
+
+ s = mfd_assert_get_seals(fd);
+ r = fcntl(fd, F_ADD_SEALS, seals);
+ if (r < 0) {
+ printf("ADD_SEALS(%d, %llu -> %llu) failed: %m\n",
+ fd, (unsigned long long)s, (unsigned long long)seals);
+ abort();
+ }
+}
+
+static void mfd_fail_add_seals(int fd, __u64 seals)
+{
+ long r;
+ __u64 s;
+
+ r = fcntl(fd, F_GET_SEALS);
+ if (r < 0)
+ s = 0;
+ else
+ s = r;
+
+ r = fcntl(fd, F_ADD_SEALS, seals);
+ if (r >= 0) {
+ printf("ADD_SEALS(%d, %llu -> %llu) didn't fail as expected\n",
+ fd, (unsigned long long)s, (unsigned long long)seals);
+ abort();
+ }
+}
+
+static void mfd_assert_size(int fd, size_t size)
+{
+ struct stat st;
+ int r;
+
+ r = fstat(fd, &st);
+ if (r < 0) {
+ printf("fstat(%d) failed: %m\n", fd);
+ abort();
+ } else if (st.st_size != size) {
+ printf("wrong file size %lld, but expected %lld\n",
+ (long long)st.st_size, (long long)size);
+ abort();
+ }
+}
+
+static int mfd_assert_dup(int fd)
+{
+ int r;
+
+ r = dup(fd);
+ if (r < 0) {
+ printf("dup(%d) failed: %m\n", fd);
+ abort();
+ }
+
+ return r;
+}
+
+static void *mfd_assert_mmap_shared(int fd)
+{
+ void *p;
+
+ p = mmap(NULL,
+ MFD_DEF_SIZE,
+ PROT_READ | PROT_WRITE,
+ MAP_SHARED,
+ fd,
+ 0);
+ if (p == MAP_FAILED) {
+ printf("mmap() failed: %m\n");
+ abort();
+ }
+
+ return p;
+}
+
+static void *mfd_assert_mmap_private(int fd)
+{
+ void *p;
+
+ p = mmap(NULL,
+ MFD_DEF_SIZE,
+ PROT_READ,
+ MAP_PRIVATE,
+ fd,
+ 0);
+ if (p == MAP_FAILED) {
+ printf("mmap() failed: %m\n");
+ abort();
+ }
+
+ return p;
+}
+
+static int mfd_assert_open(int fd, int flags, mode_t mode)
+{
+ char buf[512];
+ int r;
+
+ sprintf(buf, "/proc/self/fd/%d", fd);
+ r = open(buf, flags, mode);
+ if (r < 0) {
+ printf("open(%s) failed: %m\n", buf);
+ abort();
+ }
+
+ return r;
+}
+
+static void mfd_fail_open(int fd, int flags, mode_t mode)
+{
+ char buf[512];
+ int r;
+
+ sprintf(buf, "/proc/self/fd/%d", fd);
+ r = open(buf, flags, mode);
+ if (r >= 0) {
+ printf("open(%s) didn't fail as expected\n");
+ abort();
+ }
+}
+
+static void mfd_assert_read(int fd)
+{
+ char buf[16];
+ void *p;
+ ssize_t l;
+
+ l = read(fd, buf, sizeof(buf));
+ if (l != sizeof(buf)) {
+ printf("read() failed: %m\n");
+ abort();
+ }
+
+ /* verify PROT_READ *is* allowed */
+ p = mmap(NULL,
+ MFD_DEF_SIZE,
+ PROT_READ,
+ MAP_PRIVATE,
+ fd,
+ 0);
+ if (p == MAP_FAILED) {
+ printf("mmap() failed: %m\n");
+ abort();
+ }
+ munmap(p, MFD_DEF_SIZE);
+
+ /* verify MAP_PRIVATE is *always* allowed (even writable) */
+ p = mmap(NULL,
+ MFD_DEF_SIZE,
+ PROT_READ | PROT_WRITE,
+ MAP_PRIVATE,
+ fd,
+ 0);
+ if (p == MAP_FAILED) {
+ printf("mmap() failed: %m\n");
+ abort();
+ }
+ munmap(p, MFD_DEF_SIZE);
+}
+
+static void mfd_assert_write(int fd)
+{
+ ssize_t l;
+ void *p;
+ int r;
+
+ /* verify write() succeeds */
+ l = write(fd, "\0\0\0\0", 4);
+ if (l != 4) {
+ printf("write() failed: %m\n");
+ abort();
+ }
+
+ /* verify PROT_READ | PROT_WRITE is allowed */
+ p = mmap(NULL,
+ MFD_DEF_SIZE,
+ PROT_READ | PROT_WRITE,
+ MAP_SHARED,
+ fd,
+ 0);
+ if (p == MAP_FAILED) {
+ printf("mmap() failed: %m\n");
+ abort();
+ }
+ *(char *)p = 0;
+ munmap(p, MFD_DEF_SIZE);
+
+ /* verify PROT_WRITE is allowed */
+ p = mmap(NULL,
+ MFD_DEF_SIZE,
+ PROT_WRITE,
+ MAP_SHARED,
+ fd,
+ 0);
+ if (p == MAP_FAILED) {
+ printf("mmap() failed: %m\n");
+ abort();
+ }
+ *(char *)p = 0;
+ munmap(p, MFD_DEF_SIZE);
+
+ /* verify PROT_READ with MAP_SHARED is allowed and a following
+ * mprotect(PROT_WRITE) allows writing */
+ p = mmap(NULL,
+ MFD_DEF_SIZE,
+ PROT_READ,
+ MAP_SHARED,
+ fd,
+ 0);
+ if (p == MAP_FAILED) {
+ printf("mmap() failed: %m\n");
+ abort();
+ }
+
+ r = mprotect(p, MFD_DEF_SIZE, PROT_READ | PROT_WRITE);
+ if (r < 0) {
+ printf("mprotect() failed: %m\n");
+ abort();
+ }
+
+ *(char *)p = 0;
+ munmap(p, MFD_DEF_SIZE);
+
+ /* verify PUNCH_HOLE works */
+ r = fallocate(fd,
+ FALLOC_FL_PUNCH_HOLE | FALLOC_FL_KEEP_SIZE,
+ 0,
+ MFD_DEF_SIZE);
+ if (r < 0) {
+ printf("fallocate(PUNCH_HOLE) failed: %m\n");
+ abort();
+ }
+}
+
+static void mfd_fail_write(int fd)
+{
+ ssize_t l;
+ void *p;
+ int r;
+
+ /* verify write() fails */
+ l = write(fd, "data", 4);
+ if (l != -EPERM) {
+ printf("expected EPERM on write(), but got %d: %m\n", (int)l);
+ abort();
+ }
+
+ /* verify PROT_READ | PROT_WRITE is not allowed */
+ p = mmap(NULL,
+ MFD_DEF_SIZE,
+ PROT_READ | PROT_WRITE,
+ MAP_SHARED,
+ fd,
+ 0);
+ if (p != MAP_FAILED) {
+ printf("mmap() didn't fail as expected\n");
+ abort();
+ }
+
+ /* verify PROT_WRITE is not allowed */
+ p = mmap(NULL,
+ MFD_DEF_SIZE,
+ PROT_WRITE,
+ MAP_SHARED,
+ fd,
+ 0);
+ if (p != MAP_FAILED) {
+ printf("mmap() didn't fail as expected\n");
+ abort();
+ }
+
+ /* Verify PROT_READ with MAP_SHARED with a following mprotect is not
+ * allowed. Note that for r/w the kernel already prevents the mmap. */
+ p = mmap(NULL,
+ MFD_DEF_SIZE,
+ PROT_READ,
+ MAP_SHARED,
+ fd,
+ 0);
+ if (p != MAP_FAILED) {
+ r = mprotect(p, MFD_DEF_SIZE, PROT_READ | PROT_WRITE);
+ if (r >= 0) {
+ printf("mmap()+mprotect() didn't fail as expected\n");
+ abort();
+ }
+ }
+
+ /* verify PUNCH_HOLE fails */
+ r = fallocate(fd,
+ FALLOC_FL_PUNCH_HOLE | FALLOC_FL_KEEP_SIZE,
+ 0,
+ MFD_DEF_SIZE);
+ if (r >= 0) {
+ printf("fallocate(PUNCH_HOLE) didn't fail as expected\n");
+ abort();
+ }
+}
+
+static void mfd_assert_shrink(int fd)
+{
+ int r, fd2;
+
+ r = ftruncate(fd, MFD_DEF_SIZE / 2);
+ if (r < 0) {
+ printf("ftruncate(SHRINK) failed: %m\n");
+ abort();
+ }
+
+ mfd_assert_size(fd, MFD_DEF_SIZE / 2);
+
+ fd2 = mfd_assert_open(fd,
+ O_RDWR | O_CREAT | O_TRUNC,
+ S_IRUSR | S_IWUSR);
+ close(fd2);
+
+ mfd_assert_size(fd, 0);
+}
+
+static void mfd_fail_shrink(int fd)
+{
+ int r;
+
+ r = ftruncate(fd, MFD_DEF_SIZE / 2);
+ if (r >= 0) {
+ printf("ftruncate(SHRINK) didn't fail as expected\n");
+ abort();
+ }
+
+ mfd_fail_open(fd,
+ O_RDWR | O_CREAT | O_TRUNC,
+ S_IRUSR | S_IWUSR);
+}
+
+static void mfd_assert_grow(int fd)
+{
+ int r;
+
+ r = ftruncate(fd, MFD_DEF_SIZE * 2);
+ if (r < 0) {
+ printf("ftruncate(GROW) failed: %m\n");
+ abort();
+ }
+
+ mfd_assert_size(fd, MFD_DEF_SIZE * 2);
+
+ r = fallocate(fd,
+ 0,
+ 0,
+ MFD_DEF_SIZE * 4);
+ if (r < 0) {
+ printf("fallocate(ALLOC) failed: %m\n");
+ abort();
+ }
+
+ mfd_assert_size(fd, MFD_DEF_SIZE * 4);
+}
+
+static void mfd_fail_grow(int fd)
+{
+ int r;
+
+ r = ftruncate(fd, MFD_DEF_SIZE * 2);
+ if (r >= 0) {
+ printf("ftruncate(GROW) didn't fail as expected\n");
+ abort();
+ }
+
+ r = fallocate(fd,
+ 0,
+ 0,
+ MFD_DEF_SIZE * 4);
+ if (r >= 0) {
+ printf("fallocate(ALLOC) didn't fail as expected\n");
+ abort();
+ }
+}
+
+static void mfd_assert_grow_write(int fd)
+{
+ static char buf[MFD_DEF_SIZE * 8];
+ ssize_t l;
+
+ l = pwrite(fd, buf, sizeof(buf), 0);
+ if (l != sizeof(buf)) {
+ printf("pwrite() failed: %m\n");
+ abort();
+ }
+
+ mfd_assert_size(fd, MFD_DEF_SIZE * 8);
+}
+
+static void mfd_fail_grow_write(int fd)
+{
+ static char buf[MFD_DEF_SIZE * 8];
+ ssize_t l;
+
+ l = pwrite(fd, buf, sizeof(buf), 0);
+ if (l == sizeof(buf)) {
+ printf("pwrite() didn't fail as expected\n");
+ abort();
+ }
+}
+
+static int idle_thread_fn(void *arg)
+{
+ sigset_t set;
+ int sig;
+
+ /* dummy waiter; SIGTERM terminates us anyway */
+ sigemptyset(&set);
+ sigaddset(&set, SIGTERM);
+ sigwait(&set, &sig);
+
+ return 0;
+}
+
+static pid_t spawn_idle_thread(unsigned int flags)
+{
+ uint8_t *stack;
+ pid_t pid;
+
+ stack = malloc(STACK_SIZE);
+ if (!stack) {
+ printf("malloc(STACK_SIZE) failed: %m\n");
+ abort();
+ }
+
+ pid = clone(idle_thread_fn,
+ stack + STACK_SIZE,
+ SIGCHLD | flags,
+ NULL);
+ if (pid < 0) {
+ printf("clone() failed: %m\n");
+ abort();
+ }
+
+ return pid;
+}
+
+static void join_idle_thread(pid_t pid)
+{
+ kill(pid, SIGTERM);
+ waitpid(pid, NULL, 0);
+}
+
+/*
+ * Test memfd_create() syscall
+ * Verify syscall-argument validation, including name checks, flag validation
+ * and more.
+ */
+static void test_create(void)
+{
+ char buf[2048];
+ int fd;
+
+ /* test NULL name */
+ mfd_fail_new(NULL, 0);
+
+ /* test over-long name (not zero-terminated) */
+ memset(buf, 0xff, sizeof(buf));
+ mfd_fail_new(buf, 0);
+
+ /* test over-long zero-terminated name */
+ memset(buf, 0xff, sizeof(buf));
+ buf[sizeof(buf) - 1] = 0;
+ mfd_fail_new(buf, 0);
+
+ /* verify "" is a valid name */
+ fd = mfd_assert_new("", 0, 0);
+ close(fd);
+
+ /* verify invalid O_* open flags */
+ mfd_fail_new("", 0x0100);
+ mfd_fail_new("", ~MFD_CLOEXEC);
+ mfd_fail_new("", ~MFD_ALLOW_SEALING);
+ mfd_fail_new("", ~0);
+ mfd_fail_new("", 0x80000000U);
+
+ /* verify MFD_CLOEXEC is allowed */
+ fd = mfd_assert_new("", 0, MFD_CLOEXEC);
+ close(fd);
+
+ /* verify MFD_ALLOW_SEALING is allowed */
+ fd = mfd_assert_new("", 0, MFD_ALLOW_SEALING);
+ close(fd);
+
+ /* verify MFD_ALLOW_SEALING | MFD_CLOEXEC is allowed */
+ fd = mfd_assert_new("", 0, MFD_ALLOW_SEALING | MFD_CLOEXEC);
+ close(fd);
+}
+
+/*
+ * Test basic sealing
+ * A very basic sealing test to see whether setting/retrieving seals works.
+ */
+static void test_basic(void)
+{
+ int fd;
+
+ fd = mfd_assert_new("kern_memfd_basic",
+ MFD_DEF_SIZE,
+ MFD_CLOEXEC | MFD_ALLOW_SEALING);
+
+ /* add basic seals */
+ mfd_assert_has_seals(fd, 0);
+ mfd_assert_add_seals(fd, F_SEAL_SHRINK |
+ F_SEAL_WRITE);
+ mfd_assert_has_seals(fd, F_SEAL_SHRINK |
+ F_SEAL_WRITE);
+
+ /* add them again */
+ mfd_assert_add_seals(fd, F_SEAL_SHRINK |
+ F_SEAL_WRITE);
+ mfd_assert_has_seals(fd, F_SEAL_SHRINK |
+ F_SEAL_WRITE);
+
+ /* add more seals and seal against sealing */
+ mfd_assert_add_seals(fd, F_SEAL_GROW | F_SEAL_SEAL);
+ mfd_assert_has_seals(fd, F_SEAL_SHRINK |
+ F_SEAL_GROW |
+ F_SEAL_WRITE |
+ F_SEAL_SEAL);
+
+ /* verify that sealing no longer works */
+ mfd_fail_add_seals(fd, F_SEAL_GROW);
+ mfd_fail_add_seals(fd, 0);
+
+ close(fd);
+
+ /* verify sealing does not work without MFD_ALLOW_SEALING */
+ fd = mfd_assert_new("kern_memfd_basic",
+ MFD_DEF_SIZE,
+ MFD_CLOEXEC);
+ mfd_assert_has_seals(fd, F_SEAL_SEAL);
+ mfd_fail_add_seals(fd, F_SEAL_SHRINK |
+ F_SEAL_GROW |
+ F_SEAL_WRITE);
+ mfd_assert_has_seals(fd, F_SEAL_SEAL);
+ close(fd);
+}
+
+/*
+ * Test SEAL_WRITE
+ * Test whether SEAL_WRITE actually prevents modifications.
+ */
+static void test_seal_write(void)
+{
+ int fd;
+
+ fd = mfd_assert_new("kern_memfd_seal_write",
+ MFD_DEF_SIZE,
+ MFD_CLOEXEC | MFD_ALLOW_SEALING);
+ mfd_assert_has_seals(fd, 0);
+ mfd_assert_add_seals(fd, F_SEAL_WRITE);
+ mfd_assert_has_seals(fd, F_SEAL_WRITE);
+
+ mfd_assert_read(fd);
+ mfd_fail_write(fd);
+ mfd_assert_shrink(fd);
+ mfd_assert_grow(fd);
+ mfd_fail_grow_write(fd);
+
+ close(fd);
+}
+
+/*
+ * Test SEAL_SHRINK
+ * Test whether SEAL_SHRINK actually prevents shrinking
+ */
+static void test_seal_shrink(void)
+{
+ int fd;
+
+ fd = mfd_assert_new("kern_memfd_seal_shrink",
+ MFD_DEF_SIZE,
+ MFD_CLOEXEC | MFD_ALLOW_SEALING);
+ mfd_assert_has_seals(fd, 0);
+ mfd_assert_add_seals(fd, F_SEAL_SHRINK);
+ mfd_assert_has_seals(fd, F_SEAL_SHRINK);
+
+ mfd_assert_read(fd);
+ mfd_assert_write(fd);
+ mfd_fail_shrink(fd);
+ mfd_assert_grow(fd);
+ mfd_assert_grow_write(fd);
+
+ close(fd);
+}
+
+/*
+ * Test SEAL_GROW
+ * Test whether SEAL_GROW actually prevents growing
+ */
+static void test_seal_grow(void)
+{
+ int fd;
+
+ fd = mfd_assert_new("kern_memfd_seal_grow",
+ MFD_DEF_SIZE,
+ MFD_CLOEXEC | MFD_ALLOW_SEALING);
+ mfd_assert_has_seals(fd, 0);
+ mfd_assert_add_seals(fd, F_SEAL_GROW);
+ mfd_assert_has_seals(fd, F_SEAL_GROW);
+
+ mfd_assert_read(fd);
+ mfd_assert_write(fd);
+ mfd_assert_shrink(fd);
+ mfd_fail_grow(fd);
+ mfd_fail_grow_write(fd);
+
+ close(fd);
+}
+
+/*
+ * Test SEAL_SHRINK | SEAL_GROW
+ * Test whether SEAL_SHRINK | SEAL_GROW actually prevents resizing
+ */
+static void test_seal_resize(void)
+{
+ int fd;
+
+ fd = mfd_assert_new("kern_memfd_seal_resize",
+ MFD_DEF_SIZE,
+ MFD_CLOEXEC | MFD_ALLOW_SEALING);
+ mfd_assert_has_seals(fd, 0);
+ mfd_assert_add_seals(fd, F_SEAL_SHRINK | F_SEAL_GROW);
+ mfd_assert_has_seals(fd, F_SEAL_SHRINK | F_SEAL_GROW);
+
+ mfd_assert_read(fd);
+ mfd_assert_write(fd);
+ mfd_fail_shrink(fd);
+ mfd_fail_grow(fd);
+ mfd_fail_grow_write(fd);
+
+ close(fd);
+}
+
+/*
+ * Test sharing via dup()
+ * Test that seals are shared between dupped FDs and they're all equal.
+ */
+static void test_share_dup(void)
+{
+ int fd, fd2;
+
+ fd = mfd_assert_new("kern_memfd_share_dup",
+ MFD_DEF_SIZE,
+ MFD_CLOEXEC | MFD_ALLOW_SEALING);
+ mfd_assert_has_seals(fd, 0);
+
+ fd2 = mfd_assert_dup(fd);
+ mfd_assert_has_seals(fd2, 0);
+
+ mfd_assert_add_seals(fd, F_SEAL_WRITE);
+ mfd_assert_has_seals(fd, F_SEAL_WRITE);
+ mfd_assert_has_seals(fd2, F_SEAL_WRITE);
+
+ mfd_assert_add_seals(fd2, F_SEAL_SHRINK);
+ mfd_assert_has_seals(fd, F_SEAL_WRITE | F_SEAL_SHRINK);
+ mfd_assert_has_seals(fd2, F_SEAL_WRITE | F_SEAL_SHRINK);
+
+ mfd_assert_add_seals(fd, F_SEAL_SEAL);
+ mfd_assert_has_seals(fd, F_SEAL_WRITE | F_SEAL_SHRINK | F_SEAL_SEAL);
+ mfd_assert_has_seals(fd2, F_SEAL_WRITE | F_SEAL_SHRINK | F_SEAL_SEAL);
+
+ mfd_fail_add_seals(fd, F_SEAL_GROW);
+ mfd_fail_add_seals(fd2, F_SEAL_GROW);
+ mfd_fail_add_seals(fd, F_SEAL_SEAL);
+ mfd_fail_add_seals(fd2, F_SEAL_SEAL);
+
+ close(fd2);
+
+ mfd_fail_add_seals(fd, F_SEAL_GROW);
+ close(fd);
+}
+
+/*
+ * Test sealing with active mmap()s
+ * Modifying seals is only allowed if no other mmap() refs exist.
+ */
+static void test_share_mmap(void)
+{
+ int fd;
+ void *p;
+
+ fd = mfd_assert_new("kern_memfd_share_mmap",
+ MFD_DEF_SIZE,
+ MFD_CLOEXEC | MFD_ALLOW_SEALING);
+ mfd_assert_has_seals(fd, 0);
+
+ /* shared/writable ref prevents sealing WRITE, but allows others */
+ p = mfd_assert_mmap_shared(fd);
+ mfd_fail_add_seals(fd, F_SEAL_WRITE);
+ mfd_assert_has_seals(fd, 0);
+ mfd_assert_add_seals(fd, F_SEAL_SHRINK);
+ mfd_assert_has_seals(fd, F_SEAL_SHRINK);
+ munmap(p, MFD_DEF_SIZE);
+
+ /* readable ref allows sealing */
+ p = mfd_assert_mmap_private(fd);
+ mfd_assert_add_seals(fd, F_SEAL_WRITE);
+ mfd_assert_has_seals(fd, F_SEAL_WRITE | F_SEAL_SHRINK);
+ munmap(p, MFD_DEF_SIZE);
+
+ close(fd);
+}
+
+/*
+ * Test sealing with open(/proc/self/fd/%d)
+ * Via /proc we can get access to a separate file-context for the same memfd.
+ * This is *not* like dup(), but like a real separate open(). Make sure the
+ * semantics are as expected and we correctly check for RDONLY / WRONLY / RDWR.
+ */
+static void test_share_open(void)
+{
+ int fd, fd2;
+
+ fd = mfd_assert_new("kern_memfd_share_open",
+ MFD_DEF_SIZE,
+ MFD_CLOEXEC | MFD_ALLOW_SEALING);
+ mfd_assert_has_seals(fd, 0);
+
+ fd2 = mfd_assert_open(fd, O_RDWR, 0);
+ mfd_assert_add_seals(fd, F_SEAL_WRITE);
+ mfd_assert_has_seals(fd, F_SEAL_WRITE);
+ mfd_assert_has_seals(fd2, F_SEAL_WRITE);
+
+ mfd_assert_add_seals(fd2, F_SEAL_SHRINK);
+ mfd_assert_has_seals(fd, F_SEAL_WRITE | F_SEAL_SHRINK);
+ mfd_assert_has_seals(fd2, F_SEAL_WRITE | F_SEAL_SHRINK);
+
+ close(fd);
+ fd = mfd_assert_open(fd2, O_RDONLY, 0);
+
+ mfd_fail_add_seals(fd, F_SEAL_SEAL);
+ mfd_assert_has_seals(fd, F_SEAL_WRITE | F_SEAL_SHRINK);
+ mfd_assert_has_seals(fd2, F_SEAL_WRITE | F_SEAL_SHRINK);
+
+ close(fd2);
+ fd2 = mfd_assert_open(fd, O_RDWR, 0);
+
+ mfd_assert_add_seals(fd2, F_SEAL_SEAL);
+ mfd_assert_has_seals(fd, F_SEAL_WRITE | F_SEAL_SHRINK | F_SEAL_SEAL);
+ mfd_assert_has_seals(fd2, F_SEAL_WRITE | F_SEAL_SHRINK | F_SEAL_SEAL);
+
+ close(fd2);
+ close(fd);
+}
+
+/*
+ * Test sharing via fork()
+ * Test whether seal-modifications work as expected with forked childs.
+ */
+static void test_share_fork(void)
+{
+ int fd;
+ pid_t pid;
+
+ fd = mfd_assert_new("kern_memfd_share_fork",
+ MFD_DEF_SIZE,
+ MFD_CLOEXEC | MFD_ALLOW_SEALING);
+ mfd_assert_has_seals(fd, 0);
+
+ pid = spawn_idle_thread(0);
+ mfd_assert_add_seals(fd, F_SEAL_SEAL);
+ mfd_assert_has_seals(fd, F_SEAL_SEAL);
+
+ mfd_fail_add_seals(fd, F_SEAL_WRITE);
+ mfd_assert_has_seals(fd, F_SEAL_SEAL);
+
+ join_idle_thread(pid);
+
+ mfd_fail_add_seals(fd, F_SEAL_WRITE);
+ mfd_assert_has_seals(fd, F_SEAL_SEAL);
+
+ close(fd);
+}
+
+int main(int argc, char **argv)
+{
+ pid_t pid;
+
+ printf("memfd: CREATE\n");
+ test_create();
+ printf("memfd: BASIC\n");
+ test_basic();
+
+ printf("memfd: SEAL-WRITE\n");
+ test_seal_write();
+ printf("memfd: SEAL-SHRINK\n");
+ test_seal_shrink();
+ printf("memfd: SEAL-GROW\n");
+ test_seal_grow();
+ printf("memfd: SEAL-RESIZE\n");
+ test_seal_resize();
+
+ printf("memfd: SHARE-DUP\n");
+ test_share_dup();
+ printf("memfd: SHARE-MMAP\n");
+ test_share_mmap();
+ printf("memfd: SHARE-OPEN\n");
+ test_share_open();
+ printf("memfd: SHARE-FORK\n");
+ test_share_fork();
+
+ /* Run test-suite in a multi-threaded environment with a shared
+ * file-table. */
+ pid = spawn_idle_thread(CLONE_FILES | CLONE_FS | CLONE_VM);
+ printf("memfd: SHARE-DUP (shared file-table)\n");
+ test_share_dup();
+ printf("memfd: SHARE-MMAP (shared file-table)\n");
+ test_share_mmap();
+ printf("memfd: SHARE-OPEN (shared file-table)\n");
+ test_share_open();
+ printf("memfd: SHARE-FORK (shared file-table)\n");
+ test_share_fork();
+ join_idle_thread(pid);
+
+ printf("memfd: DONE\n");
+
+ return 0;
+}
diff --git a/memfd/run_fuse_test.sh b/memfd/run_fuse_test.sh
new file mode 100644
index 0000000..69b930e
--- /dev/null
+++ b/memfd/run_fuse_test.sh
@@ -0,0 +1,14 @@
+#!/bin/sh
+
+if test -d "./mnt" ; then
+ fusermount -u ./mnt
+ rmdir ./mnt
+fi
+
+set -e
+
+mkdir mnt
+./fuse_mnt ./mnt
+./fuse_test ./mnt/memfd
+fusermount -u ./mnt
+rmdir ./mnt
diff --git a/memory-hotplug/Makefile b/memory-hotplug/Makefile
new file mode 100644
index 0000000..d46b8d4
--- /dev/null
+++ b/memory-hotplug/Makefile
@@ -0,0 +1,9 @@
+all:
+
+run_tests:
+ @/bin/bash ./on-off-test.sh -r 2 || echo "memory-hotplug selftests: [FAIL]"
+
+run_full_test:
+ @/bin/bash ./on-off-test.sh || echo "memory-hotplug selftests: [FAIL]"
+
+clean:
diff --git a/memory-hotplug/on-off-test.sh b/memory-hotplug/on-off-test.sh
new file mode 100644
index 0000000..6cddde0
--- /dev/null
+++ b/memory-hotplug/on-off-test.sh
@@ -0,0 +1,238 @@
+#!/bin/bash
+
+SYSFS=
+
+prerequisite()
+{
+ msg="skip all tests:"
+
+ if [ $UID != 0 ]; then
+ echo $msg must be run as root >&2
+ exit 0
+ fi
+
+ SYSFS=`mount -t sysfs | head -1 | awk '{ print $3 }'`
+
+ if [ ! -d "$SYSFS" ]; then
+ echo $msg sysfs is not mounted >&2
+ exit 0
+ fi
+
+ if ! ls $SYSFS/devices/system/memory/memory* > /dev/null 2>&1; then
+ echo $msg memory hotplug is not supported >&2
+ exit 0
+ fi
+}
+
+#
+# list all hot-pluggable memory
+#
+hotpluggable_memory()
+{
+ local state=${1:-.\*}
+
+ for memory in $SYSFS/devices/system/memory/memory*; do
+ if grep -q 1 $memory/removable &&
+ grep -q $state $memory/state; then
+ echo ${memory##/*/memory}
+ fi
+ done
+}
+
+hotplaggable_offline_memory()
+{
+ hotpluggable_memory offline
+}
+
+hotpluggable_online_memory()
+{
+ hotpluggable_memory online
+}
+
+memory_is_online()
+{
+ grep -q online $SYSFS/devices/system/memory/memory$1/state
+}
+
+memory_is_offline()
+{
+ grep -q offline $SYSFS/devices/system/memory/memory$1/state
+}
+
+online_memory()
+{
+ echo online > $SYSFS/devices/system/memory/memory$1/state
+}
+
+offline_memory()
+{
+ echo offline > $SYSFS/devices/system/memory/memory$1/state
+}
+
+online_memory_expect_success()
+{
+ local memory=$1
+
+ if ! online_memory $memory; then
+ echo $FUNCNAME $memory: unexpected fail >&2
+ elif ! memory_is_online $memory; then
+ echo $FUNCNAME $memory: unexpected offline >&2
+ fi
+}
+
+online_memory_expect_fail()
+{
+ local memory=$1
+
+ if online_memory $memory 2> /dev/null; then
+ echo $FUNCNAME $memory: unexpected success >&2
+ elif ! memory_is_offline $memory; then
+ echo $FUNCNAME $memory: unexpected online >&2
+ fi
+}
+
+offline_memory_expect_success()
+{
+ local memory=$1
+
+ if ! offline_memory $memory; then
+ echo $FUNCNAME $memory: unexpected fail >&2
+ elif ! memory_is_offline $memory; then
+ echo $FUNCNAME $memory: unexpected offline >&2
+ fi
+}
+
+offline_memory_expect_fail()
+{
+ local memory=$1
+
+ if offline_memory $memory 2> /dev/null; then
+ echo $FUNCNAME $memory: unexpected success >&2
+ elif ! memory_is_online $memory; then
+ echo $FUNCNAME $memory: unexpected offline >&2
+ fi
+}
+
+error=-12
+priority=0
+ratio=10
+
+while getopts e:hp:r: opt; do
+ case $opt in
+ e)
+ error=$OPTARG
+ ;;
+ h)
+ echo "Usage $0 [ -e errno ] [ -p notifier-priority ] [ -r percent-of-memory-to-offline ]"
+ exit
+ ;;
+ p)
+ priority=$OPTARG
+ ;;
+ r)
+ ratio=$OPTARG
+ ;;
+ esac
+done
+
+if ! [ "$error" -ge -4095 -a "$error" -lt 0 ]; then
+ echo "error code must be -4095 <= errno < 0" >&2
+ exit 1
+fi
+
+prerequisite
+
+echo "Test scope: $ratio% hotplug memory"
+echo -e "\t online all hotplug memory in offline state"
+echo -e "\t offline $ratio% hotplug memory in online state"
+echo -e "\t online all hotplug memory in offline state"
+
+#
+# Online all hot-pluggable memory
+#
+for memory in `hotplaggable_offline_memory`; do
+ echo offline-online $memory
+ online_memory_expect_success $memory
+done
+
+#
+# Offline $ratio percent of hot-pluggable memory
+#
+for memory in `hotpluggable_online_memory`; do
+ if [ $((RANDOM % 100)) -lt $ratio ]; then
+ echo online-offline $memory
+ offline_memory_expect_success $memory
+ fi
+done
+
+#
+# Online all hot-pluggable memory again
+#
+for memory in `hotplaggable_offline_memory`; do
+ echo offline-online $memory
+ online_memory_expect_success $memory
+done
+
+#
+# Test with memory notifier error injection
+#
+
+DEBUGFS=`mount -t debugfs | head -1 | awk '{ print $3 }'`
+NOTIFIER_ERR_INJECT_DIR=$DEBUGFS/notifier-error-inject/memory
+
+prerequisite_extra()
+{
+ msg="skip extra tests:"
+
+ /sbin/modprobe -q -r memory-notifier-error-inject
+ /sbin/modprobe -q memory-notifier-error-inject priority=$priority
+
+ if [ ! -d "$DEBUGFS" ]; then
+ echo $msg debugfs is not mounted >&2
+ exit 0
+ fi
+
+ if [ ! -d $NOTIFIER_ERR_INJECT_DIR ]; then
+ echo $msg memory-notifier-error-inject module is not available >&2
+ exit 0
+ fi
+}
+
+prerequisite_extra
+
+#
+# Offline $ratio percent of hot-pluggable memory
+#
+echo 0 > $NOTIFIER_ERR_INJECT_DIR/actions/MEM_GOING_OFFLINE/error
+for memory in `hotpluggable_online_memory`; do
+ if [ $((RANDOM % 100)) -lt $ratio ]; then
+ offline_memory_expect_success $memory
+ fi
+done
+
+#
+# Test memory hot-add error handling (offline => online)
+#
+echo $error > $NOTIFIER_ERR_INJECT_DIR/actions/MEM_GOING_ONLINE/error
+for memory in `hotplaggable_offline_memory`; do
+ online_memory_expect_fail $memory
+done
+
+#
+# Online all hot-pluggable memory
+#
+echo 0 > $NOTIFIER_ERR_INJECT_DIR/actions/MEM_GOING_ONLINE/error
+for memory in `hotplaggable_offline_memory`; do
+ online_memory_expect_success $memory
+done
+
+#
+# Test memory hot-remove error handling (online => offline)
+#
+echo $error > $NOTIFIER_ERR_INJECT_DIR/actions/MEM_GOING_OFFLINE/error
+for memory in `hotpluggable_online_memory`; do
+ offline_memory_expect_fail $memory
+done
+
+echo 0 > $NOTIFIER_ERR_INJECT_DIR/actions/MEM_GOING_OFFLINE/error
+/sbin/modprobe -q -r memory-notifier-error-inject
diff --git a/mount/Makefile b/mount/Makefile
new file mode 100644
index 0000000..337d853
--- /dev/null
+++ b/mount/Makefile
@@ -0,0 +1,17 @@
+# Makefile for mount selftests.
+
+all: unprivileged-remount-test
+
+unprivileged-remount-test: unprivileged-remount-test.c
+ gcc -Wall -O2 unprivileged-remount-test.c -o unprivileged-remount-test
+
+# Allow specific tests to be selected.
+test_unprivileged_remount: unprivileged-remount-test
+ @if [ -f /proc/self/uid_map ] ; then ./unprivileged-remount-test ; fi
+
+run_tests: all test_unprivileged_remount
+
+clean:
+ rm -f unprivileged-remount-test
+
+.PHONY: all test_unprivileged_remount
diff --git a/mount/unprivileged-remount-test.c b/mount/unprivileged-remount-test.c
new file mode 100644
index 0000000..1b3ff2f
--- /dev/null
+++ b/mount/unprivileged-remount-test.c
@@ -0,0 +1,242 @@
+#define _GNU_SOURCE
+#include <sched.h>
+#include <stdio.h>
+#include <errno.h>
+#include <string.h>
+#include <sys/types.h>
+#include <sys/mount.h>
+#include <sys/wait.h>
+#include <stdlib.h>
+#include <unistd.h>
+#include <fcntl.h>
+#include <grp.h>
+#include <stdbool.h>
+#include <stdarg.h>
+
+#ifndef CLONE_NEWNS
+# define CLONE_NEWNS 0x00020000
+#endif
+#ifndef CLONE_NEWUTS
+# define CLONE_NEWUTS 0x04000000
+#endif
+#ifndef CLONE_NEWIPC
+# define CLONE_NEWIPC 0x08000000
+#endif
+#ifndef CLONE_NEWNET
+# define CLONE_NEWNET 0x40000000
+#endif
+#ifndef CLONE_NEWUSER
+# define CLONE_NEWUSER 0x10000000
+#endif
+#ifndef CLONE_NEWPID
+# define CLONE_NEWPID 0x20000000
+#endif
+
+#ifndef MS_RELATIME
+#define MS_RELATIME (1 << 21)
+#endif
+#ifndef MS_STRICTATIME
+#define MS_STRICTATIME (1 << 24)
+#endif
+
+static void die(char *fmt, ...)
+{
+ va_list ap;
+ va_start(ap, fmt);
+ vfprintf(stderr, fmt, ap);
+ va_end(ap);
+ exit(EXIT_FAILURE);
+}
+
+static void write_file(char *filename, char *fmt, ...)
+{
+ char buf[4096];
+ int fd;
+ ssize_t written;
+ int buf_len;
+ va_list ap;
+
+ va_start(ap, fmt);
+ buf_len = vsnprintf(buf, sizeof(buf), fmt, ap);
+ va_end(ap);
+ if (buf_len < 0) {
+ die("vsnprintf failed: %s\n",
+ strerror(errno));
+ }
+ if (buf_len >= sizeof(buf)) {
+ die("vsnprintf output truncated\n");
+ }
+
+ fd = open(filename, O_WRONLY);
+ if (fd < 0) {
+ die("open of %s failed: %s\n",
+ filename, strerror(errno));
+ }
+ written = write(fd, buf, buf_len);
+ if (written != buf_len) {
+ if (written >= 0) {
+ die("short write to %s\n", filename);
+ } else {
+ die("write to %s failed: %s\n",
+ filename, strerror(errno));
+ }
+ }
+ if (close(fd) != 0) {
+ die("close of %s failed: %s\n",
+ filename, strerror(errno));
+ }
+}
+
+static void create_and_enter_userns(void)
+{
+ uid_t uid;
+ gid_t gid;
+
+ uid = getuid();
+ gid = getgid();
+
+ if (unshare(CLONE_NEWUSER) !=0) {
+ die("unshare(CLONE_NEWUSER) failed: %s\n",
+ strerror(errno));
+ }
+
+ write_file("/proc/self/uid_map", "0 %d 1", uid);
+ write_file("/proc/self/gid_map", "0 %d 1", gid);
+
+ if (setgroups(0, NULL) != 0) {
+ die("setgroups failed: %s\n",
+ strerror(errno));
+ }
+ if (setgid(0) != 0) {
+ die ("setgid(0) failed %s\n",
+ strerror(errno));
+ }
+ if (setuid(0) != 0) {
+ die("setuid(0) failed %s\n",
+ strerror(errno));
+ }
+}
+
+static
+bool test_unpriv_remount(int mount_flags, int remount_flags, int invalid_flags)
+{
+ pid_t child;
+
+ child = fork();
+ if (child == -1) {
+ die("fork failed: %s\n",
+ strerror(errno));
+ }
+ if (child != 0) { /* parent */
+ pid_t pid;
+ int status;
+ pid = waitpid(child, &status, 0);
+ if (pid == -1) {
+ die("waitpid failed: %s\n",
+ strerror(errno));
+ }
+ if (pid != child) {
+ die("waited for %d got %d\n",
+ child, pid);
+ }
+ if (!WIFEXITED(status)) {
+ die("child did not terminate cleanly\n");
+ }
+ return WEXITSTATUS(status) == EXIT_SUCCESS ? true : false;
+ }
+
+ create_and_enter_userns();
+ if (unshare(CLONE_NEWNS) != 0) {
+ die("unshare(CLONE_NEWNS) failed: %s\n",
+ strerror(errno));
+ }
+
+ if (mount("testing", "/tmp", "ramfs", mount_flags, NULL) != 0) {
+ die("mount of /tmp failed: %s\n",
+ strerror(errno));
+ }
+
+ create_and_enter_userns();
+
+ if (unshare(CLONE_NEWNS) != 0) {
+ die("unshare(CLONE_NEWNS) failed: %s\n",
+ strerror(errno));
+ }
+
+ if (mount("/tmp", "/tmp", "none",
+ MS_REMOUNT | MS_BIND | remount_flags, NULL) != 0) {
+ /* system("cat /proc/self/mounts"); */
+ die("remount of /tmp failed: %s\n",
+ strerror(errno));
+ }
+
+ if (mount("/tmp", "/tmp", "none",
+ MS_REMOUNT | MS_BIND | invalid_flags, NULL) == 0) {
+ /* system("cat /proc/self/mounts"); */
+ die("remount of /tmp with invalid flags "
+ "succeeded unexpectedly\n");
+ }
+ exit(EXIT_SUCCESS);
+}
+
+static bool test_unpriv_remount_simple(int mount_flags)
+{
+ return test_unpriv_remount(mount_flags, mount_flags, 0);
+}
+
+static bool test_unpriv_remount_atime(int mount_flags, int invalid_flags)
+{
+ return test_unpriv_remount(mount_flags, mount_flags, invalid_flags);
+}
+
+int main(int argc, char **argv)
+{
+ if (!test_unpriv_remount_simple(MS_RDONLY|MS_NODEV)) {
+ die("MS_RDONLY malfunctions\n");
+ }
+ if (!test_unpriv_remount_simple(MS_NODEV)) {
+ die("MS_NODEV malfunctions\n");
+ }
+ if (!test_unpriv_remount_simple(MS_NOSUID|MS_NODEV)) {
+ die("MS_NOSUID malfunctions\n");
+ }
+ if (!test_unpriv_remount_simple(MS_NOEXEC|MS_NODEV)) {
+ die("MS_NOEXEC malfunctions\n");
+ }
+ if (!test_unpriv_remount_atime(MS_RELATIME|MS_NODEV,
+ MS_NOATIME|MS_NODEV))
+ {
+ die("MS_RELATIME malfunctions\n");
+ }
+ if (!test_unpriv_remount_atime(MS_STRICTATIME|MS_NODEV,
+ MS_NOATIME|MS_NODEV))
+ {
+ die("MS_STRICTATIME malfunctions\n");
+ }
+ if (!test_unpriv_remount_atime(MS_NOATIME|MS_NODEV,
+ MS_STRICTATIME|MS_NODEV))
+ {
+ die("MS_RELATIME malfunctions\n");
+ }
+ if (!test_unpriv_remount_atime(MS_RELATIME|MS_NODIRATIME|MS_NODEV,
+ MS_NOATIME|MS_NODEV))
+ {
+ die("MS_RELATIME malfunctions\n");
+ }
+ if (!test_unpriv_remount_atime(MS_STRICTATIME|MS_NODIRATIME|MS_NODEV,
+ MS_NOATIME|MS_NODEV))
+ {
+ die("MS_RELATIME malfunctions\n");
+ }
+ if (!test_unpriv_remount_atime(MS_NOATIME|MS_NODIRATIME|MS_NODEV,
+ MS_STRICTATIME|MS_NODEV))
+ {
+ die("MS_RELATIME malfunctions\n");
+ }
+ if (!test_unpriv_remount(MS_STRICTATIME|MS_NODEV, MS_NODEV,
+ MS_NOATIME|MS_NODEV))
+ {
+ die("Default atime malfunctions\n");
+ }
+ return EXIT_SUCCESS;
+}
diff --git a/mqueue/.gitignore b/mqueue/.gitignore
new file mode 100644
index 0000000..d8d4237
--- /dev/null
+++ b/mqueue/.gitignore
@@ -0,0 +1,2 @@
+mq_open_tests
+mq_perf_tests
diff --git a/mqueue/Makefile b/mqueue/Makefile
new file mode 100644
index 0000000..8056e2e
--- /dev/null
+++ b/mqueue/Makefile
@@ -0,0 +1,10 @@
+all:
+ gcc -O2 mq_open_tests.c -o mq_open_tests -lrt
+ gcc -O2 -o mq_perf_tests mq_perf_tests.c -lrt -lpthread -lpopt
+
+run_tests:
+ @./mq_open_tests /test1 || echo "mq_open_tests: [FAIL]"
+ @./mq_perf_tests || echo "mq_perf_tests: [FAIL]"
+
+clean:
+ rm -f mq_open_tests mq_perf_tests
diff --git a/mqueue/mq_open_tests.c b/mqueue/mq_open_tests.c
new file mode 100644
index 0000000..9c1a5d3
--- /dev/null
+++ b/mqueue/mq_open_tests.c
@@ -0,0 +1,500 @@
+/*
+ * This application is Copyright 2012 Red Hat, Inc.
+ * Doug Ledford <dledford@redhat.com>
+ *
+ * mq_open_tests is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation, version 3.
+ *
+ * mq_open_tests is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * For the full text of the license, see <http://www.gnu.org/licenses/>.
+ *
+ * mq_open_tests.c
+ * Tests the various situations that should either succeed or fail to
+ * open a posix message queue and then reports whether or not they
+ * did as they were supposed to.
+ *
+ */
+#include <stdio.h>
+#include <stdlib.h>
+#include <unistd.h>
+#include <fcntl.h>
+#include <string.h>
+#include <limits.h>
+#include <errno.h>
+#include <sys/types.h>
+#include <sys/time.h>
+#include <sys/resource.h>
+#include <sys/stat.h>
+#include <mqueue.h>
+
+static char *usage =
+"Usage:\n"
+" %s path\n"
+"\n"
+" path Path name of the message queue to create\n"
+"\n"
+" Note: this program must be run as root in order to enable all tests\n"
+"\n";
+
+char *DEF_MSGS = "/proc/sys/fs/mqueue/msg_default";
+char *DEF_MSGSIZE = "/proc/sys/fs/mqueue/msgsize_default";
+char *MAX_MSGS = "/proc/sys/fs/mqueue/msg_max";
+char *MAX_MSGSIZE = "/proc/sys/fs/mqueue/msgsize_max";
+
+int default_settings;
+struct rlimit saved_limits, cur_limits;
+int saved_def_msgs, saved_def_msgsize, saved_max_msgs, saved_max_msgsize;
+int cur_def_msgs, cur_def_msgsize, cur_max_msgs, cur_max_msgsize;
+FILE *def_msgs, *def_msgsize, *max_msgs, *max_msgsize;
+char *queue_path;
+mqd_t queue = -1;
+
+static inline void __set(FILE *stream, int value, char *err_msg);
+void shutdown(int exit_val, char *err_cause, int line_no);
+static inline int get(FILE *stream);
+static inline void set(FILE *stream, int value);
+static inline void getr(int type, struct rlimit *rlim);
+static inline void setr(int type, struct rlimit *rlim);
+void validate_current_settings();
+static inline void test_queue(struct mq_attr *attr, struct mq_attr *result);
+static inline int test_queue_fail(struct mq_attr *attr, struct mq_attr *result);
+
+static inline void __set(FILE *stream, int value, char *err_msg)
+{
+ rewind(stream);
+ if (fprintf(stream, "%d", value) < 0)
+ perror(err_msg);
+}
+
+
+void shutdown(int exit_val, char *err_cause, int line_no)
+{
+ static int in_shutdown = 0;
+
+ /* In case we get called recursively by a set() call below */
+ if (in_shutdown++)
+ return;
+
+ if (seteuid(0) == -1)
+ perror("seteuid() failed");
+
+ if (queue != -1)
+ if (mq_close(queue))
+ perror("mq_close() during shutdown");
+ if (queue_path)
+ /*
+ * Be silent if this fails, if we cleaned up already it's
+ * expected to fail
+ */
+ mq_unlink(queue_path);
+ if (default_settings) {
+ if (saved_def_msgs)
+ __set(def_msgs, saved_def_msgs,
+ "failed to restore saved_def_msgs");
+ if (saved_def_msgsize)
+ __set(def_msgsize, saved_def_msgsize,
+ "failed to restore saved_def_msgsize");
+ }
+ if (saved_max_msgs)
+ __set(max_msgs, saved_max_msgs,
+ "failed to restore saved_max_msgs");
+ if (saved_max_msgsize)
+ __set(max_msgsize, saved_max_msgsize,
+ "failed to restore saved_max_msgsize");
+ if (exit_val)
+ error(exit_val, errno, "%s at %d", err_cause, line_no);
+ exit(0);
+}
+
+static inline int get(FILE *stream)
+{
+ int value;
+ rewind(stream);
+ if (fscanf(stream, "%d", &value) != 1)
+ shutdown(4, "Error reading /proc entry", __LINE__ - 1);
+ return value;
+}
+
+static inline void set(FILE *stream, int value)
+{
+ int new_value;
+
+ rewind(stream);
+ if (fprintf(stream, "%d", value) < 0)
+ return shutdown(5, "Failed writing to /proc file",
+ __LINE__ - 1);
+ new_value = get(stream);
+ if (new_value != value)
+ return shutdown(5, "We didn't get what we wrote to /proc back",
+ __LINE__ - 1);
+}
+
+static inline void getr(int type, struct rlimit *rlim)
+{
+ if (getrlimit(type, rlim))
+ shutdown(6, "getrlimit()", __LINE__ - 1);
+}
+
+static inline void setr(int type, struct rlimit *rlim)
+{
+ if (setrlimit(type, rlim))
+ shutdown(7, "setrlimit()", __LINE__ - 1);
+}
+
+void validate_current_settings()
+{
+ int rlim_needed;
+
+ if (cur_limits.rlim_cur < 4096) {
+ printf("Current rlimit value for POSIX message queue bytes is "
+ "unreasonably low,\nincreasing.\n\n");
+ cur_limits.rlim_cur = 8192;
+ cur_limits.rlim_max = 16384;
+ setr(RLIMIT_MSGQUEUE, &cur_limits);
+ }
+
+ if (default_settings) {
+ rlim_needed = (cur_def_msgs + 1) * (cur_def_msgsize + 1 +
+ 2 * sizeof(void *));
+ if (rlim_needed > cur_limits.rlim_cur) {
+ printf("Temporarily lowering default queue parameters "
+ "to something that will work\n"
+ "with the current rlimit values.\n\n");
+ set(def_msgs, 10);
+ cur_def_msgs = 10;
+ set(def_msgsize, 128);
+ cur_def_msgsize = 128;
+ }
+ } else {
+ rlim_needed = (cur_max_msgs + 1) * (cur_max_msgsize + 1 +
+ 2 * sizeof(void *));
+ if (rlim_needed > cur_limits.rlim_cur) {
+ printf("Temporarily lowering maximum queue parameters "
+ "to something that will work\n"
+ "with the current rlimit values in case this is "
+ "a kernel that ties the default\n"
+ "queue parameters to the maximum queue "
+ "parameters.\n\n");
+ set(max_msgs, 10);
+ cur_max_msgs = 10;
+ set(max_msgsize, 128);
+ cur_max_msgsize = 128;
+ }
+ }
+}
+
+/*
+ * test_queue - Test opening a queue, shutdown if we fail. This should
+ * only be called in situations that should never fail. We clean up
+ * after ourselves and return the queue attributes in *result.
+ */
+static inline void test_queue(struct mq_attr *attr, struct mq_attr *result)
+{
+ int flags = O_RDWR | O_EXCL | O_CREAT;
+ int perms = DEFFILEMODE;
+
+ if ((queue = mq_open(queue_path, flags, perms, attr)) == -1)
+ shutdown(1, "mq_open()", __LINE__);
+ if (mq_getattr(queue, result))
+ shutdown(1, "mq_getattr()", __LINE__);
+ if (mq_close(queue))
+ shutdown(1, "mq_close()", __LINE__);
+ queue = -1;
+ if (mq_unlink(queue_path))
+ shutdown(1, "mq_unlink()", __LINE__);
+}
+
+/*
+ * Same as test_queue above, but failure is not fatal.
+ * Returns:
+ * 0 - Failed to create a queue
+ * 1 - Created a queue, attributes in *result
+ */
+static inline int test_queue_fail(struct mq_attr *attr, struct mq_attr *result)
+{
+ int flags = O_RDWR | O_EXCL | O_CREAT;
+ int perms = DEFFILEMODE;
+
+ if ((queue = mq_open(queue_path, flags, perms, attr)) == -1)
+ return 0;
+ if (mq_getattr(queue, result))
+ shutdown(1, "mq_getattr()", __LINE__);
+ if (mq_close(queue))
+ shutdown(1, "mq_close()", __LINE__);
+ queue = -1;
+ if (mq_unlink(queue_path))
+ shutdown(1, "mq_unlink()", __LINE__);
+ return 1;
+}
+
+int main(int argc, char *argv[])
+{
+ struct mq_attr attr, result;
+
+ if (argc != 2) {
+ fprintf(stderr, "Must pass a valid queue name\n\n");
+ fprintf(stderr, usage, argv[0]);
+ exit(1);
+ }
+
+ /*
+ * Although we can create a msg queue with a non-absolute path name,
+ * unlink will fail. So, if the name doesn't start with a /, add one
+ * when we save it.
+ */
+ if (*argv[1] == '/')
+ queue_path = strdup(argv[1]);
+ else {
+ queue_path = malloc(strlen(argv[1]) + 2);
+ if (!queue_path) {
+ perror("malloc()");
+ exit(1);
+ }
+ queue_path[0] = '/';
+ queue_path[1] = 0;
+ strcat(queue_path, argv[1]);
+ }
+
+ if (getuid() != 0) {
+ fprintf(stderr, "Not running as root, but almost all tests "
+ "require root in order to modify\nsystem settings. "
+ "Exiting.\n");
+ exit(1);
+ }
+
+ /* Find out what files there are for us to make tweaks in */
+ def_msgs = fopen(DEF_MSGS, "r+");
+ def_msgsize = fopen(DEF_MSGSIZE, "r+");
+ max_msgs = fopen(MAX_MSGS, "r+");
+ max_msgsize = fopen(MAX_MSGSIZE, "r+");
+
+ if (!max_msgs)
+ shutdown(2, "Failed to open msg_max", __LINE__);
+ if (!max_msgsize)
+ shutdown(2, "Failed to open msgsize_max", __LINE__);
+ if (def_msgs || def_msgsize)
+ default_settings = 1;
+
+ /* Load up the current system values for everything we can */
+ getr(RLIMIT_MSGQUEUE, &saved_limits);
+ cur_limits = saved_limits;
+ if (default_settings) {
+ saved_def_msgs = cur_def_msgs = get(def_msgs);
+ saved_def_msgsize = cur_def_msgsize = get(def_msgsize);
+ }
+ saved_max_msgs = cur_max_msgs = get(max_msgs);
+ saved_max_msgsize = cur_max_msgsize = get(max_msgsize);
+
+ /* Tell the user our initial state */
+ printf("\nInitial system state:\n");
+ printf("\tUsing queue path:\t\t%s\n", queue_path);
+ printf("\tRLIMIT_MSGQUEUE(soft):\t\t%ld\n",
+ (long) saved_limits.rlim_cur);
+ printf("\tRLIMIT_MSGQUEUE(hard):\t\t%ld\n",
+ (long) saved_limits.rlim_max);
+ printf("\tMaximum Message Size:\t\t%d\n", saved_max_msgsize);
+ printf("\tMaximum Queue Size:\t\t%d\n", saved_max_msgs);
+ if (default_settings) {
+ printf("\tDefault Message Size:\t\t%d\n", saved_def_msgsize);
+ printf("\tDefault Queue Size:\t\t%d\n", saved_def_msgs);
+ } else {
+ printf("\tDefault Message Size:\t\tNot Supported\n");
+ printf("\tDefault Queue Size:\t\tNot Supported\n");
+ }
+ printf("\n");
+
+ validate_current_settings();
+
+ printf("Adjusted system state for testing:\n");
+ printf("\tRLIMIT_MSGQUEUE(soft):\t\t%ld\n", (long) cur_limits.rlim_cur);
+ printf("\tRLIMIT_MSGQUEUE(hard):\t\t%ld\n", (long) cur_limits.rlim_max);
+ printf("\tMaximum Message Size:\t\t%d\n", cur_max_msgsize);
+ printf("\tMaximum Queue Size:\t\t%d\n", cur_max_msgs);
+ if (default_settings) {
+ printf("\tDefault Message Size:\t\t%d\n", cur_def_msgsize);
+ printf("\tDefault Queue Size:\t\t%d\n", cur_def_msgs);
+ }
+
+ printf("\n\nTest series 1, behavior when no attr struct "
+ "passed to mq_open:\n");
+ if (!default_settings) {
+ test_queue(NULL, &result);
+ printf("Given sane system settings, mq_open without an attr "
+ "struct succeeds:\tPASS\n");
+ if (result.mq_maxmsg != cur_max_msgs ||
+ result.mq_msgsize != cur_max_msgsize) {
+ printf("Kernel does not support setting the default "
+ "mq attributes,\nbut also doesn't tie the "
+ "defaults to the maximums:\t\t\tPASS\n");
+ } else {
+ set(max_msgs, ++cur_max_msgs);
+ set(max_msgsize, ++cur_max_msgsize);
+ test_queue(NULL, &result);
+ if (result.mq_maxmsg == cur_max_msgs &&
+ result.mq_msgsize == cur_max_msgsize)
+ printf("Kernel does not support setting the "
+ "default mq attributes and\n"
+ "also ties system wide defaults to "
+ "the system wide maximums:\t\t"
+ "FAIL\n");
+ else
+ printf("Kernel does not support setting the "
+ "default mq attributes,\n"
+ "but also doesn't tie the defaults to "
+ "the maximums:\t\t\tPASS\n");
+ }
+ } else {
+ printf("Kernel supports setting defaults separately from "
+ "maximums:\t\tPASS\n");
+ /*
+ * While we are here, go ahead and test that the kernel
+ * properly follows the default settings
+ */
+ test_queue(NULL, &result);
+ printf("Given sane values, mq_open without an attr struct "
+ "succeeds:\t\tPASS\n");
+ if (result.mq_maxmsg != cur_def_msgs ||
+ result.mq_msgsize != cur_def_msgsize)
+ printf("Kernel supports setting defaults, but does "
+ "not actually honor them:\tFAIL\n\n");
+ else {
+ set(def_msgs, ++cur_def_msgs);
+ set(def_msgsize, ++cur_def_msgsize);
+ /* In case max was the same as the default */
+ set(max_msgs, ++cur_max_msgs);
+ set(max_msgsize, ++cur_max_msgsize);
+ test_queue(NULL, &result);
+ if (result.mq_maxmsg != cur_def_msgs ||
+ result.mq_msgsize != cur_def_msgsize)
+ printf("Kernel supports setting defaults, but "
+ "does not actually honor them:\t"
+ "FAIL\n");
+ else
+ printf("Kernel properly honors default setting "
+ "knobs:\t\t\t\tPASS\n");
+ }
+ set(def_msgs, cur_max_msgs + 1);
+ cur_def_msgs = cur_max_msgs + 1;
+ set(def_msgsize, cur_max_msgsize + 1);
+ cur_def_msgsize = cur_max_msgsize + 1;
+ if (cur_def_msgs * (cur_def_msgsize + 2 * sizeof(void *)) >=
+ cur_limits.rlim_cur) {
+ cur_limits.rlim_cur = (cur_def_msgs + 2) *
+ (cur_def_msgsize + 2 * sizeof(void *));
+ cur_limits.rlim_max = 2 * cur_limits.rlim_cur;
+ setr(RLIMIT_MSGQUEUE, &cur_limits);
+ }
+ if (test_queue_fail(NULL, &result)) {
+ if (result.mq_maxmsg == cur_max_msgs &&
+ result.mq_msgsize == cur_max_msgsize)
+ printf("Kernel properly limits default values "
+ "to lesser of default/max:\t\tPASS\n");
+ else
+ printf("Kernel does not properly set default "
+ "queue parameters when\ndefaults > "
+ "max:\t\t\t\t\t\t\t\tFAIL\n");
+ } else
+ printf("Kernel fails to open mq because defaults are "
+ "greater than maximums:\tFAIL\n");
+ set(def_msgs, --cur_def_msgs);
+ set(def_msgsize, --cur_def_msgsize);
+ cur_limits.rlim_cur = cur_limits.rlim_max = cur_def_msgs *
+ cur_def_msgsize;
+ setr(RLIMIT_MSGQUEUE, &cur_limits);
+ if (test_queue_fail(NULL, &result))
+ printf("Kernel creates queue even though defaults "
+ "would exceed\nrlimit setting:"
+ "\t\t\t\t\t\t\t\tFAIL\n");
+ else
+ printf("Kernel properly fails to create queue when "
+ "defaults would\nexceed rlimit:"
+ "\t\t\t\t\t\t\t\tPASS\n");
+ }
+
+ /*
+ * Test #2 - open with an attr struct that exceeds rlimit
+ */
+ printf("\n\nTest series 2, behavior when attr struct is "
+ "passed to mq_open:\n");
+ cur_max_msgs = 32;
+ cur_max_msgsize = cur_limits.rlim_max >> 4;
+ set(max_msgs, cur_max_msgs);
+ set(max_msgsize, cur_max_msgsize);
+ attr.mq_maxmsg = cur_max_msgs;
+ attr.mq_msgsize = cur_max_msgsize;
+ if (test_queue_fail(&attr, &result))
+ printf("Queue open in excess of rlimit max when euid = 0 "
+ "succeeded:\t\tFAIL\n");
+ else
+ printf("Queue open in excess of rlimit max when euid = 0 "
+ "failed:\t\tPASS\n");
+ attr.mq_maxmsg = cur_max_msgs + 1;
+ attr.mq_msgsize = 10;
+ if (test_queue_fail(&attr, &result))
+ printf("Queue open with mq_maxmsg > limit when euid = 0 "
+ "succeeded:\t\tPASS\n");
+ else
+ printf("Queue open with mq_maxmsg > limit when euid = 0 "
+ "failed:\t\tFAIL\n");
+ attr.mq_maxmsg = 1;
+ attr.mq_msgsize = cur_max_msgsize + 1;
+ if (test_queue_fail(&attr, &result))
+ printf("Queue open with mq_msgsize > limit when euid = 0 "
+ "succeeded:\t\tPASS\n");
+ else
+ printf("Queue open with mq_msgsize > limit when euid = 0 "
+ "failed:\t\tFAIL\n");
+ attr.mq_maxmsg = 65536;
+ attr.mq_msgsize = 65536;
+ if (test_queue_fail(&attr, &result))
+ printf("Queue open with total size > 2GB when euid = 0 "
+ "succeeded:\t\tFAIL\n");
+ else
+ printf("Queue open with total size > 2GB when euid = 0 "
+ "failed:\t\t\tPASS\n");
+
+ if (seteuid(99) == -1) {
+ perror("seteuid() failed");
+ exit(1);
+ }
+
+ attr.mq_maxmsg = cur_max_msgs;
+ attr.mq_msgsize = cur_max_msgsize;
+ if (test_queue_fail(&attr, &result))
+ printf("Queue open in excess of rlimit max when euid = 99 "
+ "succeeded:\t\tFAIL\n");
+ else
+ printf("Queue open in excess of rlimit max when euid = 99 "
+ "failed:\t\tPASS\n");
+ attr.mq_maxmsg = cur_max_msgs + 1;
+ attr.mq_msgsize = 10;
+ if (test_queue_fail(&attr, &result))
+ printf("Queue open with mq_maxmsg > limit when euid = 99 "
+ "succeeded:\t\tFAIL\n");
+ else
+ printf("Queue open with mq_maxmsg > limit when euid = 99 "
+ "failed:\t\tPASS\n");
+ attr.mq_maxmsg = 1;
+ attr.mq_msgsize = cur_max_msgsize + 1;
+ if (test_queue_fail(&attr, &result))
+ printf("Queue open with mq_msgsize > limit when euid = 99 "
+ "succeeded:\t\tFAIL\n");
+ else
+ printf("Queue open with mq_msgsize > limit when euid = 99 "
+ "failed:\t\tPASS\n");
+ attr.mq_maxmsg = 65536;
+ attr.mq_msgsize = 65536;
+ if (test_queue_fail(&attr, &result))
+ printf("Queue open with total size > 2GB when euid = 99 "
+ "succeeded:\t\tFAIL\n");
+ else
+ printf("Queue open with total size > 2GB when euid = 99 "
+ "failed:\t\t\tPASS\n");
+
+ shutdown(0,"",0);
+}
diff --git a/mqueue/mq_perf_tests.c b/mqueue/mq_perf_tests.c
new file mode 100644
index 0000000..94dae65
--- /dev/null
+++ b/mqueue/mq_perf_tests.c
@@ -0,0 +1,743 @@
+/*
+ * This application is Copyright 2012 Red Hat, Inc.
+ * Doug Ledford <dledford@redhat.com>
+ *
+ * mq_perf_tests is free software: you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation, version 3.
+ *
+ * mq_perf_tests is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * For the full text of the license, see <http://www.gnu.org/licenses/>.
+ *
+ * mq_perf_tests.c
+ * Tests various types of message queue workloads, concentrating on those
+ * situations that invole large message sizes, large message queue depths,
+ * or both, and reports back useful metrics about kernel message queue
+ * performance.
+ *
+ */
+#define _GNU_SOURCE
+#include <stdio.h>
+#include <stdlib.h>
+#include <unistd.h>
+#include <fcntl.h>
+#include <string.h>
+#include <limits.h>
+#include <errno.h>
+#include <signal.h>
+#include <pthread.h>
+#include <sched.h>
+#include <sys/types.h>
+#include <sys/time.h>
+#include <sys/resource.h>
+#include <sys/stat.h>
+#include <mqueue.h>
+#include <popt.h>
+
+static char *usage =
+"Usage:\n"
+" %s [-c #[,#..] -f] path\n"
+"\n"
+" -c # Skip most tests and go straight to a high queue depth test\n"
+" and then run that test continuously (useful for running at\n"
+" the same time as some other workload to see how much the\n"
+" cache thrashing caused by adding messages to a very deep\n"
+" queue impacts the performance of other programs). The number\n"
+" indicates which CPU core we should bind the process to during\n"
+" the run. If you have more than one physical CPU, then you\n"
+" will need one copy per physical CPU package, and you should\n"
+" specify the CPU cores to pin ourself to via a comma separated\n"
+" list of CPU values.\n"
+" -f Only usable with continuous mode. Pin ourself to the CPUs\n"
+" as requested, then instead of looping doing a high mq\n"
+" workload, just busy loop. This will allow us to lock up a\n"
+" single CPU just like we normally would, but without actually\n"
+" thrashing the CPU cache. This is to make it easier to get\n"
+" comparable numbers from some other workload running on the\n"
+" other CPUs. One set of numbers with # CPUs locked up running\n"
+" an mq workload, and another set of numbers with those same\n"
+" CPUs locked away from the test workload, but not doing\n"
+" anything to trash the cache like the mq workload might.\n"
+" path Path name of the message queue to create\n"
+"\n"
+" Note: this program must be run as root in order to enable all tests\n"
+"\n";
+
+char *MAX_MSGS = "/proc/sys/fs/mqueue/msg_max";
+char *MAX_MSGSIZE = "/proc/sys/fs/mqueue/msgsize_max";
+
+#define min(a, b) ((a) < (b) ? (a) : (b))
+#define MAX_CPUS 64
+char *cpu_option_string;
+int cpus_to_pin[MAX_CPUS];
+int num_cpus_to_pin;
+pthread_t cpu_threads[MAX_CPUS];
+pthread_t main_thread;
+cpu_set_t *cpu_set;
+int cpu_set_size;
+int cpus_online;
+
+#define MSG_SIZE 16
+#define TEST1_LOOPS 10000000
+#define TEST2_LOOPS 100000
+int continuous_mode;
+int continuous_mode_fake;
+
+struct rlimit saved_limits, cur_limits;
+int saved_max_msgs, saved_max_msgsize;
+int cur_max_msgs, cur_max_msgsize;
+FILE *max_msgs, *max_msgsize;
+int cur_nice;
+char *queue_path = "/mq_perf_tests";
+mqd_t queue = -1;
+struct mq_attr result;
+int mq_prio_max;
+
+const struct poptOption options[] = {
+ {
+ .longName = "continuous",
+ .shortName = 'c',
+ .argInfo = POPT_ARG_STRING,
+ .arg = &cpu_option_string,
+ .val = 'c',
+ .descrip = "Run continuous tests at a high queue depth in "
+ "order to test the effects of cache thrashing on "
+ "other tasks on the system. This test is intended "
+ "to be run on one core of each physical CPU while "
+ "some other CPU intensive task is run on all the other "
+ "cores of that same physical CPU and the other task "
+ "is timed. It is assumed that the process of adding "
+ "messages to the message queue in a tight loop will "
+ "impact that other task to some degree. Once the "
+ "tests are performed in this way, you should then "
+ "re-run the tests using fake mode in order to check "
+ "the difference in time required to perform the CPU "
+ "intensive task",
+ .argDescrip = "cpu[,cpu]",
+ },
+ {
+ .longName = "fake",
+ .shortName = 'f',
+ .argInfo = POPT_ARG_NONE,
+ .arg = &continuous_mode_fake,
+ .val = 0,
+ .descrip = "Tie up the CPUs that we would normally tie up in"
+ "continuous mode, but don't actually do any mq stuff, "
+ "just keep the CPU busy so it can't be used to process "
+ "system level tasks as this would free up resources on "
+ "the other CPU cores and skew the comparison between "
+ "the no-mqueue work and mqueue work tests",
+ .argDescrip = NULL,
+ },
+ {
+ .longName = "path",
+ .shortName = 'p',
+ .argInfo = POPT_ARG_STRING | POPT_ARGFLAG_SHOW_DEFAULT,
+ .arg = &queue_path,
+ .val = 'p',
+ .descrip = "The name of the path to use in the mqueue "
+ "filesystem for our tests",
+ .argDescrip = "pathname",
+ },
+ POPT_AUTOHELP
+ POPT_TABLEEND
+};
+
+static inline void __set(FILE *stream, int value, char *err_msg);
+void shutdown(int exit_val, char *err_cause, int line_no);
+void sig_action_SIGUSR1(int signum, siginfo_t *info, void *context);
+void sig_action(int signum, siginfo_t *info, void *context);
+static inline int get(FILE *stream);
+static inline void set(FILE *stream, int value);
+static inline int try_set(FILE *stream, int value);
+static inline void getr(int type, struct rlimit *rlim);
+static inline void setr(int type, struct rlimit *rlim);
+static inline void open_queue(struct mq_attr *attr);
+void increase_limits(void);
+
+static inline void __set(FILE *stream, int value, char *err_msg)
+{
+ rewind(stream);
+ if (fprintf(stream, "%d", value) < 0)
+ perror(err_msg);
+}
+
+
+void shutdown(int exit_val, char *err_cause, int line_no)
+{
+ static int in_shutdown = 0;
+ int errno_at_shutdown = errno;
+ int i;
+
+ /* In case we get called by multiple threads or from an sighandler */
+ if (in_shutdown++)
+ return;
+
+ for (i = 0; i < num_cpus_to_pin; i++)
+ if (cpu_threads[i]) {
+ pthread_kill(cpu_threads[i], SIGUSR1);
+ pthread_join(cpu_threads[i], NULL);
+ }
+
+ if (queue != -1)
+ if (mq_close(queue))
+ perror("mq_close() during shutdown");
+ if (queue_path)
+ /*
+ * Be silent if this fails, if we cleaned up already it's
+ * expected to fail
+ */
+ mq_unlink(queue_path);
+ if (saved_max_msgs)
+ __set(max_msgs, saved_max_msgs,
+ "failed to restore saved_max_msgs");
+ if (saved_max_msgsize)
+ __set(max_msgsize, saved_max_msgsize,
+ "failed to restore saved_max_msgsize");
+ if (exit_val)
+ error(exit_val, errno_at_shutdown, "%s at %d",
+ err_cause, line_no);
+ exit(0);
+}
+
+void sig_action_SIGUSR1(int signum, siginfo_t *info, void *context)
+{
+ if (pthread_self() != main_thread)
+ pthread_exit(0);
+ else {
+ fprintf(stderr, "Caught signal %d in SIGUSR1 handler, "
+ "exiting\n", signum);
+ shutdown(0, "", 0);
+ fprintf(stderr, "\n\nReturned from shutdown?!?!\n\n");
+ exit(0);
+ }
+}
+
+void sig_action(int signum, siginfo_t *info, void *context)
+{
+ if (pthread_self() != main_thread)
+ pthread_kill(main_thread, signum);
+ else {
+ fprintf(stderr, "Caught signal %d, exiting\n", signum);
+ shutdown(0, "", 0);
+ fprintf(stderr, "\n\nReturned from shutdown?!?!\n\n");
+ exit(0);
+ }
+}
+
+static inline int get(FILE *stream)
+{
+ int value;
+ rewind(stream);
+ if (fscanf(stream, "%d", &value) != 1)
+ shutdown(4, "Error reading /proc entry", __LINE__);
+ return value;
+}
+
+static inline void set(FILE *stream, int value)
+{
+ int new_value;
+
+ rewind(stream);
+ if (fprintf(stream, "%d", value) < 0)
+ return shutdown(5, "Failed writing to /proc file", __LINE__);
+ new_value = get(stream);
+ if (new_value != value)
+ return shutdown(5, "We didn't get what we wrote to /proc back",
+ __LINE__);
+}
+
+static inline int try_set(FILE *stream, int value)
+{
+ int new_value;
+
+ rewind(stream);
+ fprintf(stream, "%d", value);
+ new_value = get(stream);
+ return new_value == value;
+}
+
+static inline void getr(int type, struct rlimit *rlim)
+{
+ if (getrlimit(type, rlim))
+ shutdown(6, "getrlimit()", __LINE__);
+}
+
+static inline void setr(int type, struct rlimit *rlim)
+{
+ if (setrlimit(type, rlim))
+ shutdown(7, "setrlimit()", __LINE__);
+}
+
+/**
+ * open_queue - open the global queue for testing
+ * @attr - An attr struct specifying the desired queue traits
+ * @result - An attr struct that lists the actual traits the queue has
+ *
+ * This open is not allowed to fail, failure will result in an orderly
+ * shutdown of the program. The global queue_path is used to set what
+ * queue to open, the queue descriptor is saved in the global queue
+ * variable.
+ */
+static inline void open_queue(struct mq_attr *attr)
+{
+ int flags = O_RDWR | O_EXCL | O_CREAT | O_NONBLOCK;
+ int perms = DEFFILEMODE;
+
+ queue = mq_open(queue_path, flags, perms, attr);
+ if (queue == -1)
+ shutdown(1, "mq_open()", __LINE__);
+ if (mq_getattr(queue, &result))
+ shutdown(1, "mq_getattr()", __LINE__);
+ printf("\n\tQueue %s created:\n", queue_path);
+ printf("\t\tmq_flags:\t\t\t%s\n", result.mq_flags & O_NONBLOCK ?
+ "O_NONBLOCK" : "(null)");
+ printf("\t\tmq_maxmsg:\t\t\t%lu\n", result.mq_maxmsg);
+ printf("\t\tmq_msgsize:\t\t\t%lu\n", result.mq_msgsize);
+ printf("\t\tmq_curmsgs:\t\t\t%lu\n", result.mq_curmsgs);
+}
+
+void *fake_cont_thread(void *arg)
+{
+ int i;
+
+ for (i = 0; i < num_cpus_to_pin; i++)
+ if (cpu_threads[i] == pthread_self())
+ break;
+ printf("\tStarted fake continuous mode thread %d on CPU %d\n", i,
+ cpus_to_pin[i]);
+ while (1)
+ ;
+}
+
+void *cont_thread(void *arg)
+{
+ char buff[MSG_SIZE];
+ int i, priority;
+
+ for (i = 0; i < num_cpus_to_pin; i++)
+ if (cpu_threads[i] == pthread_self())
+ break;
+ printf("\tStarted continuous mode thread %d on CPU %d\n", i,
+ cpus_to_pin[i]);
+ while (1) {
+ while (mq_send(queue, buff, sizeof(buff), 0) == 0)
+ ;
+ mq_receive(queue, buff, sizeof(buff), &priority);
+ }
+}
+
+#define drain_queue() \
+ while (mq_receive(queue, buff, MSG_SIZE, &prio_in) == MSG_SIZE)
+
+#define do_untimed_send() \
+ do { \
+ if (mq_send(queue, buff, MSG_SIZE, prio_out)) \
+ shutdown(3, "Test send failure", __LINE__); \
+ } while (0)
+
+#define do_send_recv() \
+ do { \
+ clock_gettime(clock, &start); \
+ if (mq_send(queue, buff, MSG_SIZE, prio_out)) \
+ shutdown(3, "Test send failure", __LINE__); \
+ clock_gettime(clock, &middle); \
+ if (mq_receive(queue, buff, MSG_SIZE, &prio_in) != MSG_SIZE) \
+ shutdown(3, "Test receive failure", __LINE__); \
+ clock_gettime(clock, &end); \
+ nsec = ((middle.tv_sec - start.tv_sec) * 1000000000) + \
+ (middle.tv_nsec - start.tv_nsec); \
+ send_total.tv_nsec += nsec; \
+ if (send_total.tv_nsec >= 1000000000) { \
+ send_total.tv_sec++; \
+ send_total.tv_nsec -= 1000000000; \
+ } \
+ nsec = ((end.tv_sec - middle.tv_sec) * 1000000000) + \
+ (end.tv_nsec - middle.tv_nsec); \
+ recv_total.tv_nsec += nsec; \
+ if (recv_total.tv_nsec >= 1000000000) { \
+ recv_total.tv_sec++; \
+ recv_total.tv_nsec -= 1000000000; \
+ } \
+ } while (0)
+
+struct test {
+ char *desc;
+ void (*func)(int *);
+};
+
+void const_prio(int *prio)
+{
+ return;
+}
+
+void inc_prio(int *prio)
+{
+ if (++*prio == mq_prio_max)
+ *prio = 0;
+}
+
+void dec_prio(int *prio)
+{
+ if (--*prio < 0)
+ *prio = mq_prio_max - 1;
+}
+
+void random_prio(int *prio)
+{
+ *prio = random() % mq_prio_max;
+}
+
+struct test test2[] = {
+ {"\n\tTest #2a: Time send/recv message, queue full, constant prio\n",
+ const_prio},
+ {"\n\tTest #2b: Time send/recv message, queue full, increasing prio\n",
+ inc_prio},
+ {"\n\tTest #2c: Time send/recv message, queue full, decreasing prio\n",
+ dec_prio},
+ {"\n\tTest #2d: Time send/recv message, queue full, random prio\n",
+ random_prio},
+ {NULL, NULL}
+};
+
+/**
+ * Tests to perform (all done with MSG_SIZE messages):
+ *
+ * 1) Time to add/remove message with 0 messages on queue
+ * 1a) with constant prio
+ * 2) Time to add/remove message when queue close to capacity:
+ * 2a) with constant prio
+ * 2b) with increasing prio
+ * 2c) with decreasing prio
+ * 2d) with random prio
+ * 3) Test limits of priorities honored (double check _SC_MQ_PRIO_MAX)
+ */
+void *perf_test_thread(void *arg)
+{
+ char buff[MSG_SIZE];
+ int prio_out, prio_in;
+ int i;
+ clockid_t clock;
+ pthread_t *t;
+ struct timespec res, start, middle, end, send_total, recv_total;
+ unsigned long long nsec;
+ struct test *cur_test;
+
+ t = &cpu_threads[0];
+ printf("\n\tStarted mqueue performance test thread on CPU %d\n",
+ cpus_to_pin[0]);
+ mq_prio_max = sysconf(_SC_MQ_PRIO_MAX);
+ if (mq_prio_max == -1)
+ shutdown(2, "sysconf(_SC_MQ_PRIO_MAX)", __LINE__);
+ if (pthread_getcpuclockid(cpu_threads[0], &clock) != 0)
+ shutdown(2, "pthread_getcpuclockid", __LINE__);
+
+ if (clock_getres(clock, &res))
+ shutdown(2, "clock_getres()", __LINE__);
+
+ printf("\t\tMax priorities:\t\t\t%d\n", mq_prio_max);
+ printf("\t\tClock resolution:\t\t%lu nsec%s\n", res.tv_nsec,
+ res.tv_nsec > 1 ? "s" : "");
+
+
+
+ printf("\n\tTest #1: Time send/recv message, queue empty\n");
+ printf("\t\t(%d iterations)\n", TEST1_LOOPS);
+ prio_out = 0;
+ send_total.tv_sec = 0;
+ send_total.tv_nsec = 0;
+ recv_total.tv_sec = 0;
+ recv_total.tv_nsec = 0;
+ for (i = 0; i < TEST1_LOOPS; i++)
+ do_send_recv();
+ printf("\t\tSend msg:\t\t\t%ld.%lus total time\n",
+ send_total.tv_sec, send_total.tv_nsec);
+ nsec = ((unsigned long long)send_total.tv_sec * 1000000000 +
+ send_total.tv_nsec) / TEST1_LOOPS;
+ printf("\t\t\t\t\t\t%lld nsec/msg\n", nsec);
+ printf("\t\tRecv msg:\t\t\t%ld.%lus total time\n",
+ recv_total.tv_sec, recv_total.tv_nsec);
+ nsec = ((unsigned long long)recv_total.tv_sec * 1000000000 +
+ recv_total.tv_nsec) / TEST1_LOOPS;
+ printf("\t\t\t\t\t\t%lld nsec/msg\n", nsec);
+
+
+ for (cur_test = test2; cur_test->desc != NULL; cur_test++) {
+ printf("%s:\n", cur_test->desc);
+ printf("\t\t(%d iterations)\n", TEST2_LOOPS);
+ prio_out = 0;
+ send_total.tv_sec = 0;
+ send_total.tv_nsec = 0;
+ recv_total.tv_sec = 0;
+ recv_total.tv_nsec = 0;
+ printf("\t\tFilling queue...");
+ fflush(stdout);
+ clock_gettime(clock, &start);
+ for (i = 0; i < result.mq_maxmsg - 1; i++) {
+ do_untimed_send();
+ cur_test->func(&prio_out);
+ }
+ clock_gettime(clock, &end);
+ nsec = ((unsigned long long)(end.tv_sec - start.tv_sec) *
+ 1000000000) + (end.tv_nsec - start.tv_nsec);
+ printf("done.\t\t%lld.%llds\n", nsec / 1000000000,
+ nsec % 1000000000);
+ printf("\t\tTesting...");
+ fflush(stdout);
+ for (i = 0; i < TEST2_LOOPS; i++) {
+ do_send_recv();
+ cur_test->func(&prio_out);
+ }
+ printf("done.\n");
+ printf("\t\tSend msg:\t\t\t%ld.%lus total time\n",
+ send_total.tv_sec, send_total.tv_nsec);
+ nsec = ((unsigned long long)send_total.tv_sec * 1000000000 +
+ send_total.tv_nsec) / TEST2_LOOPS;
+ printf("\t\t\t\t\t\t%lld nsec/msg\n", nsec);
+ printf("\t\tRecv msg:\t\t\t%ld.%lus total time\n",
+ recv_total.tv_sec, recv_total.tv_nsec);
+ nsec = ((unsigned long long)recv_total.tv_sec * 1000000000 +
+ recv_total.tv_nsec) / TEST2_LOOPS;
+ printf("\t\t\t\t\t\t%lld nsec/msg\n", nsec);
+ printf("\t\tDraining queue...");
+ fflush(stdout);
+ clock_gettime(clock, &start);
+ drain_queue();
+ clock_gettime(clock, &end);
+ nsec = ((unsigned long long)(end.tv_sec - start.tv_sec) *
+ 1000000000) + (end.tv_nsec - start.tv_nsec);
+ printf("done.\t\t%lld.%llds\n", nsec / 1000000000,
+ nsec % 1000000000);
+ }
+ return 0;
+}
+
+void increase_limits(void)
+{
+ cur_limits.rlim_cur = RLIM_INFINITY;
+ cur_limits.rlim_max = RLIM_INFINITY;
+ setr(RLIMIT_MSGQUEUE, &cur_limits);
+ while (try_set(max_msgs, cur_max_msgs += 10))
+ ;
+ cur_max_msgs = get(max_msgs);
+ while (try_set(max_msgsize, cur_max_msgsize += 1024))
+ ;
+ cur_max_msgsize = get(max_msgsize);
+ if (setpriority(PRIO_PROCESS, 0, -20) != 0)
+ shutdown(2, "setpriority()", __LINE__);
+ cur_nice = -20;
+}
+
+int main(int argc, char *argv[])
+{
+ struct mq_attr attr;
+ char *option, *next_option;
+ int i, cpu;
+ struct sigaction sa;
+ poptContext popt_context;
+ char rc;
+ void *retval;
+
+ main_thread = pthread_self();
+ num_cpus_to_pin = 0;
+
+ if (sysconf(_SC_NPROCESSORS_ONLN) == -1) {
+ perror("sysconf(_SC_NPROCESSORS_ONLN)");
+ exit(1);
+ }
+ cpus_online = min(MAX_CPUS, sysconf(_SC_NPROCESSORS_ONLN));
+ cpu_set = CPU_ALLOC(cpus_online);
+ if (cpu_set == NULL) {
+ perror("CPU_ALLOC()");
+ exit(1);
+ }
+ cpu_set_size = CPU_ALLOC_SIZE(cpus_online);
+ CPU_ZERO_S(cpu_set_size, cpu_set);
+
+ popt_context = poptGetContext(NULL, argc, (const char **)argv,
+ options, 0);
+
+ while ((rc = poptGetNextOpt(popt_context)) > 0) {
+ switch (rc) {
+ case 'c':
+ continuous_mode = 1;
+ option = cpu_option_string;
+ do {
+ next_option = strchr(option, ',');
+ if (next_option)
+ *next_option = '\0';
+ cpu = atoi(option);
+ if (cpu >= cpus_online)
+ fprintf(stderr, "CPU %d exceeds "
+ "cpus online, ignoring.\n",
+ cpu);
+ else
+ cpus_to_pin[num_cpus_to_pin++] = cpu;
+ if (next_option)
+ option = ++next_option;
+ } while (next_option && num_cpus_to_pin < MAX_CPUS);
+ /* Double check that they didn't give us the same CPU
+ * more than once */
+ for (cpu = 0; cpu < num_cpus_to_pin; cpu++) {
+ if (CPU_ISSET_S(cpus_to_pin[cpu], cpu_set_size,
+ cpu_set)) {
+ fprintf(stderr, "Any given CPU may "
+ "only be given once.\n");
+ exit(1);
+ } else
+ CPU_SET_S(cpus_to_pin[cpu],
+ cpu_set_size, cpu_set);
+ }
+ break;
+ case 'p':
+ /*
+ * Although we can create a msg queue with a
+ * non-absolute path name, unlink will fail. So,
+ * if the name doesn't start with a /, add one
+ * when we save it.
+ */
+ option = queue_path;
+ if (*option != '/') {
+ queue_path = malloc(strlen(option) + 2);
+ if (!queue_path) {
+ perror("malloc()");
+ exit(1);
+ }
+ queue_path[0] = '/';
+ queue_path[1] = 0;
+ strcat(queue_path, option);
+ free(option);
+ }
+ break;
+ }
+ }
+
+ if (continuous_mode && num_cpus_to_pin == 0) {
+ fprintf(stderr, "Must pass at least one CPU to continuous "
+ "mode.\n");
+ poptPrintUsage(popt_context, stderr, 0);
+ exit(1);
+ } else if (!continuous_mode) {
+ num_cpus_to_pin = 1;
+ cpus_to_pin[0] = cpus_online - 1;
+ }
+
+ if (getuid() != 0) {
+ fprintf(stderr, "Not running as root, but almost all tests "
+ "require root in order to modify\nsystem settings. "
+ "Exiting.\n");
+ exit(1);
+ }
+
+ max_msgs = fopen(MAX_MSGS, "r+");
+ max_msgsize = fopen(MAX_MSGSIZE, "r+");
+ if (!max_msgs)
+ shutdown(2, "Failed to open msg_max", __LINE__);
+ if (!max_msgsize)
+ shutdown(2, "Failed to open msgsize_max", __LINE__);
+
+ /* Load up the current system values for everything we can */
+ getr(RLIMIT_MSGQUEUE, &saved_limits);
+ cur_limits = saved_limits;
+ saved_max_msgs = cur_max_msgs = get(max_msgs);
+ saved_max_msgsize = cur_max_msgsize = get(max_msgsize);
+ errno = 0;
+ cur_nice = getpriority(PRIO_PROCESS, 0);
+ if (errno)
+ shutdown(2, "getpriority()", __LINE__);
+
+ /* Tell the user our initial state */
+ printf("\nInitial system state:\n");
+ printf("\tUsing queue path:\t\t\t%s\n", queue_path);
+ printf("\tRLIMIT_MSGQUEUE(soft):\t\t\t%ld\n",
+ (long) saved_limits.rlim_cur);
+ printf("\tRLIMIT_MSGQUEUE(hard):\t\t\t%ld\n",
+ (long) saved_limits.rlim_max);
+ printf("\tMaximum Message Size:\t\t\t%d\n", saved_max_msgsize);
+ printf("\tMaximum Queue Size:\t\t\t%d\n", saved_max_msgs);
+ printf("\tNice value:\t\t\t\t%d\n", cur_nice);
+ printf("\n");
+
+ increase_limits();
+
+ printf("Adjusted system state for testing:\n");
+ if (cur_limits.rlim_cur == RLIM_INFINITY) {
+ printf("\tRLIMIT_MSGQUEUE(soft):\t\t\t(unlimited)\n");
+ printf("\tRLIMIT_MSGQUEUE(hard):\t\t\t(unlimited)\n");
+ } else {
+ printf("\tRLIMIT_MSGQUEUE(soft):\t\t\t%ld\n",
+ (long) cur_limits.rlim_cur);
+ printf("\tRLIMIT_MSGQUEUE(hard):\t\t\t%ld\n",
+ (long) cur_limits.rlim_max);
+ }
+ printf("\tMaximum Message Size:\t\t\t%d\n", cur_max_msgsize);
+ printf("\tMaximum Queue Size:\t\t\t%d\n", cur_max_msgs);
+ printf("\tNice value:\t\t\t\t%d\n", cur_nice);
+ printf("\tContinuous mode:\t\t\t(%s)\n", continuous_mode ?
+ (continuous_mode_fake ? "fake mode" : "enabled") :
+ "disabled");
+ printf("\tCPUs to pin:\t\t\t\t%d", cpus_to_pin[0]);
+ for (cpu = 1; cpu < num_cpus_to_pin; cpu++)
+ printf(",%d", cpus_to_pin[cpu]);
+ printf("\n");
+
+ sa.sa_sigaction = sig_action_SIGUSR1;
+ sigemptyset(&sa.sa_mask);
+ sigaddset(&sa.sa_mask, SIGHUP);
+ sigaddset(&sa.sa_mask, SIGINT);
+ sigaddset(&sa.sa_mask, SIGQUIT);
+ sigaddset(&sa.sa_mask, SIGTERM);
+ sa.sa_flags = SA_SIGINFO;
+ if (sigaction(SIGUSR1, &sa, NULL) == -1)
+ shutdown(1, "sigaction(SIGUSR1)", __LINE__);
+ sa.sa_sigaction = sig_action;
+ if (sigaction(SIGHUP, &sa, NULL) == -1)
+ shutdown(1, "sigaction(SIGHUP)", __LINE__);
+ if (sigaction(SIGINT, &sa, NULL) == -1)
+ shutdown(1, "sigaction(SIGINT)", __LINE__);
+ if (sigaction(SIGQUIT, &sa, NULL) == -1)
+ shutdown(1, "sigaction(SIGQUIT)", __LINE__);
+ if (sigaction(SIGTERM, &sa, NULL) == -1)
+ shutdown(1, "sigaction(SIGTERM)", __LINE__);
+
+ if (!continuous_mode_fake) {
+ attr.mq_flags = O_NONBLOCK;
+ attr.mq_maxmsg = cur_max_msgs;
+ attr.mq_msgsize = MSG_SIZE;
+ open_queue(&attr);
+ }
+ for (i = 0; i < num_cpus_to_pin; i++) {
+ pthread_attr_t thread_attr;
+ void *thread_func;
+
+ if (continuous_mode_fake)
+ thread_func = &fake_cont_thread;
+ else if (continuous_mode)
+ thread_func = &cont_thread;
+ else
+ thread_func = &perf_test_thread;
+
+ CPU_ZERO_S(cpu_set_size, cpu_set);
+ CPU_SET_S(cpus_to_pin[i], cpu_set_size, cpu_set);
+ pthread_attr_init(&thread_attr);
+ pthread_attr_setaffinity_np(&thread_attr, cpu_set_size,
+ cpu_set);
+ if (pthread_create(&cpu_threads[i], &thread_attr, thread_func,
+ NULL))
+ shutdown(1, "pthread_create()", __LINE__);
+ pthread_attr_destroy(&thread_attr);
+ }
+
+ if (!continuous_mode) {
+ pthread_join(cpu_threads[0], &retval);
+ shutdown((long)retval, "perf_test_thread()", __LINE__);
+ } else {
+ while (1)
+ sleep(1);
+ }
+ shutdown(0, "", 0);
+}
diff --git a/net/.gitignore b/net/.gitignore
new file mode 100644
index 0000000..0032662
--- /dev/null
+++ b/net/.gitignore
@@ -0,0 +1,3 @@
+socket
+psock_fanout
+psock_tpacket
diff --git a/net/Makefile b/net/Makefile
new file mode 100644
index 0000000..c7493b8
--- /dev/null
+++ b/net/Makefile
@@ -0,0 +1,25 @@
+# Makefile for net selftests
+
+CC = $(CROSS_COMPILE)gcc
+CFLAGS = -Wall -O2 -g
+
+CFLAGS += -I../../../../usr/include/
+
+NET_PROGS = socket psock_fanout psock_tpacket
+
+all: $(NET_PROGS)
+%: %.c
+ $(CC) $(CFLAGS) -o $@ $^
+
+run_tests: all
+ @/bin/sh ./run_netsocktests || echo "sockettests: [FAIL]"
+ @/bin/sh ./run_afpackettests || echo "afpackettests: [FAIL]"
+ @if /sbin/modprobe test_bpf ; then \
+ /sbin/rmmod test_bpf; \
+ echo "test_bpf: ok"; \
+ else \
+ echo "test_bpf: [FAIL]"; \
+ exit 1; \
+ fi
+clean:
+ $(RM) $(NET_PROGS)
diff --git a/net/psock_fanout.c b/net/psock_fanout.c
new file mode 100644
index 0000000..57b9c2b
--- /dev/null
+++ b/net/psock_fanout.c
@@ -0,0 +1,312 @@
+/*
+ * Copyright 2013 Google Inc.
+ * Author: Willem de Bruijn (willemb@google.com)
+ *
+ * A basic test of packet socket fanout behavior.
+ *
+ * Control:
+ * - create fanout fails as expected with illegal flag combinations
+ * - join fanout fails as expected with diverging types or flags
+ *
+ * Datapath:
+ * Open a pair of packet sockets and a pair of INET sockets, send a known
+ * number of packets across the two INET sockets and count the number of
+ * packets enqueued onto the two packet sockets.
+ *
+ * The test currently runs for
+ * - PACKET_FANOUT_HASH
+ * - PACKET_FANOUT_HASH with PACKET_FANOUT_FLAG_ROLLOVER
+ * - PACKET_FANOUT_LB
+ * - PACKET_FANOUT_CPU
+ * - PACKET_FANOUT_ROLLOVER
+ *
+ * Todo:
+ * - functionality: PACKET_FANOUT_FLAG_DEFRAG
+ *
+ * License (GPLv2):
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. * See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+
+#define _GNU_SOURCE /* for sched_setaffinity */
+
+#include <arpa/inet.h>
+#include <errno.h>
+#include <fcntl.h>
+#include <linux/filter.h>
+#include <linux/if_packet.h>
+#include <net/ethernet.h>
+#include <netinet/ip.h>
+#include <netinet/udp.h>
+#include <poll.h>
+#include <sched.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/mman.h>
+#include <sys/socket.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <unistd.h>
+
+#include "psock_lib.h"
+
+#define RING_NUM_FRAMES 20
+
+/* Open a socket in a given fanout mode.
+ * @return -1 if mode is bad, a valid socket otherwise */
+static int sock_fanout_open(uint16_t typeflags, int num_packets)
+{
+ int fd, val;
+
+ fd = socket(PF_PACKET, SOCK_DGRAM, htons(ETH_P_IP));
+ if (fd < 0) {
+ perror("socket packet");
+ exit(1);
+ }
+
+ /* fanout group ID is always 0: tests whether old groups are deleted */
+ val = ((int) typeflags) << 16;
+ if (setsockopt(fd, SOL_PACKET, PACKET_FANOUT, &val, sizeof(val))) {
+ if (close(fd)) {
+ perror("close packet");
+ exit(1);
+ }
+ return -1;
+ }
+
+ pair_udp_setfilter(fd);
+ return fd;
+}
+
+static char *sock_fanout_open_ring(int fd)
+{
+ struct tpacket_req req = {
+ .tp_block_size = getpagesize(),
+ .tp_frame_size = getpagesize(),
+ .tp_block_nr = RING_NUM_FRAMES,
+ .tp_frame_nr = RING_NUM_FRAMES,
+ };
+ char *ring;
+ int val = TPACKET_V2;
+
+ if (setsockopt(fd, SOL_PACKET, PACKET_VERSION, (void *) &val,
+ sizeof(val))) {
+ perror("packetsock ring setsockopt version");
+ exit(1);
+ }
+ if (setsockopt(fd, SOL_PACKET, PACKET_RX_RING, (void *) &req,
+ sizeof(req))) {
+ perror("packetsock ring setsockopt");
+ exit(1);
+ }
+
+ ring = mmap(0, req.tp_block_size * req.tp_block_nr,
+ PROT_READ | PROT_WRITE, MAP_SHARED, fd, 0);
+ if (!ring) {
+ fprintf(stderr, "packetsock ring mmap\n");
+ exit(1);
+ }
+
+ return ring;
+}
+
+static int sock_fanout_read_ring(int fd, void *ring)
+{
+ struct tpacket2_hdr *header = ring;
+ int count = 0;
+
+ while (header->tp_status & TP_STATUS_USER && count < RING_NUM_FRAMES) {
+ count++;
+ header = ring + (count * getpagesize());
+ }
+
+ return count;
+}
+
+static int sock_fanout_read(int fds[], char *rings[], const int expect[])
+{
+ int ret[2];
+
+ ret[0] = sock_fanout_read_ring(fds[0], rings[0]);
+ ret[1] = sock_fanout_read_ring(fds[1], rings[1]);
+
+ fprintf(stderr, "info: count=%d,%d, expect=%d,%d\n",
+ ret[0], ret[1], expect[0], expect[1]);
+
+ if ((!(ret[0] == expect[0] && ret[1] == expect[1])) &&
+ (!(ret[0] == expect[1] && ret[1] == expect[0]))) {
+ fprintf(stderr, "ERROR: incorrect queue lengths\n");
+ return 1;
+ }
+
+ return 0;
+}
+
+/* Test illegal mode + flag combination */
+static void test_control_single(void)
+{
+ fprintf(stderr, "test: control single socket\n");
+
+ if (sock_fanout_open(PACKET_FANOUT_ROLLOVER |
+ PACKET_FANOUT_FLAG_ROLLOVER, 0) != -1) {
+ fprintf(stderr, "ERROR: opened socket with dual rollover\n");
+ exit(1);
+ }
+}
+
+/* Test illegal group with different modes or flags */
+static void test_control_group(void)
+{
+ int fds[2];
+
+ fprintf(stderr, "test: control multiple sockets\n");
+
+ fds[0] = sock_fanout_open(PACKET_FANOUT_HASH, 20);
+ if (fds[0] == -1) {
+ fprintf(stderr, "ERROR: failed to open HASH socket\n");
+ exit(1);
+ }
+ if (sock_fanout_open(PACKET_FANOUT_HASH |
+ PACKET_FANOUT_FLAG_DEFRAG, 10) != -1) {
+ fprintf(stderr, "ERROR: joined group with wrong flag defrag\n");
+ exit(1);
+ }
+ if (sock_fanout_open(PACKET_FANOUT_HASH |
+ PACKET_FANOUT_FLAG_ROLLOVER, 10) != -1) {
+ fprintf(stderr, "ERROR: joined group with wrong flag ro\n");
+ exit(1);
+ }
+ if (sock_fanout_open(PACKET_FANOUT_CPU, 10) != -1) {
+ fprintf(stderr, "ERROR: joined group with wrong mode\n");
+ exit(1);
+ }
+ fds[1] = sock_fanout_open(PACKET_FANOUT_HASH, 20);
+ if (fds[1] == -1) {
+ fprintf(stderr, "ERROR: failed to join group\n");
+ exit(1);
+ }
+ if (close(fds[1]) || close(fds[0])) {
+ fprintf(stderr, "ERROR: closing sockets\n");
+ exit(1);
+ }
+}
+
+static int test_datapath(uint16_t typeflags, int port_off,
+ const int expect1[], const int expect2[])
+{
+ const int expect0[] = { 0, 0 };
+ char *rings[2];
+ int fds[2], fds_udp[2][2], ret;
+
+ fprintf(stderr, "test: datapath 0x%hx\n", typeflags);
+
+ fds[0] = sock_fanout_open(typeflags, 20);
+ fds[1] = sock_fanout_open(typeflags, 20);
+ if (fds[0] == -1 || fds[1] == -1) {
+ fprintf(stderr, "ERROR: failed open\n");
+ exit(1);
+ }
+ rings[0] = sock_fanout_open_ring(fds[0]);
+ rings[1] = sock_fanout_open_ring(fds[1]);
+ pair_udp_open(fds_udp[0], PORT_BASE);
+ pair_udp_open(fds_udp[1], PORT_BASE + port_off);
+ sock_fanout_read(fds, rings, expect0);
+
+ /* Send data, but not enough to overflow a queue */
+ pair_udp_send(fds_udp[0], 15);
+ pair_udp_send(fds_udp[1], 5);
+ ret = sock_fanout_read(fds, rings, expect1);
+
+ /* Send more data, overflow the queue */
+ pair_udp_send(fds_udp[0], 15);
+ /* TODO: ensure consistent order between expect1 and expect2 */
+ ret |= sock_fanout_read(fds, rings, expect2);
+
+ if (munmap(rings[1], RING_NUM_FRAMES * getpagesize()) ||
+ munmap(rings[0], RING_NUM_FRAMES * getpagesize())) {
+ fprintf(stderr, "close rings\n");
+ exit(1);
+ }
+ if (close(fds_udp[1][1]) || close(fds_udp[1][0]) ||
+ close(fds_udp[0][1]) || close(fds_udp[0][0]) ||
+ close(fds[1]) || close(fds[0])) {
+ fprintf(stderr, "close datapath\n");
+ exit(1);
+ }
+
+ return ret;
+}
+
+static int set_cpuaffinity(int cpuid)
+{
+ cpu_set_t mask;
+
+ CPU_ZERO(&mask);
+ CPU_SET(cpuid, &mask);
+ if (sched_setaffinity(0, sizeof(mask), &mask)) {
+ if (errno != EINVAL) {
+ fprintf(stderr, "setaffinity %d\n", cpuid);
+ exit(1);
+ }
+ return 1;
+ }
+
+ return 0;
+}
+
+int main(int argc, char **argv)
+{
+ const int expect_hash[2][2] = { { 15, 5 }, { 20, 5 } };
+ const int expect_hash_rb[2][2] = { { 15, 5 }, { 20, 15 } };
+ const int expect_lb[2][2] = { { 10, 10 }, { 18, 17 } };
+ const int expect_rb[2][2] = { { 20, 0 }, { 20, 15 } };
+ const int expect_cpu0[2][2] = { { 20, 0 }, { 20, 0 } };
+ const int expect_cpu1[2][2] = { { 0, 20 }, { 0, 20 } };
+ int port_off = 2, tries = 5, ret;
+
+ test_control_single();
+ test_control_group();
+
+ /* find a set of ports that do not collide onto the same socket */
+ ret = test_datapath(PACKET_FANOUT_HASH, port_off,
+ expect_hash[0], expect_hash[1]);
+ while (ret && tries--) {
+ fprintf(stderr, "info: trying alternate ports (%d)\n", tries);
+ ret = test_datapath(PACKET_FANOUT_HASH, ++port_off,
+ expect_hash[0], expect_hash[1]);
+ }
+
+ ret |= test_datapath(PACKET_FANOUT_HASH | PACKET_FANOUT_FLAG_ROLLOVER,
+ port_off, expect_hash_rb[0], expect_hash_rb[1]);
+ ret |= test_datapath(PACKET_FANOUT_LB,
+ port_off, expect_lb[0], expect_lb[1]);
+ ret |= test_datapath(PACKET_FANOUT_ROLLOVER,
+ port_off, expect_rb[0], expect_rb[1]);
+
+ set_cpuaffinity(0);
+ ret |= test_datapath(PACKET_FANOUT_CPU, port_off,
+ expect_cpu0[0], expect_cpu0[1]);
+ if (!set_cpuaffinity(1))
+ /* TODO: test that choice alternates with previous */
+ ret |= test_datapath(PACKET_FANOUT_CPU, port_off,
+ expect_cpu1[0], expect_cpu1[1]);
+
+ if (ret)
+ return 1;
+
+ printf("OK. All tests passed\n");
+ return 0;
+}
diff --git a/net/psock_lib.h b/net/psock_lib.h
new file mode 100644
index 0000000..37da54a
--- /dev/null
+++ b/net/psock_lib.h
@@ -0,0 +1,127 @@
+/*
+ * Copyright 2013 Google Inc.
+ * Author: Willem de Bruijn <willemb@google.com>
+ * Daniel Borkmann <dborkman@redhat.com>
+ *
+ * License (GPLv2):
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. * See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+
+#ifndef PSOCK_LIB_H
+#define PSOCK_LIB_H
+
+#include <sys/types.h>
+#include <sys/socket.h>
+#include <string.h>
+#include <arpa/inet.h>
+#include <unistd.h>
+
+#define DATA_LEN 100
+#define DATA_CHAR 'a'
+
+#define PORT_BASE 8000
+
+#ifndef __maybe_unused
+# define __maybe_unused __attribute__ ((__unused__))
+#endif
+
+static __maybe_unused void pair_udp_setfilter(int fd)
+{
+ struct sock_filter bpf_filter[] = {
+ { 0x80, 0, 0, 0x00000000 }, /* LD pktlen */
+ { 0x35, 0, 5, DATA_LEN }, /* JGE DATA_LEN [f goto nomatch]*/
+ { 0x30, 0, 0, 0x00000050 }, /* LD ip[80] */
+ { 0x15, 0, 3, DATA_CHAR }, /* JEQ DATA_CHAR [f goto nomatch]*/
+ { 0x30, 0, 0, 0x00000051 }, /* LD ip[81] */
+ { 0x15, 0, 1, DATA_CHAR }, /* JEQ DATA_CHAR [f goto nomatch]*/
+ { 0x06, 0, 0, 0x00000060 }, /* RET match */
+ { 0x06, 0, 0, 0x00000000 }, /* RET no match */
+ };
+ struct sock_fprog bpf_prog;
+
+ bpf_prog.filter = bpf_filter;
+ bpf_prog.len = sizeof(bpf_filter) / sizeof(struct sock_filter);
+ if (setsockopt(fd, SOL_SOCKET, SO_ATTACH_FILTER, &bpf_prog,
+ sizeof(bpf_prog))) {
+ perror("setsockopt SO_ATTACH_FILTER");
+ exit(1);
+ }
+}
+
+static __maybe_unused void pair_udp_open(int fds[], uint16_t port)
+{
+ struct sockaddr_in saddr, daddr;
+
+ fds[0] = socket(PF_INET, SOCK_DGRAM, 0);
+ fds[1] = socket(PF_INET, SOCK_DGRAM, 0);
+ if (fds[0] == -1 || fds[1] == -1) {
+ fprintf(stderr, "ERROR: socket dgram\n");
+ exit(1);
+ }
+
+ memset(&saddr, 0, sizeof(saddr));
+ saddr.sin_family = AF_INET;
+ saddr.sin_port = htons(port);
+ saddr.sin_addr.s_addr = htonl(INADDR_LOOPBACK);
+
+ memset(&daddr, 0, sizeof(daddr));
+ daddr.sin_family = AF_INET;
+ daddr.sin_port = htons(port + 1);
+ daddr.sin_addr.s_addr = htonl(INADDR_LOOPBACK);
+
+ /* must bind both to get consistent hash result */
+ if (bind(fds[1], (void *) &daddr, sizeof(daddr))) {
+ perror("bind");
+ exit(1);
+ }
+ if (bind(fds[0], (void *) &saddr, sizeof(saddr))) {
+ perror("bind");
+ exit(1);
+ }
+ if (connect(fds[0], (void *) &daddr, sizeof(daddr))) {
+ perror("connect");
+ exit(1);
+ }
+}
+
+static __maybe_unused void pair_udp_send(int fds[], int num)
+{
+ char buf[DATA_LEN], rbuf[DATA_LEN];
+
+ memset(buf, DATA_CHAR, sizeof(buf));
+ while (num--) {
+ /* Should really handle EINTR and EAGAIN */
+ if (write(fds[0], buf, sizeof(buf)) != sizeof(buf)) {
+ fprintf(stderr, "ERROR: send failed left=%d\n", num);
+ exit(1);
+ }
+ if (read(fds[1], rbuf, sizeof(rbuf)) != sizeof(rbuf)) {
+ fprintf(stderr, "ERROR: recv failed left=%d\n", num);
+ exit(1);
+ }
+ if (memcmp(buf, rbuf, sizeof(buf))) {
+ fprintf(stderr, "ERROR: data failed left=%d\n", num);
+ exit(1);
+ }
+ }
+}
+
+static __maybe_unused void pair_udp_close(int fds[])
+{
+ close(fds[0]);
+ close(fds[1]);
+}
+
+#endif /* PSOCK_LIB_H */
diff --git a/net/psock_tpacket.c b/net/psock_tpacket.c
new file mode 100644
index 0000000..24adf70
--- /dev/null
+++ b/net/psock_tpacket.c
@@ -0,0 +1,805 @@
+/*
+ * Copyright 2013 Red Hat, Inc.
+ * Author: Daniel Borkmann <dborkman@redhat.com>
+ * Chetan Loke <loke.chetan@gmail.com> (TPACKET_V3 usage example)
+ *
+ * A basic test of packet socket's TPACKET_V1/TPACKET_V2/TPACKET_V3 behavior.
+ *
+ * Control:
+ * Test the setup of the TPACKET socket with different patterns that are
+ * known to fail (TODO) resp. succeed (OK).
+ *
+ * Datapath:
+ * Open a pair of packet sockets and send resp. receive an a priori known
+ * packet pattern accross the sockets and check if it was received resp.
+ * sent correctly. Fanout in combination with RX_RING is currently not
+ * tested here.
+ *
+ * The test currently runs for
+ * - TPACKET_V1: RX_RING, TX_RING
+ * - TPACKET_V2: RX_RING, TX_RING
+ * - TPACKET_V3: RX_RING
+ *
+ * License (GPLv2):
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. * See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <sys/socket.h>
+#include <sys/mman.h>
+#include <linux/if_packet.h>
+#include <linux/filter.h>
+#include <ctype.h>
+#include <fcntl.h>
+#include <unistd.h>
+#include <bits/wordsize.h>
+#include <net/ethernet.h>
+#include <netinet/ip.h>
+#include <arpa/inet.h>
+#include <stdint.h>
+#include <string.h>
+#include <assert.h>
+#include <net/if.h>
+#include <inttypes.h>
+#include <poll.h>
+
+#include "psock_lib.h"
+
+#ifndef bug_on
+# define bug_on(cond) assert(!(cond))
+#endif
+
+#ifndef __aligned_tpacket
+# define __aligned_tpacket __attribute__((aligned(TPACKET_ALIGNMENT)))
+#endif
+
+#ifndef __align_tpacket
+# define __align_tpacket(x) __attribute__((aligned(TPACKET_ALIGN(x))))
+#endif
+
+#define NUM_PACKETS 100
+#define ALIGN_8(x) (((x) + 8 - 1) & ~(8 - 1))
+
+struct ring {
+ struct iovec *rd;
+ uint8_t *mm_space;
+ size_t mm_len, rd_len;
+ struct sockaddr_ll ll;
+ void (*walk)(int sock, struct ring *ring);
+ int type, rd_num, flen, version;
+ union {
+ struct tpacket_req req;
+ struct tpacket_req3 req3;
+ };
+};
+
+struct block_desc {
+ uint32_t version;
+ uint32_t offset_to_priv;
+ struct tpacket_hdr_v1 h1;
+};
+
+union frame_map {
+ struct {
+ struct tpacket_hdr tp_h __aligned_tpacket;
+ struct sockaddr_ll s_ll __align_tpacket(sizeof(struct tpacket_hdr));
+ } *v1;
+ struct {
+ struct tpacket2_hdr tp_h __aligned_tpacket;
+ struct sockaddr_ll s_ll __align_tpacket(sizeof(struct tpacket2_hdr));
+ } *v2;
+ void *raw;
+};
+
+static unsigned int total_packets, total_bytes;
+
+static int pfsocket(int ver)
+{
+ int ret, sock = socket(PF_PACKET, SOCK_RAW, htons(ETH_P_ALL));
+ if (sock == -1) {
+ perror("socket");
+ exit(1);
+ }
+
+ ret = setsockopt(sock, SOL_PACKET, PACKET_VERSION, &ver, sizeof(ver));
+ if (ret == -1) {
+ perror("setsockopt");
+ exit(1);
+ }
+
+ return sock;
+}
+
+static void status_bar_update(void)
+{
+ if (total_packets % 10 == 0) {
+ fprintf(stderr, ".");
+ fflush(stderr);
+ }
+}
+
+static void test_payload(void *pay, size_t len)
+{
+ struct ethhdr *eth = pay;
+
+ if (len < sizeof(struct ethhdr)) {
+ fprintf(stderr, "test_payload: packet too "
+ "small: %zu bytes!\n", len);
+ exit(1);
+ }
+
+ if (eth->h_proto != htons(ETH_P_IP)) {
+ fprintf(stderr, "test_payload: wrong ethernet "
+ "type: 0x%x!\n", ntohs(eth->h_proto));
+ exit(1);
+ }
+}
+
+static void create_payload(void *pay, size_t *len)
+{
+ int i;
+ struct ethhdr *eth = pay;
+ struct iphdr *ip = pay + sizeof(*eth);
+
+ /* Lets create some broken crap, that still passes
+ * our BPF filter.
+ */
+
+ *len = DATA_LEN + 42;
+
+ memset(pay, 0xff, ETH_ALEN * 2);
+ eth->h_proto = htons(ETH_P_IP);
+
+ for (i = 0; i < sizeof(*ip); ++i)
+ ((uint8_t *) pay)[i + sizeof(*eth)] = (uint8_t) rand();
+
+ ip->ihl = 5;
+ ip->version = 4;
+ ip->protocol = 0x11;
+ ip->frag_off = 0;
+ ip->ttl = 64;
+ ip->tot_len = htons((uint16_t) *len - sizeof(*eth));
+
+ ip->saddr = htonl(INADDR_LOOPBACK);
+ ip->daddr = htonl(INADDR_LOOPBACK);
+
+ memset(pay + sizeof(*eth) + sizeof(*ip),
+ DATA_CHAR, DATA_LEN);
+}
+
+static inline int __v1_rx_kernel_ready(struct tpacket_hdr *hdr)
+{
+ return ((hdr->tp_status & TP_STATUS_USER) == TP_STATUS_USER);
+}
+
+static inline void __v1_rx_user_ready(struct tpacket_hdr *hdr)
+{
+ hdr->tp_status = TP_STATUS_KERNEL;
+ __sync_synchronize();
+}
+
+static inline int __v2_rx_kernel_ready(struct tpacket2_hdr *hdr)
+{
+ return ((hdr->tp_status & TP_STATUS_USER) == TP_STATUS_USER);
+}
+
+static inline void __v2_rx_user_ready(struct tpacket2_hdr *hdr)
+{
+ hdr->tp_status = TP_STATUS_KERNEL;
+ __sync_synchronize();
+}
+
+static inline int __v1_v2_rx_kernel_ready(void *base, int version)
+{
+ switch (version) {
+ case TPACKET_V1:
+ return __v1_rx_kernel_ready(base);
+ case TPACKET_V2:
+ return __v2_rx_kernel_ready(base);
+ default:
+ bug_on(1);
+ return 0;
+ }
+}
+
+static inline void __v1_v2_rx_user_ready(void *base, int version)
+{
+ switch (version) {
+ case TPACKET_V1:
+ __v1_rx_user_ready(base);
+ break;
+ case TPACKET_V2:
+ __v2_rx_user_ready(base);
+ break;
+ }
+}
+
+static void walk_v1_v2_rx(int sock, struct ring *ring)
+{
+ struct pollfd pfd;
+ int udp_sock[2];
+ union frame_map ppd;
+ unsigned int frame_num = 0;
+
+ bug_on(ring->type != PACKET_RX_RING);
+
+ pair_udp_open(udp_sock, PORT_BASE);
+ pair_udp_setfilter(sock);
+
+ memset(&pfd, 0, sizeof(pfd));
+ pfd.fd = sock;
+ pfd.events = POLLIN | POLLERR;
+ pfd.revents = 0;
+
+ pair_udp_send(udp_sock, NUM_PACKETS);
+
+ while (total_packets < NUM_PACKETS * 2) {
+ while (__v1_v2_rx_kernel_ready(ring->rd[frame_num].iov_base,
+ ring->version)) {
+ ppd.raw = ring->rd[frame_num].iov_base;
+
+ switch (ring->version) {
+ case TPACKET_V1:
+ test_payload((uint8_t *) ppd.raw + ppd.v1->tp_h.tp_mac,
+ ppd.v1->tp_h.tp_snaplen);
+ total_bytes += ppd.v1->tp_h.tp_snaplen;
+ break;
+
+ case TPACKET_V2:
+ test_payload((uint8_t *) ppd.raw + ppd.v2->tp_h.tp_mac,
+ ppd.v2->tp_h.tp_snaplen);
+ total_bytes += ppd.v2->tp_h.tp_snaplen;
+ break;
+ }
+
+ status_bar_update();
+ total_packets++;
+
+ __v1_v2_rx_user_ready(ppd.raw, ring->version);
+
+ frame_num = (frame_num + 1) % ring->rd_num;
+ }
+
+ poll(&pfd, 1, 1);
+ }
+
+ pair_udp_close(udp_sock);
+
+ if (total_packets != 2 * NUM_PACKETS) {
+ fprintf(stderr, "walk_v%d_rx: received %u out of %u pkts\n",
+ ring->version, total_packets, NUM_PACKETS);
+ exit(1);
+ }
+
+ fprintf(stderr, " %u pkts (%u bytes)", NUM_PACKETS, total_bytes >> 1);
+}
+
+static inline int __v1_tx_kernel_ready(struct tpacket_hdr *hdr)
+{
+ return !(hdr->tp_status & (TP_STATUS_SEND_REQUEST | TP_STATUS_SENDING));
+}
+
+static inline void __v1_tx_user_ready(struct tpacket_hdr *hdr)
+{
+ hdr->tp_status = TP_STATUS_SEND_REQUEST;
+ __sync_synchronize();
+}
+
+static inline int __v2_tx_kernel_ready(struct tpacket2_hdr *hdr)
+{
+ return !(hdr->tp_status & (TP_STATUS_SEND_REQUEST | TP_STATUS_SENDING));
+}
+
+static inline void __v2_tx_user_ready(struct tpacket2_hdr *hdr)
+{
+ hdr->tp_status = TP_STATUS_SEND_REQUEST;
+ __sync_synchronize();
+}
+
+static inline int __v1_v2_tx_kernel_ready(void *base, int version)
+{
+ switch (version) {
+ case TPACKET_V1:
+ return __v1_tx_kernel_ready(base);
+ case TPACKET_V2:
+ return __v2_tx_kernel_ready(base);
+ default:
+ bug_on(1);
+ return 0;
+ }
+}
+
+static inline void __v1_v2_tx_user_ready(void *base, int version)
+{
+ switch (version) {
+ case TPACKET_V1:
+ __v1_tx_user_ready(base);
+ break;
+ case TPACKET_V2:
+ __v2_tx_user_ready(base);
+ break;
+ }
+}
+
+static void __v1_v2_set_packet_loss_discard(int sock)
+{
+ int ret, discard = 1;
+
+ ret = setsockopt(sock, SOL_PACKET, PACKET_LOSS, (void *) &discard,
+ sizeof(discard));
+ if (ret == -1) {
+ perror("setsockopt");
+ exit(1);
+ }
+}
+
+static void walk_v1_v2_tx(int sock, struct ring *ring)
+{
+ struct pollfd pfd;
+ int rcv_sock, ret;
+ size_t packet_len;
+ union frame_map ppd;
+ char packet[1024];
+ unsigned int frame_num = 0, got = 0;
+ struct sockaddr_ll ll = {
+ .sll_family = PF_PACKET,
+ .sll_halen = ETH_ALEN,
+ };
+
+ bug_on(ring->type != PACKET_TX_RING);
+ bug_on(ring->rd_num < NUM_PACKETS);
+
+ rcv_sock = socket(PF_PACKET, SOCK_RAW, htons(ETH_P_ALL));
+ if (rcv_sock == -1) {
+ perror("socket");
+ exit(1);
+ }
+
+ pair_udp_setfilter(rcv_sock);
+
+ ll.sll_ifindex = if_nametoindex("lo");
+ ret = bind(rcv_sock, (struct sockaddr *) &ll, sizeof(ll));
+ if (ret == -1) {
+ perror("bind");
+ exit(1);
+ }
+
+ memset(&pfd, 0, sizeof(pfd));
+ pfd.fd = sock;
+ pfd.events = POLLOUT | POLLERR;
+ pfd.revents = 0;
+
+ total_packets = NUM_PACKETS;
+ create_payload(packet, &packet_len);
+
+ while (total_packets > 0) {
+ while (__v1_v2_tx_kernel_ready(ring->rd[frame_num].iov_base,
+ ring->version) &&
+ total_packets > 0) {
+ ppd.raw = ring->rd[frame_num].iov_base;
+
+ switch (ring->version) {
+ case TPACKET_V1:
+ ppd.v1->tp_h.tp_snaplen = packet_len;
+ ppd.v1->tp_h.tp_len = packet_len;
+
+ memcpy((uint8_t *) ppd.raw + TPACKET_HDRLEN -
+ sizeof(struct sockaddr_ll), packet,
+ packet_len);
+ total_bytes += ppd.v1->tp_h.tp_snaplen;
+ break;
+
+ case TPACKET_V2:
+ ppd.v2->tp_h.tp_snaplen = packet_len;
+ ppd.v2->tp_h.tp_len = packet_len;
+
+ memcpy((uint8_t *) ppd.raw + TPACKET2_HDRLEN -
+ sizeof(struct sockaddr_ll), packet,
+ packet_len);
+ total_bytes += ppd.v2->tp_h.tp_snaplen;
+ break;
+ }
+
+ status_bar_update();
+ total_packets--;
+
+ __v1_v2_tx_user_ready(ppd.raw, ring->version);
+
+ frame_num = (frame_num + 1) % ring->rd_num;
+ }
+
+ poll(&pfd, 1, 1);
+ }
+
+ bug_on(total_packets != 0);
+
+ ret = sendto(sock, NULL, 0, 0, NULL, 0);
+ if (ret == -1) {
+ perror("sendto");
+ exit(1);
+ }
+
+ while ((ret = recvfrom(rcv_sock, packet, sizeof(packet),
+ 0, NULL, NULL)) > 0 &&
+ total_packets < NUM_PACKETS) {
+ got += ret;
+ test_payload(packet, ret);
+
+ status_bar_update();
+ total_packets++;
+ }
+
+ close(rcv_sock);
+
+ if (total_packets != NUM_PACKETS) {
+ fprintf(stderr, "walk_v%d_rx: received %u out of %u pkts\n",
+ ring->version, total_packets, NUM_PACKETS);
+ exit(1);
+ }
+
+ fprintf(stderr, " %u pkts (%u bytes)", NUM_PACKETS, got);
+}
+
+static void walk_v1_v2(int sock, struct ring *ring)
+{
+ if (ring->type == PACKET_RX_RING)
+ walk_v1_v2_rx(sock, ring);
+ else
+ walk_v1_v2_tx(sock, ring);
+}
+
+static uint64_t __v3_prev_block_seq_num = 0;
+
+void __v3_test_block_seq_num(struct block_desc *pbd)
+{
+ if (__v3_prev_block_seq_num + 1 != pbd->h1.seq_num) {
+ fprintf(stderr, "\nprev_block_seq_num:%"PRIu64", expected "
+ "seq:%"PRIu64" != actual seq:%"PRIu64"\n",
+ __v3_prev_block_seq_num, __v3_prev_block_seq_num + 1,
+ (uint64_t) pbd->h1.seq_num);
+ exit(1);
+ }
+
+ __v3_prev_block_seq_num = pbd->h1.seq_num;
+}
+
+static void __v3_test_block_len(struct block_desc *pbd, uint32_t bytes, int block_num)
+{
+ if (pbd->h1.num_pkts && bytes != pbd->h1.blk_len) {
+ fprintf(stderr, "\nblock:%u with %upackets, expected "
+ "len:%u != actual len:%u\n", block_num,
+ pbd->h1.num_pkts, bytes, pbd->h1.blk_len);
+ exit(1);
+ }
+}
+
+static void __v3_test_block_header(struct block_desc *pbd, const int block_num)
+{
+ if ((pbd->h1.block_status & TP_STATUS_USER) == 0) {
+ fprintf(stderr, "\nblock %u: not in TP_STATUS_USER\n", block_num);
+ exit(1);
+ }
+
+ __v3_test_block_seq_num(pbd);
+}
+
+static void __v3_walk_block(struct block_desc *pbd, const int block_num)
+{
+ int num_pkts = pbd->h1.num_pkts, i;
+ unsigned long bytes = 0, bytes_with_padding = ALIGN_8(sizeof(*pbd));
+ struct tpacket3_hdr *ppd;
+
+ __v3_test_block_header(pbd, block_num);
+
+ ppd = (struct tpacket3_hdr *) ((uint8_t *) pbd +
+ pbd->h1.offset_to_first_pkt);
+
+ for (i = 0; i < num_pkts; ++i) {
+ bytes += ppd->tp_snaplen;
+
+ if (ppd->tp_next_offset)
+ bytes_with_padding += ppd->tp_next_offset;
+ else
+ bytes_with_padding += ALIGN_8(ppd->tp_snaplen + ppd->tp_mac);
+
+ test_payload((uint8_t *) ppd + ppd->tp_mac, ppd->tp_snaplen);
+
+ status_bar_update();
+ total_packets++;
+
+ ppd = (struct tpacket3_hdr *) ((uint8_t *) ppd + ppd->tp_next_offset);
+ __sync_synchronize();
+ }
+
+ __v3_test_block_len(pbd, bytes_with_padding, block_num);
+ total_bytes += bytes;
+}
+
+void __v3_flush_block(struct block_desc *pbd)
+{
+ pbd->h1.block_status = TP_STATUS_KERNEL;
+ __sync_synchronize();
+}
+
+static void walk_v3_rx(int sock, struct ring *ring)
+{
+ unsigned int block_num = 0;
+ struct pollfd pfd;
+ struct block_desc *pbd;
+ int udp_sock[2];
+
+ bug_on(ring->type != PACKET_RX_RING);
+
+ pair_udp_open(udp_sock, PORT_BASE);
+ pair_udp_setfilter(sock);
+
+ memset(&pfd, 0, sizeof(pfd));
+ pfd.fd = sock;
+ pfd.events = POLLIN | POLLERR;
+ pfd.revents = 0;
+
+ pair_udp_send(udp_sock, NUM_PACKETS);
+
+ while (total_packets < NUM_PACKETS * 2) {
+ pbd = (struct block_desc *) ring->rd[block_num].iov_base;
+
+ while ((pbd->h1.block_status & TP_STATUS_USER) == 0)
+ poll(&pfd, 1, 1);
+
+ __v3_walk_block(pbd, block_num);
+ __v3_flush_block(pbd);
+
+ block_num = (block_num + 1) % ring->rd_num;
+ }
+
+ pair_udp_close(udp_sock);
+
+ if (total_packets != 2 * NUM_PACKETS) {
+ fprintf(stderr, "walk_v3_rx: received %u out of %u pkts\n",
+ total_packets, NUM_PACKETS);
+ exit(1);
+ }
+
+ fprintf(stderr, " %u pkts (%u bytes)", NUM_PACKETS, total_bytes >> 1);
+}
+
+static void walk_v3(int sock, struct ring *ring)
+{
+ if (ring->type == PACKET_RX_RING)
+ walk_v3_rx(sock, ring);
+ else
+ bug_on(1);
+}
+
+static void __v1_v2_fill(struct ring *ring, unsigned int blocks)
+{
+ ring->req.tp_block_size = getpagesize() << 2;
+ ring->req.tp_frame_size = TPACKET_ALIGNMENT << 7;
+ ring->req.tp_block_nr = blocks;
+
+ ring->req.tp_frame_nr = ring->req.tp_block_size /
+ ring->req.tp_frame_size *
+ ring->req.tp_block_nr;
+
+ ring->mm_len = ring->req.tp_block_size * ring->req.tp_block_nr;
+ ring->walk = walk_v1_v2;
+ ring->rd_num = ring->req.tp_frame_nr;
+ ring->flen = ring->req.tp_frame_size;
+}
+
+static void __v3_fill(struct ring *ring, unsigned int blocks)
+{
+ ring->req3.tp_retire_blk_tov = 64;
+ ring->req3.tp_sizeof_priv = 0;
+ ring->req3.tp_feature_req_word = TP_FT_REQ_FILL_RXHASH;
+
+ ring->req3.tp_block_size = getpagesize() << 2;
+ ring->req3.tp_frame_size = TPACKET_ALIGNMENT << 7;
+ ring->req3.tp_block_nr = blocks;
+
+ ring->req3.tp_frame_nr = ring->req3.tp_block_size /
+ ring->req3.tp_frame_size *
+ ring->req3.tp_block_nr;
+
+ ring->mm_len = ring->req3.tp_block_size * ring->req3.tp_block_nr;
+ ring->walk = walk_v3;
+ ring->rd_num = ring->req3.tp_block_nr;
+ ring->flen = ring->req3.tp_block_size;
+}
+
+static void setup_ring(int sock, struct ring *ring, int version, int type)
+{
+ int ret = 0;
+ unsigned int blocks = 256;
+
+ ring->type = type;
+ ring->version = version;
+
+ switch (version) {
+ case TPACKET_V1:
+ case TPACKET_V2:
+ if (type == PACKET_TX_RING)
+ __v1_v2_set_packet_loss_discard(sock);
+ __v1_v2_fill(ring, blocks);
+ ret = setsockopt(sock, SOL_PACKET, type, &ring->req,
+ sizeof(ring->req));
+ break;
+
+ case TPACKET_V3:
+ __v3_fill(ring, blocks);
+ ret = setsockopt(sock, SOL_PACKET, type, &ring->req3,
+ sizeof(ring->req3));
+ break;
+ }
+
+ if (ret == -1) {
+ perror("setsockopt");
+ exit(1);
+ }
+
+ ring->rd_len = ring->rd_num * sizeof(*ring->rd);
+ ring->rd = malloc(ring->rd_len);
+ if (ring->rd == NULL) {
+ perror("malloc");
+ exit(1);
+ }
+
+ total_packets = 0;
+ total_bytes = 0;
+}
+
+static void mmap_ring(int sock, struct ring *ring)
+{
+ int i;
+
+ ring->mm_space = mmap(0, ring->mm_len, PROT_READ | PROT_WRITE,
+ MAP_SHARED | MAP_LOCKED | MAP_POPULATE, sock, 0);
+ if (ring->mm_space == MAP_FAILED) {
+ perror("mmap");
+ exit(1);
+ }
+
+ memset(ring->rd, 0, ring->rd_len);
+ for (i = 0; i < ring->rd_num; ++i) {
+ ring->rd[i].iov_base = ring->mm_space + (i * ring->flen);
+ ring->rd[i].iov_len = ring->flen;
+ }
+}
+
+static void bind_ring(int sock, struct ring *ring)
+{
+ int ret;
+
+ ring->ll.sll_family = PF_PACKET;
+ ring->ll.sll_protocol = htons(ETH_P_ALL);
+ ring->ll.sll_ifindex = if_nametoindex("lo");
+ ring->ll.sll_hatype = 0;
+ ring->ll.sll_pkttype = 0;
+ ring->ll.sll_halen = 0;
+
+ ret = bind(sock, (struct sockaddr *) &ring->ll, sizeof(ring->ll));
+ if (ret == -1) {
+ perror("bind");
+ exit(1);
+ }
+}
+
+static void walk_ring(int sock, struct ring *ring)
+{
+ ring->walk(sock, ring);
+}
+
+static void unmap_ring(int sock, struct ring *ring)
+{
+ munmap(ring->mm_space, ring->mm_len);
+ free(ring->rd);
+}
+
+static int test_kernel_bit_width(void)
+{
+ char in[512], *ptr;
+ int num = 0, fd;
+ ssize_t ret;
+
+ fd = open("/proc/kallsyms", O_RDONLY);
+ if (fd == -1) {
+ perror("open");
+ exit(1);
+ }
+
+ ret = read(fd, in, sizeof(in));
+ if (ret <= 0) {
+ perror("read");
+ exit(1);
+ }
+
+ close(fd);
+
+ ptr = in;
+ while(!isspace(*ptr)) {
+ num++;
+ ptr++;
+ }
+
+ return num * 4;
+}
+
+static int test_user_bit_width(void)
+{
+ return __WORDSIZE;
+}
+
+static const char *tpacket_str[] = {
+ [TPACKET_V1] = "TPACKET_V1",
+ [TPACKET_V2] = "TPACKET_V2",
+ [TPACKET_V3] = "TPACKET_V3",
+};
+
+static const char *type_str[] = {
+ [PACKET_RX_RING] = "PACKET_RX_RING",
+ [PACKET_TX_RING] = "PACKET_TX_RING",
+};
+
+static int test_tpacket(int version, int type)
+{
+ int sock;
+ struct ring ring;
+
+ fprintf(stderr, "test: %s with %s ", tpacket_str[version],
+ type_str[type]);
+ fflush(stderr);
+
+ if (version == TPACKET_V1 &&
+ test_kernel_bit_width() != test_user_bit_width()) {
+ fprintf(stderr, "test: skip %s %s since user and kernel "
+ "space have different bit width\n",
+ tpacket_str[version], type_str[type]);
+ return 0;
+ }
+
+ sock = pfsocket(version);
+ memset(&ring, 0, sizeof(ring));
+ setup_ring(sock, &ring, version, type);
+ mmap_ring(sock, &ring);
+ bind_ring(sock, &ring);
+ walk_ring(sock, &ring);
+ unmap_ring(sock, &ring);
+ close(sock);
+
+ fprintf(stderr, "\n");
+ return 0;
+}
+
+int main(void)
+{
+ int ret = 0;
+
+ ret |= test_tpacket(TPACKET_V1, PACKET_RX_RING);
+ ret |= test_tpacket(TPACKET_V1, PACKET_TX_RING);
+
+ ret |= test_tpacket(TPACKET_V2, PACKET_RX_RING);
+ ret |= test_tpacket(TPACKET_V2, PACKET_TX_RING);
+
+ ret |= test_tpacket(TPACKET_V3, PACKET_RX_RING);
+
+ if (ret)
+ return 1;
+
+ printf("OK. All tests passed\n");
+ return 0;
+}
diff --git a/net/run_afpackettests b/net/run_afpackettests
new file mode 100644
index 0000000..5246e78
--- /dev/null
+++ b/net/run_afpackettests
@@ -0,0 +1,26 @@
+#!/bin/sh
+
+if [ $(id -u) != 0 ]; then
+ echo $msg must be run as root >&2
+ exit 0
+fi
+
+echo "--------------------"
+echo "running psock_fanout test"
+echo "--------------------"
+./psock_fanout
+if [ $? -ne 0 ]; then
+ echo "[FAIL]"
+else
+ echo "[PASS]"
+fi
+
+echo "--------------------"
+echo "running psock_tpacket test"
+echo "--------------------"
+./psock_tpacket
+if [ $? -ne 0 ]; then
+ echo "[FAIL]"
+else
+ echo "[PASS]"
+fi
diff --git a/net/run_netsocktests b/net/run_netsocktests
new file mode 100644
index 0000000..c09a682
--- /dev/null
+++ b/net/run_netsocktests
@@ -0,0 +1,12 @@
+#!/bin/bash
+
+echo "--------------------"
+echo "running socket test"
+echo "--------------------"
+./socket
+if [ $? -ne 0 ]; then
+ echo "[FAIL]"
+else
+ echo "[PASS]"
+fi
+
diff --git a/net/socket.c b/net/socket.c
new file mode 100644
index 0000000..0f227f2
--- /dev/null
+++ b/net/socket.c
@@ -0,0 +1,92 @@
+#include <stdio.h>
+#include <errno.h>
+#include <unistd.h>
+#include <string.h>
+#include <sys/types.h>
+#include <sys/socket.h>
+#include <netinet/in.h>
+
+struct socket_testcase {
+ int domain;
+ int type;
+ int protocol;
+
+ /* 0 = valid file descriptor
+ * -foo = error foo
+ */
+ int expect;
+
+ /* If non-zero, accept EAFNOSUPPORT to handle the case
+ * of the protocol not being configured into the kernel.
+ */
+ int nosupport_ok;
+};
+
+static struct socket_testcase tests[] = {
+ { AF_MAX, 0, 0, -EAFNOSUPPORT, 0 },
+ { AF_INET, SOCK_STREAM, IPPROTO_TCP, 0, 1 },
+ { AF_INET, SOCK_DGRAM, IPPROTO_TCP, -EPROTONOSUPPORT, 1 },
+ { AF_INET, SOCK_DGRAM, IPPROTO_UDP, 0, 1 },
+ { AF_INET, SOCK_STREAM, IPPROTO_UDP, -EPROTONOSUPPORT, 1 },
+};
+
+#define ARRAY_SIZE(arr) (sizeof(arr) / sizeof((arr)[0]))
+#define ERR_STRING_SZ 64
+
+static int run_tests(void)
+{
+ char err_string1[ERR_STRING_SZ];
+ char err_string2[ERR_STRING_SZ];
+ int i, err;
+
+ err = 0;
+ for (i = 0; i < ARRAY_SIZE(tests); i++) {
+ struct socket_testcase *s = &tests[i];
+ int fd;
+
+ fd = socket(s->domain, s->type, s->protocol);
+ if (fd < 0) {
+ if (s->nosupport_ok &&
+ errno == EAFNOSUPPORT)
+ continue;
+
+ if (s->expect < 0 &&
+ errno == -s->expect)
+ continue;
+
+ strerror_r(-s->expect, err_string1, ERR_STRING_SZ);
+ strerror_r(errno, err_string2, ERR_STRING_SZ);
+
+ fprintf(stderr, "socket(%d, %d, %d) expected "
+ "err (%s) got (%s)\n",
+ s->domain, s->type, s->protocol,
+ err_string1, err_string2);
+
+ err = -1;
+ break;
+ } else {
+ close(fd);
+
+ if (s->expect < 0) {
+ strerror_r(errno, err_string1, ERR_STRING_SZ);
+
+ fprintf(stderr, "socket(%d, %d, %d) expected "
+ "success got err (%s)\n",
+ s->domain, s->type, s->protocol,
+ err_string1);
+
+ err = -1;
+ break;
+ }
+ }
+ }
+
+ return err;
+}
+
+int main(void)
+{
+ int err = run_tests();
+
+ return err;
+}
diff --git a/powerpc/Makefile b/powerpc/Makefile
new file mode 100644
index 0000000..74a78ce
--- /dev/null
+++ b/powerpc/Makefile
@@ -0,0 +1,39 @@
+# Makefile for powerpc selftests
+
+# ARCH can be overridden by the user for cross compiling
+ARCH ?= $(shell uname -m)
+ARCH := $(shell echo $(ARCH) | sed -e s/ppc.*/powerpc/)
+
+ifeq ($(ARCH),powerpc)
+
+GIT_VERSION = $(shell git describe --always --long --dirty || echo "unknown")
+
+CC := $(CROSS_COMPILE)$(CC)
+CFLAGS := -Wall -O2 -flto -Wall -Werror -DGIT_VERSION='"$(GIT_VERSION)"' -I$(CURDIR) $(CFLAGS)
+
+export CC CFLAGS
+
+TARGETS = pmu copyloops mm tm
+
+endif
+
+all: $(TARGETS)
+
+$(TARGETS):
+ $(MAKE) -k -C $@ all
+
+run_tests: all
+ @for TARGET in $(TARGETS); do \
+ $(MAKE) -C $$TARGET run_tests; \
+ done;
+
+clean:
+ @for TARGET in $(TARGETS); do \
+ $(MAKE) -C $$TARGET clean; \
+ done;
+ rm -f tags
+
+tags:
+ find . -name '*.c' -o -name '*.h' | xargs ctags
+
+.PHONY: all run_tests clean tags $(TARGETS)
diff --git a/powerpc/copyloops/Makefile b/powerpc/copyloops/Makefile
new file mode 100644
index 0000000..6f2d3be
--- /dev/null
+++ b/powerpc/copyloops/Makefile
@@ -0,0 +1,29 @@
+# The loops are all 64-bit code
+CFLAGS += -m64
+CFLAGS += -I$(CURDIR)
+CFLAGS += -D SELFTEST
+
+# Use our CFLAGS for the implicit .S rule
+ASFLAGS = $(CFLAGS)
+
+PROGS := copyuser_64 copyuser_power7 memcpy_64 memcpy_power7
+EXTRA_SOURCES := validate.c ../harness.c
+
+all: $(PROGS)
+
+copyuser_64: CPPFLAGS += -D COPY_LOOP=test___copy_tofrom_user_base
+copyuser_power7: CPPFLAGS += -D COPY_LOOP=test___copy_tofrom_user_power7
+memcpy_64: CPPFLAGS += -D COPY_LOOP=test_memcpy
+memcpy_power7: CPPFLAGS += -D COPY_LOOP=test_memcpy_power7
+
+$(PROGS): $(EXTRA_SOURCES)
+
+run_tests: all
+ @-for PROG in $(PROGS); do \
+ ./$$PROG; \
+ done;
+
+clean:
+ rm -f $(PROGS) *.o
+
+.PHONY: all run_tests clean
diff --git a/powerpc/copyloops/asm/ppc_asm.h b/powerpc/copyloops/asm/ppc_asm.h
new file mode 100644
index 0000000..d1dc374
--- /dev/null
+++ b/powerpc/copyloops/asm/ppc_asm.h
@@ -0,0 +1,89 @@
+#include <ppc-asm.h>
+
+#define CONFIG_ALTIVEC
+
+#define r1 1
+
+#define vr0 0
+#define vr1 1
+#define vr2 2
+#define vr3 3
+#define vr4 4
+#define vr5 5
+#define vr6 6
+#define vr7 7
+#define vr8 8
+#define vr9 9
+#define vr10 10
+#define vr11 11
+#define vr12 12
+#define vr13 13
+#define vr14 14
+#define vr15 15
+#define vr16 16
+#define vr17 17
+#define vr18 18
+#define vr19 19
+#define vr20 20
+#define vr21 21
+#define vr22 22
+#define vr23 23
+#define vr24 24
+#define vr25 25
+#define vr26 26
+#define vr27 27
+#define vr28 28
+#define vr29 29
+#define vr30 30
+#define vr31 31
+
+#define R14 r14
+#define R15 r15
+#define R16 r16
+#define R17 r17
+#define R18 r18
+#define R19 r19
+#define R20 r20
+#define R21 r21
+#define R22 r22
+#define R29 r29
+#define R30 r30
+#define R31 r31
+
+#define STACKFRAMESIZE 256
+#define STK_REG(i) (112 + ((i)-14)*8)
+
+#define _GLOBAL(A) FUNC_START(test_ ## A)
+#define _GLOBAL_TOC(A) _GLOBAL(A)
+
+#define PPC_MTOCRF(A, B) mtocrf A, B
+
+FUNC_START(enter_vmx_usercopy)
+ li r3,1
+ blr
+
+FUNC_START(exit_vmx_usercopy)
+ li r3,0
+ blr
+
+FUNC_START(enter_vmx_copy)
+ li r3,1
+ blr
+
+FUNC_START(exit_vmx_copy)
+ blr
+
+FUNC_START(memcpy_power7)
+ blr
+
+FUNC_START(__copy_tofrom_user_power7)
+ blr
+
+FUNC_START(__copy_tofrom_user_base)
+ blr
+
+#define BEGIN_FTR_SECTION
+#define FTR_SECTION_ELSE
+#define ALT_FTR_SECTION_END_IFCLR(x)
+#define ALT_FTR_SECTION_END(x, y)
+#define END_FTR_SECTION_IFCLR(x)
diff --git a/powerpc/copyloops/asm/processor.h b/powerpc/copyloops/asm/processor.h
new file mode 100644
index 0000000..e69de29
--- /dev/null
+++ b/powerpc/copyloops/asm/processor.h
diff --git a/powerpc/copyloops/copyuser_64.S b/powerpc/copyloops/copyuser_64.S
new file mode 120000
index 0000000..f1c418a
--- /dev/null
+++ b/powerpc/copyloops/copyuser_64.S
@@ -0,0 +1 @@
+../../../../../arch/powerpc/lib/copyuser_64.S \ No newline at end of file
diff --git a/powerpc/copyloops/copyuser_power7.S b/powerpc/copyloops/copyuser_power7.S
new file mode 120000
index 0000000..4786895
--- /dev/null
+++ b/powerpc/copyloops/copyuser_power7.S
@@ -0,0 +1 @@
+../../../../../arch/powerpc/lib/copyuser_power7.S \ No newline at end of file
diff --git a/powerpc/copyloops/memcpy_64.S b/powerpc/copyloops/memcpy_64.S
new file mode 120000
index 0000000..cce33fb
--- /dev/null
+++ b/powerpc/copyloops/memcpy_64.S
@@ -0,0 +1 @@
+../../../../../arch/powerpc/lib/memcpy_64.S \ No newline at end of file
diff --git a/powerpc/copyloops/memcpy_power7.S b/powerpc/copyloops/memcpy_power7.S
new file mode 120000
index 0000000..0d6fbfa
--- /dev/null
+++ b/powerpc/copyloops/memcpy_power7.S
@@ -0,0 +1 @@
+../../../../../arch/powerpc/lib/memcpy_power7.S \ No newline at end of file
diff --git a/powerpc/copyloops/validate.c b/powerpc/copyloops/validate.c
new file mode 100644
index 0000000..1750ff5
--- /dev/null
+++ b/powerpc/copyloops/validate.c
@@ -0,0 +1,99 @@
+#include <malloc.h>
+#include <string.h>
+#include <stdlib.h>
+#include <stdbool.h>
+
+#include "../utils.h"
+
+#define MAX_LEN 8192
+#define MAX_OFFSET 16
+#define MIN_REDZONE 128
+#define BUFLEN (MAX_LEN+MAX_OFFSET+2*MIN_REDZONE)
+#define POISON 0xa5
+
+unsigned long COPY_LOOP(void *to, const void *from, unsigned long size);
+
+static void do_one(char *src, char *dst, unsigned long src_off,
+ unsigned long dst_off, unsigned long len, void *redzone,
+ void *fill)
+{
+ char *srcp, *dstp;
+ unsigned long ret;
+ unsigned long i;
+
+ srcp = src + MIN_REDZONE + src_off;
+ dstp = dst + MIN_REDZONE + dst_off;
+
+ memset(src, POISON, BUFLEN);
+ memset(dst, POISON, BUFLEN);
+ memcpy(srcp, fill, len);
+
+ ret = COPY_LOOP(dstp, srcp, len);
+ if (ret && ret != (unsigned long)dstp) {
+ printf("(%p,%p,%ld) returned %ld\n", dstp, srcp, len, ret);
+ abort();
+ }
+
+ if (memcmp(dstp, srcp, len)) {
+ printf("(%p,%p,%ld) miscompare\n", dstp, srcp, len);
+ printf("src: ");
+ for (i = 0; i < len; i++)
+ printf("%02x ", srcp[i]);
+ printf("\ndst: ");
+ for (i = 0; i < len; i++)
+ printf("%02x ", dstp[i]);
+ printf("\n");
+ abort();
+ }
+
+ if (memcmp(dst, redzone, dstp - dst)) {
+ printf("(%p,%p,%ld) redzone before corrupted\n",
+ dstp, srcp, len);
+ abort();
+ }
+
+ if (memcmp(dstp+len, redzone, dst+BUFLEN-(dstp+len))) {
+ printf("(%p,%p,%ld) redzone after corrupted\n",
+ dstp, srcp, len);
+ abort();
+ }
+}
+
+int test_copy_loop(void)
+{
+ char *src, *dst, *redzone, *fill;
+ unsigned long len, src_off, dst_off;
+ unsigned long i;
+
+ src = memalign(BUFLEN, BUFLEN);
+ dst = memalign(BUFLEN, BUFLEN);
+ redzone = malloc(BUFLEN);
+ fill = malloc(BUFLEN);
+
+ if (!src || !dst || !redzone || !fill) {
+ fprintf(stderr, "malloc failed\n");
+ exit(1);
+ }
+
+ memset(redzone, POISON, BUFLEN);
+
+ /* Fill with sequential bytes */
+ for (i = 0; i < BUFLEN; i++)
+ fill[i] = i & 0xff;
+
+ for (len = 1; len < MAX_LEN; len++) {
+ for (src_off = 0; src_off < MAX_OFFSET; src_off++) {
+ for (dst_off = 0; dst_off < MAX_OFFSET; dst_off++) {
+ do_one(src, dst, src_off, dst_off, len,
+ redzone, fill);
+ }
+ }
+ }
+
+ return 0;
+}
+
+int main(void)
+{
+ return test_harness(test_copy_loop, str(COPY_LOOP));
+}
diff --git a/powerpc/harness.c b/powerpc/harness.c
new file mode 100644
index 0000000..8ebc58a
--- /dev/null
+++ b/powerpc/harness.c
@@ -0,0 +1,114 @@
+/*
+ * Copyright 2013, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <errno.h>
+#include <signal.h>
+#include <stdbool.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <unistd.h>
+
+#include "subunit.h"
+#include "utils.h"
+
+#define TIMEOUT 120
+#define KILL_TIMEOUT 5
+
+
+int run_test(int (test_function)(void), char *name)
+{
+ bool terminated;
+ int rc, status;
+ pid_t pid;
+
+ /* Make sure output is flushed before forking */
+ fflush(stdout);
+
+ pid = fork();
+ if (pid == 0) {
+ setpgid(0, 0);
+ exit(test_function());
+ } else if (pid == -1) {
+ perror("fork");
+ return 1;
+ }
+
+ setpgid(pid, pid);
+
+ /* Wake us up in timeout seconds */
+ alarm(TIMEOUT);
+ terminated = false;
+
+wait:
+ rc = waitpid(pid, &status, 0);
+ if (rc == -1) {
+ if (errno != EINTR) {
+ printf("unknown error from waitpid\n");
+ return 1;
+ }
+
+ if (terminated) {
+ printf("!! force killing %s\n", name);
+ kill(-pid, SIGKILL);
+ return 1;
+ } else {
+ printf("!! killing %s\n", name);
+ kill(-pid, SIGTERM);
+ terminated = true;
+ alarm(KILL_TIMEOUT);
+ goto wait;
+ }
+ }
+
+ /* Kill anything else in the process group that is still running */
+ kill(-pid, SIGTERM);
+
+ if (WIFEXITED(status))
+ status = WEXITSTATUS(status);
+ else {
+ if (WIFSIGNALED(status))
+ printf("!! child died by signal %d\n", WTERMSIG(status));
+ else
+ printf("!! child died by unknown cause\n");
+
+ status = 1; /* Signal or other */
+ }
+
+ return status;
+}
+
+static void alarm_handler(int signum)
+{
+ /* Jut wake us up from waitpid */
+}
+
+static struct sigaction alarm_action = {
+ .sa_handler = alarm_handler,
+};
+
+int test_harness(int (test_function)(void), char *name)
+{
+ int rc;
+
+ test_start(name);
+ test_set_git_version(GIT_VERSION);
+
+ if (sigaction(SIGALRM, &alarm_action, NULL)) {
+ perror("sigaction");
+ test_error(name);
+ return 1;
+ }
+
+ rc = run_test(test_function, name);
+
+ if (rc == MAGIC_SKIP_RETURN_VALUE)
+ test_skip(name);
+ else
+ test_finish(name, rc);
+
+ return rc;
+}
diff --git a/powerpc/mm/Makefile b/powerpc/mm/Makefile
new file mode 100644
index 0000000..357ccbd
--- /dev/null
+++ b/powerpc/mm/Makefile
@@ -0,0 +1,18 @@
+noarg:
+ $(MAKE) -C ../
+
+PROGS := hugetlb_vs_thp_test
+
+all: $(PROGS)
+
+$(PROGS): ../harness.c
+
+run_tests: all
+ @-for PROG in $(PROGS); do \
+ ./$$PROG; \
+ done;
+
+clean:
+ rm -f $(PROGS)
+
+.PHONY: all run_tests clean
diff --git a/powerpc/mm/hugetlb_vs_thp_test.c b/powerpc/mm/hugetlb_vs_thp_test.c
new file mode 100644
index 0000000..3d8e5b0
--- /dev/null
+++ b/powerpc/mm/hugetlb_vs_thp_test.c
@@ -0,0 +1,72 @@
+#include <stdio.h>
+#include <sys/mman.h>
+#include <unistd.h>
+
+#include "utils.h"
+
+/* This must match the huge page & THP size */
+#define SIZE (16 * 1024 * 1024)
+
+static int test_body(void)
+{
+ void *addr;
+ char *p;
+
+ addr = (void *)0xa0000000;
+
+ p = mmap(addr, SIZE, PROT_READ | PROT_WRITE,
+ MAP_HUGETLB | MAP_ANONYMOUS | MAP_PRIVATE, -1, 0);
+ if (p != MAP_FAILED) {
+ /*
+ * Typically the mmap will fail because no huge pages are
+ * allocated on the system. But if there are huge pages
+ * allocated the mmap will succeed. That's fine too, we just
+ * munmap here before continuing.
+ */
+ munmap(addr, SIZE);
+ }
+
+ p = mmap(addr, SIZE, PROT_READ | PROT_WRITE,
+ MAP_ANONYMOUS | MAP_PRIVATE, -1, 0);
+ if (p == MAP_FAILED) {
+ printf("Mapping failed @ %p\n", addr);
+ perror("mmap");
+ return 1;
+ }
+
+ /*
+ * Either a user or kernel access is sufficient to trigger the bug.
+ * A kernel access is easier to spot & debug, as it will trigger the
+ * softlockup or RCU stall detectors, and when the system is kicked
+ * into xmon we get a backtrace in the kernel.
+ *
+ * A good option is:
+ * getcwd(p, SIZE);
+ *
+ * For the purposes of this testcase it's preferable to spin in
+ * userspace, so the harness can kill us if we get stuck. That way we
+ * see a test failure rather than a dead system.
+ */
+ *p = 0xf;
+
+ munmap(addr, SIZE);
+
+ return 0;
+}
+
+static int test_main(void)
+{
+ int i;
+
+ /* 10,000 because it's a "bunch", and completes reasonably quickly */
+ for (i = 0; i < 10000; i++)
+ if (test_body())
+ return 1;
+
+ return 0;
+}
+
+int main(void)
+{
+ return test_harness(test_main, "hugetlb_vs_thp");
+}
diff --git a/powerpc/pmu/Makefile b/powerpc/pmu/Makefile
new file mode 100644
index 0000000..c9f4263
--- /dev/null
+++ b/powerpc/pmu/Makefile
@@ -0,0 +1,38 @@
+noarg:
+ $(MAKE) -C ../
+
+PROGS := count_instructions l3_bank_test per_event_excludes
+EXTRA_SOURCES := ../harness.c event.c lib.c
+
+SUB_TARGETS = ebb
+
+all: $(PROGS) $(SUB_TARGETS)
+
+$(PROGS): $(EXTRA_SOURCES)
+
+# loop.S can only be built 64-bit
+count_instructions: loop.S count_instructions.c $(EXTRA_SOURCES)
+ $(CC) $(CFLAGS) -m64 -o $@ $^
+
+run_tests: all sub_run_tests
+ @-for PROG in $(PROGS); do \
+ ./$$PROG; \
+ done;
+
+clean: sub_clean
+ rm -f $(PROGS) loop.o
+
+$(SUB_TARGETS):
+ $(MAKE) -k -C $@ all
+
+sub_run_tests: all
+ @for TARGET in $(SUB_TARGETS); do \
+ $(MAKE) -C $$TARGET run_tests; \
+ done;
+
+sub_clean:
+ @for TARGET in $(SUB_TARGETS); do \
+ $(MAKE) -C $$TARGET clean; \
+ done;
+
+.PHONY: all run_tests clean sub_run_tests sub_clean $(SUB_TARGETS)
diff --git a/powerpc/pmu/count_instructions.c b/powerpc/pmu/count_instructions.c
new file mode 100644
index 0000000..4622117
--- /dev/null
+++ b/powerpc/pmu/count_instructions.c
@@ -0,0 +1,147 @@
+/*
+ * Copyright 2013, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#define _GNU_SOURCE
+
+#include <stdio.h>
+#include <stdbool.h>
+#include <string.h>
+#include <sys/prctl.h>
+
+#include "event.h"
+#include "utils.h"
+#include "lib.h"
+
+extern void thirty_two_instruction_loop(u64 loops);
+
+static void setup_event(struct event *e, u64 config, char *name)
+{
+ event_init_opts(e, config, PERF_TYPE_HARDWARE, name);
+
+ e->attr.disabled = 1;
+ e->attr.exclude_kernel = 1;
+ e->attr.exclude_hv = 1;
+ e->attr.exclude_idle = 1;
+}
+
+static int do_count_loop(struct event *events, u64 instructions,
+ u64 overhead, bool report)
+{
+ s64 difference, expected;
+ double percentage;
+
+ prctl(PR_TASK_PERF_EVENTS_ENABLE);
+
+ /* Run for 1M instructions */
+ thirty_two_instruction_loop(instructions >> 5);
+
+ prctl(PR_TASK_PERF_EVENTS_DISABLE);
+
+ event_read(&events[0]);
+ event_read(&events[1]);
+
+ expected = instructions + overhead;
+ difference = events[0].result.value - expected;
+ percentage = (double)difference / events[0].result.value * 100;
+
+ if (report) {
+ event_report(&events[0]);
+ event_report(&events[1]);
+
+ printf("Looped for %llu instructions, overhead %llu\n", instructions, overhead);
+ printf("Expected %llu\n", expected);
+ printf("Actual %llu\n", events[0].result.value);
+ printf("Delta %lld, %f%%\n", difference, percentage);
+ }
+
+ event_reset(&events[0]);
+ event_reset(&events[1]);
+
+ if (difference < 0)
+ difference = -difference;
+
+ /* Tolerate a difference below 0.0001 % */
+ difference *= 10000 * 100;
+ if (difference / events[0].result.value)
+ return -1;
+
+ return 0;
+}
+
+/* Count how many instructions it takes to do a null loop */
+static u64 determine_overhead(struct event *events)
+{
+ u64 current, overhead;
+ int i;
+
+ do_count_loop(events, 0, 0, false);
+ overhead = events[0].result.value;
+
+ for (i = 0; i < 100; i++) {
+ do_count_loop(events, 0, 0, false);
+ current = events[0].result.value;
+ if (current < overhead) {
+ printf("Replacing overhead %llu with %llu\n", overhead, current);
+ overhead = current;
+ }
+ }
+
+ return overhead;
+}
+
+static int test_body(void)
+{
+ struct event events[2];
+ u64 overhead;
+
+ setup_event(&events[0], PERF_COUNT_HW_INSTRUCTIONS, "instructions");
+ setup_event(&events[1], PERF_COUNT_HW_CPU_CYCLES, "cycles");
+
+ if (event_open(&events[0])) {
+ perror("perf_event_open");
+ return -1;
+ }
+
+ if (event_open_with_group(&events[1], events[0].fd)) {
+ perror("perf_event_open");
+ return -1;
+ }
+
+ overhead = determine_overhead(events);
+ printf("Overhead of null loop: %llu instructions\n", overhead);
+
+ /* Run for 1Mi instructions */
+ FAIL_IF(do_count_loop(events, 1000000, overhead, true));
+
+ /* Run for 10Mi instructions */
+ FAIL_IF(do_count_loop(events, 10000000, overhead, true));
+
+ /* Run for 100Mi instructions */
+ FAIL_IF(do_count_loop(events, 100000000, overhead, true));
+
+ /* Run for 1Bi instructions */
+ FAIL_IF(do_count_loop(events, 1000000000, overhead, true));
+
+ /* Run for 16Bi instructions */
+ FAIL_IF(do_count_loop(events, 16000000000, overhead, true));
+
+ /* Run for 64Bi instructions */
+ FAIL_IF(do_count_loop(events, 64000000000, overhead, true));
+
+ event_close(&events[0]);
+ event_close(&events[1]);
+
+ return 0;
+}
+
+static int count_instructions(void)
+{
+ return eat_cpu(test_body);
+}
+
+int main(void)
+{
+ return test_harness(count_instructions, "count_instructions");
+}
diff --git a/powerpc/pmu/ebb/Makefile b/powerpc/pmu/ebb/Makefile
new file mode 100644
index 0000000..3dc4332
--- /dev/null
+++ b/powerpc/pmu/ebb/Makefile
@@ -0,0 +1,33 @@
+noarg:
+ $(MAKE) -C ../../
+
+# The EBB handler is 64-bit code and everything links against it
+CFLAGS += -m64
+
+PROGS := reg_access_test event_attributes_test cycles_test \
+ cycles_with_freeze_test pmc56_overflow_test \
+ ebb_vs_cpu_event_test cpu_event_vs_ebb_test \
+ cpu_event_pinned_vs_ebb_test task_event_vs_ebb_test \
+ task_event_pinned_vs_ebb_test multi_ebb_procs_test \
+ multi_counter_test pmae_handling_test \
+ close_clears_pmcc_test instruction_count_test \
+ fork_cleanup_test ebb_on_child_test \
+ ebb_on_willing_child_test back_to_back_ebbs_test \
+ lost_exception_test no_handler_test \
+ cycles_with_mmcr2_test
+
+all: $(PROGS)
+
+$(PROGS): ../../harness.c ../event.c ../lib.c ebb.c ebb_handler.S trace.c busy_loop.S
+
+instruction_count_test: ../loop.S
+
+lost_exception_test: ../lib.c
+
+run_tests: all
+ @-for PROG in $(PROGS); do \
+ ./$$PROG; \
+ done;
+
+clean:
+ rm -f $(PROGS)
diff --git a/powerpc/pmu/ebb/back_to_back_ebbs_test.c b/powerpc/pmu/ebb/back_to_back_ebbs_test.c
new file mode 100644
index 0000000..66ea765
--- /dev/null
+++ b/powerpc/pmu/ebb/back_to_back_ebbs_test.c
@@ -0,0 +1,106 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <stdbool.h>
+#include <stdio.h>
+#include <stdlib.h>
+
+#include "ebb.h"
+
+
+#define NUMBER_OF_EBBS 50
+
+/*
+ * Test that if we overflow the counter while in the EBB handler, we take
+ * another EBB on exiting from the handler.
+ *
+ * We do this by counting with a stupidly low sample period, causing us to
+ * overflow the PMU while we're still in the EBB handler, leading to another
+ * EBB.
+ *
+ * We get out of what would otherwise be an infinite loop by leaving the
+ * counter frozen once we've taken enough EBBs.
+ */
+
+static void ebb_callee(void)
+{
+ uint64_t siar, val;
+
+ val = mfspr(SPRN_BESCR);
+ if (!(val & BESCR_PMEO)) {
+ ebb_state.stats.spurious++;
+ goto out;
+ }
+
+ ebb_state.stats.ebb_count++;
+ trace_log_counter(ebb_state.trace, ebb_state.stats.ebb_count);
+
+ /* Resets the PMC */
+ count_pmc(1, sample_period);
+
+out:
+ if (ebb_state.stats.ebb_count == NUMBER_OF_EBBS)
+ /* Reset but leave counters frozen */
+ reset_ebb_with_clear_mask(MMCR0_PMAO);
+ else
+ /* Unfreezes */
+ reset_ebb();
+
+ /* Do some stuff to chew some cycles and pop the counter */
+ siar = mfspr(SPRN_SIAR);
+ trace_log_reg(ebb_state.trace, SPRN_SIAR, siar);
+
+ val = mfspr(SPRN_PMC1);
+ trace_log_reg(ebb_state.trace, SPRN_PMC1, val);
+
+ val = mfspr(SPRN_MMCR0);
+ trace_log_reg(ebb_state.trace, SPRN_MMCR0, val);
+}
+
+int back_to_back_ebbs(void)
+{
+ struct event event;
+
+ event_init_named(&event, 0x1001e, "cycles");
+ event_leader_ebb_init(&event);
+
+ event.attr.exclude_kernel = 1;
+ event.attr.exclude_hv = 1;
+ event.attr.exclude_idle = 1;
+
+ FAIL_IF(event_open(&event));
+
+ setup_ebb_handler(ebb_callee);
+
+ FAIL_IF(ebb_event_enable(&event));
+
+ sample_period = 5;
+
+ ebb_freeze_pmcs();
+ mtspr(SPRN_PMC1, pmc_sample_period(sample_period));
+ ebb_global_enable();
+ ebb_unfreeze_pmcs();
+
+ while (ebb_state.stats.ebb_count < NUMBER_OF_EBBS)
+ FAIL_IF(core_busy_loop());
+
+ ebb_global_disable();
+ ebb_freeze_pmcs();
+
+ count_pmc(1, sample_period);
+
+ dump_ebb_state();
+
+ event_close(&event);
+
+ FAIL_IF(ebb_state.stats.ebb_count != NUMBER_OF_EBBS);
+
+ return 0;
+}
+
+int main(void)
+{
+ return test_harness(back_to_back_ebbs, "back_to_back_ebbs");
+}
diff --git a/powerpc/pmu/ebb/busy_loop.S b/powerpc/pmu/ebb/busy_loop.S
new file mode 100644
index 0000000..c7e4093
--- /dev/null
+++ b/powerpc/pmu/ebb/busy_loop.S
@@ -0,0 +1,271 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <ppc-asm.h>
+
+ .text
+
+FUNC_START(core_busy_loop)
+ stdu %r1, -168(%r1)
+ std r14, 160(%r1)
+ std r15, 152(%r1)
+ std r16, 144(%r1)
+ std r17, 136(%r1)
+ std r18, 128(%r1)
+ std r19, 120(%r1)
+ std r20, 112(%r1)
+ std r21, 104(%r1)
+ std r22, 96(%r1)
+ std r23, 88(%r1)
+ std r24, 80(%r1)
+ std r25, 72(%r1)
+ std r26, 64(%r1)
+ std r27, 56(%r1)
+ std r28, 48(%r1)
+ std r29, 40(%r1)
+ std r30, 32(%r1)
+ std r31, 24(%r1)
+
+ li r3, 0x3030
+ std r3, -96(%r1)
+ li r4, 0x4040
+ std r4, -104(%r1)
+ li r5, 0x5050
+ std r5, -112(%r1)
+ li r6, 0x6060
+ std r6, -120(%r1)
+ li r7, 0x7070
+ std r7, -128(%r1)
+ li r8, 0x0808
+ std r8, -136(%r1)
+ li r9, 0x0909
+ std r9, -144(%r1)
+ li r10, 0x1010
+ std r10, -152(%r1)
+ li r11, 0x1111
+ std r11, -160(%r1)
+ li r14, 0x1414
+ std r14, -168(%r1)
+ li r15, 0x1515
+ std r15, -176(%r1)
+ li r16, 0x1616
+ std r16, -184(%r1)
+ li r17, 0x1717
+ std r17, -192(%r1)
+ li r18, 0x1818
+ std r18, -200(%r1)
+ li r19, 0x1919
+ std r19, -208(%r1)
+ li r20, 0x2020
+ std r20, -216(%r1)
+ li r21, 0x2121
+ std r21, -224(%r1)
+ li r22, 0x2222
+ std r22, -232(%r1)
+ li r23, 0x2323
+ std r23, -240(%r1)
+ li r24, 0x2424
+ std r24, -248(%r1)
+ li r25, 0x2525
+ std r25, -256(%r1)
+ li r26, 0x2626
+ std r26, -264(%r1)
+ li r27, 0x2727
+ std r27, -272(%r1)
+ li r28, 0x2828
+ std r28, -280(%r1)
+ li r29, 0x2929
+ std r29, -288(%r1)
+ li r30, 0x3030
+ li r31, 0x3131
+
+ li r3, 0
+0: addi r3, r3, 1
+ cmpwi r3, 100
+ blt 0b
+
+ /* Return 1 (fail) unless we get through all the checks */
+ li r3, 1
+
+ /* Check none of our registers have been corrupted */
+ cmpwi r4, 0x4040
+ bne 1f
+ cmpwi r5, 0x5050
+ bne 1f
+ cmpwi r6, 0x6060
+ bne 1f
+ cmpwi r7, 0x7070
+ bne 1f
+ cmpwi r8, 0x0808
+ bne 1f
+ cmpwi r9, 0x0909
+ bne 1f
+ cmpwi r10, 0x1010
+ bne 1f
+ cmpwi r11, 0x1111
+ bne 1f
+ cmpwi r14, 0x1414
+ bne 1f
+ cmpwi r15, 0x1515
+ bne 1f
+ cmpwi r16, 0x1616
+ bne 1f
+ cmpwi r17, 0x1717
+ bne 1f
+ cmpwi r18, 0x1818
+ bne 1f
+ cmpwi r19, 0x1919
+ bne 1f
+ cmpwi r20, 0x2020
+ bne 1f
+ cmpwi r21, 0x2121
+ bne 1f
+ cmpwi r22, 0x2222
+ bne 1f
+ cmpwi r23, 0x2323
+ bne 1f
+ cmpwi r24, 0x2424
+ bne 1f
+ cmpwi r25, 0x2525
+ bne 1f
+ cmpwi r26, 0x2626
+ bne 1f
+ cmpwi r27, 0x2727
+ bne 1f
+ cmpwi r28, 0x2828
+ bne 1f
+ cmpwi r29, 0x2929
+ bne 1f
+ cmpwi r30, 0x3030
+ bne 1f
+ cmpwi r31, 0x3131
+ bne 1f
+
+ /* Load junk into all our registers before we reload them from the stack. */
+ li r3, 0xde
+ li r4, 0xad
+ li r5, 0xbe
+ li r6, 0xef
+ li r7, 0xde
+ li r8, 0xad
+ li r9, 0xbe
+ li r10, 0xef
+ li r11, 0xde
+ li r14, 0xad
+ li r15, 0xbe
+ li r16, 0xef
+ li r17, 0xde
+ li r18, 0xad
+ li r19, 0xbe
+ li r20, 0xef
+ li r21, 0xde
+ li r22, 0xad
+ li r23, 0xbe
+ li r24, 0xef
+ li r25, 0xde
+ li r26, 0xad
+ li r27, 0xbe
+ li r28, 0xef
+ li r29, 0xdd
+
+ ld r3, -96(%r1)
+ cmpwi r3, 0x3030
+ bne 1f
+ ld r4, -104(%r1)
+ cmpwi r4, 0x4040
+ bne 1f
+ ld r5, -112(%r1)
+ cmpwi r5, 0x5050
+ bne 1f
+ ld r6, -120(%r1)
+ cmpwi r6, 0x6060
+ bne 1f
+ ld r7, -128(%r1)
+ cmpwi r7, 0x7070
+ bne 1f
+ ld r8, -136(%r1)
+ cmpwi r8, 0x0808
+ bne 1f
+ ld r9, -144(%r1)
+ cmpwi r9, 0x0909
+ bne 1f
+ ld r10, -152(%r1)
+ cmpwi r10, 0x1010
+ bne 1f
+ ld r11, -160(%r1)
+ cmpwi r11, 0x1111
+ bne 1f
+ ld r14, -168(%r1)
+ cmpwi r14, 0x1414
+ bne 1f
+ ld r15, -176(%r1)
+ cmpwi r15, 0x1515
+ bne 1f
+ ld r16, -184(%r1)
+ cmpwi r16, 0x1616
+ bne 1f
+ ld r17, -192(%r1)
+ cmpwi r17, 0x1717
+ bne 1f
+ ld r18, -200(%r1)
+ cmpwi r18, 0x1818
+ bne 1f
+ ld r19, -208(%r1)
+ cmpwi r19, 0x1919
+ bne 1f
+ ld r20, -216(%r1)
+ cmpwi r20, 0x2020
+ bne 1f
+ ld r21, -224(%r1)
+ cmpwi r21, 0x2121
+ bne 1f
+ ld r22, -232(%r1)
+ cmpwi r22, 0x2222
+ bne 1f
+ ld r23, -240(%r1)
+ cmpwi r23, 0x2323
+ bne 1f
+ ld r24, -248(%r1)
+ cmpwi r24, 0x2424
+ bne 1f
+ ld r25, -256(%r1)
+ cmpwi r25, 0x2525
+ bne 1f
+ ld r26, -264(%r1)
+ cmpwi r26, 0x2626
+ bne 1f
+ ld r27, -272(%r1)
+ cmpwi r27, 0x2727
+ bne 1f
+ ld r28, -280(%r1)
+ cmpwi r28, 0x2828
+ bne 1f
+ ld r29, -288(%r1)
+ cmpwi r29, 0x2929
+ bne 1f
+
+ /* Load 0 (success) to return */
+ li r3, 0
+
+1: ld r14, 160(%r1)
+ ld r15, 152(%r1)
+ ld r16, 144(%r1)
+ ld r17, 136(%r1)
+ ld r18, 128(%r1)
+ ld r19, 120(%r1)
+ ld r20, 112(%r1)
+ ld r21, 104(%r1)
+ ld r22, 96(%r1)
+ ld r23, 88(%r1)
+ ld r24, 80(%r1)
+ ld r25, 72(%r1)
+ ld r26, 64(%r1)
+ ld r27, 56(%r1)
+ ld r28, 48(%r1)
+ ld r29, 40(%r1)
+ ld r30, 32(%r1)
+ ld r31, 24(%r1)
+ addi %r1, %r1, 168
+ blr
diff --git a/powerpc/pmu/ebb/close_clears_pmcc_test.c b/powerpc/pmu/ebb/close_clears_pmcc_test.c
new file mode 100644
index 0000000..0f0423d
--- /dev/null
+++ b/powerpc/pmu/ebb/close_clears_pmcc_test.c
@@ -0,0 +1,59 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <setjmp.h>
+#include <signal.h>
+
+#include "ebb.h"
+
+
+/*
+ * Test that closing the EBB event clears MMCR0_PMCC, preventing further access
+ * by userspace to the PMU hardware.
+ */
+
+int close_clears_pmcc(void)
+{
+ struct event event;
+
+ event_init_named(&event, 0x1001e, "cycles");
+ event_leader_ebb_init(&event);
+
+ FAIL_IF(event_open(&event));
+
+ ebb_enable_pmc_counting(1);
+ setup_ebb_handler(standard_ebb_callee);
+ ebb_global_enable();
+ FAIL_IF(ebb_event_enable(&event));
+
+ mtspr(SPRN_PMC1, pmc_sample_period(sample_period));
+
+ while (ebb_state.stats.ebb_count < 1)
+ FAIL_IF(core_busy_loop());
+
+ ebb_global_disable();
+ event_close(&event);
+
+ FAIL_IF(ebb_state.stats.ebb_count == 0);
+
+ /* The real test is here, do we take a SIGILL when writing PMU regs now
+ * that we have closed the event. We expect that we will. */
+
+ FAIL_IF(catch_sigill(write_pmc1));
+
+ /* We should still be able to read EBB regs though */
+ mfspr(SPRN_EBBHR);
+ mfspr(SPRN_EBBRR);
+ mfspr(SPRN_BESCR);
+
+ return 0;
+}
+
+int main(void)
+{
+ return test_harness(close_clears_pmcc, "close_clears_pmcc");
+}
diff --git a/powerpc/pmu/ebb/cpu_event_pinned_vs_ebb_test.c b/powerpc/pmu/ebb/cpu_event_pinned_vs_ebb_test.c
new file mode 100644
index 0000000..d3ed64d
--- /dev/null
+++ b/powerpc/pmu/ebb/cpu_event_pinned_vs_ebb_test.c
@@ -0,0 +1,93 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdbool.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <unistd.h>
+
+#include "ebb.h"
+
+
+/*
+ * Tests a pinned cpu event vs an EBB - in that order. The pinned cpu event
+ * should remain and the EBB event should fail to enable.
+ */
+
+static int setup_cpu_event(struct event *event, int cpu)
+{
+ event_init_named(event, 0x400FA, "PM_RUN_INST_CMPL");
+
+ event->attr.pinned = 1;
+
+ event->attr.exclude_kernel = 1;
+ event->attr.exclude_hv = 1;
+ event->attr.exclude_idle = 1;
+
+ SKIP_IF(require_paranoia_below(1));
+ FAIL_IF(event_open_with_cpu(event, cpu));
+ FAIL_IF(event_enable(event));
+
+ return 0;
+}
+
+int cpu_event_pinned_vs_ebb(void)
+{
+ union pipe read_pipe, write_pipe;
+ struct event event;
+ int cpu, rc;
+ pid_t pid;
+
+ cpu = pick_online_cpu();
+ FAIL_IF(cpu < 0);
+ FAIL_IF(bind_to_cpu(cpu));
+
+ FAIL_IF(pipe(read_pipe.fds) == -1);
+ FAIL_IF(pipe(write_pipe.fds) == -1);
+
+ pid = fork();
+ if (pid == 0) {
+ /* NB order of pipes looks reversed */
+ exit(ebb_child(write_pipe, read_pipe));
+ }
+
+ /* We setup the cpu event first */
+ rc = setup_cpu_event(&event, cpu);
+ if (rc) {
+ kill_child_and_wait(pid);
+ return rc;
+ }
+
+ /* Signal the child to install its EBB event and wait */
+ if (sync_with_child(read_pipe, write_pipe))
+ /* If it fails, wait for it to exit */
+ goto wait;
+
+ /* Signal the child to run */
+ FAIL_IF(sync_with_child(read_pipe, write_pipe));
+
+wait:
+ /* We expect it to fail to read the event */
+ FAIL_IF(wait_for_child(pid) != 2);
+
+ FAIL_IF(event_disable(&event));
+ FAIL_IF(event_read(&event));
+
+ event_report(&event);
+
+ /* The cpu event should have run */
+ FAIL_IF(event.result.value == 0);
+ FAIL_IF(event.result.enabled != event.result.running);
+
+ return 0;
+}
+
+int main(void)
+{
+ return test_harness(cpu_event_pinned_vs_ebb, "cpu_event_pinned_vs_ebb");
+}
diff --git a/powerpc/pmu/ebb/cpu_event_vs_ebb_test.c b/powerpc/pmu/ebb/cpu_event_vs_ebb_test.c
new file mode 100644
index 0000000..8b972c2
--- /dev/null
+++ b/powerpc/pmu/ebb/cpu_event_vs_ebb_test.c
@@ -0,0 +1,89 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdbool.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <unistd.h>
+
+#include "ebb.h"
+
+
+/*
+ * Tests a cpu event vs an EBB - in that order. The EBB should force the cpu
+ * event off the PMU.
+ */
+
+static int setup_cpu_event(struct event *event, int cpu)
+{
+ event_init_named(event, 0x400FA, "PM_RUN_INST_CMPL");
+
+ event->attr.exclude_kernel = 1;
+ event->attr.exclude_hv = 1;
+ event->attr.exclude_idle = 1;
+
+ SKIP_IF(require_paranoia_below(1));
+ FAIL_IF(event_open_with_cpu(event, cpu));
+ FAIL_IF(event_enable(event));
+
+ return 0;
+}
+
+int cpu_event_vs_ebb(void)
+{
+ union pipe read_pipe, write_pipe;
+ struct event event;
+ int cpu, rc;
+ pid_t pid;
+
+ cpu = pick_online_cpu();
+ FAIL_IF(cpu < 0);
+ FAIL_IF(bind_to_cpu(cpu));
+
+ FAIL_IF(pipe(read_pipe.fds) == -1);
+ FAIL_IF(pipe(write_pipe.fds) == -1);
+
+ pid = fork();
+ if (pid == 0) {
+ /* NB order of pipes looks reversed */
+ exit(ebb_child(write_pipe, read_pipe));
+ }
+
+ /* We setup the cpu event first */
+ rc = setup_cpu_event(&event, cpu);
+ if (rc) {
+ kill_child_and_wait(pid);
+ return rc;
+ }
+
+ /* Signal the child to install its EBB event and wait */
+ if (sync_with_child(read_pipe, write_pipe))
+ /* If it fails, wait for it to exit */
+ goto wait;
+
+ /* Signal the child to run */
+ FAIL_IF(sync_with_child(read_pipe, write_pipe));
+
+wait:
+ /* We expect the child to succeed */
+ FAIL_IF(wait_for_child(pid));
+
+ FAIL_IF(event_disable(&event));
+ FAIL_IF(event_read(&event));
+
+ event_report(&event);
+
+ /* The cpu event may have run */
+
+ return 0;
+}
+
+int main(void)
+{
+ return test_harness(cpu_event_vs_ebb, "cpu_event_vs_ebb");
+}
diff --git a/powerpc/pmu/ebb/cycles_test.c b/powerpc/pmu/ebb/cycles_test.c
new file mode 100644
index 0000000..8590fc1
--- /dev/null
+++ b/powerpc/pmu/ebb/cycles_test.c
@@ -0,0 +1,58 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+
+#include "ebb.h"
+
+
+/*
+ * Basic test that counts user cycles and takes EBBs.
+ */
+int cycles(void)
+{
+ struct event event;
+
+ event_init_named(&event, 0x1001e, "cycles");
+ event_leader_ebb_init(&event);
+
+ event.attr.exclude_kernel = 1;
+ event.attr.exclude_hv = 1;
+ event.attr.exclude_idle = 1;
+
+ FAIL_IF(event_open(&event));
+
+ ebb_enable_pmc_counting(1);
+ setup_ebb_handler(standard_ebb_callee);
+ ebb_global_enable();
+ FAIL_IF(ebb_event_enable(&event));
+
+ mtspr(SPRN_PMC1, pmc_sample_period(sample_period));
+
+ while (ebb_state.stats.ebb_count < 10) {
+ FAIL_IF(core_busy_loop());
+ FAIL_IF(ebb_check_mmcr0());
+ }
+
+ ebb_global_disable();
+ ebb_freeze_pmcs();
+
+ count_pmc(1, sample_period);
+
+ dump_ebb_state();
+
+ event_close(&event);
+
+ FAIL_IF(ebb_state.stats.ebb_count == 0);
+ FAIL_IF(!ebb_check_count(1, sample_period, 100));
+
+ return 0;
+}
+
+int main(void)
+{
+ return test_harness(cycles, "cycles");
+}
diff --git a/powerpc/pmu/ebb/cycles_with_freeze_test.c b/powerpc/pmu/ebb/cycles_with_freeze_test.c
new file mode 100644
index 0000000..754b3f2
--- /dev/null
+++ b/powerpc/pmu/ebb/cycles_with_freeze_test.c
@@ -0,0 +1,117 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdbool.h>
+
+#include "ebb.h"
+
+
+/*
+ * Test of counting cycles while using MMCR0_FC (freeze counters) to only count
+ * parts of the code. This is complicated by the fact that FC is set by the
+ * hardware when the event overflows. We may take the EBB after we have set FC,
+ * so we have to be careful about whether we clear FC at the end of the EBB
+ * handler or not.
+ */
+
+static bool counters_frozen = false;
+static int ebbs_while_frozen = 0;
+
+static void ebb_callee(void)
+{
+ uint64_t mask, val;
+
+ mask = MMCR0_PMAO | MMCR0_FC;
+
+ val = mfspr(SPRN_BESCR);
+ if (!(val & BESCR_PMEO)) {
+ ebb_state.stats.spurious++;
+ goto out;
+ }
+
+ ebb_state.stats.ebb_count++;
+ trace_log_counter(ebb_state.trace, ebb_state.stats.ebb_count);
+
+ val = mfspr(SPRN_MMCR0);
+ trace_log_reg(ebb_state.trace, SPRN_MMCR0, val);
+
+ if (counters_frozen) {
+ trace_log_string(ebb_state.trace, "frozen");
+ ebbs_while_frozen++;
+ mask &= ~MMCR0_FC;
+ }
+
+ count_pmc(1, sample_period);
+out:
+ reset_ebb_with_clear_mask(mask);
+}
+
+int cycles_with_freeze(void)
+{
+ struct event event;
+ uint64_t val;
+ bool fc_cleared;
+
+ event_init_named(&event, 0x1001e, "cycles");
+ event_leader_ebb_init(&event);
+
+ event.attr.exclude_kernel = 1;
+ event.attr.exclude_hv = 1;
+ event.attr.exclude_idle = 1;
+
+ FAIL_IF(event_open(&event));
+
+ setup_ebb_handler(ebb_callee);
+ ebb_global_enable();
+ FAIL_IF(ebb_event_enable(&event));
+
+ mtspr(SPRN_PMC1, pmc_sample_period(sample_period));
+
+ fc_cleared = false;
+
+ /* Make sure we loop until we take at least one EBB */
+ while ((ebb_state.stats.ebb_count < 20 && !fc_cleared) ||
+ ebb_state.stats.ebb_count < 1)
+ {
+ counters_frozen = false;
+ mb();
+ mtspr(SPRN_MMCR0, mfspr(SPRN_MMCR0) & ~MMCR0_FC);
+
+ FAIL_IF(core_busy_loop());
+
+ counters_frozen = true;
+ mb();
+ mtspr(SPRN_MMCR0, mfspr(SPRN_MMCR0) | MMCR0_FC);
+
+ val = mfspr(SPRN_MMCR0);
+ if (! (val & MMCR0_FC)) {
+ printf("Outside of loop, FC NOT set MMCR0 0x%lx\n", val);
+ fc_cleared = true;
+ }
+ }
+
+ ebb_global_disable();
+ ebb_freeze_pmcs();
+
+ count_pmc(1, sample_period);
+
+ dump_ebb_state();
+
+ printf("EBBs while frozen %d\n", ebbs_while_frozen);
+
+ event_close(&event);
+
+ FAIL_IF(ebb_state.stats.ebb_count == 0);
+ FAIL_IF(fc_cleared);
+
+ return 0;
+}
+
+int main(void)
+{
+ return test_harness(cycles_with_freeze, "cycles_with_freeze");
+}
diff --git a/powerpc/pmu/ebb/cycles_with_mmcr2_test.c b/powerpc/pmu/ebb/cycles_with_mmcr2_test.c
new file mode 100644
index 0000000..d43029b
--- /dev/null
+++ b/powerpc/pmu/ebb/cycles_with_mmcr2_test.c
@@ -0,0 +1,91 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdbool.h>
+
+#include "ebb.h"
+
+
+/*
+ * Test of counting cycles while manipulating the user accessible bits in MMCR2.
+ */
+
+/* We use two values because the first freezes PMC1 and so we would get no EBBs */
+#define MMCR2_EXPECTED_1 0x4020100804020000UL /* (FC1P|FC2P|FC3P|FC4P|FC5P|FC6P) */
+#define MMCR2_EXPECTED_2 0x0020100804020000UL /* ( FC2P|FC3P|FC4P|FC5P|FC6P) */
+
+
+int cycles_with_mmcr2(void)
+{
+ struct event event;
+ uint64_t val, expected[2], actual;
+ int i;
+ bool bad_mmcr2;
+
+ event_init_named(&event, 0x1001e, "cycles");
+ event_leader_ebb_init(&event);
+
+ event.attr.exclude_kernel = 1;
+ event.attr.exclude_hv = 1;
+ event.attr.exclude_idle = 1;
+
+ FAIL_IF(event_open(&event));
+
+ ebb_enable_pmc_counting(1);
+ setup_ebb_handler(standard_ebb_callee);
+ ebb_global_enable();
+
+ FAIL_IF(ebb_event_enable(&event));
+
+ mtspr(SPRN_PMC1, pmc_sample_period(sample_period));
+
+ /* XXX Set of MMCR2 must be after enable */
+ expected[0] = MMCR2_EXPECTED_1;
+ expected[1] = MMCR2_EXPECTED_2;
+ i = 0;
+ bad_mmcr2 = false;
+
+ /* Make sure we loop until we take at least one EBB */
+ while ((ebb_state.stats.ebb_count < 20 && !bad_mmcr2) ||
+ ebb_state.stats.ebb_count < 1)
+ {
+ mtspr(SPRN_MMCR2, expected[i % 2]);
+
+ FAIL_IF(core_busy_loop());
+
+ val = mfspr(SPRN_MMCR2);
+ if (val != expected[i % 2]) {
+ bad_mmcr2 = true;
+ actual = val;
+ }
+
+ i++;
+ }
+
+ ebb_global_disable();
+ ebb_freeze_pmcs();
+
+ count_pmc(1, sample_period);
+
+ dump_ebb_state();
+
+ event_close(&event);
+
+ FAIL_IF(ebb_state.stats.ebb_count == 0);
+
+ if (bad_mmcr2)
+ printf("Bad MMCR2 value seen is 0x%lx\n", actual);
+
+ FAIL_IF(bad_mmcr2);
+
+ return 0;
+}
+
+int main(void)
+{
+ return test_harness(cycles_with_mmcr2, "cycles_with_mmcr2");
+}
diff --git a/powerpc/pmu/ebb/ebb.c b/powerpc/pmu/ebb/ebb.c
new file mode 100644
index 0000000..d7a72ce
--- /dev/null
+++ b/powerpc/pmu/ebb/ebb.c
@@ -0,0 +1,478 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#define _GNU_SOURCE /* For CPU_ZERO etc. */
+
+#include <sched.h>
+#include <sys/wait.h>
+#include <setjmp.h>
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/ioctl.h>
+
+#include "trace.h"
+#include "reg.h"
+#include "ebb.h"
+
+
+void (*ebb_user_func)(void);
+
+void ebb_hook(void)
+{
+ if (ebb_user_func)
+ ebb_user_func();
+}
+
+struct ebb_state ebb_state;
+
+u64 sample_period = 0x40000000ull;
+
+void reset_ebb_with_clear_mask(unsigned long mmcr0_clear_mask)
+{
+ u64 val;
+
+ /* 2) clear MMCR0[PMAO] - docs say BESCR[PMEO] should do this */
+ /* 3) set MMCR0[PMAE] - docs say BESCR[PME] should do this */
+ val = mfspr(SPRN_MMCR0);
+ mtspr(SPRN_MMCR0, (val & ~mmcr0_clear_mask) | MMCR0_PMAE);
+
+ /* 4) clear BESCR[PMEO] */
+ mtspr(SPRN_BESCRR, BESCR_PMEO);
+
+ /* 5) set BESCR[PME] */
+ mtspr(SPRN_BESCRS, BESCR_PME);
+
+ /* 6) rfebb 1 - done in our caller */
+}
+
+void reset_ebb(void)
+{
+ reset_ebb_with_clear_mask(MMCR0_PMAO | MMCR0_FC);
+}
+
+/* Called outside of the EBB handler to check MMCR0 is sane */
+int ebb_check_mmcr0(void)
+{
+ u64 val;
+
+ val = mfspr(SPRN_MMCR0);
+ if ((val & (MMCR0_FC | MMCR0_PMAO)) == MMCR0_FC) {
+ /* It's OK if we see FC & PMAO, but not FC by itself */
+ printf("Outside of loop, only FC set 0x%llx\n", val);
+ return 1;
+ }
+
+ return 0;
+}
+
+bool ebb_check_count(int pmc, u64 sample_period, int fudge)
+{
+ u64 count, upper, lower;
+
+ count = ebb_state.stats.pmc_count[PMC_INDEX(pmc)];
+
+ lower = ebb_state.stats.ebb_count * (sample_period - fudge);
+
+ if (count < lower) {
+ printf("PMC%d count (0x%llx) below lower limit 0x%llx (-0x%llx)\n",
+ pmc, count, lower, lower - count);
+ return false;
+ }
+
+ upper = ebb_state.stats.ebb_count * (sample_period + fudge);
+
+ if (count > upper) {
+ printf("PMC%d count (0x%llx) above upper limit 0x%llx (+0x%llx)\n",
+ pmc, count, upper, count - upper);
+ return false;
+ }
+
+ printf("PMC%d count (0x%llx) is between 0x%llx and 0x%llx delta +0x%llx/-0x%llx\n",
+ pmc, count, lower, upper, count - lower, upper - count);
+
+ return true;
+}
+
+void standard_ebb_callee(void)
+{
+ int found, i;
+ u64 val;
+
+ val = mfspr(SPRN_BESCR);
+ if (!(val & BESCR_PMEO)) {
+ ebb_state.stats.spurious++;
+ goto out;
+ }
+
+ ebb_state.stats.ebb_count++;
+ trace_log_counter(ebb_state.trace, ebb_state.stats.ebb_count);
+
+ val = mfspr(SPRN_MMCR0);
+ trace_log_reg(ebb_state.trace, SPRN_MMCR0, val);
+
+ found = 0;
+ for (i = 1; i <= 6; i++) {
+ if (ebb_state.pmc_enable[PMC_INDEX(i)])
+ found += count_pmc(i, sample_period);
+ }
+
+ if (!found)
+ ebb_state.stats.no_overflow++;
+
+out:
+ reset_ebb();
+}
+
+extern void ebb_handler(void);
+
+void setup_ebb_handler(void (*callee)(void))
+{
+ u64 entry;
+
+#if defined(_CALL_ELF) && _CALL_ELF == 2
+ entry = (u64)ebb_handler;
+#else
+ struct opd
+ {
+ u64 entry;
+ u64 toc;
+ } *opd;
+
+ opd = (struct opd *)ebb_handler;
+ entry = opd->entry;
+#endif
+ printf("EBB Handler is at %#llx\n", entry);
+
+ ebb_user_func = callee;
+
+ /* Ensure ebb_user_func is set before we set the handler */
+ mb();
+ mtspr(SPRN_EBBHR, entry);
+
+ /* Make sure the handler is set before we return */
+ mb();
+}
+
+void clear_ebb_stats(void)
+{
+ memset(&ebb_state.stats, 0, sizeof(ebb_state.stats));
+}
+
+void dump_summary_ebb_state(void)
+{
+ printf("ebb_state:\n" \
+ " ebb_count = %d\n" \
+ " spurious = %d\n" \
+ " negative = %d\n" \
+ " no_overflow = %d\n" \
+ " pmc[1] count = 0x%llx\n" \
+ " pmc[2] count = 0x%llx\n" \
+ " pmc[3] count = 0x%llx\n" \
+ " pmc[4] count = 0x%llx\n" \
+ " pmc[5] count = 0x%llx\n" \
+ " pmc[6] count = 0x%llx\n",
+ ebb_state.stats.ebb_count, ebb_state.stats.spurious,
+ ebb_state.stats.negative, ebb_state.stats.no_overflow,
+ ebb_state.stats.pmc_count[0], ebb_state.stats.pmc_count[1],
+ ebb_state.stats.pmc_count[2], ebb_state.stats.pmc_count[3],
+ ebb_state.stats.pmc_count[4], ebb_state.stats.pmc_count[5]);
+}
+
+static char *decode_mmcr0(u32 value)
+{
+ static char buf[16];
+
+ buf[0] = '\0';
+
+ if (value & (1 << 31))
+ strcat(buf, "FC ");
+ if (value & (1 << 26))
+ strcat(buf, "PMAE ");
+ if (value & (1 << 7))
+ strcat(buf, "PMAO ");
+
+ return buf;
+}
+
+static char *decode_bescr(u64 value)
+{
+ static char buf[16];
+
+ buf[0] = '\0';
+
+ if (value & (1ull << 63))
+ strcat(buf, "GE ");
+ if (value & (1ull << 32))
+ strcat(buf, "PMAE ");
+ if (value & 1)
+ strcat(buf, "PMAO ");
+
+ return buf;
+}
+
+void dump_ebb_hw_state(void)
+{
+ u64 bescr;
+ u32 mmcr0;
+
+ mmcr0 = mfspr(SPRN_MMCR0);
+ bescr = mfspr(SPRN_BESCR);
+
+ printf("HW state:\n" \
+ "MMCR0 0x%016x %s\n" \
+ "MMCR2 0x%016lx\n" \
+ "EBBHR 0x%016lx\n" \
+ "BESCR 0x%016llx %s\n" \
+ "PMC1 0x%016lx\n" \
+ "PMC2 0x%016lx\n" \
+ "PMC3 0x%016lx\n" \
+ "PMC4 0x%016lx\n" \
+ "PMC5 0x%016lx\n" \
+ "PMC6 0x%016lx\n" \
+ "SIAR 0x%016lx\n",
+ mmcr0, decode_mmcr0(mmcr0), mfspr(SPRN_MMCR2),
+ mfspr(SPRN_EBBHR), bescr, decode_bescr(bescr),
+ mfspr(SPRN_PMC1), mfspr(SPRN_PMC2), mfspr(SPRN_PMC3),
+ mfspr(SPRN_PMC4), mfspr(SPRN_PMC5), mfspr(SPRN_PMC6),
+ mfspr(SPRN_SIAR));
+}
+
+void dump_ebb_state(void)
+{
+ dump_summary_ebb_state();
+
+ dump_ebb_hw_state();
+
+ trace_buffer_print(ebb_state.trace);
+}
+
+int count_pmc(int pmc, uint32_t sample_period)
+{
+ uint32_t start_value;
+ u64 val;
+
+ /* 0) Read PMC */
+ start_value = pmc_sample_period(sample_period);
+
+ val = read_pmc(pmc);
+ if (val < start_value)
+ ebb_state.stats.negative++;
+ else
+ ebb_state.stats.pmc_count[PMC_INDEX(pmc)] += val - start_value;
+
+ trace_log_reg(ebb_state.trace, SPRN_PMC1 + pmc - 1, val);
+
+ /* 1) Reset PMC */
+ write_pmc(pmc, start_value);
+
+ /* Report if we overflowed */
+ return val >= COUNTER_OVERFLOW;
+}
+
+int ebb_event_enable(struct event *e)
+{
+ int rc;
+
+ /* Ensure any SPR writes are ordered vs us */
+ mb();
+
+ rc = ioctl(e->fd, PERF_EVENT_IOC_ENABLE);
+ if (rc)
+ return rc;
+
+ rc = event_read(e);
+
+ /* Ditto */
+ mb();
+
+ return rc;
+}
+
+void ebb_freeze_pmcs(void)
+{
+ mtspr(SPRN_MMCR0, mfspr(SPRN_MMCR0) | MMCR0_FC);
+ mb();
+}
+
+void ebb_unfreeze_pmcs(void)
+{
+ /* Unfreeze counters */
+ mtspr(SPRN_MMCR0, mfspr(SPRN_MMCR0) & ~MMCR0_FC);
+ mb();
+}
+
+void ebb_global_enable(void)
+{
+ /* Enable EBBs globally and PMU EBBs */
+ mtspr(SPRN_BESCR, 0x8000000100000000ull);
+ mb();
+}
+
+void ebb_global_disable(void)
+{
+ /* Disable EBBs & freeze counters, events are still scheduled */
+ mtspr(SPRN_BESCRR, BESCR_PME);
+ mb();
+}
+
+void event_ebb_init(struct event *e)
+{
+ e->attr.config |= (1ull << 63);
+}
+
+void event_bhrb_init(struct event *e, unsigned ifm)
+{
+ e->attr.config |= (1ull << 62) | ((u64)ifm << 60);
+}
+
+void event_leader_ebb_init(struct event *e)
+{
+ event_ebb_init(e);
+
+ e->attr.exclusive = 1;
+ e->attr.pinned = 1;
+}
+
+int ebb_child(union pipe read_pipe, union pipe write_pipe)
+{
+ struct event event;
+ uint64_t val;
+
+ FAIL_IF(wait_for_parent(read_pipe));
+
+ event_init_named(&event, 0x1001e, "cycles");
+ event_leader_ebb_init(&event);
+
+ event.attr.exclude_kernel = 1;
+ event.attr.exclude_hv = 1;
+ event.attr.exclude_idle = 1;
+
+ FAIL_IF(event_open(&event));
+
+ ebb_enable_pmc_counting(1);
+ setup_ebb_handler(standard_ebb_callee);
+ ebb_global_enable();
+
+ FAIL_IF(event_enable(&event));
+
+ if (event_read(&event)) {
+ /*
+ * Some tests expect to fail here, so don't report an error on
+ * this line, and return a distinguisable error code. Tell the
+ * parent an error happened.
+ */
+ notify_parent_of_error(write_pipe);
+ return 2;
+ }
+
+ mtspr(SPRN_PMC1, pmc_sample_period(sample_period));
+
+ FAIL_IF(notify_parent(write_pipe));
+ FAIL_IF(wait_for_parent(read_pipe));
+ FAIL_IF(notify_parent(write_pipe));
+
+ while (ebb_state.stats.ebb_count < 20) {
+ FAIL_IF(core_busy_loop());
+
+ /* To try and hit SIGILL case */
+ val = mfspr(SPRN_MMCRA);
+ val |= mfspr(SPRN_MMCR2);
+ val |= mfspr(SPRN_MMCR0);
+ }
+
+ ebb_global_disable();
+ ebb_freeze_pmcs();
+
+ count_pmc(1, sample_period);
+
+ dump_ebb_state();
+
+ event_close(&event);
+
+ FAIL_IF(ebb_state.stats.ebb_count == 0);
+
+ return 0;
+}
+
+static jmp_buf setjmp_env;
+
+static void sigill_handler(int signal)
+{
+ printf("Took sigill\n");
+ longjmp(setjmp_env, 1);
+}
+
+static struct sigaction sigill_action = {
+ .sa_handler = sigill_handler,
+};
+
+int catch_sigill(void (*func)(void))
+{
+ if (sigaction(SIGILL, &sigill_action, NULL)) {
+ perror("sigaction");
+ return 1;
+ }
+
+ if (setjmp(setjmp_env) == 0) {
+ func();
+ return 1;
+ }
+
+ return 0;
+}
+
+void write_pmc1(void)
+{
+ mtspr(SPRN_PMC1, 0);
+}
+
+void write_pmc(int pmc, u64 value)
+{
+ switch (pmc) {
+ case 1: mtspr(SPRN_PMC1, value); break;
+ case 2: mtspr(SPRN_PMC2, value); break;
+ case 3: mtspr(SPRN_PMC3, value); break;
+ case 4: mtspr(SPRN_PMC4, value); break;
+ case 5: mtspr(SPRN_PMC5, value); break;
+ case 6: mtspr(SPRN_PMC6, value); break;
+ }
+}
+
+u64 read_pmc(int pmc)
+{
+ switch (pmc) {
+ case 1: return mfspr(SPRN_PMC1);
+ case 2: return mfspr(SPRN_PMC2);
+ case 3: return mfspr(SPRN_PMC3);
+ case 4: return mfspr(SPRN_PMC4);
+ case 5: return mfspr(SPRN_PMC5);
+ case 6: return mfspr(SPRN_PMC6);
+ }
+
+ return 0;
+}
+
+static void term_handler(int signal)
+{
+ dump_summary_ebb_state();
+ dump_ebb_hw_state();
+ abort();
+}
+
+struct sigaction term_action = {
+ .sa_handler = term_handler,
+};
+
+static void __attribute__((constructor)) ebb_init(void)
+{
+ clear_ebb_stats();
+
+ if (sigaction(SIGTERM, &term_action, NULL))
+ perror("sigaction");
+
+ ebb_state.trace = trace_buffer_allocate(1 * 1024 * 1024);
+}
diff --git a/powerpc/pmu/ebb/ebb.h b/powerpc/pmu/ebb/ebb.h
new file mode 100644
index 0000000..e44eee5
--- /dev/null
+++ b/powerpc/pmu/ebb/ebb.h
@@ -0,0 +1,77 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#ifndef _SELFTESTS_POWERPC_PMU_EBB_EBB_H
+#define _SELFTESTS_POWERPC_PMU_EBB_EBB_H
+
+#include "../event.h"
+#include "../lib.h"
+#include "trace.h"
+#include "reg.h"
+
+#define PMC_INDEX(pmc) ((pmc)-1)
+
+#define NUM_PMC_VALUES 128
+
+struct ebb_state
+{
+ struct {
+ u64 pmc_count[6];
+ volatile int ebb_count;
+ int spurious;
+ int negative;
+ int no_overflow;
+ } stats;
+
+ bool pmc_enable[6];
+ struct trace_buffer *trace;
+};
+
+extern struct ebb_state ebb_state;
+
+#define COUNTER_OVERFLOW 0x80000000ull
+
+static inline uint32_t pmc_sample_period(uint32_t value)
+{
+ return COUNTER_OVERFLOW - value;
+}
+
+static inline void ebb_enable_pmc_counting(int pmc)
+{
+ ebb_state.pmc_enable[PMC_INDEX(pmc)] = true;
+}
+
+bool ebb_check_count(int pmc, u64 sample_period, int fudge);
+void event_leader_ebb_init(struct event *e);
+void event_ebb_init(struct event *e);
+void event_bhrb_init(struct event *e, unsigned ifm);
+void setup_ebb_handler(void (*callee)(void));
+void standard_ebb_callee(void);
+int ebb_event_enable(struct event *e);
+void ebb_global_enable(void);
+void ebb_global_disable(void);
+void ebb_freeze_pmcs(void);
+void ebb_unfreeze_pmcs(void);
+void event_ebb_init(struct event *e);
+void event_leader_ebb_init(struct event *e);
+int count_pmc(int pmc, uint32_t sample_period);
+void dump_ebb_state(void);
+void dump_summary_ebb_state(void);
+void dump_ebb_hw_state(void);
+void clear_ebb_stats(void);
+void write_pmc(int pmc, u64 value);
+u64 read_pmc(int pmc);
+void reset_ebb_with_clear_mask(unsigned long mmcr0_clear_mask);
+void reset_ebb(void);
+int ebb_check_mmcr0(void);
+
+extern u64 sample_period;
+
+int core_busy_loop(void);
+int ebb_child(union pipe read_pipe, union pipe write_pipe);
+int catch_sigill(void (*func)(void));
+void write_pmc1(void);
+
+#endif /* _SELFTESTS_POWERPC_PMU_EBB_EBB_H */
diff --git a/powerpc/pmu/ebb/ebb_handler.S b/powerpc/pmu/ebb/ebb_handler.S
new file mode 100644
index 0000000..14274ea
--- /dev/null
+++ b/powerpc/pmu/ebb/ebb_handler.S
@@ -0,0 +1,365 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <ppc-asm.h>
+#include "reg.h"
+
+
+/* ppc-asm.h defines most of the reg aliases, but not r1/r2. */
+#define r1 1
+#define r2 2
+
+#define RFEBB .long 0x4c000924
+
+/* Stack layout:
+ *
+ * ^
+ * User stack |
+ * Back chain ------+ <- r1 <-------+
+ * ... |
+ * Red zone / ABI Gap |
+ * ... |
+ * vr63 <+ |
+ * vr0 | |
+ * VSCR | |
+ * FSCR | |
+ * r31 | Save area |
+ * r0 | |
+ * XER | |
+ * CTR | |
+ * LR | |
+ * CCR <+ |
+ * ... <+ |
+ * LR | Caller frame |
+ * CCR | |
+ * Back chain <+ <- updated r1 --------+
+ *
+ */
+
+#if defined(_CALL_ELF) && _CALL_ELF == 2
+#define ABIGAP 512
+#else
+#define ABIGAP 288
+#endif
+
+#define NR_GPR 32
+#define NR_SPR 6
+#define NR_VSR 64
+
+#define SAVE_AREA ((NR_GPR + NR_SPR) * 8 + (NR_VSR * 16))
+#define CALLER_FRAME 112
+
+#define STACK_FRAME (ABIGAP + SAVE_AREA + CALLER_FRAME)
+
+#define CCR_SAVE (CALLER_FRAME)
+#define LR_SAVE (CCR_SAVE + 8)
+#define CTR_SAVE (LR_SAVE + 8)
+#define XER_SAVE (CTR_SAVE + 8)
+#define GPR_SAVE(n) (XER_SAVE + 8 + (8 * n))
+#define FSCR_SAVE (GPR_SAVE(31) + 8)
+#define VSCR_SAVE (FSCR_SAVE + 8)
+#define VSR_SAVE(n) (VSCR_SAVE + 8 + (16 * n))
+
+#define SAVE_GPR(n) std n,GPR_SAVE(n)(r1)
+#define REST_GPR(n) ld n,GPR_SAVE(n)(r1)
+#define TRASH_GPR(n) lis n,0xaaaa
+
+#define SAVE_VSR(n, b) li b, VSR_SAVE(n); stxvd2x n,b,r1
+#define LOAD_VSR(n, b) li b, VSR_SAVE(n); lxvd2x n,b,r1
+
+#define LOAD_REG_IMMEDIATE(reg,expr) \
+ lis reg,(expr)@highest; \
+ ori reg,reg,(expr)@higher; \
+ rldicr reg,reg,32,31; \
+ oris reg,reg,(expr)@h; \
+ ori reg,reg,(expr)@l;
+
+
+#if defined(_CALL_ELF) && _CALL_ELF == 2
+#define ENTRY_POINT(name) \
+ .type FUNC_NAME(name),@function; \
+ .globl FUNC_NAME(name); \
+ FUNC_NAME(name):
+
+#define RESTORE_TOC(name) \
+ /* Restore our TOC pointer using our entry point */ \
+ LOAD_REG_IMMEDIATE(r12, name) \
+0: addis r2,r12,(.TOC.-0b)@ha; \
+ addi r2,r2,(.TOC.-0b)@l;
+
+#else
+#define ENTRY_POINT(name) FUNC_START(name)
+#define RESTORE_TOC(name) \
+ /* Restore our TOC pointer via our opd entry */ \
+ LOAD_REG_IMMEDIATE(r2, name) \
+ ld r2,8(r2);
+#endif
+
+ .text
+
+ENTRY_POINT(ebb_handler)
+ stdu r1,-STACK_FRAME(r1)
+ SAVE_GPR(0)
+ mflr r0
+ std r0,LR_SAVE(r1)
+ mfcr r0
+ std r0,CCR_SAVE(r1)
+ mfctr r0
+ std r0,CTR_SAVE(r1)
+ mfxer r0
+ std r0,XER_SAVE(r1)
+ SAVE_GPR(2)
+ SAVE_GPR(3)
+ SAVE_GPR(4)
+ SAVE_GPR(5)
+ SAVE_GPR(6)
+ SAVE_GPR(7)
+ SAVE_GPR(8)
+ SAVE_GPR(9)
+ SAVE_GPR(10)
+ SAVE_GPR(11)
+ SAVE_GPR(12)
+ SAVE_GPR(13)
+ SAVE_GPR(14)
+ SAVE_GPR(15)
+ SAVE_GPR(16)
+ SAVE_GPR(17)
+ SAVE_GPR(18)
+ SAVE_GPR(19)
+ SAVE_GPR(20)
+ SAVE_GPR(21)
+ SAVE_GPR(22)
+ SAVE_GPR(23)
+ SAVE_GPR(24)
+ SAVE_GPR(25)
+ SAVE_GPR(26)
+ SAVE_GPR(27)
+ SAVE_GPR(28)
+ SAVE_GPR(29)
+ SAVE_GPR(30)
+ SAVE_GPR(31)
+ SAVE_VSR(0, r3)
+ mffs f0
+ stfd f0, FSCR_SAVE(r1)
+ mfvscr f0
+ stfd f0, VSCR_SAVE(r1)
+ SAVE_VSR(1, r3)
+ SAVE_VSR(2, r3)
+ SAVE_VSR(3, r3)
+ SAVE_VSR(4, r3)
+ SAVE_VSR(5, r3)
+ SAVE_VSR(6, r3)
+ SAVE_VSR(7, r3)
+ SAVE_VSR(8, r3)
+ SAVE_VSR(9, r3)
+ SAVE_VSR(10, r3)
+ SAVE_VSR(11, r3)
+ SAVE_VSR(12, r3)
+ SAVE_VSR(13, r3)
+ SAVE_VSR(14, r3)
+ SAVE_VSR(15, r3)
+ SAVE_VSR(16, r3)
+ SAVE_VSR(17, r3)
+ SAVE_VSR(18, r3)
+ SAVE_VSR(19, r3)
+ SAVE_VSR(20, r3)
+ SAVE_VSR(21, r3)
+ SAVE_VSR(22, r3)
+ SAVE_VSR(23, r3)
+ SAVE_VSR(24, r3)
+ SAVE_VSR(25, r3)
+ SAVE_VSR(26, r3)
+ SAVE_VSR(27, r3)
+ SAVE_VSR(28, r3)
+ SAVE_VSR(29, r3)
+ SAVE_VSR(30, r3)
+ SAVE_VSR(31, r3)
+ SAVE_VSR(32, r3)
+ SAVE_VSR(33, r3)
+ SAVE_VSR(34, r3)
+ SAVE_VSR(35, r3)
+ SAVE_VSR(36, r3)
+ SAVE_VSR(37, r3)
+ SAVE_VSR(38, r3)
+ SAVE_VSR(39, r3)
+ SAVE_VSR(40, r3)
+ SAVE_VSR(41, r3)
+ SAVE_VSR(42, r3)
+ SAVE_VSR(43, r3)
+ SAVE_VSR(44, r3)
+ SAVE_VSR(45, r3)
+ SAVE_VSR(46, r3)
+ SAVE_VSR(47, r3)
+ SAVE_VSR(48, r3)
+ SAVE_VSR(49, r3)
+ SAVE_VSR(50, r3)
+ SAVE_VSR(51, r3)
+ SAVE_VSR(52, r3)
+ SAVE_VSR(53, r3)
+ SAVE_VSR(54, r3)
+ SAVE_VSR(55, r3)
+ SAVE_VSR(56, r3)
+ SAVE_VSR(57, r3)
+ SAVE_VSR(58, r3)
+ SAVE_VSR(59, r3)
+ SAVE_VSR(60, r3)
+ SAVE_VSR(61, r3)
+ SAVE_VSR(62, r3)
+ SAVE_VSR(63, r3)
+
+ TRASH_GPR(2)
+ TRASH_GPR(3)
+ TRASH_GPR(4)
+ TRASH_GPR(5)
+ TRASH_GPR(6)
+ TRASH_GPR(7)
+ TRASH_GPR(8)
+ TRASH_GPR(9)
+ TRASH_GPR(10)
+ TRASH_GPR(11)
+ TRASH_GPR(12)
+ TRASH_GPR(14)
+ TRASH_GPR(15)
+ TRASH_GPR(16)
+ TRASH_GPR(17)
+ TRASH_GPR(18)
+ TRASH_GPR(19)
+ TRASH_GPR(20)
+ TRASH_GPR(21)
+ TRASH_GPR(22)
+ TRASH_GPR(23)
+ TRASH_GPR(24)
+ TRASH_GPR(25)
+ TRASH_GPR(26)
+ TRASH_GPR(27)
+ TRASH_GPR(28)
+ TRASH_GPR(29)
+ TRASH_GPR(30)
+ TRASH_GPR(31)
+
+ RESTORE_TOC(ebb_handler)
+
+ /*
+ * r13 is our TLS pointer. We leave whatever value was in there when the
+ * EBB fired. That seems to be OK because once set the TLS pointer is not
+ * changed - but presumably that could change in future.
+ */
+
+ bl ebb_hook
+ nop
+
+ /* r2 may be changed here but we don't care */
+
+ lfd f0, FSCR_SAVE(r1)
+ mtfsf 0xff,f0
+ lfd f0, VSCR_SAVE(r1)
+ mtvscr f0
+ LOAD_VSR(0, r3)
+ LOAD_VSR(1, r3)
+ LOAD_VSR(2, r3)
+ LOAD_VSR(3, r3)
+ LOAD_VSR(4, r3)
+ LOAD_VSR(5, r3)
+ LOAD_VSR(6, r3)
+ LOAD_VSR(7, r3)
+ LOAD_VSR(8, r3)
+ LOAD_VSR(9, r3)
+ LOAD_VSR(10, r3)
+ LOAD_VSR(11, r3)
+ LOAD_VSR(12, r3)
+ LOAD_VSR(13, r3)
+ LOAD_VSR(14, r3)
+ LOAD_VSR(15, r3)
+ LOAD_VSR(16, r3)
+ LOAD_VSR(17, r3)
+ LOAD_VSR(18, r3)
+ LOAD_VSR(19, r3)
+ LOAD_VSR(20, r3)
+ LOAD_VSR(21, r3)
+ LOAD_VSR(22, r3)
+ LOAD_VSR(23, r3)
+ LOAD_VSR(24, r3)
+ LOAD_VSR(25, r3)
+ LOAD_VSR(26, r3)
+ LOAD_VSR(27, r3)
+ LOAD_VSR(28, r3)
+ LOAD_VSR(29, r3)
+ LOAD_VSR(30, r3)
+ LOAD_VSR(31, r3)
+ LOAD_VSR(32, r3)
+ LOAD_VSR(33, r3)
+ LOAD_VSR(34, r3)
+ LOAD_VSR(35, r3)
+ LOAD_VSR(36, r3)
+ LOAD_VSR(37, r3)
+ LOAD_VSR(38, r3)
+ LOAD_VSR(39, r3)
+ LOAD_VSR(40, r3)
+ LOAD_VSR(41, r3)
+ LOAD_VSR(42, r3)
+ LOAD_VSR(43, r3)
+ LOAD_VSR(44, r3)
+ LOAD_VSR(45, r3)
+ LOAD_VSR(46, r3)
+ LOAD_VSR(47, r3)
+ LOAD_VSR(48, r3)
+ LOAD_VSR(49, r3)
+ LOAD_VSR(50, r3)
+ LOAD_VSR(51, r3)
+ LOAD_VSR(52, r3)
+ LOAD_VSR(53, r3)
+ LOAD_VSR(54, r3)
+ LOAD_VSR(55, r3)
+ LOAD_VSR(56, r3)
+ LOAD_VSR(57, r3)
+ LOAD_VSR(58, r3)
+ LOAD_VSR(59, r3)
+ LOAD_VSR(60, r3)
+ LOAD_VSR(61, r3)
+ LOAD_VSR(62, r3)
+ LOAD_VSR(63, r3)
+
+ ld r0,XER_SAVE(r1)
+ mtxer r0
+ ld r0,CTR_SAVE(r1)
+ mtctr r0
+ ld r0,LR_SAVE(r1)
+ mtlr r0
+ ld r0,CCR_SAVE(r1)
+ mtcr r0
+ REST_GPR(0)
+ REST_GPR(2)
+ REST_GPR(3)
+ REST_GPR(4)
+ REST_GPR(5)
+ REST_GPR(6)
+ REST_GPR(7)
+ REST_GPR(8)
+ REST_GPR(9)
+ REST_GPR(10)
+ REST_GPR(11)
+ REST_GPR(12)
+ REST_GPR(13)
+ REST_GPR(14)
+ REST_GPR(15)
+ REST_GPR(16)
+ REST_GPR(17)
+ REST_GPR(18)
+ REST_GPR(19)
+ REST_GPR(20)
+ REST_GPR(21)
+ REST_GPR(22)
+ REST_GPR(23)
+ REST_GPR(24)
+ REST_GPR(25)
+ REST_GPR(26)
+ REST_GPR(27)
+ REST_GPR(28)
+ REST_GPR(29)
+ REST_GPR(30)
+ REST_GPR(31)
+ addi r1,r1,STACK_FRAME
+ RFEBB
+FUNC_END(ebb_handler)
diff --git a/powerpc/pmu/ebb/ebb_on_child_test.c b/powerpc/pmu/ebb/ebb_on_child_test.c
new file mode 100644
index 0000000..c45f948
--- /dev/null
+++ b/powerpc/pmu/ebb/ebb_on_child_test.c
@@ -0,0 +1,86 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdbool.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <unistd.h>
+
+#include "ebb.h"
+
+
+/*
+ * Tests we can setup an EBB on our child. Nothing interesting happens, because
+ * even though the event is enabled and running the child hasn't enabled the
+ * actual delivery of the EBBs.
+ */
+
+static int victim_child(union pipe read_pipe, union pipe write_pipe)
+{
+ int i;
+
+ FAIL_IF(wait_for_parent(read_pipe));
+ FAIL_IF(notify_parent(write_pipe));
+
+ /* Parent creates EBB event */
+
+ FAIL_IF(wait_for_parent(read_pipe));
+ FAIL_IF(notify_parent(write_pipe));
+
+ /* Check the EBB is enabled by writing PMC1 */
+ write_pmc1();
+
+ /* EBB event is enabled here */
+ for (i = 0; i < 1000000; i++) ;
+
+ return 0;
+}
+
+int ebb_on_child(void)
+{
+ union pipe read_pipe, write_pipe;
+ struct event event;
+ pid_t pid;
+
+ FAIL_IF(pipe(read_pipe.fds) == -1);
+ FAIL_IF(pipe(write_pipe.fds) == -1);
+
+ pid = fork();
+ if (pid == 0) {
+ /* NB order of pipes looks reversed */
+ exit(victim_child(write_pipe, read_pipe));
+ }
+
+ FAIL_IF(sync_with_child(read_pipe, write_pipe));
+
+ /* Child is running now */
+
+ event_init_named(&event, 0x1001e, "cycles");
+ event_leader_ebb_init(&event);
+
+ event.attr.exclude_kernel = 1;
+ event.attr.exclude_hv = 1;
+ event.attr.exclude_idle = 1;
+
+ FAIL_IF(event_open_with_pid(&event, pid));
+ FAIL_IF(ebb_event_enable(&event));
+
+ FAIL_IF(sync_with_child(read_pipe, write_pipe));
+
+ /* Child should just exit happily */
+ FAIL_IF(wait_for_child(pid));
+
+ event_close(&event);
+
+ return 0;
+}
+
+int main(void)
+{
+ return test_harness(ebb_on_child, "ebb_on_child");
+}
diff --git a/powerpc/pmu/ebb/ebb_on_willing_child_test.c b/powerpc/pmu/ebb/ebb_on_willing_child_test.c
new file mode 100644
index 0000000..11acf1d
--- /dev/null
+++ b/powerpc/pmu/ebb/ebb_on_willing_child_test.c
@@ -0,0 +1,92 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdbool.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <unistd.h>
+
+#include "ebb.h"
+
+
+/*
+ * Tests we can setup an EBB on our child. The child expects this and enables
+ * EBBs, which are then delivered to the child, even though the event is
+ * created by the parent.
+ */
+
+static int victim_child(union pipe read_pipe, union pipe write_pipe)
+{
+ FAIL_IF(wait_for_parent(read_pipe));
+
+ /* Setup our EBB handler, before the EBB event is created */
+ ebb_enable_pmc_counting(1);
+ setup_ebb_handler(standard_ebb_callee);
+ ebb_global_enable();
+
+ FAIL_IF(notify_parent(write_pipe));
+
+ while (ebb_state.stats.ebb_count < 20) {
+ FAIL_IF(core_busy_loop());
+ }
+
+ ebb_global_disable();
+ ebb_freeze_pmcs();
+
+ count_pmc(1, sample_period);
+
+ dump_ebb_state();
+
+ FAIL_IF(ebb_state.stats.ebb_count == 0);
+
+ return 0;
+}
+
+/* Tests we can setup an EBB on our child - if it's expecting it */
+int ebb_on_willing_child(void)
+{
+ union pipe read_pipe, write_pipe;
+ struct event event;
+ pid_t pid;
+
+ FAIL_IF(pipe(read_pipe.fds) == -1);
+ FAIL_IF(pipe(write_pipe.fds) == -1);
+
+ pid = fork();
+ if (pid == 0) {
+ /* NB order of pipes looks reversed */
+ exit(victim_child(write_pipe, read_pipe));
+ }
+
+ /* Signal the child to setup its EBB handler */
+ FAIL_IF(sync_with_child(read_pipe, write_pipe));
+
+ /* Child is running now */
+
+ event_init_named(&event, 0x1001e, "cycles");
+ event_leader_ebb_init(&event);
+
+ event.attr.exclude_kernel = 1;
+ event.attr.exclude_hv = 1;
+ event.attr.exclude_idle = 1;
+
+ FAIL_IF(event_open_with_pid(&event, pid));
+ FAIL_IF(ebb_event_enable(&event));
+
+ /* Child show now take EBBs and then exit */
+ FAIL_IF(wait_for_child(pid));
+
+ event_close(&event);
+
+ return 0;
+}
+
+int main(void)
+{
+ return test_harness(ebb_on_willing_child, "ebb_on_willing_child");
+}
diff --git a/powerpc/pmu/ebb/ebb_vs_cpu_event_test.c b/powerpc/pmu/ebb/ebb_vs_cpu_event_test.c
new file mode 100644
index 0000000..be4dd5a
--- /dev/null
+++ b/powerpc/pmu/ebb/ebb_vs_cpu_event_test.c
@@ -0,0 +1,86 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdbool.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <unistd.h>
+
+#include "ebb.h"
+
+
+/*
+ * Tests an EBB vs a cpu event - in that order. The EBB should force the cpu
+ * event off the PMU.
+ */
+
+static int setup_cpu_event(struct event *event, int cpu)
+{
+ event_init_named(event, 0x400FA, "PM_RUN_INST_CMPL");
+
+ event->attr.exclude_kernel = 1;
+ event->attr.exclude_hv = 1;
+ event->attr.exclude_idle = 1;
+
+ SKIP_IF(require_paranoia_below(1));
+ FAIL_IF(event_open_with_cpu(event, cpu));
+ FAIL_IF(event_enable(event));
+
+ return 0;
+}
+
+int ebb_vs_cpu_event(void)
+{
+ union pipe read_pipe, write_pipe;
+ struct event event;
+ int cpu, rc;
+ pid_t pid;
+
+ cpu = pick_online_cpu();
+ FAIL_IF(cpu < 0);
+ FAIL_IF(bind_to_cpu(cpu));
+
+ FAIL_IF(pipe(read_pipe.fds) == -1);
+ FAIL_IF(pipe(write_pipe.fds) == -1);
+
+ pid = fork();
+ if (pid == 0) {
+ /* NB order of pipes looks reversed */
+ exit(ebb_child(write_pipe, read_pipe));
+ }
+
+ /* Signal the child to install its EBB event and wait */
+ FAIL_IF(sync_with_child(read_pipe, write_pipe));
+
+ /* Now try to install our CPU event */
+ rc = setup_cpu_event(&event, cpu);
+ if (rc) {
+ kill_child_and_wait(pid);
+ return rc;
+ }
+
+ /* Signal the child to run */
+ FAIL_IF(sync_with_child(read_pipe, write_pipe));
+
+ /* .. and wait for it to complete */
+ FAIL_IF(wait_for_child(pid));
+ FAIL_IF(event_disable(&event));
+ FAIL_IF(event_read(&event));
+
+ event_report(&event);
+
+ /* The cpu event may have run, but we don't expect 100% */
+ FAIL_IF(event.result.enabled >= event.result.running);
+
+ return 0;
+}
+
+int main(void)
+{
+ return test_harness(ebb_vs_cpu_event, "ebb_vs_cpu_event");
+}
diff --git a/powerpc/pmu/ebb/event_attributes_test.c b/powerpc/pmu/ebb/event_attributes_test.c
new file mode 100644
index 0000000..7e78153
--- /dev/null
+++ b/powerpc/pmu/ebb/event_attributes_test.c
@@ -0,0 +1,131 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+
+#include "ebb.h"
+
+
+/*
+ * Test various attributes of the EBB event are enforced.
+ */
+int event_attributes(void)
+{
+ struct event event, leader;
+
+ event_init(&event, 0x1001e);
+ event_leader_ebb_init(&event);
+ /* Expected to succeed */
+ FAIL_IF(event_open(&event));
+ event_close(&event);
+
+
+ event_init(&event, 0x001e); /* CYCLES - no PMC specified */
+ event_leader_ebb_init(&event);
+ /* Expected to fail, no PMC specified */
+ FAIL_IF(event_open(&event) == 0);
+
+
+ event_init(&event, 0x2001e);
+ event_leader_ebb_init(&event);
+ event.attr.exclusive = 0;
+ /* Expected to fail, not exclusive */
+ FAIL_IF(event_open(&event) == 0);
+
+
+ event_init(&event, 0x3001e);
+ event_leader_ebb_init(&event);
+ event.attr.freq = 1;
+ /* Expected to fail, sets freq */
+ FAIL_IF(event_open(&event) == 0);
+
+
+ event_init(&event, 0x4001e);
+ event_leader_ebb_init(&event);
+ event.attr.sample_period = 1;
+ /* Expected to fail, sets sample_period */
+ FAIL_IF(event_open(&event) == 0);
+
+
+ event_init(&event, 0x1001e);
+ event_leader_ebb_init(&event);
+ event.attr.enable_on_exec = 1;
+ /* Expected to fail, sets enable_on_exec */
+ FAIL_IF(event_open(&event) == 0);
+
+
+ event_init(&event, 0x1001e);
+ event_leader_ebb_init(&event);
+ event.attr.inherit = 1;
+ /* Expected to fail, sets inherit */
+ FAIL_IF(event_open(&event) == 0);
+
+
+ event_init(&leader, 0x1001e);
+ event_leader_ebb_init(&leader);
+ FAIL_IF(event_open(&leader));
+
+ event_init(&event, 0x20002);
+ event_ebb_init(&event);
+
+ /* Expected to succeed */
+ FAIL_IF(event_open_with_group(&event, leader.fd));
+ event_close(&leader);
+ event_close(&event);
+
+
+ event_init(&leader, 0x1001e);
+ event_leader_ebb_init(&leader);
+ FAIL_IF(event_open(&leader));
+
+ event_init(&event, 0x20002);
+
+ /* Expected to fail, event doesn't request EBB, leader does */
+ FAIL_IF(event_open_with_group(&event, leader.fd) == 0);
+ event_close(&leader);
+
+
+ event_init(&leader, 0x1001e);
+ event_leader_ebb_init(&leader);
+ /* Clear the EBB flag */
+ leader.attr.config &= ~(1ull << 63);
+
+ FAIL_IF(event_open(&leader));
+
+ event_init(&event, 0x20002);
+ event_ebb_init(&event);
+
+ /* Expected to fail, leader doesn't request EBB */
+ FAIL_IF(event_open_with_group(&event, leader.fd) == 0);
+ event_close(&leader);
+
+
+ event_init(&leader, 0x1001e);
+ event_leader_ebb_init(&leader);
+ leader.attr.exclusive = 0;
+ /* Expected to fail, leader isn't exclusive */
+ FAIL_IF(event_open(&leader) == 0);
+
+
+ event_init(&leader, 0x1001e);
+ event_leader_ebb_init(&leader);
+ leader.attr.pinned = 0;
+ /* Expected to fail, leader isn't pinned */
+ FAIL_IF(event_open(&leader) == 0);
+
+ event_init(&event, 0x1001e);
+ event_leader_ebb_init(&event);
+ /* Expected to fail, not a task event */
+ SKIP_IF(require_paranoia_below(1));
+ FAIL_IF(event_open_with_cpu(&event, 0) == 0);
+
+ return 0;
+}
+
+int main(void)
+{
+ return test_harness(event_attributes, "event_attributes");
+}
diff --git a/powerpc/pmu/ebb/fixed_instruction_loop.S b/powerpc/pmu/ebb/fixed_instruction_loop.S
new file mode 100644
index 0000000..b866a05
--- /dev/null
+++ b/powerpc/pmu/ebb/fixed_instruction_loop.S
@@ -0,0 +1,43 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <ppc-asm.h>
+
+ .text
+
+FUNC_START(thirty_two_instruction_loop)
+ cmpwi r3,0
+ beqlr
+ addi r4,r3,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1 # 28 addi's
+ subi r3,r3,1
+ b FUNC_NAME(thirty_two_instruction_loop)
+FUNC_END(thirty_two_instruction_loop)
diff --git a/powerpc/pmu/ebb/fork_cleanup_test.c b/powerpc/pmu/ebb/fork_cleanup_test.c
new file mode 100644
index 0000000..9e7af6e
--- /dev/null
+++ b/powerpc/pmu/ebb/fork_cleanup_test.c
@@ -0,0 +1,79 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdbool.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <unistd.h>
+#include <setjmp.h>
+#include <signal.h>
+
+#include "ebb.h"
+
+
+/*
+ * Test that a fork clears the PMU state of the child. eg. BESCR/EBBHR/EBBRR
+ * are cleared, and MMCR0_PMCC is reset, preventing the child from accessing
+ * the PMU.
+ */
+
+static struct event event;
+
+static int child(void)
+{
+ /* Even though we have EBE=0 we can still see the EBB regs */
+ FAIL_IF(mfspr(SPRN_BESCR) != 0);
+ FAIL_IF(mfspr(SPRN_EBBHR) != 0);
+ FAIL_IF(mfspr(SPRN_EBBRR) != 0);
+
+ FAIL_IF(catch_sigill(write_pmc1));
+
+ /* We can still read from the event, though it is on our parent */
+ FAIL_IF(event_read(&event));
+
+ return 0;
+}
+
+/* Tests that fork clears EBB state */
+int fork_cleanup(void)
+{
+ pid_t pid;
+
+ event_init_named(&event, 0x1001e, "cycles");
+ event_leader_ebb_init(&event);
+
+ FAIL_IF(event_open(&event));
+
+ ebb_enable_pmc_counting(1);
+ setup_ebb_handler(standard_ebb_callee);
+ ebb_global_enable();
+
+ FAIL_IF(ebb_event_enable(&event));
+
+ mtspr(SPRN_MMCR0, MMCR0_FC);
+ mtspr(SPRN_PMC1, pmc_sample_period(sample_period));
+
+ /* Don't need to actually take any EBBs */
+
+ pid = fork();
+ if (pid == 0)
+ exit(child());
+
+ /* Child does the actual testing */
+ FAIL_IF(wait_for_child(pid));
+
+ /* After fork */
+ event_close(&event);
+
+ return 0;
+}
+
+int main(void)
+{
+ return test_harness(fork_cleanup, "fork_cleanup");
+}
diff --git a/powerpc/pmu/ebb/instruction_count_test.c b/powerpc/pmu/ebb/instruction_count_test.c
new file mode 100644
index 0000000..f8190fa
--- /dev/null
+++ b/powerpc/pmu/ebb/instruction_count_test.c
@@ -0,0 +1,164 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#define _GNU_SOURCE
+
+#include <stdio.h>
+#include <stdbool.h>
+#include <string.h>
+#include <sys/prctl.h>
+
+#include "ebb.h"
+
+
+/*
+ * Run a calibrated instruction loop and count instructions executed using
+ * EBBs. Make sure the counts look right.
+ */
+
+extern void thirty_two_instruction_loop(uint64_t loops);
+
+static bool counters_frozen = true;
+
+static int do_count_loop(struct event *event, uint64_t instructions,
+ uint64_t overhead, bool report)
+{
+ int64_t difference, expected;
+ double percentage;
+
+ clear_ebb_stats();
+
+ counters_frozen = false;
+ mb();
+ mtspr(SPRN_MMCR0, mfspr(SPRN_MMCR0) & ~MMCR0_FC);
+
+ thirty_two_instruction_loop(instructions >> 5);
+
+ counters_frozen = true;
+ mb();
+ mtspr(SPRN_MMCR0, mfspr(SPRN_MMCR0) | MMCR0_FC);
+
+ count_pmc(4, sample_period);
+
+ event->result.value = ebb_state.stats.pmc_count[4-1];
+ expected = instructions + overhead;
+ difference = event->result.value - expected;
+ percentage = (double)difference / event->result.value * 100;
+
+ if (report) {
+ printf("Looped for %lu instructions, overhead %lu\n", instructions, overhead);
+ printf("Expected %lu\n", expected);
+ printf("Actual %llu\n", event->result.value);
+ printf("Error %ld, %f%%\n", difference, percentage);
+ printf("Took %d EBBs\n", ebb_state.stats.ebb_count);
+ }
+
+ if (difference < 0)
+ difference = -difference;
+
+ /* Tolerate a difference of up to 0.0001 % */
+ difference *= 10000 * 100;
+ if (difference / event->result.value)
+ return -1;
+
+ return 0;
+}
+
+/* Count how many instructions it takes to do a null loop */
+static uint64_t determine_overhead(struct event *event)
+{
+ uint64_t current, overhead;
+ int i;
+
+ do_count_loop(event, 0, 0, false);
+ overhead = event->result.value;
+
+ for (i = 0; i < 100; i++) {
+ do_count_loop(event, 0, 0, false);
+ current = event->result.value;
+ if (current < overhead) {
+ printf("Replacing overhead %lu with %lu\n", overhead, current);
+ overhead = current;
+ }
+ }
+
+ return overhead;
+}
+
+static void pmc4_ebb_callee(void)
+{
+ uint64_t val;
+
+ val = mfspr(SPRN_BESCR);
+ if (!(val & BESCR_PMEO)) {
+ ebb_state.stats.spurious++;
+ goto out;
+ }
+
+ ebb_state.stats.ebb_count++;
+ count_pmc(4, sample_period);
+out:
+ if (counters_frozen)
+ reset_ebb_with_clear_mask(MMCR0_PMAO);
+ else
+ reset_ebb();
+}
+
+int instruction_count(void)
+{
+ struct event event;
+ uint64_t overhead;
+
+ event_init_named(&event, 0x400FA, "PM_RUN_INST_CMPL");
+ event_leader_ebb_init(&event);
+ event.attr.exclude_kernel = 1;
+ event.attr.exclude_hv = 1;
+ event.attr.exclude_idle = 1;
+
+ FAIL_IF(event_open(&event));
+ FAIL_IF(ebb_event_enable(&event));
+
+ sample_period = COUNTER_OVERFLOW;
+
+ setup_ebb_handler(pmc4_ebb_callee);
+ mtspr(SPRN_MMCR0, mfspr(SPRN_MMCR0) & ~MMCR0_FC);
+ ebb_global_enable();
+
+ overhead = determine_overhead(&event);
+ printf("Overhead of null loop: %lu instructions\n", overhead);
+
+ /* Run for 1M instructions */
+ FAIL_IF(do_count_loop(&event, 0x100000, overhead, true));
+
+ /* Run for 10M instructions */
+ FAIL_IF(do_count_loop(&event, 0xa00000, overhead, true));
+
+ /* Run for 100M instructions */
+ FAIL_IF(do_count_loop(&event, 0x6400000, overhead, true));
+
+ /* Run for 1G instructions */
+ FAIL_IF(do_count_loop(&event, 0x40000000, overhead, true));
+
+ /* Run for 16G instructions */
+ FAIL_IF(do_count_loop(&event, 0x400000000, overhead, true));
+
+ /* Run for 64G instructions */
+ FAIL_IF(do_count_loop(&event, 0x1000000000, overhead, true));
+
+ /* Run for 128G instructions */
+ FAIL_IF(do_count_loop(&event, 0x2000000000, overhead, true));
+
+ ebb_global_disable();
+ event_close(&event);
+
+ printf("Finished OK\n");
+
+ return 0;
+}
+
+int main(void)
+{
+ return test_harness(instruction_count, "instruction_count");
+}
diff --git a/powerpc/pmu/ebb/lost_exception_test.c b/powerpc/pmu/ebb/lost_exception_test.c
new file mode 100644
index 0000000..0c9dd9b
--- /dev/null
+++ b/powerpc/pmu/ebb/lost_exception_test.c
@@ -0,0 +1,100 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <sched.h>
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <sys/mman.h>
+
+#include "ebb.h"
+
+
+/*
+ * Test that tries to trigger CPU_FTR_PMAO_BUG. Which is a hardware defect
+ * where an exception triggers but we context switch before it is delivered and
+ * lose the exception.
+ */
+
+static int test_body(void)
+{
+ int i, orig_period, max_period;
+ struct event event;
+
+ /* We use PMC4 to make sure the kernel switches all counters correctly */
+ event_init_named(&event, 0x40002, "instructions");
+ event_leader_ebb_init(&event);
+
+ event.attr.exclude_kernel = 1;
+ event.attr.exclude_hv = 1;
+ event.attr.exclude_idle = 1;
+
+ FAIL_IF(event_open(&event));
+
+ ebb_enable_pmc_counting(4);
+ setup_ebb_handler(standard_ebb_callee);
+ ebb_global_enable();
+ FAIL_IF(ebb_event_enable(&event));
+
+ /*
+ * We want a low sample period, but we also want to get out of the EBB
+ * handler without tripping up again.
+ *
+ * This value picked after much experimentation.
+ */
+ orig_period = max_period = sample_period = 400;
+
+ mtspr(SPRN_PMC4, pmc_sample_period(sample_period));
+
+ while (ebb_state.stats.ebb_count < 1000000) {
+ /*
+ * We are trying to get the EBB exception to race exactly with
+ * us entering the kernel to do the syscall. We then need the
+ * kernel to decide our timeslice is up and context switch to
+ * the other thread. When we come back our EBB will have been
+ * lost and we'll spin in this while loop forever.
+ */
+
+ for (i = 0; i < 100000; i++)
+ sched_yield();
+
+ /* Change the sample period slightly to try and hit the race */
+ if (sample_period >= (orig_period + 200))
+ sample_period = orig_period;
+ else
+ sample_period++;
+
+ if (sample_period > max_period)
+ max_period = sample_period;
+ }
+
+ ebb_freeze_pmcs();
+ ebb_global_disable();
+
+ count_pmc(4, sample_period);
+ mtspr(SPRN_PMC4, 0xdead);
+
+ dump_summary_ebb_state();
+ dump_ebb_hw_state();
+
+ event_close(&event);
+
+ FAIL_IF(ebb_state.stats.ebb_count == 0);
+
+ /* We vary our sample period so we need extra fudge here */
+ FAIL_IF(!ebb_check_count(4, orig_period, 2 * (max_period - orig_period)));
+
+ return 0;
+}
+
+static int lost_exception(void)
+{
+ return eat_cpu(test_body);
+}
+
+int main(void)
+{
+ return test_harness(lost_exception, "lost_exception");
+}
diff --git a/powerpc/pmu/ebb/multi_counter_test.c b/powerpc/pmu/ebb/multi_counter_test.c
new file mode 100644
index 0000000..67d78af
--- /dev/null
+++ b/powerpc/pmu/ebb/multi_counter_test.c
@@ -0,0 +1,91 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <sys/ioctl.h>
+
+#include "ebb.h"
+
+
+/*
+ * Test counting multiple events using EBBs.
+ */
+int multi_counter(void)
+{
+ struct event events[6];
+ int i, group_fd;
+
+ event_init_named(&events[0], 0x1001C, "PM_CMPLU_STALL_THRD");
+ event_init_named(&events[1], 0x2D016, "PM_CMPLU_STALL_FXU");
+ event_init_named(&events[2], 0x30006, "PM_CMPLU_STALL_OTHER_CMPL");
+ event_init_named(&events[3], 0x4000A, "PM_CMPLU_STALL");
+ event_init_named(&events[4], 0x600f4, "PM_RUN_CYC");
+ event_init_named(&events[5], 0x500fa, "PM_RUN_INST_CMPL");
+
+ event_leader_ebb_init(&events[0]);
+ for (i = 1; i < 6; i++)
+ event_ebb_init(&events[i]);
+
+ group_fd = -1;
+ for (i = 0; i < 6; i++) {
+ events[i].attr.exclude_kernel = 1;
+ events[i].attr.exclude_hv = 1;
+ events[i].attr.exclude_idle = 1;
+
+ FAIL_IF(event_open_with_group(&events[i], group_fd));
+ if (group_fd == -1)
+ group_fd = events[0].fd;
+ }
+
+ ebb_enable_pmc_counting(1);
+ ebb_enable_pmc_counting(2);
+ ebb_enable_pmc_counting(3);
+ ebb_enable_pmc_counting(4);
+ ebb_enable_pmc_counting(5);
+ ebb_enable_pmc_counting(6);
+ setup_ebb_handler(standard_ebb_callee);
+
+ FAIL_IF(ioctl(events[0].fd, PERF_EVENT_IOC_ENABLE, PERF_IOC_FLAG_GROUP));
+ FAIL_IF(event_read(&events[0]));
+
+ ebb_global_enable();
+
+ mtspr(SPRN_PMC1, pmc_sample_period(sample_period));
+ mtspr(SPRN_PMC2, pmc_sample_period(sample_period));
+ mtspr(SPRN_PMC3, pmc_sample_period(sample_period));
+ mtspr(SPRN_PMC4, pmc_sample_period(sample_period));
+ mtspr(SPRN_PMC5, pmc_sample_period(sample_period));
+ mtspr(SPRN_PMC6, pmc_sample_period(sample_period));
+
+ while (ebb_state.stats.ebb_count < 50) {
+ FAIL_IF(core_busy_loop());
+ FAIL_IF(ebb_check_mmcr0());
+ }
+
+ ebb_global_disable();
+ ebb_freeze_pmcs();
+
+ count_pmc(1, sample_period);
+ count_pmc(2, sample_period);
+ count_pmc(3, sample_period);
+ count_pmc(4, sample_period);
+ count_pmc(5, sample_period);
+ count_pmc(6, sample_period);
+
+ dump_ebb_state();
+
+ for (i = 0; i < 6; i++)
+ event_close(&events[i]);
+
+ FAIL_IF(ebb_state.stats.ebb_count == 0);
+
+ return 0;
+}
+
+int main(void)
+{
+ return test_harness(multi_counter, "multi_counter");
+}
diff --git a/powerpc/pmu/ebb/multi_ebb_procs_test.c b/powerpc/pmu/ebb/multi_ebb_procs_test.c
new file mode 100644
index 0000000..b8dc371
--- /dev/null
+++ b/powerpc/pmu/ebb/multi_ebb_procs_test.c
@@ -0,0 +1,109 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <stdbool.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <signal.h>
+
+#include "ebb.h"
+
+
+/*
+ * Test running multiple EBB using processes at once on a single CPU. They
+ * should all run happily without interfering with each other.
+ */
+
+static bool child_should_exit;
+
+static void sigint_handler(int signal)
+{
+ child_should_exit = true;
+}
+
+struct sigaction sigint_action = {
+ .sa_handler = sigint_handler,
+};
+
+static int cycles_child(void)
+{
+ struct event event;
+
+ if (sigaction(SIGINT, &sigint_action, NULL)) {
+ perror("sigaction");
+ return 1;
+ }
+
+ event_init_named(&event, 0x1001e, "cycles");
+ event_leader_ebb_init(&event);
+
+ event.attr.exclude_kernel = 1;
+ event.attr.exclude_hv = 1;
+ event.attr.exclude_idle = 1;
+
+ FAIL_IF(event_open(&event));
+
+ ebb_enable_pmc_counting(1);
+ setup_ebb_handler(standard_ebb_callee);
+ ebb_global_enable();
+
+ FAIL_IF(ebb_event_enable(&event));
+
+ mtspr(SPRN_PMC1, pmc_sample_period(sample_period));
+
+ while (!child_should_exit) {
+ FAIL_IF(core_busy_loop());
+ FAIL_IF(ebb_check_mmcr0());
+ }
+
+ ebb_global_disable();
+ ebb_freeze_pmcs();
+
+ count_pmc(1, sample_period);
+
+ dump_summary_ebb_state();
+
+ event_close(&event);
+
+ FAIL_IF(ebb_state.stats.ebb_count == 0);
+
+ return 0;
+}
+
+#define NR_CHILDREN 4
+
+int multi_ebb_procs(void)
+{
+ pid_t pids[NR_CHILDREN];
+ int cpu, rc, i;
+
+ cpu = pick_online_cpu();
+ FAIL_IF(cpu < 0);
+ FAIL_IF(bind_to_cpu(cpu));
+
+ for (i = 0; i < NR_CHILDREN; i++) {
+ pids[i] = fork();
+ if (pids[i] == 0)
+ exit(cycles_child());
+ }
+
+ /* Have them all run for "a while" */
+ sleep(10);
+
+ rc = 0;
+ for (i = 0; i < NR_CHILDREN; i++) {
+ /* Tell them to stop */
+ kill(pids[i], SIGINT);
+ /* And wait */
+ rc |= wait_for_child(pids[i]);
+ }
+
+ return rc;
+}
+
+int main(void)
+{
+ return test_harness(multi_ebb_procs, "multi_ebb_procs");
+}
diff --git a/powerpc/pmu/ebb/no_handler_test.c b/powerpc/pmu/ebb/no_handler_test.c
new file mode 100644
index 0000000..2f9bf8e
--- /dev/null
+++ b/powerpc/pmu/ebb/no_handler_test.c
@@ -0,0 +1,61 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <setjmp.h>
+#include <signal.h>
+
+#include "ebb.h"
+
+
+/* Test that things work sanely if we have no handler */
+
+static int no_handler_test(void)
+{
+ struct event event;
+ u64 val;
+ int i;
+
+ event_init_named(&event, 0x1001e, "cycles");
+ event_leader_ebb_init(&event);
+
+ event.attr.exclude_kernel = 1;
+ event.attr.exclude_hv = 1;
+ event.attr.exclude_idle = 1;
+
+ FAIL_IF(event_open(&event));
+ FAIL_IF(ebb_event_enable(&event));
+
+ val = mfspr(SPRN_EBBHR);
+ FAIL_IF(val != 0);
+
+ /* Make sure it overflows quickly */
+ sample_period = 1000;
+ mtspr(SPRN_PMC1, pmc_sample_period(sample_period));
+
+ /* Spin to make sure the event has time to overflow */
+ for (i = 0; i < 1000; i++)
+ mb();
+
+ dump_ebb_state();
+
+ /* We expect to see the PMU frozen & PMAO set */
+ val = mfspr(SPRN_MMCR0);
+ FAIL_IF(val != 0x0000000080000080);
+
+ event_close(&event);
+
+ dump_ebb_state();
+
+ /* The real test is that we never took an EBB at 0x0 */
+
+ return 0;
+}
+
+int main(void)
+{
+ return test_harness(no_handler_test,"no_handler_test");
+}
diff --git a/powerpc/pmu/ebb/pmae_handling_test.c b/powerpc/pmu/ebb/pmae_handling_test.c
new file mode 100644
index 0000000..986500f
--- /dev/null
+++ b/powerpc/pmu/ebb/pmae_handling_test.c
@@ -0,0 +1,106 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <sched.h>
+#include <signal.h>
+#include <stdbool.h>
+#include <stdio.h>
+#include <stdlib.h>
+
+#include "ebb.h"
+
+
+/*
+ * Test that the kernel properly handles PMAE across context switches.
+ *
+ * We test this by calling into the kernel inside our EBB handler, where PMAE
+ * is clear. A cpu eater companion thread is running on the same CPU as us to
+ * encourage the scheduler to switch us.
+ *
+ * The kernel must make sure that when it context switches us back in, it
+ * honours the fact that we had PMAE clear.
+ *
+ * Observed to hit the failing case on the first EBB with a broken kernel.
+ */
+
+static bool mmcr0_mismatch;
+static uint64_t before, after;
+
+static void syscall_ebb_callee(void)
+{
+ uint64_t val;
+
+ val = mfspr(SPRN_BESCR);
+ if (!(val & BESCR_PMEO)) {
+ ebb_state.stats.spurious++;
+ goto out;
+ }
+
+ ebb_state.stats.ebb_count++;
+ count_pmc(1, sample_period);
+
+ before = mfspr(SPRN_MMCR0);
+
+ /* Try and get ourselves scheduled, to force a PMU context switch */
+ sched_yield();
+
+ after = mfspr(SPRN_MMCR0);
+ if (before != after)
+ mmcr0_mismatch = true;
+
+out:
+ reset_ebb();
+}
+
+static int test_body(void)
+{
+ struct event event;
+
+ event_init_named(&event, 0x1001e, "cycles");
+ event_leader_ebb_init(&event);
+
+ event.attr.exclude_kernel = 1;
+ event.attr.exclude_hv = 1;
+ event.attr.exclude_idle = 1;
+
+ FAIL_IF(event_open(&event));
+
+ setup_ebb_handler(syscall_ebb_callee);
+ ebb_global_enable();
+
+ FAIL_IF(ebb_event_enable(&event));
+
+ mtspr(SPRN_PMC1, pmc_sample_period(sample_period));
+
+ while (ebb_state.stats.ebb_count < 20 && !mmcr0_mismatch)
+ FAIL_IF(core_busy_loop());
+
+ ebb_global_disable();
+ ebb_freeze_pmcs();
+
+ count_pmc(1, sample_period);
+
+ dump_ebb_state();
+
+ if (mmcr0_mismatch)
+ printf("Saw MMCR0 before 0x%lx after 0x%lx\n", before, after);
+
+ event_close(&event);
+
+ FAIL_IF(ebb_state.stats.ebb_count == 0);
+ FAIL_IF(mmcr0_mismatch);
+
+ return 0;
+}
+
+int pmae_handling(void)
+{
+ return eat_cpu(test_body);
+}
+
+int main(void)
+{
+ return test_harness(pmae_handling, "pmae_handling");
+}
diff --git a/powerpc/pmu/ebb/pmc56_overflow_test.c b/powerpc/pmu/ebb/pmc56_overflow_test.c
new file mode 100644
index 0000000..a503fa7
--- /dev/null
+++ b/powerpc/pmu/ebb/pmc56_overflow_test.c
@@ -0,0 +1,93 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+
+#include "ebb.h"
+
+
+/*
+ * Test that PMC5 & 6 are frozen (ie. don't overflow) when they are not being
+ * used. Tests the MMCR0_FC56 logic in the kernel.
+ */
+
+static int pmc56_overflowed;
+
+static void ebb_callee(void)
+{
+ uint64_t val;
+
+ val = mfspr(SPRN_BESCR);
+ if (!(val & BESCR_PMEO)) {
+ ebb_state.stats.spurious++;
+ goto out;
+ }
+
+ ebb_state.stats.ebb_count++;
+ count_pmc(2, sample_period);
+
+ val = mfspr(SPRN_PMC5);
+ if (val >= COUNTER_OVERFLOW)
+ pmc56_overflowed++;
+
+ count_pmc(5, COUNTER_OVERFLOW);
+
+ val = mfspr(SPRN_PMC6);
+ if (val >= COUNTER_OVERFLOW)
+ pmc56_overflowed++;
+
+ count_pmc(6, COUNTER_OVERFLOW);
+
+out:
+ reset_ebb();
+}
+
+int pmc56_overflow(void)
+{
+ struct event event;
+
+ /* Use PMC2 so we set PMCjCE, which enables PMC5/6 */
+ event_init(&event, 0x2001e);
+ event_leader_ebb_init(&event);
+
+ event.attr.exclude_kernel = 1;
+ event.attr.exclude_hv = 1;
+ event.attr.exclude_idle = 1;
+
+ FAIL_IF(event_open(&event));
+
+ setup_ebb_handler(ebb_callee);
+ ebb_global_enable();
+
+ FAIL_IF(ebb_event_enable(&event));
+
+ mtspr(SPRN_PMC1, pmc_sample_period(sample_period));
+ mtspr(SPRN_PMC5, 0);
+ mtspr(SPRN_PMC6, 0);
+
+ while (ebb_state.stats.ebb_count < 10)
+ FAIL_IF(core_busy_loop());
+
+ ebb_global_disable();
+ ebb_freeze_pmcs();
+
+ count_pmc(2, sample_period);
+
+ dump_ebb_state();
+
+ printf("PMC5/6 overflow %d\n", pmc56_overflowed);
+
+ event_close(&event);
+
+ FAIL_IF(ebb_state.stats.ebb_count == 0 || pmc56_overflowed != 0);
+
+ return 0;
+}
+
+int main(void)
+{
+ return test_harness(pmc56_overflow, "pmc56_overflow");
+}
diff --git a/powerpc/pmu/ebb/reg.h b/powerpc/pmu/ebb/reg.h
new file mode 100644
index 0000000..5921b0d
--- /dev/null
+++ b/powerpc/pmu/ebb/reg.h
@@ -0,0 +1,49 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#ifndef _SELFTESTS_POWERPC_REG_H
+#define _SELFTESTS_POWERPC_REG_H
+
+#define __stringify_1(x) #x
+#define __stringify(x) __stringify_1(x)
+
+#define mfspr(rn) ({unsigned long rval; \
+ asm volatile("mfspr %0," __stringify(rn) \
+ : "=r" (rval)); rval; })
+#define mtspr(rn, v) asm volatile("mtspr " __stringify(rn) ",%0" : \
+ : "r" ((unsigned long)(v)) \
+ : "memory")
+
+#define mb() asm volatile("sync" : : : "memory");
+
+#define SPRN_MMCR2 769
+#define SPRN_MMCRA 770
+#define SPRN_MMCR0 779
+#define MMCR0_PMAO 0x00000080
+#define MMCR0_PMAE 0x04000000
+#define MMCR0_FC 0x80000000
+#define SPRN_EBBHR 804
+#define SPRN_EBBRR 805
+#define SPRN_BESCR 806 /* Branch event status & control register */
+#define SPRN_BESCRS 800 /* Branch event status & control set (1 bits set to 1) */
+#define SPRN_BESCRSU 801 /* Branch event status & control set upper */
+#define SPRN_BESCRR 802 /* Branch event status & control REset (1 bits set to 0) */
+#define SPRN_BESCRRU 803 /* Branch event status & control REset upper */
+
+#define BESCR_PMEO 0x1 /* PMU Event-based exception Occurred */
+#define BESCR_PME (0x1ul << 32) /* PMU Event-based exception Enable */
+
+#define SPRN_PMC1 771
+#define SPRN_PMC2 772
+#define SPRN_PMC3 773
+#define SPRN_PMC4 774
+#define SPRN_PMC5 775
+#define SPRN_PMC6 776
+
+#define SPRN_SIAR 780
+#define SPRN_SDAR 781
+#define SPRN_SIER 768
+
+#endif /* _SELFTESTS_POWERPC_REG_H */
diff --git a/powerpc/pmu/ebb/reg_access_test.c b/powerpc/pmu/ebb/reg_access_test.c
new file mode 100644
index 0000000..0cae66f
--- /dev/null
+++ b/powerpc/pmu/ebb/reg_access_test.c
@@ -0,0 +1,39 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+
+#include "ebb.h"
+#include "reg.h"
+
+
+/*
+ * Test basic access to the EBB regs, they should be user accessible with no
+ * kernel interaction required.
+ */
+int reg_access(void)
+{
+ uint64_t val, expected;
+
+ expected = 0x8000000100000000ull;
+ mtspr(SPRN_BESCR, expected);
+ val = mfspr(SPRN_BESCR);
+
+ FAIL_IF(val != expected);
+
+ expected = 0x0000000001000000ull;
+ mtspr(SPRN_EBBHR, expected);
+ val = mfspr(SPRN_EBBHR);
+
+ FAIL_IF(val != expected);
+
+ return 0;
+}
+
+int main(void)
+{
+ return test_harness(reg_access, "reg_access");
+}
diff --git a/powerpc/pmu/ebb/task_event_pinned_vs_ebb_test.c b/powerpc/pmu/ebb/task_event_pinned_vs_ebb_test.c
new file mode 100644
index 0000000..d56607e
--- /dev/null
+++ b/powerpc/pmu/ebb/task_event_pinned_vs_ebb_test.c
@@ -0,0 +1,91 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdbool.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <unistd.h>
+
+#include "ebb.h"
+
+
+/*
+ * Tests a pinned per-task event vs an EBB - in that order. The pinned per-task
+ * event should prevent the EBB event from being enabled.
+ */
+
+static int setup_child_event(struct event *event, pid_t child_pid)
+{
+ event_init_named(event, 0x400FA, "PM_RUN_INST_CMPL");
+
+ event->attr.pinned = 1;
+
+ event->attr.exclude_kernel = 1;
+ event->attr.exclude_hv = 1;
+ event->attr.exclude_idle = 1;
+
+ FAIL_IF(event_open_with_pid(event, child_pid));
+ FAIL_IF(event_enable(event));
+
+ return 0;
+}
+
+int task_event_pinned_vs_ebb(void)
+{
+ union pipe read_pipe, write_pipe;
+ struct event event;
+ pid_t pid;
+ int rc;
+
+ FAIL_IF(pipe(read_pipe.fds) == -1);
+ FAIL_IF(pipe(write_pipe.fds) == -1);
+
+ pid = fork();
+ if (pid == 0) {
+ /* NB order of pipes looks reversed */
+ exit(ebb_child(write_pipe, read_pipe));
+ }
+
+ /* We setup the task event first */
+ rc = setup_child_event(&event, pid);
+ if (rc) {
+ kill_child_and_wait(pid);
+ return rc;
+ }
+
+ /* Signal the child to install its EBB event and wait */
+ if (sync_with_child(read_pipe, write_pipe))
+ /* If it fails, wait for it to exit */
+ goto wait;
+
+ /* Signal the child to run */
+ FAIL_IF(sync_with_child(read_pipe, write_pipe));
+
+wait:
+ /* We expect it to fail to read the event */
+ FAIL_IF(wait_for_child(pid) != 2);
+ FAIL_IF(event_disable(&event));
+ FAIL_IF(event_read(&event));
+
+ event_report(&event);
+
+ FAIL_IF(event.result.value == 0);
+ /*
+ * For reasons I don't understand enabled is usually just slightly
+ * lower than running. Would be good to confirm why.
+ */
+ FAIL_IF(event.result.enabled == 0);
+ FAIL_IF(event.result.running == 0);
+
+ return 0;
+}
+
+int main(void)
+{
+ return test_harness(task_event_pinned_vs_ebb, "task_event_pinned_vs_ebb");
+}
diff --git a/powerpc/pmu/ebb/task_event_vs_ebb_test.c b/powerpc/pmu/ebb/task_event_vs_ebb_test.c
new file mode 100644
index 0000000..eba3219
--- /dev/null
+++ b/powerpc/pmu/ebb/task_event_vs_ebb_test.c
@@ -0,0 +1,83 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdbool.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <unistd.h>
+
+#include "ebb.h"
+
+
+/*
+ * Tests a per-task event vs an EBB - in that order. The EBB should push the
+ * per-task event off the PMU.
+ */
+
+static int setup_child_event(struct event *event, pid_t child_pid)
+{
+ event_init_named(event, 0x400FA, "PM_RUN_INST_CMPL");
+
+ event->attr.exclude_kernel = 1;
+ event->attr.exclude_hv = 1;
+ event->attr.exclude_idle = 1;
+
+ FAIL_IF(event_open_with_pid(event, child_pid));
+ FAIL_IF(event_enable(event));
+
+ return 0;
+}
+
+int task_event_vs_ebb(void)
+{
+ union pipe read_pipe, write_pipe;
+ struct event event;
+ pid_t pid;
+ int rc;
+
+ FAIL_IF(pipe(read_pipe.fds) == -1);
+ FAIL_IF(pipe(write_pipe.fds) == -1);
+
+ pid = fork();
+ if (pid == 0) {
+ /* NB order of pipes looks reversed */
+ exit(ebb_child(write_pipe, read_pipe));
+ }
+
+ /* We setup the task event first */
+ rc = setup_child_event(&event, pid);
+ if (rc) {
+ kill_child_and_wait(pid);
+ return rc;
+ }
+
+ /* Signal the child to install its EBB event and wait */
+ if (sync_with_child(read_pipe, write_pipe))
+ /* If it fails, wait for it to exit */
+ goto wait;
+
+ /* Signal the child to run */
+ FAIL_IF(sync_with_child(read_pipe, write_pipe));
+
+wait:
+ /* The EBB event should push the task event off so the child should succeed */
+ FAIL_IF(wait_for_child(pid));
+ FAIL_IF(event_disable(&event));
+ FAIL_IF(event_read(&event));
+
+ event_report(&event);
+
+ /* The task event may have run, or not so we can't assert anything about it */
+
+ return 0;
+}
+
+int main(void)
+{
+ return test_harness(task_event_vs_ebb, "task_event_vs_ebb");
+}
diff --git a/powerpc/pmu/ebb/trace.c b/powerpc/pmu/ebb/trace.c
new file mode 100644
index 0000000..251e66a
--- /dev/null
+++ b/powerpc/pmu/ebb/trace.c
@@ -0,0 +1,300 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <errno.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/mman.h>
+
+#include "trace.h"
+
+
+struct trace_buffer *trace_buffer_allocate(u64 size)
+{
+ struct trace_buffer *tb;
+
+ if (size < sizeof(*tb)) {
+ fprintf(stderr, "Error: trace buffer too small\n");
+ return NULL;
+ }
+
+ tb = mmap(NULL, size, PROT_READ | PROT_WRITE,
+ MAP_ANONYMOUS | MAP_PRIVATE, -1, 0);
+ if (tb == MAP_FAILED) {
+ perror("mmap");
+ return NULL;
+ }
+
+ tb->size = size;
+ tb->tail = tb->data;
+ tb->overflow = false;
+
+ return tb;
+}
+
+static bool trace_check_bounds(struct trace_buffer *tb, void *p)
+{
+ return p < ((void *)tb + tb->size);
+}
+
+static bool trace_check_alloc(struct trace_buffer *tb, void *p)
+{
+ /*
+ * If we ever overflowed don't allow any more input. This prevents us
+ * from dropping a large item and then later logging a small one. The
+ * buffer should just stop when overflow happened, not be patchy. If
+ * you're overflowing, make your buffer bigger.
+ */
+ if (tb->overflow)
+ return false;
+
+ if (!trace_check_bounds(tb, p)) {
+ tb->overflow = true;
+ return false;
+ }
+
+ return true;
+}
+
+static void *trace_alloc(struct trace_buffer *tb, int bytes)
+{
+ void *p, *newtail;
+
+ p = tb->tail;
+ newtail = tb->tail + bytes;
+ if (!trace_check_alloc(tb, newtail))
+ return NULL;
+
+ tb->tail = newtail;
+
+ return p;
+}
+
+static struct trace_entry *trace_alloc_entry(struct trace_buffer *tb, int payload_size)
+{
+ struct trace_entry *e;
+
+ e = trace_alloc(tb, sizeof(*e) + payload_size);
+ if (e)
+ e->length = payload_size;
+
+ return e;
+}
+
+int trace_log_reg(struct trace_buffer *tb, u64 reg, u64 value)
+{
+ struct trace_entry *e;
+ u64 *p;
+
+ e = trace_alloc_entry(tb, sizeof(reg) + sizeof(value));
+ if (!e)
+ return -ENOSPC;
+
+ e->type = TRACE_TYPE_REG;
+ p = (u64 *)e->data;
+ *p++ = reg;
+ *p++ = value;
+
+ return 0;
+}
+
+int trace_log_counter(struct trace_buffer *tb, u64 value)
+{
+ struct trace_entry *e;
+ u64 *p;
+
+ e = trace_alloc_entry(tb, sizeof(value));
+ if (!e)
+ return -ENOSPC;
+
+ e->type = TRACE_TYPE_COUNTER;
+ p = (u64 *)e->data;
+ *p++ = value;
+
+ return 0;
+}
+
+int trace_log_string(struct trace_buffer *tb, char *str)
+{
+ struct trace_entry *e;
+ char *p;
+ int len;
+
+ len = strlen(str);
+
+ /* We NULL terminate to make printing easier */
+ e = trace_alloc_entry(tb, len + 1);
+ if (!e)
+ return -ENOSPC;
+
+ e->type = TRACE_TYPE_STRING;
+ p = (char *)e->data;
+ memcpy(p, str, len);
+ p += len;
+ *p = '\0';
+
+ return 0;
+}
+
+int trace_log_indent(struct trace_buffer *tb)
+{
+ struct trace_entry *e;
+
+ e = trace_alloc_entry(tb, 0);
+ if (!e)
+ return -ENOSPC;
+
+ e->type = TRACE_TYPE_INDENT;
+
+ return 0;
+}
+
+int trace_log_outdent(struct trace_buffer *tb)
+{
+ struct trace_entry *e;
+
+ e = trace_alloc_entry(tb, 0);
+ if (!e)
+ return -ENOSPC;
+
+ e->type = TRACE_TYPE_OUTDENT;
+
+ return 0;
+}
+
+static void trace_print_header(int seq, int prefix)
+{
+ printf("%*s[%d]: ", prefix, "", seq);
+}
+
+static char *trace_decode_reg(int reg)
+{
+ switch (reg) {
+ case 769: return "SPRN_MMCR2"; break;
+ case 770: return "SPRN_MMCRA"; break;
+ case 779: return "SPRN_MMCR0"; break;
+ case 804: return "SPRN_EBBHR"; break;
+ case 805: return "SPRN_EBBRR"; break;
+ case 806: return "SPRN_BESCR"; break;
+ case 800: return "SPRN_BESCRS"; break;
+ case 801: return "SPRN_BESCRSU"; break;
+ case 802: return "SPRN_BESCRR"; break;
+ case 803: return "SPRN_BESCRRU"; break;
+ case 771: return "SPRN_PMC1"; break;
+ case 772: return "SPRN_PMC2"; break;
+ case 773: return "SPRN_PMC3"; break;
+ case 774: return "SPRN_PMC4"; break;
+ case 775: return "SPRN_PMC5"; break;
+ case 776: return "SPRN_PMC6"; break;
+ case 780: return "SPRN_SIAR"; break;
+ case 781: return "SPRN_SDAR"; break;
+ case 768: return "SPRN_SIER"; break;
+ }
+
+ return NULL;
+}
+
+static void trace_print_reg(struct trace_entry *e)
+{
+ u64 *p, *reg, *value;
+ char *name;
+
+ p = (u64 *)e->data;
+ reg = p++;
+ value = p;
+
+ name = trace_decode_reg(*reg);
+ if (name)
+ printf("register %-10s = 0x%016llx\n", name, *value);
+ else
+ printf("register %lld = 0x%016llx\n", *reg, *value);
+}
+
+static void trace_print_counter(struct trace_entry *e)
+{
+ u64 *value;
+
+ value = (u64 *)e->data;
+ printf("counter = %lld\n", *value);
+}
+
+static void trace_print_string(struct trace_entry *e)
+{
+ char *str;
+
+ str = (char *)e->data;
+ puts(str);
+}
+
+#define BASE_PREFIX 2
+#define PREFIX_DELTA 8
+
+static void trace_print_entry(struct trace_entry *e, int seq, int *prefix)
+{
+ switch (e->type) {
+ case TRACE_TYPE_REG:
+ trace_print_header(seq, *prefix);
+ trace_print_reg(e);
+ break;
+ case TRACE_TYPE_COUNTER:
+ trace_print_header(seq, *prefix);
+ trace_print_counter(e);
+ break;
+ case TRACE_TYPE_STRING:
+ trace_print_header(seq, *prefix);
+ trace_print_string(e);
+ break;
+ case TRACE_TYPE_INDENT:
+ trace_print_header(seq, *prefix);
+ puts("{");
+ *prefix += PREFIX_DELTA;
+ break;
+ case TRACE_TYPE_OUTDENT:
+ *prefix -= PREFIX_DELTA;
+ if (*prefix < BASE_PREFIX)
+ *prefix = BASE_PREFIX;
+ trace_print_header(seq, *prefix);
+ puts("}");
+ break;
+ default:
+ trace_print_header(seq, *prefix);
+ printf("entry @ %p type %d\n", e, e->type);
+ break;
+ }
+}
+
+void trace_buffer_print(struct trace_buffer *tb)
+{
+ struct trace_entry *e;
+ int i, prefix;
+ void *p;
+
+ printf("Trace buffer dump:\n");
+ printf(" address %p \n", tb);
+ printf(" tail %p\n", tb->tail);
+ printf(" size %llu\n", tb->size);
+ printf(" overflow %s\n", tb->overflow ? "TRUE" : "false");
+ printf(" Content:\n");
+
+ p = tb->data;
+
+ i = 0;
+ prefix = BASE_PREFIX;
+
+ while (trace_check_bounds(tb, p) && p < tb->tail) {
+ e = p;
+
+ trace_print_entry(e, i, &prefix);
+
+ i++;
+ p = (void *)e + sizeof(*e) + e->length;
+ }
+}
+
+void trace_print_location(struct trace_buffer *tb)
+{
+ printf("Trace buffer 0x%llx bytes @ %p\n", tb->size, tb);
+}
diff --git a/powerpc/pmu/ebb/trace.h b/powerpc/pmu/ebb/trace.h
new file mode 100644
index 0000000..926458e
--- /dev/null
+++ b/powerpc/pmu/ebb/trace.h
@@ -0,0 +1,41 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#ifndef _SELFTESTS_POWERPC_PMU_EBB_TRACE_H
+#define _SELFTESTS_POWERPC_PMU_EBB_TRACE_H
+
+#include "utils.h"
+
+#define TRACE_TYPE_REG 1
+#define TRACE_TYPE_COUNTER 2
+#define TRACE_TYPE_STRING 3
+#define TRACE_TYPE_INDENT 4
+#define TRACE_TYPE_OUTDENT 5
+
+struct trace_entry
+{
+ u8 type;
+ u8 length;
+ u8 data[0];
+};
+
+struct trace_buffer
+{
+ u64 size;
+ bool overflow;
+ void *tail;
+ u8 data[0];
+};
+
+struct trace_buffer *trace_buffer_allocate(u64 size);
+int trace_log_reg(struct trace_buffer *tb, u64 reg, u64 value);
+int trace_log_counter(struct trace_buffer *tb, u64 value);
+int trace_log_string(struct trace_buffer *tb, char *str);
+int trace_log_indent(struct trace_buffer *tb);
+int trace_log_outdent(struct trace_buffer *tb);
+void trace_buffer_print(struct trace_buffer *tb);
+void trace_print_location(struct trace_buffer *tb);
+
+#endif /* _SELFTESTS_POWERPC_PMU_EBB_TRACE_H */
diff --git a/powerpc/pmu/event.c b/powerpc/pmu/event.c
new file mode 100644
index 0000000..184b368
--- /dev/null
+++ b/powerpc/pmu/event.c
@@ -0,0 +1,131 @@
+/*
+ * Copyright 2013, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#define _GNU_SOURCE
+#include <unistd.h>
+#include <sys/syscall.h>
+#include <string.h>
+#include <stdio.h>
+#include <sys/ioctl.h>
+
+#include "event.h"
+
+
+int perf_event_open(struct perf_event_attr *attr, pid_t pid, int cpu,
+ int group_fd, unsigned long flags)
+{
+ return syscall(__NR_perf_event_open, attr, pid, cpu,
+ group_fd, flags);
+}
+
+void event_init_opts(struct event *e, u64 config, int type, char *name)
+{
+ memset(e, 0, sizeof(*e));
+
+ e->name = name;
+
+ e->attr.type = type;
+ e->attr.config = config;
+ e->attr.size = sizeof(e->attr);
+ /* This has to match the structure layout in the header */
+ e->attr.read_format = PERF_FORMAT_TOTAL_TIME_ENABLED | \
+ PERF_FORMAT_TOTAL_TIME_RUNNING;
+}
+
+void event_init_named(struct event *e, u64 config, char *name)
+{
+ event_init_opts(e, config, PERF_TYPE_RAW, name);
+}
+
+void event_init(struct event *e, u64 config)
+{
+ event_init_opts(e, config, PERF_TYPE_RAW, "event");
+}
+
+#define PERF_CURRENT_PID 0
+#define PERF_NO_PID -1
+#define PERF_NO_CPU -1
+#define PERF_NO_GROUP -1
+
+int event_open_with_options(struct event *e, pid_t pid, int cpu, int group_fd)
+{
+ e->fd = perf_event_open(&e->attr, pid, cpu, group_fd, 0);
+ if (e->fd == -1) {
+ perror("perf_event_open");
+ return -1;
+ }
+
+ return 0;
+}
+
+int event_open_with_group(struct event *e, int group_fd)
+{
+ return event_open_with_options(e, PERF_CURRENT_PID, PERF_NO_CPU, group_fd);
+}
+
+int event_open_with_pid(struct event *e, pid_t pid)
+{
+ return event_open_with_options(e, pid, PERF_NO_CPU, PERF_NO_GROUP);
+}
+
+int event_open_with_cpu(struct event *e, int cpu)
+{
+ return event_open_with_options(e, PERF_NO_PID, cpu, PERF_NO_GROUP);
+}
+
+int event_open(struct event *e)
+{
+ return event_open_with_options(e, PERF_CURRENT_PID, PERF_NO_CPU, PERF_NO_GROUP);
+}
+
+void event_close(struct event *e)
+{
+ close(e->fd);
+}
+
+int event_enable(struct event *e)
+{
+ return ioctl(e->fd, PERF_EVENT_IOC_ENABLE);
+}
+
+int event_disable(struct event *e)
+{
+ return ioctl(e->fd, PERF_EVENT_IOC_DISABLE);
+}
+
+int event_reset(struct event *e)
+{
+ return ioctl(e->fd, PERF_EVENT_IOC_RESET);
+}
+
+int event_read(struct event *e)
+{
+ int rc;
+
+ rc = read(e->fd, &e->result, sizeof(e->result));
+ if (rc != sizeof(e->result)) {
+ fprintf(stderr, "read error on event %p!\n", e);
+ return -1;
+ }
+
+ return 0;
+}
+
+void event_report_justified(struct event *e, int name_width, int result_width)
+{
+ printf("%*s: result %*llu ", name_width, e->name, result_width,
+ e->result.value);
+
+ if (e->result.running == e->result.enabled)
+ printf("running/enabled %llu\n", e->result.running);
+ else
+ printf("running %llu enabled %llu\n", e->result.running,
+ e->result.enabled);
+}
+
+void event_report(struct event *e)
+{
+ event_report_justified(e, 0, 0);
+}
diff --git a/powerpc/pmu/event.h b/powerpc/pmu/event.h
new file mode 100644
index 0000000..a0ea6b1
--- /dev/null
+++ b/powerpc/pmu/event.h
@@ -0,0 +1,43 @@
+/*
+ * Copyright 2013, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#ifndef _SELFTESTS_POWERPC_PMU_EVENT_H
+#define _SELFTESTS_POWERPC_PMU_EVENT_H
+
+#include <unistd.h>
+#include <linux/perf_event.h>
+
+#include "utils.h"
+
+
+struct event {
+ struct perf_event_attr attr;
+ char *name;
+ int fd;
+ /* This must match the read_format we use */
+ struct {
+ u64 value;
+ u64 running;
+ u64 enabled;
+ } result;
+};
+
+void event_init(struct event *e, u64 config);
+void event_init_named(struct event *e, u64 config, char *name);
+void event_init_opts(struct event *e, u64 config, int type, char *name);
+int event_open_with_options(struct event *e, pid_t pid, int cpu, int group_fd);
+int event_open_with_group(struct event *e, int group_fd);
+int event_open_with_pid(struct event *e, pid_t pid);
+int event_open_with_cpu(struct event *e, int cpu);
+int event_open(struct event *e);
+void event_close(struct event *e);
+int event_enable(struct event *e);
+int event_disable(struct event *e);
+int event_reset(struct event *e);
+int event_read(struct event *e);
+void event_report_justified(struct event *e, int name_width, int result_width);
+void event_report(struct event *e);
+
+#endif /* _SELFTESTS_POWERPC_PMU_EVENT_H */
diff --git a/powerpc/pmu/l3_bank_test.c b/powerpc/pmu/l3_bank_test.c
new file mode 100644
index 0000000..77472f3
--- /dev/null
+++ b/powerpc/pmu/l3_bank_test.c
@@ -0,0 +1,48 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+
+#include "event.h"
+#include "utils.h"
+
+#define MALLOC_SIZE (0x10000 * 10) /* Ought to be enough .. */
+
+/*
+ * Tests that the L3 bank handling is correct. We fixed it in commit e9aaac1.
+ */
+static int l3_bank_test(void)
+{
+ struct event event;
+ char *p;
+ int i;
+
+ p = malloc(MALLOC_SIZE);
+ FAIL_IF(!p);
+
+ event_init(&event, 0x84918F);
+
+ FAIL_IF(event_open(&event));
+
+ for (i = 0; i < MALLOC_SIZE; i += 0x10000)
+ p[i] = i;
+
+ event_read(&event);
+ event_report(&event);
+
+ FAIL_IF(event.result.running == 0);
+ FAIL_IF(event.result.enabled == 0);
+
+ event_close(&event);
+ free(p);
+
+ return 0;
+}
+
+int main(void)
+{
+ return test_harness(l3_bank_test, "l3_bank_test");
+}
diff --git a/powerpc/pmu/lib.c b/powerpc/pmu/lib.c
new file mode 100644
index 0000000..9768dea
--- /dev/null
+++ b/powerpc/pmu/lib.c
@@ -0,0 +1,300 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#define _GNU_SOURCE /* For CPU_ZERO etc. */
+
+#include <elf.h>
+#include <errno.h>
+#include <fcntl.h>
+#include <link.h>
+#include <sched.h>
+#include <setjmp.h>
+#include <stdlib.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+
+#include "utils.h"
+#include "lib.h"
+
+
+int pick_online_cpu(void)
+{
+ cpu_set_t mask;
+ int cpu;
+
+ CPU_ZERO(&mask);
+
+ if (sched_getaffinity(0, sizeof(mask), &mask)) {
+ perror("sched_getaffinity");
+ return -1;
+ }
+
+ /* We prefer a primary thread, but skip 0 */
+ for (cpu = 8; cpu < CPU_SETSIZE; cpu += 8)
+ if (CPU_ISSET(cpu, &mask))
+ return cpu;
+
+ /* Search for anything, but in reverse */
+ for (cpu = CPU_SETSIZE - 1; cpu >= 0; cpu--)
+ if (CPU_ISSET(cpu, &mask))
+ return cpu;
+
+ printf("No cpus in affinity mask?!\n");
+ return -1;
+}
+
+int bind_to_cpu(int cpu)
+{
+ cpu_set_t mask;
+
+ printf("Binding to cpu %d\n", cpu);
+
+ CPU_ZERO(&mask);
+ CPU_SET(cpu, &mask);
+
+ return sched_setaffinity(0, sizeof(mask), &mask);
+}
+
+#define PARENT_TOKEN 0xAA
+#define CHILD_TOKEN 0x55
+
+int sync_with_child(union pipe read_pipe, union pipe write_pipe)
+{
+ char c = PARENT_TOKEN;
+
+ FAIL_IF(write(write_pipe.write_fd, &c, 1) != 1);
+ FAIL_IF(read(read_pipe.read_fd, &c, 1) != 1);
+ if (c != CHILD_TOKEN) /* sometimes expected */
+ return 1;
+
+ return 0;
+}
+
+int wait_for_parent(union pipe read_pipe)
+{
+ char c;
+
+ FAIL_IF(read(read_pipe.read_fd, &c, 1) != 1);
+ FAIL_IF(c != PARENT_TOKEN);
+
+ return 0;
+}
+
+int notify_parent(union pipe write_pipe)
+{
+ char c = CHILD_TOKEN;
+
+ FAIL_IF(write(write_pipe.write_fd, &c, 1) != 1);
+
+ return 0;
+}
+
+int notify_parent_of_error(union pipe write_pipe)
+{
+ char c = ~CHILD_TOKEN;
+
+ FAIL_IF(write(write_pipe.write_fd, &c, 1) != 1);
+
+ return 0;
+}
+
+int wait_for_child(pid_t child_pid)
+{
+ int rc;
+
+ if (waitpid(child_pid, &rc, 0) == -1) {
+ perror("waitpid");
+ return 1;
+ }
+
+ if (WIFEXITED(rc))
+ rc = WEXITSTATUS(rc);
+ else
+ rc = 1; /* Signal or other */
+
+ return rc;
+}
+
+int kill_child_and_wait(pid_t child_pid)
+{
+ kill(child_pid, SIGTERM);
+
+ return wait_for_child(child_pid);
+}
+
+static int eat_cpu_child(union pipe read_pipe, union pipe write_pipe)
+{
+ volatile int i = 0;
+
+ /*
+ * We are just here to eat cpu and die. So make sure we can be killed,
+ * and also don't do any custom SIGTERM handling.
+ */
+ signal(SIGTERM, SIG_DFL);
+
+ notify_parent(write_pipe);
+ wait_for_parent(read_pipe);
+
+ /* Soak up cpu forever */
+ while (1) i++;
+
+ return 0;
+}
+
+pid_t eat_cpu(int (test_function)(void))
+{
+ union pipe read_pipe, write_pipe;
+ int cpu, rc;
+ pid_t pid;
+
+ cpu = pick_online_cpu();
+ FAIL_IF(cpu < 0);
+ FAIL_IF(bind_to_cpu(cpu));
+
+ if (pipe(read_pipe.fds) == -1)
+ return -1;
+
+ if (pipe(write_pipe.fds) == -1)
+ return -1;
+
+ pid = fork();
+ if (pid == 0)
+ exit(eat_cpu_child(write_pipe, read_pipe));
+
+ if (sync_with_child(read_pipe, write_pipe)) {
+ rc = -1;
+ goto out;
+ }
+
+ printf("main test running as pid %d\n", getpid());
+
+ rc = test_function();
+out:
+ kill(pid, SIGKILL);
+
+ return rc;
+}
+
+struct addr_range libc, vdso;
+
+int parse_proc_maps(void)
+{
+ unsigned long start, end;
+ char execute, name[128];
+ FILE *f;
+ int rc;
+
+ f = fopen("/proc/self/maps", "r");
+ if (!f) {
+ perror("fopen");
+ return -1;
+ }
+
+ do {
+ /* This skips line with no executable which is what we want */
+ rc = fscanf(f, "%lx-%lx %*c%*c%c%*c %*x %*d:%*d %*d %127s\n",
+ &start, &end, &execute, name);
+ if (rc <= 0)
+ break;
+
+ if (execute != 'x')
+ continue;
+
+ if (strstr(name, "libc")) {
+ libc.first = start;
+ libc.last = end - 1;
+ } else if (strstr(name, "[vdso]")) {
+ vdso.first = start;
+ vdso.last = end - 1;
+ }
+ } while(1);
+
+ fclose(f);
+
+ return 0;
+}
+
+#define PARANOID_PATH "/proc/sys/kernel/perf_event_paranoid"
+
+bool require_paranoia_below(int level)
+{
+ unsigned long current;
+ char *end, buf[16];
+ FILE *f;
+ int rc;
+
+ rc = -1;
+
+ f = fopen(PARANOID_PATH, "r");
+ if (!f) {
+ perror("fopen");
+ goto out;
+ }
+
+ if (!fgets(buf, sizeof(buf), f)) {
+ printf("Couldn't read " PARANOID_PATH "?\n");
+ goto out_close;
+ }
+
+ current = strtoul(buf, &end, 10);
+
+ if (end == buf) {
+ printf("Couldn't parse " PARANOID_PATH "?\n");
+ goto out_close;
+ }
+
+ if (current >= level)
+ goto out;
+
+ rc = 0;
+out_close:
+ fclose(f);
+out:
+ return rc;
+}
+
+static char auxv[4096];
+
+void *get_auxv_entry(int type)
+{
+ ElfW(auxv_t) *p;
+ void *result;
+ ssize_t num;
+ int fd;
+
+ fd = open("/proc/self/auxv", O_RDONLY);
+ if (fd == -1) {
+ perror("open");
+ return NULL;
+ }
+
+ result = NULL;
+
+ num = read(fd, auxv, sizeof(auxv));
+ if (num < 0) {
+ perror("read");
+ goto out;
+ }
+
+ if (num > sizeof(auxv)) {
+ printf("Overflowed auxv buffer\n");
+ goto out;
+ }
+
+ p = (ElfW(auxv_t) *)auxv;
+
+ while (p->a_type != AT_NULL) {
+ if (p->a_type == type) {
+ result = (void *)p->a_un.a_val;
+ break;
+ }
+
+ p++;
+ }
+out:
+ close(fd);
+ return result;
+}
diff --git a/powerpc/pmu/lib.h b/powerpc/pmu/lib.h
new file mode 100644
index 0000000..0f0339c
--- /dev/null
+++ b/powerpc/pmu/lib.h
@@ -0,0 +1,42 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#ifndef __SELFTESTS_POWERPC_PMU_LIB_H
+#define __SELFTESTS_POWERPC_PMU_LIB_H
+
+#include <stdio.h>
+#include <stdint.h>
+#include <string.h>
+#include <unistd.h>
+
+union pipe {
+ struct {
+ int read_fd;
+ int write_fd;
+ };
+ int fds[2];
+};
+
+extern int pick_online_cpu(void);
+extern int bind_to_cpu(int cpu);
+extern int kill_child_and_wait(pid_t child_pid);
+extern int wait_for_child(pid_t child_pid);
+extern int sync_with_child(union pipe read_pipe, union pipe write_pipe);
+extern int wait_for_parent(union pipe read_pipe);
+extern int notify_parent(union pipe write_pipe);
+extern int notify_parent_of_error(union pipe write_pipe);
+extern pid_t eat_cpu(int (test_function)(void));
+extern bool require_paranoia_below(int level);
+extern void *get_auxv_entry(int type);
+
+struct addr_range {
+ uint64_t first, last;
+};
+
+extern struct addr_range libc, vdso;
+
+int parse_proc_maps(void);
+
+#endif /* __SELFTESTS_POWERPC_PMU_LIB_H */
diff --git a/powerpc/pmu/loop.S b/powerpc/pmu/loop.S
new file mode 100644
index 0000000..20c1f08
--- /dev/null
+++ b/powerpc/pmu/loop.S
@@ -0,0 +1,43 @@
+/*
+ * Copyright 2013, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <ppc-asm.h>
+
+ .text
+
+FUNC_START(thirty_two_instruction_loop)
+ cmpdi r3,0
+ beqlr
+ addi r4,r3,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1
+ addi r4,r4,1 # 28 addi's
+ subi r3,r3,1
+ b FUNC_NAME(thirty_two_instruction_loop)
+FUNC_END(thirty_two_instruction_loop)
diff --git a/powerpc/pmu/per_event_excludes.c b/powerpc/pmu/per_event_excludes.c
new file mode 100644
index 0000000..fddbbc9
--- /dev/null
+++ b/powerpc/pmu/per_event_excludes.c
@@ -0,0 +1,114 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#define _GNU_SOURCE
+
+#include <elf.h>
+#include <limits.h>
+#include <stdio.h>
+#include <stdbool.h>
+#include <string.h>
+#include <sys/prctl.h>
+
+#include "event.h"
+#include "lib.h"
+#include "utils.h"
+
+/*
+ * Test that per-event excludes work.
+ */
+
+static int per_event_excludes(void)
+{
+ struct event *e, events[4];
+ char *platform;
+ int i;
+
+ platform = (char *)get_auxv_entry(AT_BASE_PLATFORM);
+ FAIL_IF(!platform);
+ SKIP_IF(strcmp(platform, "power8") != 0);
+
+ /*
+ * We need to create the events disabled, otherwise the running/enabled
+ * counts don't match up.
+ */
+ e = &events[0];
+ event_init_opts(e, PERF_COUNT_HW_INSTRUCTIONS,
+ PERF_TYPE_HARDWARE, "instructions");
+ e->attr.disabled = 1;
+
+ e = &events[1];
+ event_init_opts(e, PERF_COUNT_HW_INSTRUCTIONS,
+ PERF_TYPE_HARDWARE, "instructions(k)");
+ e->attr.disabled = 1;
+ e->attr.exclude_user = 1;
+ e->attr.exclude_hv = 1;
+
+ e = &events[2];
+ event_init_opts(e, PERF_COUNT_HW_INSTRUCTIONS,
+ PERF_TYPE_HARDWARE, "instructions(h)");
+ e->attr.disabled = 1;
+ e->attr.exclude_user = 1;
+ e->attr.exclude_kernel = 1;
+
+ e = &events[3];
+ event_init_opts(e, PERF_COUNT_HW_INSTRUCTIONS,
+ PERF_TYPE_HARDWARE, "instructions(u)");
+ e->attr.disabled = 1;
+ e->attr.exclude_hv = 1;
+ e->attr.exclude_kernel = 1;
+
+ FAIL_IF(event_open(&events[0]));
+
+ /*
+ * The open here will fail if we don't have per event exclude support,
+ * because the second event has an incompatible set of exclude settings
+ * and we're asking for the events to be in a group.
+ */
+ for (i = 1; i < 4; i++)
+ FAIL_IF(event_open_with_group(&events[i], events[0].fd));
+
+ /*
+ * Even though the above will fail without per-event excludes we keep
+ * testing in order to be thorough.
+ */
+ prctl(PR_TASK_PERF_EVENTS_ENABLE);
+
+ /* Spin for a while */
+ for (i = 0; i < INT_MAX; i++)
+ asm volatile("" : : : "memory");
+
+ prctl(PR_TASK_PERF_EVENTS_DISABLE);
+
+ for (i = 0; i < 4; i++) {
+ FAIL_IF(event_read(&events[i]));
+ event_report(&events[i]);
+ }
+
+ /*
+ * We should see that all events have enabled == running. That
+ * shows that they were all on the PMU at once.
+ */
+ for (i = 0; i < 4; i++)
+ FAIL_IF(events[i].result.running != events[i].result.enabled);
+
+ /*
+ * We can also check that the result for instructions is >= all the
+ * other counts. That's because it is counting all instructions while
+ * the others are counting a subset.
+ */
+ for (i = 1; i < 4; i++)
+ FAIL_IF(events[0].result.value < events[i].result.value);
+
+ for (i = 0; i < 4; i++)
+ event_close(&events[i]);
+
+ return 0;
+}
+
+int main(void)
+{
+ return test_harness(per_event_excludes, "per_event_excludes");
+}
diff --git a/powerpc/subunit.h b/powerpc/subunit.h
new file mode 100644
index 0000000..9c6c4e9
--- /dev/null
+++ b/powerpc/subunit.h
@@ -0,0 +1,52 @@
+/*
+ * Copyright 2013, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#ifndef _SELFTESTS_POWERPC_SUBUNIT_H
+#define _SELFTESTS_POWERPC_SUBUNIT_H
+
+static inline void test_start(char *name)
+{
+ printf("test: %s\n", name);
+}
+
+static inline void test_failure_detail(char *name, char *detail)
+{
+ printf("failure: %s [%s]\n", name, detail);
+}
+
+static inline void test_failure(char *name)
+{
+ printf("failure: %s\n", name);
+}
+
+static inline void test_error(char *name)
+{
+ printf("error: %s\n", name);
+}
+
+static inline void test_skip(char *name)
+{
+ printf("skip: %s\n", name);
+}
+
+static inline void test_success(char *name)
+{
+ printf("success: %s\n", name);
+}
+
+static inline void test_finish(char *name, int status)
+{
+ if (status)
+ test_failure(name);
+ else
+ test_success(name);
+}
+
+static inline void test_set_git_version(char *value)
+{
+ printf("tags: git_version:%s\n", value);
+}
+
+#endif /* _SELFTESTS_POWERPC_SUBUNIT_H */
diff --git a/powerpc/tm/Makefile b/powerpc/tm/Makefile
new file mode 100644
index 0000000..2cede23
--- /dev/null
+++ b/powerpc/tm/Makefile
@@ -0,0 +1,15 @@
+PROGS := tm-resched-dscr
+
+all: $(PROGS)
+
+$(PROGS): ../harness.c
+
+run_tests: all
+ @-for PROG in $(PROGS); do \
+ ./$$PROG; \
+ done;
+
+clean:
+ rm -f $(PROGS) *.o
+
+.PHONY: all run_tests clean
diff --git a/powerpc/tm/tm-resched-dscr.c b/powerpc/tm/tm-resched-dscr.c
new file mode 100644
index 0000000..42d4c8c
--- /dev/null
+++ b/powerpc/tm/tm-resched-dscr.c
@@ -0,0 +1,98 @@
+/* Test context switching to see if the DSCR SPR is correctly preserved
+ * when within a transaction.
+ *
+ * Note: We assume that the DSCR has been left at the default value (0)
+ * for all CPUs.
+ *
+ * Method:
+ *
+ * Set a value into the DSCR.
+ *
+ * Start a transaction, and suspend it (*).
+ *
+ * Hard loop checking to see if the transaction has become doomed.
+ *
+ * Now that we *may* have been preempted, record the DSCR and TEXASR SPRS.
+ *
+ * If the abort was because of a context switch, check the DSCR value.
+ * Otherwise, try again.
+ *
+ * (*) If the transaction is not suspended we can't see the problem because
+ * the transaction abort handler will restore the DSCR to it's checkpointed
+ * value before we regain control.
+ */
+
+#include <inttypes.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <assert.h>
+#include <asm/tm.h>
+
+#include "utils.h"
+
+#define TBEGIN ".long 0x7C00051D ;"
+#define TEND ".long 0x7C00055D ;"
+#define TCHECK ".long 0x7C00059C ;"
+#define TSUSPEND ".long 0x7C0005DD ;"
+#define TRESUME ".long 0x7C2005DD ;"
+#define SPRN_TEXASR 0x82
+#define SPRN_DSCR 0x03
+
+int test_body(void)
+{
+ uint64_t rv, dscr1 = 1, dscr2, texasr;
+
+ printf("Check DSCR TM context switch: ");
+ fflush(stdout);
+ for (;;) {
+ rv = 1;
+ asm __volatile__ (
+ /* set a known value into the DSCR */
+ "ld 3, %[dscr1];"
+ "mtspr %[sprn_dscr], 3;"
+
+ /* start and suspend a transaction */
+ TBEGIN
+ "beq 1f;"
+ TSUSPEND
+
+ /* hard loop until the transaction becomes doomed */
+ "2: ;"
+ TCHECK
+ "bc 4, 0, 2b;"
+
+ /* record DSCR and TEXASR */
+ "mfspr 3, %[sprn_dscr];"
+ "std 3, %[dscr2];"
+ "mfspr 3, %[sprn_texasr];"
+ "std 3, %[texasr];"
+
+ TRESUME
+ TEND
+ "li %[rv], 0;"
+ "1: ;"
+ : [rv]"=r"(rv), [dscr2]"=m"(dscr2), [texasr]"=m"(texasr)
+ : [dscr1]"m"(dscr1)
+ , [sprn_dscr]"i"(SPRN_DSCR), [sprn_texasr]"i"(SPRN_TEXASR)
+ : "memory", "r3"
+ );
+ assert(rv); /* make sure the transaction aborted */
+ if ((texasr >> 56) != TM_CAUSE_RESCHED) {
+ putchar('.');
+ fflush(stdout);
+ continue;
+ }
+ if (dscr2 != dscr1) {
+ printf(" FAIL\n");
+ return 1;
+ } else {
+ printf(" OK\n");
+ return 0;
+ }
+ }
+}
+
+int main(void)
+{
+ return test_harness(test_body, "tm_resched_dscr");
+}
diff --git a/powerpc/utils.h b/powerpc/utils.h
new file mode 100644
index 0000000..a93777a
--- /dev/null
+++ b/powerpc/utils.h
@@ -0,0 +1,49 @@
+/*
+ * Copyright 2013, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#ifndef _SELFTESTS_POWERPC_UTILS_H
+#define _SELFTESTS_POWERPC_UTILS_H
+
+#include <stdint.h>
+#include <stdbool.h>
+
+/* Avoid headaches with PRI?64 - just use %ll? always */
+typedef unsigned long long u64;
+typedef signed long long s64;
+
+/* Just for familiarity */
+typedef uint32_t u32;
+typedef uint8_t u8;
+
+
+int test_harness(int (test_function)(void), char *name);
+
+
+/* Yes, this is evil */
+#define FAIL_IF(x) \
+do { \
+ if ((x)) { \
+ fprintf(stderr, \
+ "[FAIL] Test FAILED on line %d\n", __LINE__); \
+ return 1; \
+ } \
+} while (0)
+
+/* The test harness uses this, yes it's gross */
+#define MAGIC_SKIP_RETURN_VALUE 99
+
+#define SKIP_IF(x) \
+do { \
+ if ((x)) { \
+ fprintf(stderr, \
+ "[SKIP] Test skipped on line %d\n", __LINE__); \
+ return MAGIC_SKIP_RETURN_VALUE; \
+ } \
+} while (0)
+
+#define _str(s) #s
+#define str(s) _str(s)
+
+#endif /* _SELFTESTS_POWERPC_UTILS_H */
diff --git a/ptrace/Makefile b/ptrace/Makefile
new file mode 100644
index 0000000..47ae2d3
--- /dev/null
+++ b/ptrace/Makefile
@@ -0,0 +1,10 @@
+CFLAGS += -iquote../../../../include/uapi -Wall
+peeksiginfo: peeksiginfo.c
+
+all: peeksiginfo
+
+clean:
+ rm -f peeksiginfo
+
+run_tests: all
+ @./peeksiginfo || echo "peeksiginfo selftests: [FAIL]"
diff --git a/ptrace/peeksiginfo.c b/ptrace/peeksiginfo.c
new file mode 100644
index 0000000..c34cd8a
--- /dev/null
+++ b/ptrace/peeksiginfo.c
@@ -0,0 +1,218 @@
+#define _GNU_SOURCE
+#include <stdio.h>
+#include <signal.h>
+#include <unistd.h>
+#include <errno.h>
+#include <linux/types.h>
+#include <sys/wait.h>
+#include <sys/syscall.h>
+#include <sys/user.h>
+#include <sys/mman.h>
+
+#include "linux/ptrace.h"
+
+static int sys_rt_sigqueueinfo(pid_t tgid, int sig, siginfo_t *uinfo)
+{
+ return syscall(SYS_rt_sigqueueinfo, tgid, sig, uinfo);
+}
+
+static int sys_rt_tgsigqueueinfo(pid_t tgid, pid_t tid,
+ int sig, siginfo_t *uinfo)
+{
+ return syscall(SYS_rt_tgsigqueueinfo, tgid, tid, sig, uinfo);
+}
+
+static int sys_ptrace(int request, pid_t pid, void *addr, void *data)
+{
+ return syscall(SYS_ptrace, request, pid, addr, data);
+}
+
+#define SIGNR 10
+#define TEST_SICODE_PRIV -1
+#define TEST_SICODE_SHARE -2
+
+#ifndef PAGE_SIZE
+#define PAGE_SIZE sysconf(_SC_PAGESIZE)
+#endif
+
+#define err(fmt, ...) \
+ fprintf(stderr, \
+ "Error (%s:%d): " fmt, \
+ __FILE__, __LINE__, ##__VA_ARGS__)
+
+static int check_error_paths(pid_t child)
+{
+ struct ptrace_peeksiginfo_args arg;
+ int ret, exit_code = -1;
+ void *addr_rw, *addr_ro;
+
+ /*
+ * Allocate two contiguous pages. The first one is for read-write,
+ * another is for read-only.
+ */
+ addr_rw = mmap(NULL, 2 * PAGE_SIZE, PROT_READ | PROT_WRITE,
+ MAP_PRIVATE | MAP_ANONYMOUS, -1, 0);
+ if (addr_rw == MAP_FAILED) {
+ err("mmap() failed: %m\n");
+ return 1;
+ }
+
+ addr_ro = mmap(addr_rw + PAGE_SIZE, PAGE_SIZE, PROT_READ,
+ MAP_PRIVATE | MAP_ANONYMOUS | MAP_FIXED, -1, 0);
+ if (addr_ro == MAP_FAILED) {
+ err("mmap() failed: %m\n");
+ goto out;
+ }
+
+ arg.nr = SIGNR;
+ arg.off = 0;
+
+ /* Unsupported flags */
+ arg.flags = ~0;
+ ret = sys_ptrace(PTRACE_PEEKSIGINFO, child, &arg, addr_rw);
+ if (ret != -1 || errno != EINVAL) {
+ err("sys_ptrace() returns %d (expected -1),"
+ " errno %d (expected %d): %m\n",
+ ret, errno, EINVAL);
+ goto out;
+ }
+ arg.flags = 0;
+
+ /* A part of the buffer is read-only */
+ ret = sys_ptrace(PTRACE_PEEKSIGINFO, child, &arg,
+ addr_ro - sizeof(siginfo_t) * 2);
+ if (ret != 2) {
+ err("sys_ptrace() returns %d (expected 2): %m\n", ret);
+ goto out;
+ }
+
+ /* Read-only buffer */
+ ret = sys_ptrace(PTRACE_PEEKSIGINFO, child, &arg, addr_ro);
+ if (ret != -1 && errno != EFAULT) {
+ err("sys_ptrace() returns %d (expected -1),"
+ " errno %d (expected %d): %m\n",
+ ret, errno, EFAULT);
+ goto out;
+ }
+
+ exit_code = 0;
+out:
+ munmap(addr_rw, 2 * PAGE_SIZE);
+ return exit_code;
+}
+
+int check_direct_path(pid_t child, int shared, int nr)
+{
+ struct ptrace_peeksiginfo_args arg = {.flags = 0, .nr = nr, .off = 0};
+ int i, j, ret, exit_code = -1;
+ siginfo_t siginfo[SIGNR];
+ int si_code;
+
+ if (shared == 1) {
+ arg.flags = PTRACE_PEEKSIGINFO_SHARED;
+ si_code = TEST_SICODE_SHARE;
+ } else {
+ arg.flags = 0;
+ si_code = TEST_SICODE_PRIV;
+ }
+
+ for (i = 0; i < SIGNR; ) {
+ arg.off = i;
+ ret = sys_ptrace(PTRACE_PEEKSIGINFO, child, &arg, siginfo);
+ if (ret == -1) {
+ err("ptrace() failed: %m\n");
+ goto out;
+ }
+
+ if (ret == 0)
+ break;
+
+ for (j = 0; j < ret; j++, i++) {
+ if (siginfo[j].si_code == si_code &&
+ siginfo[j].si_int == i)
+ continue;
+
+ err("%d: Wrong siginfo i=%d si_code=%d si_int=%d\n",
+ shared, i, siginfo[j].si_code, siginfo[j].si_int);
+ goto out;
+ }
+ }
+
+ if (i != SIGNR) {
+ err("Only %d signals were read\n", i);
+ goto out;
+ }
+
+ exit_code = 0;
+out:
+ return exit_code;
+}
+
+int main(int argc, char *argv[])
+{
+ siginfo_t siginfo[SIGNR];
+ int i, exit_code = 1;
+ sigset_t blockmask;
+ pid_t child;
+
+ sigemptyset(&blockmask);
+ sigaddset(&blockmask, SIGRTMIN);
+ sigprocmask(SIG_BLOCK, &blockmask, NULL);
+
+ child = fork();
+ if (child == -1) {
+ err("fork() failed: %m");
+ return 1;
+ } else if (child == 0) {
+ pid_t ppid = getppid();
+ while (1) {
+ if (ppid != getppid())
+ break;
+ sleep(1);
+ }
+ return 1;
+ }
+
+ /* Send signals in process-wide and per-thread queues */
+ for (i = 0; i < SIGNR; i++) {
+ siginfo->si_code = TEST_SICODE_SHARE;
+ siginfo->si_int = i;
+ sys_rt_sigqueueinfo(child, SIGRTMIN, siginfo);
+
+ siginfo->si_code = TEST_SICODE_PRIV;
+ siginfo->si_int = i;
+ sys_rt_tgsigqueueinfo(child, child, SIGRTMIN, siginfo);
+ }
+
+ if (sys_ptrace(PTRACE_ATTACH, child, NULL, NULL) == -1)
+ return 1;
+
+ waitpid(child, NULL, 0);
+
+ /* Dump signals one by one*/
+ if (check_direct_path(child, 0, 1))
+ goto out;
+ /* Dump all signals for one call */
+ if (check_direct_path(child, 0, SIGNR))
+ goto out;
+
+ /*
+ * Dump signal from the process-wide queue.
+ * The number of signals is not multible to the buffer size
+ */
+ if (check_direct_path(child, 1, 3))
+ goto out;
+
+ if (check_error_paths(child))
+ goto out;
+
+ printf("PASS\n");
+ exit_code = 0;
+out:
+ if (sys_ptrace(PTRACE_KILL, child, NULL, NULL) == -1)
+ return 1;
+
+ waitpid(child, NULL, 0);
+
+ return exit_code;
+}
diff --git a/rcutorture/.gitignore b/rcutorture/.gitignore
new file mode 100644
index 0000000..05838f6
--- /dev/null
+++ b/rcutorture/.gitignore
@@ -0,0 +1,6 @@
+initrd
+linux-2.6
+b[0-9]*
+rcu-test-image
+res
+*.swp
diff --git a/rcutorture/bin/config2frag.sh b/rcutorture/bin/config2frag.sh
new file mode 100644
index 0000000..9f9ffcd
--- /dev/null
+++ b/rcutorture/bin/config2frag.sh
@@ -0,0 +1,25 @@
+#!/bin/sh
+# Usage: sh config2frag.sh < .config > configfrag
+#
+# Converts the "# CONFIG_XXX is not set" to "CONFIG_XXX=n" so that the
+# resulting file becomes a legitimate Kconfig fragment.
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, you can access it online at
+# http://www.gnu.org/licenses/gpl-2.0.html.
+#
+# Copyright (C) IBM Corporation, 2013
+#
+# Authors: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
+
+LANG=C sed -e 's/^# CONFIG_\([a-zA-Z0-9_]*\) is not set$/CONFIG_\1=n/'
diff --git a/rcutorture/bin/configNR_CPUS.sh b/rcutorture/bin/configNR_CPUS.sh
new file mode 100755
index 0000000..43540f1
--- /dev/null
+++ b/rcutorture/bin/configNR_CPUS.sh
@@ -0,0 +1,45 @@
+#!/bin/bash
+#
+# Extract the number of CPUs expected from the specified Kconfig-file
+# fragment by checking CONFIG_SMP and CONFIG_NR_CPUS. If the specified
+# file gives no clue, base the number on the number of idle CPUs on
+# the system.
+#
+# Usage: configNR_CPUS.sh config-frag
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, you can access it online at
+# http://www.gnu.org/licenses/gpl-2.0.html.
+#
+# Copyright (C) IBM Corporation, 2013
+#
+# Authors: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
+
+cf=$1
+if test ! -r $cf
+then
+ echo Unreadable config fragment $cf 1>&2
+ exit -1
+fi
+if grep -q '^CONFIG_SMP=n$' $cf
+then
+ echo 1
+ exit 0
+fi
+if grep -q '^CONFIG_NR_CPUS=' $cf
+then
+ grep '^CONFIG_NR_CPUS=' $cf |
+ sed -e 's/^CONFIG_NR_CPUS=\([0-9]*\).*$/\1/'
+ exit 0
+fi
+cpus2use.sh
diff --git a/rcutorture/bin/configcheck.sh b/rcutorture/bin/configcheck.sh
new file mode 100755
index 0000000..d686537
--- /dev/null
+++ b/rcutorture/bin/configcheck.sh
@@ -0,0 +1,54 @@
+#!/bin/sh
+# Usage: sh configcheck.sh .config .config-template
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, you can access it online at
+# http://www.gnu.org/licenses/gpl-2.0.html.
+#
+# Copyright (C) IBM Corporation, 2011
+#
+# Authors: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
+
+T=/tmp/abat-chk-config.sh.$$
+trap 'rm -rf $T' 0
+mkdir $T
+
+cat $1 > $T/.config
+
+cat $2 | sed -e 's/\(.*\)=n/# \1 is not set/' -e 's/^#CHECK#//' |
+awk '
+BEGIN {
+ print "if grep -q \"" $0 "\" < '"$T/.config"'";
+ print "then";
+ print "\t:";
+ print "else";
+ if ($1 == "#") {
+ print "\tif grep -q \"" $2 "\" < '"$T/.config"'";
+ print "\tthen";
+ print "\t\tif test \"$firsttime\" = \"\""
+ print "\t\tthen"
+ print "\t\t\tfirsttime=1"
+ print "\t\tfi"
+ print "\t\techo \":" $2 ": improperly set\"";
+ print "\telse";
+ print "\t\t:";
+ print "\tfi";
+ } else {
+ print "\tif test \"$firsttime\" = \"\""
+ print "\tthen"
+ print "\t\tfirsttime=1"
+ print "\tfi"
+ print "\techo \":" $0 ": improperly set\"";
+ }
+ print "fi";
+ }' | sh
diff --git a/rcutorture/bin/configinit.sh b/rcutorture/bin/configinit.sh
new file mode 100755
index 0000000..9c3f3d3
--- /dev/null
+++ b/rcutorture/bin/configinit.sh
@@ -0,0 +1,74 @@
+#!/bin/sh
+#
+# sh configinit.sh config-spec-file [ build output dir ]
+#
+# Create a .config file from the spec file. Run from the kernel source tree.
+# Exits with 0 if all went well, with 1 if all went well but the config
+# did not match, and some other number for other failures.
+#
+# The first argument is the .config specification file, which contains
+# desired settings, for example, "CONFIG_NO_HZ=y". For best results,
+# this should be a full pathname.
+#
+# The second argument is a optional path to a build output directory,
+# for example, "O=/tmp/foo". If this argument is omitted, the .config
+# file will be generated directly in the current directory.
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, you can access it online at
+# http://www.gnu.org/licenses/gpl-2.0.html.
+#
+# Copyright (C) IBM Corporation, 2013
+#
+# Authors: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
+
+T=/tmp/configinit.sh.$$
+trap 'rm -rf $T' 0
+mkdir $T
+
+# Capture config spec file.
+
+c=$1
+buildloc=$2
+builddir=
+if test -n $buildloc
+then
+ if echo $buildloc | grep -q '^O='
+ then
+ builddir=`echo $buildloc | sed -e 's/^O=//'`
+ if test ! -d $builddir
+ then
+ mkdir $builddir
+ fi
+ else
+ echo Bad build directory: \"$builddir\"
+ exit 2
+ fi
+fi
+
+sed -e 's/^\(CONFIG[0-9A-Z_]*\)=.*$/grep -v "^# \1" |/' < $c > $T/u.sh
+sed -e 's/^\(CONFIG[0-9A-Z_]*=\).*$/grep -v \1 |/' < $c >> $T/u.sh
+grep '^grep' < $T/u.sh > $T/upd.sh
+echo "cat - $c" >> $T/upd.sh
+make mrproper
+make $buildloc distclean > $builddir/Make.distclean 2>&1
+make $buildloc $TORTURE_DEFCONFIG > $builddir/Make.defconfig.out 2>&1
+mv $builddir/.config $builddir/.config.sav
+sh $T/upd.sh < $builddir/.config.sav > $builddir/.config
+cp $builddir/.config $builddir/.config.new
+yes '' | make $buildloc oldconfig > $builddir/Make.modconfig.out 2>&1
+
+# verify new config matches specification.
+configcheck.sh $builddir/.config $c
+
+exit 0
diff --git a/rcutorture/bin/cpus2use.sh b/rcutorture/bin/cpus2use.sh
new file mode 100755
index 0000000..abe14b7
--- /dev/null
+++ b/rcutorture/bin/cpus2use.sh
@@ -0,0 +1,41 @@
+#!/bin/bash
+#
+# Get an estimate of how CPU-hoggy to be.
+#
+# Usage: cpus2use.sh
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, you can access it online at
+# http://www.gnu.org/licenses/gpl-2.0.html.
+#
+# Copyright (C) IBM Corporation, 2013
+#
+# Authors: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
+
+ncpus=`grep '^processor' /proc/cpuinfo | wc -l`
+idlecpus=`mpstat | tail -1 | \
+ awk -v ncpus=$ncpus '{ print ncpus * ($7 + $12) / 100 }'`
+awk -v ncpus=$ncpus -v idlecpus=$idlecpus < /dev/null '
+BEGIN {
+ cpus2use = idlecpus;
+ if (cpus2use < 1)
+ cpus2use = 1;
+ if (cpus2use < ncpus / 10)
+ cpus2use = ncpus / 10;
+ if (cpus2use == int(cpus2use))
+ cpus2use = int(cpus2use)
+ else
+ cpus2use = int(cpus2use) + 1
+ print cpus2use;
+}'
+
diff --git a/rcutorture/bin/functions.sh b/rcutorture/bin/functions.sh
new file mode 100644
index 0000000..d01b865
--- /dev/null
+++ b/rcutorture/bin/functions.sh
@@ -0,0 +1,223 @@
+#!/bin/bash
+#
+# Shell functions for the rest of the scripts.
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, you can access it online at
+# http://www.gnu.org/licenses/gpl-2.0.html.
+#
+# Copyright (C) IBM Corporation, 2013
+#
+# Authors: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
+
+# bootparam_hotplug_cpu bootparam-string
+#
+# Returns 1 if the specified boot-parameter string tells rcutorture to
+# test CPU-hotplug operations.
+bootparam_hotplug_cpu () {
+ echo "$1" | grep -q "rcutorture\.onoff_"
+}
+
+# checkarg --argname argtype $# arg mustmatch cannotmatch
+#
+# Checks the specified argument "arg" against the mustmatch and cannotmatch
+# patterns.
+checkarg () {
+ if test $3 -le 1
+ then
+ echo $1 needs argument $2 matching \"$5\"
+ usage
+ fi
+ if echo "$4" | grep -q -e "$5"
+ then
+ :
+ else
+ echo $1 $2 \"$4\" must match \"$5\"
+ usage
+ fi
+ if echo "$4" | grep -q -e "$6"
+ then
+ echo $1 $2 \"$4\" must not match \"$6\"
+ usage
+ fi
+}
+
+# configfrag_boot_params bootparam-string config-fragment-file
+#
+# Adds boot parameters from the .boot file, if any.
+configfrag_boot_params () {
+ if test -r "$2.boot"
+ then
+ echo $1 `grep -v '^#' "$2.boot" | tr '\012' ' '`
+ else
+ echo $1
+ fi
+}
+
+# configfrag_hotplug_cpu config-fragment-file
+#
+# Returns 1 if the config fragment specifies hotplug CPU.
+configfrag_hotplug_cpu () {
+ if test ! -r "$1"
+ then
+ echo Unreadable config fragment "$1" 1>&2
+ exit -1
+ fi
+ grep -q '^CONFIG_HOTPLUG_CPU=y$' "$1"
+}
+
+# identify_boot_image qemu-cmd
+#
+# Returns the relative path to the kernel build image. This will be
+# arch/<arch>/boot/bzImage unless overridden with the TORTURE_BOOT_IMAGE
+# environment variable.
+identify_boot_image () {
+ if test -n "$TORTURE_BOOT_IMAGE"
+ then
+ echo $TORTURE_BOOT_IMAGE
+ else
+ case "$1" in
+ qemu-system-x86_64|qemu-system-i386)
+ echo arch/x86/boot/bzImage
+ ;;
+ qemu-system-ppc64)
+ echo arch/powerpc/boot/bzImage
+ ;;
+ *)
+ echo ""
+ ;;
+ esac
+ fi
+}
+
+# identify_qemu builddir
+#
+# Returns our best guess as to which qemu command is appropriate for
+# the kernel at hand. Override with the TORTURE_QEMU_CMD environment variable.
+identify_qemu () {
+ local u="`file "$1"`"
+ if test -n "$TORTURE_QEMU_CMD"
+ then
+ echo $TORTURE_QEMU_CMD
+ elif echo $u | grep -q x86-64
+ then
+ echo qemu-system-x86_64
+ elif echo $u | grep -q "Intel 80386"
+ then
+ echo qemu-system-i386
+ elif uname -a | grep -q ppc64
+ then
+ echo qemu-system-ppc64
+ else
+ echo Cannot figure out what qemu command to use! 1>&2
+ echo file $1 output: $u
+ # Usually this will be one of /usr/bin/qemu-system-*
+ # Use TORTURE_QEMU_CMD environment variable or appropriate
+ # argument to top-level script.
+ exit 1
+ fi
+}
+
+# identify_qemu_append qemu-cmd
+#
+# Output arguments for the qemu "-append" string based on CPU type
+# and the TORTURE_QEMU_INTERACTIVE environment variable.
+identify_qemu_append () {
+ case "$1" in
+ qemu-system-x86_64|qemu-system-i386)
+ echo noapic selinux=0 initcall_debug debug
+ ;;
+ esac
+ if test -n "$TORTURE_QEMU_INTERACTIVE"
+ then
+ echo root=/dev/sda
+ else
+ echo console=ttyS0
+ fi
+}
+
+# identify_qemu_args qemu-cmd serial-file
+#
+# Output arguments for qemu arguments based on the TORTURE_QEMU_MAC
+# and TORTURE_QEMU_INTERACTIVE environment variables.
+identify_qemu_args () {
+ case "$1" in
+ qemu-system-x86_64|qemu-system-i386)
+ ;;
+ qemu-system-ppc64)
+ echo -enable-kvm -M pseries -cpu POWER7 -nodefaults
+ echo -device spapr-vscsi
+ if test -n "$TORTURE_QEMU_INTERACTIVE" -a -n "$TORTURE_QEMU_MAC"
+ then
+ echo -device spapr-vlan,netdev=net0,mac=$TORTURE_QEMU_MAC
+ echo -netdev bridge,br=br0,id=net0
+ elif test -n "$TORTURE_QEMU_INTERACTIVE"
+ then
+ echo -net nic -net user
+ fi
+ ;;
+ esac
+ if test -n "$TORTURE_QEMU_INTERACTIVE"
+ then
+ echo -monitor stdio -serial pty -S
+ else
+ echo -serial file:$2
+ fi
+}
+
+# identify_qemu_vcpus
+#
+# Returns the number of virtual CPUs available to the aggregate of the
+# guest OSes.
+identify_qemu_vcpus () {
+ lscpu | grep '^CPU(s):' | sed -e 's/CPU(s)://'
+}
+
+# print_bug
+#
+# Prints "BUG: " in red followed by remaining arguments
+print_bug () {
+ printf '\033[031mBUG: \033[m'
+ echo $*
+}
+
+# print_warning
+#
+# Prints "WARNING: " in yellow followed by remaining arguments
+print_warning () {
+ printf '\033[033mWARNING: \033[m'
+ echo $*
+}
+
+# specify_qemu_cpus qemu-cmd qemu-args #cpus
+#
+# Appends a string containing "-smp XXX" to qemu-args, unless the incoming
+# qemu-args already contains "-smp".
+specify_qemu_cpus () {
+ local nt;
+
+ if echo $2 | grep -q -e -smp
+ then
+ echo $2
+ else
+ case "$1" in
+ qemu-system-x86_64|qemu-system-i386)
+ echo $2 -smp $3
+ ;;
+ qemu-system-ppc64)
+ nt="`lscpu | grep '^NUMA node0' | sed -e 's/^[^,]*,\([0-9]*\),.*$/\1/'`"
+ echo $2 -smp cores=`expr \( $3 + $nt - 1 \) / $nt`,threads=$nt
+ ;;
+ esac
+ fi
+}
diff --git a/rcutorture/bin/kvm-build.sh b/rcutorture/bin/kvm-build.sh
new file mode 100755
index 0000000..7c1e56b
--- /dev/null
+++ b/rcutorture/bin/kvm-build.sh
@@ -0,0 +1,71 @@
+#!/bin/bash
+#
+# Build a kvm-ready Linux kernel from the tree in the current directory.
+#
+# Usage: sh kvm-build.sh config-template build-dir more-configs
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, you can access it online at
+# http://www.gnu.org/licenses/gpl-2.0.html.
+#
+# Copyright (C) IBM Corporation, 2011
+#
+# Authors: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
+
+config_template=${1}
+if test -z "$config_template" -o ! -f "$config_template" -o ! -r "$config_template"
+then
+ echo "kvm-build.sh :$config_template: Not a readable file"
+ exit 1
+fi
+builddir=${2}
+if test -z "$builddir" -o ! -d "$builddir" -o ! -w "$builddir"
+then
+ echo "kvm-build.sh :$builddir: Not a writable directory, cannot build into it"
+ exit 1
+fi
+moreconfigs=${3}
+if test -z "$moreconfigs" -o ! -r "$moreconfigs"
+then
+ echo "kvm-build.sh :$moreconfigs: Not a readable file"
+ exit 1
+fi
+
+T=/tmp/test-linux.sh.$$
+trap 'rm -rf $T' 0
+mkdir $T
+
+grep -v 'CONFIG_[A-Z]*_TORTURE_TEST' < ${config_template} > $T/config
+cat << ___EOF___ >> $T/config
+CONFIG_INITRAMFS_SOURCE="$TORTURE_INITRD"
+CONFIG_VIRTIO_PCI=y
+CONFIG_VIRTIO_CONSOLE=y
+___EOF___
+cat $moreconfigs >> $T/config
+
+configinit.sh $T/config O=$builddir
+retval=$?
+if test $retval -gt 1
+then
+ exit 2
+fi
+ncpus=`cpus2use.sh`
+make O=$builddir -j$ncpus $TORTURE_KMAKE_ARG > $builddir/Make.out 2>&1
+retval=$?
+if test $retval -ne 0 || grep "rcu[^/]*": < $builddir/Make.out | egrep -q "Stop|Error|error:|warning:" || egrep -q "Stop|Error|error:" < $builddir/Make.out
+then
+ echo Kernel build error
+ egrep "Stop|Error|error:|warning:" < $builddir/Make.out
+ echo Run aborted.
+ exit 3
+fi
diff --git a/rcutorture/bin/kvm-recheck-lock.sh b/rcutorture/bin/kvm-recheck-lock.sh
new file mode 100755
index 0000000..7f1ff1a
--- /dev/null
+++ b/rcutorture/bin/kvm-recheck-lock.sh
@@ -0,0 +1,51 @@
+#!/bin/bash
+#
+# Analyze a given results directory for locktorture progress.
+#
+# Usage: sh kvm-recheck-lock.sh resdir
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, you can access it online at
+# http://www.gnu.org/licenses/gpl-2.0.html.
+#
+# Copyright (C) IBM Corporation, 2014
+#
+# Authors: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
+
+i="$1"
+if test -d $i
+then
+ :
+else
+ echo Unreadable results directory: $i
+ exit 1
+fi
+
+configfile=`echo $i | sed -e 's/^.*\///'`
+ncs=`grep "Writes: Total:" $i/console.log 2> /dev/null | tail -1 | sed -e 's/^.* Total: //' -e 's/ .*$//'`
+if test -z "$ncs"
+then
+ echo "$configfile -------"
+else
+ title="$configfile ------- $ncs acquisitions/releases"
+ dur=`sed -e 's/^.* locktorture.shutdown_secs=//' -e 's/ .*$//' < $i/qemu-cmd 2> /dev/null`
+ if test -z "$dur"
+ then
+ :
+ else
+ ncsps=`awk -v ncs=$ncs -v dur=$dur '
+ BEGIN { print ncs / dur }' < /dev/null`
+ title="$title ($ncsps per second)"
+ fi
+ echo $title
+fi
diff --git a/rcutorture/bin/kvm-recheck-rcu.sh b/rcutorture/bin/kvm-recheck-rcu.sh
new file mode 100755
index 0000000..307c4b9
--- /dev/null
+++ b/rcutorture/bin/kvm-recheck-rcu.sh
@@ -0,0 +1,51 @@
+#!/bin/bash
+#
+# Analyze a given results directory for rcutorture progress.
+#
+# Usage: sh kvm-recheck-rcu.sh resdir
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, you can access it online at
+# http://www.gnu.org/licenses/gpl-2.0.html.
+#
+# Copyright (C) IBM Corporation, 2014
+#
+# Authors: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
+
+i="$1"
+if test -d $i
+then
+ :
+else
+ echo Unreadable results directory: $i
+ exit 1
+fi
+
+configfile=`echo $i | sed -e 's/^.*\///'`
+ngps=`grep ver: $i/console.log 2> /dev/null | tail -1 | sed -e 's/^.* ver: //' -e 's/ .*$//'`
+if test -z "$ngps"
+then
+ echo "$configfile -------"
+else
+ title="$configfile ------- $ngps grace periods"
+ dur=`sed -e 's/^.* rcutorture.shutdown_secs=//' -e 's/ .*$//' < $i/qemu-cmd 2> /dev/null`
+ if test -z "$dur"
+ then
+ :
+ else
+ ngpsps=`awk -v ngps=$ngps -v dur=$dur '
+ BEGIN { print ngps / dur }' < /dev/null`
+ title="$title ($ngpsps per second)"
+ fi
+ echo $title
+fi
diff --git a/rcutorture/bin/kvm-recheck.sh b/rcutorture/bin/kvm-recheck.sh
new file mode 100755
index 0000000..3f6c9b7
--- /dev/null
+++ b/rcutorture/bin/kvm-recheck.sh
@@ -0,0 +1,69 @@
+#!/bin/bash
+#
+# Given the results directories for previous KVM-based torture runs,
+# check the build and console output for errors. Given a directory
+# containing results directories, this recursively checks them all.
+#
+# Usage: sh kvm-recheck.sh resdir ...
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, you can access it online at
+# http://www.gnu.org/licenses/gpl-2.0.html.
+#
+# Copyright (C) IBM Corporation, 2011
+#
+# Authors: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
+
+PATH=`pwd`/tools/testing/selftests/rcutorture/bin:$PATH; export PATH
+. tools/testing/selftests/rcutorture/bin/functions.sh
+for rd in "$@"
+do
+ firsttime=1
+ dirs=`find $rd -name Make.defconfig.out -print | sort | sed -e 's,/[^/]*$,,' | sort -u`
+ for i in $dirs
+ do
+ if test -n "$firsttime"
+ then
+ firsttime=""
+ resdir=`echo $i | sed -e 's,/$,,' -e 's,/[^/]*$,,'`
+ head -1 $resdir/log
+ fi
+ TORTURE_SUITE="`cat $i/../TORTURE_SUITE`"
+ kvm-recheck-${TORTURE_SUITE}.sh $i
+ if test -f "$i/console.log"
+ then
+ configcheck.sh $i/.config $i/ConfigFragment
+ parse-build.sh $i/Make.out $configfile
+ parse-torture.sh $i/console.log $configfile
+ parse-console.sh $i/console.log $configfile
+ if test -r $i/Warnings
+ then
+ cat $i/Warnings
+ fi
+ else
+ if test -f "$i/qemu-cmd"
+ then
+ print_bug qemu failed
+ echo " $i"
+ elif test -f "$i/buildonly"
+ then
+ echo Build-only run, no boot/test
+ configcheck.sh $i/.config $i/ConfigFragment
+ parse-build.sh $i/Make.out $configfile
+ else
+ print_bug Build failed
+ echo " $i"
+ fi
+ fi
+ done
+done
diff --git a/rcutorture/bin/kvm-test-1-run.sh b/rcutorture/bin/kvm-test-1-run.sh
new file mode 100755
index 0000000..0f69dcb
--- /dev/null
+++ b/rcutorture/bin/kvm-test-1-run.sh
@@ -0,0 +1,227 @@
+#!/bin/bash
+#
+# Run a kvm-based test of the specified tree on the specified configs.
+# Fully automated run and error checking, no graphics console.
+#
+# Execute this in the source tree. Do not run it as a background task
+# because qemu does not seem to like that much.
+#
+# Usage: sh kvm-test-1-run.sh config builddir resdir minutes qemu-args boot_args
+#
+# qemu-args defaults to "-nographic", along with arguments specifying the
+# number of CPUs and other options generated from
+# the underlying CPU architecture.
+# boot_args defaults to value returned by the per_version_boot_params
+# shell function.
+#
+# Anything you specify for either qemu-args or boot_args is appended to
+# the default values. The "-smp" value is deduced from the contents of
+# the config fragment.
+#
+# More sophisticated argument parsing is clearly needed.
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, you can access it online at
+# http://www.gnu.org/licenses/gpl-2.0.html.
+#
+# Copyright (C) IBM Corporation, 2011
+#
+# Authors: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
+
+grace=120
+
+T=/tmp/kvm-test-1-run.sh.$$
+trap 'rm -rf $T' 0
+touch $T
+
+. $KVM/bin/functions.sh
+. $KVPATH/ver_functions.sh
+
+config_template=${1}
+config_dir=`echo $config_template | sed -e 's,/[^/]*$,,'`
+title=`echo $config_template | sed -e 's/^.*\///'`
+builddir=${2}
+if test -z "$builddir" -o ! -d "$builddir" -o ! -w "$builddir"
+then
+ echo "kvm-test-1-run.sh :$builddir: Not a writable directory, cannot build into it"
+ exit 1
+fi
+resdir=${3}
+if test -z "$resdir" -o ! -d "$resdir" -o ! -w "$resdir"
+then
+ echo "kvm-test-1-run.sh :$resdir: Not a writable directory, cannot store results into it"
+ exit 1
+fi
+cp $config_template $resdir/ConfigFragment
+echo ' ---' `date`: Starting build
+echo ' ---' Kconfig fragment at: $config_template >> $resdir/log
+if test -r "$config_dir/CFcommon"
+then
+ cat < $config_dir/CFcommon >> $T
+fi
+# Optimizations below this point
+# CONFIG_USB=n
+# CONFIG_SECURITY=n
+# CONFIG_NFS_FS=n
+# CONFIG_SOUND=n
+# CONFIG_INPUT_JOYSTICK=n
+# CONFIG_INPUT_TABLET=n
+# CONFIG_INPUT_TOUCHSCREEN=n
+# CONFIG_INPUT_MISC=n
+# CONFIG_INPUT_MOUSE=n
+# # CONFIG_NET=n # disables console access, so accept the slower build.
+# CONFIG_SCSI=n
+# CONFIG_ATA=n
+# CONFIG_FAT_FS=n
+# CONFIG_MSDOS_FS=n
+# CONFIG_VFAT_FS=n
+# CONFIG_ISO9660_FS=n
+# CONFIG_QUOTA=n
+# CONFIG_HID=n
+# CONFIG_CRYPTO=n
+# CONFIG_PCCARD=n
+# CONFIG_PCMCIA=n
+# CONFIG_CARDBUS=n
+# CONFIG_YENTA=n
+if kvm-build.sh $config_template $builddir $T
+then
+ QEMU="`identify_qemu $builddir/vmlinux`"
+ BOOT_IMAGE="`identify_boot_image $QEMU`"
+ cp $builddir/Make*.out $resdir
+ cp $builddir/.config $resdir
+ if test -n "$BOOT_IMAGE"
+ then
+ cp $builddir/$BOOT_IMAGE $resdir
+ else
+ echo No identifiable boot image, not running KVM, see $resdir.
+ echo Do the torture scripts know about your architecture?
+ fi
+ parse-build.sh $resdir/Make.out $title
+ if test -f $builddir.wait
+ then
+ mv $builddir.wait $builddir.ready
+ fi
+else
+ cp $builddir/Make*.out $resdir
+ cp $builddir/.config $resdir || :
+ echo Build failed, not running KVM, see $resdir.
+ if test -f $builddir.wait
+ then
+ mv $builddir.wait $builddir.ready
+ fi
+ exit 1
+fi
+while test -f $builddir.ready
+do
+ sleep 1
+done
+minutes=$4
+seconds=$(($minutes * 60))
+qemu_args=$5
+boot_args=$6
+
+cd $KVM
+kstarttime=`awk 'BEGIN { print systime() }' < /dev/null`
+if test -z "$TORTURE_BUILDONLY"
+then
+ echo ' ---' `date`: Starting kernel
+fi
+
+# Generate -smp qemu argument.
+qemu_args="-nographic $qemu_args"
+cpu_count=`configNR_CPUS.sh $config_template`
+vcpus=`identify_qemu_vcpus`
+if test $cpu_count -gt $vcpus
+then
+ echo CPU count limited from $cpu_count to $vcpus
+ touch $resdir/Warnings
+ echo CPU count limited from $cpu_count to $vcpus >> $resdir/Warnings
+ cpu_count=$vcpus
+fi
+qemu_args="`specify_qemu_cpus "$QEMU" "$qemu_args" "$cpu_count"`"
+
+# Generate architecture-specific and interaction-specific qemu arguments
+qemu_args="$qemu_args `identify_qemu_args "$QEMU" "$builddir/console.log"`"
+
+# Generate qemu -append arguments
+qemu_append="`identify_qemu_append "$QEMU"`"
+
+# Pull in Kconfig-fragment boot parameters
+boot_args="`configfrag_boot_params "$boot_args" "$config_template"`"
+# Generate kernel-version-specific boot parameters
+boot_args="`per_version_boot_params "$boot_args" $builddir/.config $seconds`"
+
+if test -n "$TORTURE_BUILDONLY"
+then
+ echo Build-only run specified, boot/test omitted.
+ touch $resdir/buildonly
+ exit 0
+fi
+echo $QEMU $qemu_args -m 512 -kernel $builddir/$BOOT_IMAGE -append \"$qemu_append $boot_args\" > $resdir/qemu-cmd
+( $QEMU $qemu_args -m 512 -kernel $builddir/$BOOT_IMAGE -append "$qemu_append $boot_args"; echo $? > $resdir/qemu-retval ) &
+qemu_pid=$!
+commandcompleted=0
+echo Monitoring qemu job at pid $qemu_pid
+while :
+do
+ kruntime=`awk 'BEGIN { print systime() - '"$kstarttime"' }' < /dev/null`
+ if kill -0 $qemu_pid > /dev/null 2>&1
+ then
+ if test $kruntime -ge $seconds
+ then
+ break;
+ fi
+ sleep 1
+ else
+ commandcompleted=1
+ if test $kruntime -lt $seconds
+ then
+ echo Completed in $kruntime vs. $seconds >> $resdir/Warnings 2>&1
+ grep "^(qemu) qemu:" $resdir/kvm-test-1-run.sh.out >> $resdir/Warnings 2>&1
+ killpid="`sed -n "s/^(qemu) qemu: terminating on signal [0-9]* from pid \([0-9]*\).*$/\1/p" $resdir/Warnings`"
+ if test -n "$killpid"
+ then
+ echo "ps -fp $killpid" >> $resdir/Warnings 2>&1
+ ps -fp $killpid >> $resdir/Warnings 2>&1
+ fi
+ else
+ echo ' ---' `date`: Kernel done
+ fi
+ break
+ fi
+done
+if test $commandcompleted -eq 0
+then
+ echo Grace period for qemu job at pid $qemu_pid
+ while :
+ do
+ kruntime=`awk 'BEGIN { print systime() - '"$kstarttime"' }' < /dev/null`
+ if kill -0 $qemu_pid > /dev/null 2>&1
+ then
+ :
+ else
+ break
+ fi
+ if test $kruntime -ge $((seconds + grace))
+ then
+ echo "!!! Hang at $kruntime vs. $seconds seconds" >> $resdir/Warnings 2>&1
+ kill -KILL $qemu_pid
+ break
+ fi
+ sleep 1
+ done
+fi
+
+cp $builddir/console.log $resdir
+parse-torture.sh $resdir/console.log $title
+parse-console.sh $resdir/console.log $title
diff --git a/rcutorture/bin/kvm.sh b/rcutorture/bin/kvm.sh
new file mode 100644
index 0000000..589e9c3
--- /dev/null
+++ b/rcutorture/bin/kvm.sh
@@ -0,0 +1,413 @@
+#!/bin/bash
+#
+# Run a series of 14 tests under KVM. These are not particularly
+# well-selected or well-tuned, but are the current set. Run from the
+# top level of the source tree.
+#
+# Edit the definitions below to set the locations of the various directories,
+# as well as the test duration.
+#
+# Usage: sh kvm.sh [ options ]
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, you can access it online at
+# http://www.gnu.org/licenses/gpl-2.0.html.
+#
+# Copyright (C) IBM Corporation, 2011
+#
+# Authors: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
+
+scriptname=$0
+args="$*"
+
+T=/tmp/kvm.sh.$$
+trap 'rm -rf $T' 0
+mkdir $T
+
+dur=30
+dryrun=""
+KVM="`pwd`/tools/testing/selftests/rcutorture"; export KVM
+PATH=${KVM}/bin:$PATH; export PATH
+TORTURE_DEFCONFIG=defconfig
+TORTURE_BOOT_IMAGE=""
+TORTURE_INITRD="$KVM/initrd"; export TORTURE_INITRD
+TORTURE_KMAKE_ARG=""
+TORTURE_SUITE=rcu
+resdir=""
+configs=""
+cpus=0
+ds=`date +%Y.%m.%d-%H:%M:%S`
+kversion=""
+
+. functions.sh
+
+usage () {
+ echo "Usage: $scriptname optional arguments:"
+ echo " --bootargs kernel-boot-arguments"
+ echo " --bootimage relative-path-to-kernel-boot-image"
+ echo " --buildonly"
+ echo " --configs \"config-file list\""
+ echo " --cpus N"
+ echo " --datestamp string"
+ echo " --defconfig string"
+ echo " --dryrun sched|script"
+ echo " --duration minutes"
+ echo " --interactive"
+ echo " --kmake-arg kernel-make-arguments"
+ echo " --kversion vN.NN"
+ echo " --mac nn:nn:nn:nn:nn:nn"
+ echo " --no-initrd"
+ echo " --qemu-args qemu-system-..."
+ echo " --qemu-cmd qemu-system-..."
+ echo " --results absolute-pathname"
+ echo " --torture rcu"
+ exit 1
+}
+
+while test $# -gt 0
+do
+ case "$1" in
+ --bootargs)
+ checkarg --bootargs "(list of kernel boot arguments)" "$#" "$2" '.*' '^--'
+ TORTURE_BOOTARGS="$2"
+ shift
+ ;;
+ --bootimage)
+ checkarg --bootimage "(relative path to kernel boot image)" "$#" "$2" '[a-zA-Z0-9][a-zA-Z0-9_]*' '^--'
+ TORTURE_BOOT_IMAGE="$2"
+ shift
+ ;;
+ --buildonly)
+ TORTURE_BUILDONLY=1
+ ;;
+ --configs)
+ checkarg --configs "(list of config files)" "$#" "$2" '^[^/]*$' '^--'
+ configs="$2"
+ shift
+ ;;
+ --cpus)
+ checkarg --cpus "(number)" "$#" "$2" '^[0-9]*$' '^--'
+ cpus=$2
+ shift
+ ;;
+ --datestamp)
+ checkarg --datestamp "(relative pathname)" "$#" "$2" '^[^/]*$' '^--'
+ ds=$2
+ shift
+ ;;
+ --defconfig)
+ checkarg --defconfig "defconfigtype" "$#" "$2" '^[^/][^/]*$' '^--'
+ TORTURE_DEFCONFIG=$2
+ shift
+ ;;
+ --dryrun)
+ checkarg --dryrun "sched|script" $# "$2" 'sched\|script' '^--'
+ dryrun=$2
+ shift
+ ;;
+ --duration)
+ checkarg --duration "(minutes)" $# "$2" '^[0-9]*$' '^error'
+ dur=$2
+ shift
+ ;;
+ --interactive)
+ TORTURE_QEMU_INTERACTIVE=1; export TORTURE_QEMU_INTERACTIVE
+ ;;
+ --kmake-arg)
+ checkarg --kmake-arg "(kernel make arguments)" $# "$2" '.*' '^error$'
+ TORTURE_KMAKE_ARG="$2"
+ shift
+ ;;
+ --kversion)
+ checkarg --kversion "(kernel version)" $# "$2" '^v[0-9.]*$' '^error'
+ kversion=$2
+ shift
+ ;;
+ --mac)
+ checkarg --mac "(MAC address)" $# "$2" '^\([0-9a-fA-F]\{2\}:\)\{5\}[0-9a-fA-F]\{2\}$' error
+ TORTURE_QEMU_MAC=$2
+ shift
+ ;;
+ --no-initrd)
+ TORTURE_INITRD=""; export TORTURE_INITRD
+ ;;
+ --qemu-args)
+ checkarg --qemu-args "-qemu args" $# "$2" '^-' '^error'
+ TORTURE_QEMU_ARG="$2"
+ shift
+ ;;
+ --qemu-cmd)
+ checkarg --qemu-cmd "(qemu-system-...)" $# "$2" 'qemu-system-' '^--'
+ TORTURE_QEMU_CMD="$2"
+ shift
+ ;;
+ --results)
+ checkarg --results "(absolute pathname)" "$#" "$2" '^/' '^error'
+ resdir=$2
+ shift
+ ;;
+ --torture)
+ checkarg --torture "(suite name)" "$#" "$2" '^\(lock\|rcu\)$' '^--'
+ TORTURE_SUITE=$2
+ shift
+ ;;
+ *)
+ echo Unknown argument $1
+ usage
+ ;;
+ esac
+ shift
+done
+
+CONFIGFRAG=${KVM}/configs/${TORTURE_SUITE}; export CONFIGFRAG
+KVPATH=${CONFIGFRAG}/$kversion; export KVPATH
+
+if test -z "$configs"
+then
+ configs="`cat $CONFIGFRAG/$kversion/CFLIST`"
+fi
+
+if test -z "$resdir"
+then
+ resdir=$KVM/res
+fi
+
+# Create a file of test-name/#cpus pairs, sorted by decreasing #cpus.
+touch $T/cfgcpu
+for CF in $configs
+do
+ if test -f "$CONFIGFRAG/$kversion/$CF"
+ then
+ echo $CF `configNR_CPUS.sh $CONFIGFRAG/$kversion/$CF` >> $T/cfgcpu
+ else
+ echo "The --configs file $CF does not exist, terminating."
+ exit 1
+ fi
+done
+sort -k2nr $T/cfgcpu > $T/cfgcpu.sort
+
+# Use a greedy bin-packing algorithm, sorting the list accordingly.
+awk < $T/cfgcpu.sort > $T/cfgcpu.pack -v ncpus=$cpus '
+BEGIN {
+ njobs = 0;
+}
+
+{
+ # Read file of tests and corresponding required numbers of CPUs.
+ cf[njobs] = $1;
+ cpus[njobs] = $2;
+ njobs++;
+}
+
+END {
+ alldone = 0;
+ batch = 0;
+ nc = -1;
+
+ # Each pass through the following loop creates on test batch
+ # that can be executed concurrently given ncpus. Note that a
+ # given test that requires more than the available CPUs will run in
+ # their own batch. Such tests just have to make do with what
+ # is available.
+ while (nc != ncpus) {
+ batch++;
+ nc = ncpus;
+
+ # Each pass through the following loop considers one
+ # test for inclusion in the current batch.
+ for (i = 0; i < njobs; i++) {
+ if (done[i])
+ continue; # Already part of a batch.
+ if (nc >= cpus[i] || nc == ncpus) {
+
+ # This test fits into the current batch.
+ done[i] = batch;
+ nc -= cpus[i];
+ if (nc <= 0)
+ break; # Too-big test in its own batch.
+ }
+ }
+ }
+
+ # Dump out the tests in batch order.
+ for (b = 1; b <= batch; b++)
+ for (i = 0; i < njobs; i++)
+ if (done[i] == b)
+ print cf[i], cpus[i];
+}'
+
+# Generate a script to execute the tests in appropriate batches.
+cat << ___EOF___ > $T/script
+CONFIGFRAG="$CONFIGFRAG"; export CONFIGFRAG
+KVM="$KVM"; export KVM
+KVPATH="$KVPATH"; export KVPATH
+PATH="$PATH"; export PATH
+TORTURE_BOOT_IMAGE="$TORTURE_BOOT_IMAGE"; export TORTURE_BOOT_IMAGE
+TORTURE_BUILDONLY="$TORTURE_BUILDONLY"; export TORTURE_BUILDONLY
+TORTURE_DEFCONFIG="$TORTURE_DEFCONFIG"; export TORTURE_DEFCONFIG
+TORTURE_INITRD="$TORTURE_INITRD"; export TORTURE_INITRD
+TORTURE_KMAKE_ARG="$TORTURE_KMAKE_ARG"; export TORTURE_KMAKE_ARG
+TORTURE_QEMU_CMD="$TORTURE_QEMU_CMD"; export TORTURE_QEMU_CMD
+TORTURE_QEMU_INTERACTIVE="$TORTURE_QEMU_INTERACTIVE"; export TORTURE_QEMU_INTERACTIVE
+TORTURE_QEMU_MAC="$TORTURE_QEMU_MAC"; export TORTURE_QEMU_MAC
+TORTURE_SUITE="$TORTURE_SUITE"; export TORTURE_SUITE
+if ! test -e $resdir
+then
+ mkdir -p "$resdir" || :
+fi
+mkdir $resdir/$ds
+echo Results directory: $resdir/$ds
+echo $scriptname $args
+touch $resdir/$ds/log
+echo $scriptname $args >> $resdir/$ds/log
+echo ${TORTURE_SUITE} > $resdir/$ds/TORTURE_SUITE
+pwd > $resdir/$ds/testid.txt
+if test -d .git
+then
+ git status >> $resdir/$ds/testid.txt
+ git rev-parse HEAD >> $resdir/$ds/testid.txt
+ if ! git diff HEAD > $T/git-diff 2>&1
+ then
+ cp $T/git-diff $resdir/$ds
+ fi
+fi
+___EOF___
+awk < $T/cfgcpu.pack \
+ -v CONFIGDIR="$CONFIGFRAG/$kversion/" \
+ -v KVM="$KVM" \
+ -v ncpus=$cpus \
+ -v rd=$resdir/$ds/ \
+ -v dur=$dur \
+ -v TORTURE_QEMU_ARG="$TORTURE_QEMU_ARG" \
+ -v TORTURE_BOOTARGS="$TORTURE_BOOTARGS" \
+'BEGIN {
+ i = 0;
+}
+
+{
+ cf[i] = $1;
+ cpus[i] = $2;
+ i++;
+}
+
+# Dump out the scripting required to run one test batch.
+function dump(first, pastlast)
+{
+ print "echo ----Start batch: `date`";
+ print "echo ----Start batch: `date` >> " rd "/log";
+ jn=1
+ for (j = first; j < pastlast; j++) {
+ builddir=KVM "/b" jn
+ cpusr[jn] = cpus[j];
+ if (cfrep[cf[j]] == "") {
+ cfr[jn] = cf[j];
+ cfrep[cf[j]] = 1;
+ } else {
+ cfrep[cf[j]]++;
+ cfr[jn] = cf[j] "." cfrep[cf[j]];
+ }
+ if (cpusr[jn] > ncpus && ncpus != 0)
+ ovf = "(!)";
+ else
+ ovf = "";
+ print "echo ", cfr[jn], cpusr[jn] ovf ": Starting build. `date`";
+ print "echo ", cfr[jn], cpusr[jn] ovf ": Starting build. `date` >> " rd "/log";
+ print "rm -f " builddir ".*";
+ print "touch " builddir ".wait";
+ print "mkdir " builddir " > /dev/null 2>&1 || :";
+ print "mkdir " rd cfr[jn] " || :";
+ print "kvm-test-1-run.sh " CONFIGDIR cf[j], builddir, rd cfr[jn], dur " \"" TORTURE_QEMU_ARG "\" \"" TORTURE_BOOTARGS "\" > " rd cfr[jn] "/kvm-test-1-run.sh.out 2>&1 &"
+ print "echo ", cfr[jn], cpusr[jn] ovf ": Waiting for build to complete. `date`";
+ print "echo ", cfr[jn], cpusr[jn] ovf ": Waiting for build to complete. `date` >> " rd "/log";
+ print "while test -f " builddir ".wait"
+ print "do"
+ print "\tsleep 1"
+ print "done"
+ print "echo ", cfr[jn], cpusr[jn] ovf ": Build complete. `date`";
+ print "echo ", cfr[jn], cpusr[jn] ovf ": Build complete. `date` >> " rd "/log";
+ jn++;
+ }
+ for (j = 1; j < jn; j++) {
+ builddir=KVM "/b" j
+ print "rm -f " builddir ".ready"
+ print "if test -z \"$TORTURE_BUILDONLY\""
+ print "then"
+ print "\techo ----", cfr[j], cpusr[j] ovf ": Starting kernel. `date`";
+ print "\techo ----", cfr[j], cpusr[j] ovf ": Starting kernel. `date` >> " rd "/log";
+ print "fi"
+ }
+ print "wait"
+ print "if test -z \"$TORTURE_BUILDONLY\""
+ print "then"
+ print "\techo ---- All kernel runs complete. `date`";
+ print "\techo ---- All kernel runs complete. `date` >> " rd "/log";
+ print "fi"
+ for (j = 1; j < jn; j++) {
+ builddir=KVM "/b" j
+ print "echo ----", cfr[j], cpusr[j] ovf ": Build/run results:";
+ print "echo ----", cfr[j], cpusr[j] ovf ": Build/run results: >> " rd "/log";
+ print "cat " rd cfr[j] "/kvm-test-1-run.sh.out";
+ print "cat " rd cfr[j] "/kvm-test-1-run.sh.out >> " rd "/log";
+ }
+}
+
+END {
+ njobs = i;
+ nc = ncpus;
+ first = 0;
+
+ # Each pass through the following loop considers one test.
+ for (i = 0; i < njobs; i++) {
+ if (ncpus == 0) {
+ # Sequential test specified, each test its own batch.
+ dump(i, i + 1);
+ first = i;
+ } else if (nc < cpus[i] && i != 0) {
+ # Out of CPUs, dump out a batch.
+ dump(first, i);
+ first = i;
+ nc = ncpus;
+ }
+ # Account for the CPUs needed by the current test.
+ nc -= cpus[i];
+ }
+ # Dump the last batch.
+ if (ncpus != 0)
+ dump(first, i);
+}' >> $T/script
+
+cat << ___EOF___ >> $T/script
+echo
+echo
+echo " --- `date` Test summary:"
+echo Results directory: $resdir/$ds
+kvm-recheck.sh $resdir/$ds
+___EOF___
+
+if test "$dryrun" = script
+then
+ cat $T/script
+ exit 0
+elif test "$dryrun" = sched
+then
+ # Extract the test run schedule from the script.
+ egrep 'Start batch|Starting build\.' $T/script |
+ grep -v ">>" |
+ sed -e 's/:.*$//' -e 's/^echo //'
+ exit 0
+else
+ # Not a dryrun, so run the script.
+ sh $T/script
+fi
+
+# Tracing: trace_event=rcu:rcu_grace_period,rcu:rcu_future_grace_period,rcu:rcu_grace_period_init,rcu:rcu_nocb_wake,rcu:rcu_preempt_task,rcu:rcu_unlock_preempted_task,rcu:rcu_quiescent_state_report,rcu:rcu_fqs,rcu:rcu_callback,rcu:rcu_kfree_callback,rcu:rcu_batch_start,rcu:rcu_invoke_callback,rcu:rcu_invoke_kfree_callback,rcu:rcu_batch_end,rcu:rcu_torture_read,rcu:rcu_barrier
diff --git a/rcutorture/bin/parse-build.sh b/rcutorture/bin/parse-build.sh
new file mode 100755
index 0000000..5432309
--- /dev/null
+++ b/rcutorture/bin/parse-build.sh
@@ -0,0 +1,57 @@
+#!/bin/sh
+#
+# Check the build output from an rcutorture run for goodness.
+# The "file" is a pathname on the local system, and "title" is
+# a text string for error-message purposes.
+#
+# The file must contain kernel build output.
+#
+# Usage:
+# sh parse-build.sh file title
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, you can access it online at
+# http://www.gnu.org/licenses/gpl-2.0.html.
+#
+# Copyright (C) IBM Corporation, 2011
+#
+# Authors: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
+
+T=$1
+title=$2
+
+. functions.sh
+
+if grep -q CC < $T
+then
+ :
+else
+ print_bug $title no build
+ exit 1
+fi
+
+if grep -q "error:" < $T
+then
+ print_bug $title build errors:
+ grep "error:" < $T
+ exit 2
+fi
+exit 0
+
+if egrep -q "rcu[^/]*\.c.*warning:|rcu.*\.h.*warning:" < $T
+then
+ print_warning $title build errors:
+ egrep "rcu[^/]*\.c.*warning:|rcu.*\.h.*warning:" < $T
+ exit 2
+fi
+exit 0
diff --git a/rcutorture/bin/parse-console.sh b/rcutorture/bin/parse-console.sh
new file mode 100755
index 0000000..4185d4c
--- /dev/null
+++ b/rcutorture/bin/parse-console.sh
@@ -0,0 +1,41 @@
+#!/bin/sh
+#
+# Check the console output from an rcutorture run for oopses.
+# The "file" is a pathname on the local system, and "title" is
+# a text string for error-message purposes.
+#
+# Usage:
+# sh parse-console.sh file title
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, you can access it online at
+# http://www.gnu.org/licenses/gpl-2.0.html.
+#
+# Copyright (C) IBM Corporation, 2011
+#
+# Authors: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
+
+T=/tmp/abat-chk-badness.sh.$$
+trap 'rm -f $T' 0
+
+file="$1"
+title="$2"
+
+. functions.sh
+
+egrep 'Badness|WARNING:|Warn|BUG|===========|Call Trace:|Oops:' < $file | grep -v 'ODEBUG: ' | grep -v 'Warning: unable to open an initial console' > $T
+if test -s $T
+then
+ print_warning Assertion failure in $file $title
+ cat $T
+fi
diff --git a/rcutorture/bin/parse-torture.sh b/rcutorture/bin/parse-torture.sh
new file mode 100755
index 0000000..3455560
--- /dev/null
+++ b/rcutorture/bin/parse-torture.sh
@@ -0,0 +1,106 @@
+#!/bin/sh
+#
+# Check the console output from a torture run for goodness.
+# The "file" is a pathname on the local system, and "title" is
+# a text string for error-message purposes.
+#
+# The file must contain torture output, but can be interspersed
+# with other dmesg text, as in console-log output.
+#
+# Usage:
+# sh parse-torture.sh file title
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, you can access it online at
+# http://www.gnu.org/licenses/gpl-2.0.html.
+#
+# Copyright (C) IBM Corporation, 2011
+#
+# Authors: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
+
+T=/tmp/parse-torture.sh.$$
+file="$1"
+title="$2"
+
+trap 'rm -f $T.seq' 0
+
+. functions.sh
+
+# check for presence of torture output file.
+
+if test -f "$file" -a -r "$file"
+then
+ :
+else
+ echo $title unreadable torture output file: $file
+ exit 1
+fi
+
+# check for abject failure
+
+if grep -q FAILURE $file || grep -q -e '-torture.*!!!' $file
+then
+ nerrs=`grep --binary-files=text '!!!' $file | tail -1 | awk '{for (i=NF-8;i<=NF;i++) sum+=$i; } END {print sum}'`
+ print_bug $title FAILURE, $nerrs instances
+ echo " " $url
+ exit
+fi
+
+grep --binary-files=text 'torture:.*ver:' $file | grep --binary-files=text -v '(null)' | sed -e 's/^(initramfs)[^]]*] //' -e 's/^\[[^]]*] //' |
+awk '
+BEGIN {
+ ver = 0;
+ badseq = 0;
+ }
+
+ {
+ if (!badseq && ($5 + 0 != $5 || $5 <= ver)) {
+ badseqno1 = ver;
+ badseqno2 = $5;
+ badseqnr = NR;
+ badseq = 1;
+ }
+ ver = $5
+ }
+
+END {
+ if (badseq) {
+ if (badseqno1 == badseqno2 && badseqno2 == ver)
+ print "GP HANG at " ver " torture stat " badseqnr;
+ else
+ print "BAD SEQ " badseqno1 ":" badseqno2 " last:" ver " version " badseqnr;
+ }
+ }' > $T.seq
+
+if grep -q SUCCESS $file
+then
+ if test -s $T.seq
+ then
+ print_warning $title $title `cat $T.seq`
+ echo " " $file
+ exit 2
+ fi
+else
+ if grep -q "_HOTPLUG:" $file
+ then
+ print_warning HOTPLUG FAILURES $title `cat $T.seq`
+ echo " " $file
+ exit 3
+ fi
+ echo $title no success message, `grep --binary-files=text 'ver:' $file | wc -l` successful version messages
+ if test -s $T.seq
+ then
+ print_warning $title `cat $T.seq`
+ fi
+ exit 2
+fi
diff --git a/rcutorture/configs/lock/BUSTED b/rcutorture/configs/lock/BUSTED
new file mode 100644
index 0000000..1d1da14
--- /dev/null
+++ b/rcutorture/configs/lock/BUSTED
@@ -0,0 +1,6 @@
+CONFIG_SMP=y
+CONFIG_NR_CPUS=4
+CONFIG_HOTPLUG_CPU=y
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
diff --git a/rcutorture/configs/lock/BUSTED.boot b/rcutorture/configs/lock/BUSTED.boot
new file mode 100644
index 0000000..6386c15
--- /dev/null
+++ b/rcutorture/configs/lock/BUSTED.boot
@@ -0,0 +1 @@
+locktorture.torture_type=lock_busted
diff --git a/rcutorture/configs/lock/CFLIST b/rcutorture/configs/lock/CFLIST
new file mode 100644
index 0000000..a061b22
--- /dev/null
+++ b/rcutorture/configs/lock/CFLIST
@@ -0,0 +1 @@
+LOCK01
diff --git a/rcutorture/configs/lock/CFcommon b/rcutorture/configs/lock/CFcommon
new file mode 100644
index 0000000..e372dc2
--- /dev/null
+++ b/rcutorture/configs/lock/CFcommon
@@ -0,0 +1,2 @@
+CONFIG_LOCK_TORTURE_TEST=y
+CONFIG_PRINTK_TIME=y
diff --git a/rcutorture/configs/lock/LOCK01 b/rcutorture/configs/lock/LOCK01
new file mode 100644
index 0000000..a9625e3
--- /dev/null
+++ b/rcutorture/configs/lock/LOCK01
@@ -0,0 +1,6 @@
+CONFIG_SMP=y
+CONFIG_NR_CPUS=8
+CONFIG_HOTPLUG_CPU=y
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
diff --git a/rcutorture/configs/lock/ver_functions.sh b/rcutorture/configs/lock/ver_functions.sh
new file mode 100644
index 0000000..9746ea1
--- /dev/null
+++ b/rcutorture/configs/lock/ver_functions.sh
@@ -0,0 +1,43 @@
+#!/bin/bash
+#
+# Kernel-version-dependent shell functions for the rest of the scripts.
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, you can access it online at
+# http://www.gnu.org/licenses/gpl-2.0.html.
+#
+# Copyright (C) IBM Corporation, 2014
+#
+# Authors: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
+
+# locktorture_param_onoff bootparam-string config-file
+#
+# Adds onoff locktorture module parameters to kernels having it.
+locktorture_param_onoff () {
+ if ! bootparam_hotplug_cpu "$1" && configfrag_hotplug_cpu "$2"
+ then
+ echo CPU-hotplug kernel, adding locktorture onoff. 1>&2
+ echo locktorture.onoff_interval=3 locktorture.onoff_holdoff=30
+ fi
+}
+
+# per_version_boot_params bootparam-string config-file seconds
+#
+# Adds per-version torture-module parameters to kernels supporting them.
+per_version_boot_params () {
+ echo $1 `locktorture_param_onoff "$1" "$2"` \
+ locktorture.stat_interval=15 \
+ locktorture.shutdown_secs=$3 \
+ locktorture.locktorture_runnable=1 \
+ locktorture.verbose=1
+}
diff --git a/rcutorture/configs/rcu/BUSTED b/rcutorture/configs/rcu/BUSTED
new file mode 100644
index 0000000..48d8a24
--- /dev/null
+++ b/rcutorture/configs/rcu/BUSTED
@@ -0,0 +1,7 @@
+CONFIG_RCU_TRACE=n
+CONFIG_SMP=y
+CONFIG_NR_CPUS=4
+CONFIG_HOTPLUG_CPU=y
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
diff --git a/rcutorture/configs/rcu/BUSTED.boot b/rcutorture/configs/rcu/BUSTED.boot
new file mode 100644
index 0000000..6804f9d
--- /dev/null
+++ b/rcutorture/configs/rcu/BUSTED.boot
@@ -0,0 +1 @@
+rcutorture.torture_type=rcu_busted
diff --git a/rcutorture/configs/rcu/CFLIST b/rcutorture/configs/rcu/CFLIST
new file mode 100644
index 0000000..cd3d29c
--- /dev/null
+++ b/rcutorture/configs/rcu/CFLIST
@@ -0,0 +1,13 @@
+TREE01
+TREE02
+TREE03
+TREE04
+TREE05
+TREE06
+TREE07
+TREE08
+TREE09
+SRCU-N
+SRCU-P
+TINY01
+TINY02
diff --git a/rcutorture/configs/rcu/CFcommon b/rcutorture/configs/rcu/CFcommon
new file mode 100644
index 0000000..d2d2a86
--- /dev/null
+++ b/rcutorture/configs/rcu/CFcommon
@@ -0,0 +1,2 @@
+CONFIG_RCU_TORTURE_TEST=y
+CONFIG_PRINTK_TIME=y
diff --git a/rcutorture/configs/rcu/SRCU-N b/rcutorture/configs/rcu/SRCU-N
new file mode 100644
index 0000000..9fbb41b
--- /dev/null
+++ b/rcutorture/configs/rcu/SRCU-N
@@ -0,0 +1,7 @@
+CONFIG_RCU_TRACE=n
+CONFIG_SMP=y
+CONFIG_NR_CPUS=4
+CONFIG_HOTPLUG_CPU=y
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
diff --git a/rcutorture/configs/rcu/SRCU-N.boot b/rcutorture/configs/rcu/SRCU-N.boot
new file mode 100644
index 0000000..238bfe3
--- /dev/null
+++ b/rcutorture/configs/rcu/SRCU-N.boot
@@ -0,0 +1 @@
+rcutorture.torture_type=srcu
diff --git a/rcutorture/configs/rcu/SRCU-P b/rcutorture/configs/rcu/SRCU-P
new file mode 100644
index 0000000..4b6f272
--- /dev/null
+++ b/rcutorture/configs/rcu/SRCU-P
@@ -0,0 +1,7 @@
+CONFIG_RCU_TRACE=n
+CONFIG_SMP=y
+CONFIG_NR_CPUS=8
+CONFIG_HOTPLUG_CPU=y
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
diff --git a/rcutorture/configs/rcu/SRCU-P.boot b/rcutorture/configs/rcu/SRCU-P.boot
new file mode 100644
index 0000000..238bfe3
--- /dev/null
+++ b/rcutorture/configs/rcu/SRCU-P.boot
@@ -0,0 +1 @@
+rcutorture.torture_type=srcu
diff --git a/rcutorture/configs/rcu/TINY01 b/rcutorture/configs/rcu/TINY01
new file mode 100644
index 0000000..0a63e07
--- /dev/null
+++ b/rcutorture/configs/rcu/TINY01
@@ -0,0 +1,12 @@
+CONFIG_SMP=n
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TINY_RCU=y
+CONFIG_HZ_PERIODIC=n
+CONFIG_NO_HZ_IDLE=y
+CONFIG_NO_HZ_FULL=n
+CONFIG_RCU_TRACE=n
+CONFIG_DEBUG_LOCK_ALLOC=n
+CONFIG_DEBUG_OBJECTS_RCU_HEAD=n
+CONFIG_PREEMPT_COUNT=n
diff --git a/rcutorture/configs/rcu/TINY02 b/rcutorture/configs/rcu/TINY02
new file mode 100644
index 0000000..f4feaee
--- /dev/null
+++ b/rcutorture/configs/rcu/TINY02
@@ -0,0 +1,12 @@
+CONFIG_SMP=n
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TINY_RCU=y
+CONFIG_HZ_PERIODIC=y
+CONFIG_NO_HZ_IDLE=n
+CONFIG_NO_HZ_FULL=n
+CONFIG_RCU_TRACE=y
+CONFIG_DEBUG_LOCK_ALLOC=y
+CONFIG_DEBUG_OBJECTS_RCU_HEAD=n
+CONFIG_PREEMPT_COUNT=y
diff --git a/rcutorture/configs/rcu/TREE01 b/rcutorture/configs/rcu/TREE01
new file mode 100644
index 0000000..063b707
--- /dev/null
+++ b/rcutorture/configs/rcu/TREE01
@@ -0,0 +1,21 @@
+CONFIG_SMP=y
+CONFIG_NR_CPUS=8
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_HZ_PERIODIC=n
+CONFIG_NO_HZ_IDLE=y
+CONFIG_NO_HZ_FULL=n
+CONFIG_RCU_FAST_NO_HZ=y
+CONFIG_RCU_TRACE=y
+CONFIG_HOTPLUG_CPU=y
+CONFIG_RCU_FANOUT=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_RCU_NOCB_CPU=y
+CONFIG_RCU_NOCB_CPU_ZERO=y
+CONFIG_DEBUG_LOCK_ALLOC=n
+CONFIG_RCU_CPU_STALL_INFO=n
+CONFIG_RCU_CPU_STALL_VERBOSE=n
+CONFIG_RCU_BOOST=n
+CONFIG_DEBUG_OBJECTS_RCU_HEAD=n
diff --git a/rcutorture/configs/rcu/TREE01.boot b/rcutorture/configs/rcu/TREE01.boot
new file mode 100644
index 0000000..0fc8a34
--- /dev/null
+++ b/rcutorture/configs/rcu/TREE01.boot
@@ -0,0 +1 @@
+rcutorture.torture_type=rcu_bh
diff --git a/rcutorture/configs/rcu/TREE02 b/rcutorture/configs/rcu/TREE02
new file mode 100644
index 0000000..ea119ba
--- /dev/null
+++ b/rcutorture/configs/rcu/TREE02
@@ -0,0 +1,24 @@
+CONFIG_SMP=y
+CONFIG_NR_CPUS=8
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_HZ_PERIODIC=n
+CONFIG_NO_HZ_IDLE=y
+CONFIG_NO_HZ_FULL=n
+CONFIG_RCU_FAST_NO_HZ=n
+CONFIG_RCU_TRACE=n
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_RCU_FANOUT=3
+CONFIG_RCU_FANOUT_LEAF=3
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_RCU_NOCB_CPU=n
+CONFIG_DEBUG_LOCK_ALLOC=y
+CONFIG_PROVE_LOCKING=n
+CONFIG_RCU_CPU_STALL_INFO=n
+CONFIG_RCU_CPU_STALL_VERBOSE=y
+CONFIG_RCU_BOOST=n
+CONFIG_DEBUG_OBJECTS_RCU_HEAD=n
diff --git a/rcutorture/configs/rcu/TREE02-T b/rcutorture/configs/rcu/TREE02-T
new file mode 100644
index 0000000..19cf948
--- /dev/null
+++ b/rcutorture/configs/rcu/TREE02-T
@@ -0,0 +1,24 @@
+CONFIG_SMP=y
+CONFIG_NR_CPUS=8
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_HZ_PERIODIC=n
+CONFIG_NO_HZ_IDLE=y
+CONFIG_NO_HZ_FULL=n
+CONFIG_RCU_FAST_NO_HZ=n
+CONFIG_RCU_TRACE=y
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_RCU_FANOUT=3
+CONFIG_RCU_FANOUT_LEAF=3
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_RCU_NOCB_CPU=n
+CONFIG_DEBUG_LOCK_ALLOC=y
+CONFIG_PROVE_LOCKING=n
+CONFIG_RCU_CPU_STALL_INFO=n
+CONFIG_RCU_CPU_STALL_VERBOSE=y
+CONFIG_RCU_BOOST=n
+CONFIG_DEBUG_OBJECTS_RCU_HEAD=n
diff --git a/rcutorture/configs/rcu/TREE03 b/rcutorture/configs/rcu/TREE03
new file mode 100644
index 0000000..f4567fb
--- /dev/null
+++ b/rcutorture/configs/rcu/TREE03
@@ -0,0 +1,21 @@
+CONFIG_SMP=y
+CONFIG_NR_CPUS=8
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_HZ_PERIODIC=y
+CONFIG_NO_HZ_IDLE=n
+CONFIG_NO_HZ_FULL=n
+CONFIG_RCU_TRACE=y
+CONFIG_HOTPLUG_CPU=y
+CONFIG_RCU_FANOUT=4
+CONFIG_RCU_FANOUT_LEAF=4
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_RCU_NOCB_CPU=n
+CONFIG_DEBUG_LOCK_ALLOC=n
+CONFIG_RCU_CPU_STALL_INFO=n
+CONFIG_RCU_CPU_STALL_VERBOSE=n
+CONFIG_RCU_BOOST=y
+CONFIG_RCU_BOOST_PRIO=2
+CONFIG_DEBUG_OBJECTS_RCU_HEAD=n
diff --git a/rcutorture/configs/rcu/TREE04 b/rcutorture/configs/rcu/TREE04
new file mode 100644
index 0000000..0a262fb
--- /dev/null
+++ b/rcutorture/configs/rcu/TREE04
@@ -0,0 +1,23 @@
+CONFIG_SMP=y
+CONFIG_NR_CPUS=8
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_HZ_PERIODIC=n
+CONFIG_NO_HZ_IDLE=n
+CONFIG_NO_HZ_FULL=y
+CONFIG_NO_HZ_FULL_ALL=y
+CONFIG_RCU_FAST_NO_HZ=y
+CONFIG_RCU_TRACE=y
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_RCU_FANOUT=2
+CONFIG_RCU_FANOUT_LEAF=2
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_RCU_NOCB_CPU=n
+CONFIG_DEBUG_LOCK_ALLOC=n
+CONFIG_RCU_CPU_STALL_INFO=y
+CONFIG_RCU_CPU_STALL_VERBOSE=y
+CONFIG_DEBUG_OBJECTS_RCU_HEAD=n
diff --git a/rcutorture/configs/rcu/TREE04.boot b/rcutorture/configs/rcu/TREE04.boot
new file mode 100644
index 0000000..0fc8a34
--- /dev/null
+++ b/rcutorture/configs/rcu/TREE04.boot
@@ -0,0 +1 @@
+rcutorture.torture_type=rcu_bh
diff --git a/rcutorture/configs/rcu/TREE05 b/rcutorture/configs/rcu/TREE05
new file mode 100644
index 0000000..3a06b97
--- /dev/null
+++ b/rcutorture/configs/rcu/TREE05
@@ -0,0 +1,23 @@
+CONFIG_SMP=y
+CONFIG_NR_CPUS=8
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_HZ_PERIODIC=n
+CONFIG_NO_HZ_IDLE=y
+CONFIG_NO_HZ_FULL=n
+CONFIG_RCU_FAST_NO_HZ=n
+CONFIG_RCU_TRACE=n
+CONFIG_HOTPLUG_CPU=y
+CONFIG_RCU_FANOUT=6
+CONFIG_RCU_FANOUT_LEAF=6
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_RCU_NOCB_CPU=y
+CONFIG_RCU_NOCB_CPU_NONE=y
+CONFIG_DEBUG_LOCK_ALLOC=y
+CONFIG_PROVE_LOCKING=y
+CONFIG_PROVE_RCU=y
+CONFIG_RCU_CPU_STALL_INFO=n
+CONFIG_RCU_CPU_STALL_VERBOSE=n
+CONFIG_DEBUG_OBJECTS_RCU_HEAD=n
diff --git a/rcutorture/configs/rcu/TREE05.boot b/rcutorture/configs/rcu/TREE05.boot
new file mode 100644
index 0000000..3b42b8b
--- /dev/null
+++ b/rcutorture/configs/rcu/TREE05.boot
@@ -0,0 +1 @@
+rcutorture.torture_type=sched
diff --git a/rcutorture/configs/rcu/TREE06 b/rcutorture/configs/rcu/TREE06
new file mode 100644
index 0000000..8f084cc
--- /dev/null
+++ b/rcutorture/configs/rcu/TREE06
@@ -0,0 +1,24 @@
+CONFIG_SMP=y
+CONFIG_NR_CPUS=8
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_HZ_PERIODIC=n
+CONFIG_NO_HZ_IDLE=y
+CONFIG_NO_HZ_FULL=n
+CONFIG_RCU_FAST_NO_HZ=n
+CONFIG_RCU_TRACE=n
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_RCU_FANOUT=6
+CONFIG_RCU_FANOUT_LEAF=6
+CONFIG_RCU_FANOUT_EXACT=y
+CONFIG_RCU_NOCB_CPU=n
+CONFIG_DEBUG_LOCK_ALLOC=y
+CONFIG_PROVE_LOCKING=y
+CONFIG_PROVE_RCU=y
+CONFIG_RCU_CPU_STALL_INFO=n
+CONFIG_RCU_CPU_STALL_VERBOSE=n
+CONFIG_DEBUG_OBJECTS_RCU_HEAD=y
diff --git a/rcutorture/configs/rcu/TREE07 b/rcutorture/configs/rcu/TREE07
new file mode 100644
index 0000000..ab62255
--- /dev/null
+++ b/rcutorture/configs/rcu/TREE07
@@ -0,0 +1,22 @@
+CONFIG_SMP=y
+CONFIG_NR_CPUS=16
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_HZ_PERIODIC=n
+CONFIG_NO_HZ_IDLE=n
+CONFIG_NO_HZ_FULL=y
+CONFIG_NO_HZ_FULL_ALL=y
+CONFIG_NO_HZ_FULL_SYSIDLE=y
+CONFIG_RCU_FAST_NO_HZ=n
+CONFIG_RCU_TRACE=y
+CONFIG_HOTPLUG_CPU=y
+CONFIG_RCU_FANOUT=2
+CONFIG_RCU_FANOUT_LEAF=2
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_RCU_NOCB_CPU=n
+CONFIG_DEBUG_LOCK_ALLOC=n
+CONFIG_RCU_CPU_STALL_INFO=y
+CONFIG_RCU_CPU_STALL_VERBOSE=n
+CONFIG_DEBUG_OBJECTS_RCU_HEAD=n
diff --git a/rcutorture/configs/rcu/TREE08 b/rcutorture/configs/rcu/TREE08
new file mode 100644
index 0000000..69a2e25
--- /dev/null
+++ b/rcutorture/configs/rcu/TREE08
@@ -0,0 +1,24 @@
+CONFIG_SMP=y
+CONFIG_NR_CPUS=16
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_HZ_PERIODIC=n
+CONFIG_NO_HZ_IDLE=y
+CONFIG_NO_HZ_FULL=n
+CONFIG_RCU_FAST_NO_HZ=n
+CONFIG_RCU_TRACE=n
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_RCU_FANOUT=3
+CONFIG_RCU_FANOUT_EXACT=y
+CONFIG_RCU_FANOUT_LEAF=2
+CONFIG_RCU_NOCB_CPU=y
+CONFIG_RCU_NOCB_CPU_ALL=y
+CONFIG_DEBUG_LOCK_ALLOC=n
+CONFIG_RCU_CPU_STALL_INFO=n
+CONFIG_RCU_CPU_STALL_VERBOSE=n
+CONFIG_RCU_BOOST=n
+CONFIG_DEBUG_OBJECTS_RCU_HEAD=n
diff --git a/rcutorture/configs/rcu/TREE08-T b/rcutorture/configs/rcu/TREE08-T
new file mode 100644
index 0000000..a0f32fb
--- /dev/null
+++ b/rcutorture/configs/rcu/TREE08-T
@@ -0,0 +1,24 @@
+CONFIG_SMP=y
+CONFIG_NR_CPUS=16
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_HZ_PERIODIC=n
+CONFIG_NO_HZ_IDLE=y
+CONFIG_NO_HZ_FULL=n
+CONFIG_RCU_FAST_NO_HZ=n
+CONFIG_RCU_TRACE=y
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_RCU_FANOUT=3
+CONFIG_RCU_FANOUT_EXACT=y
+CONFIG_RCU_FANOUT_LEAF=2
+CONFIG_RCU_NOCB_CPU=y
+CONFIG_RCU_NOCB_CPU_ALL=y
+CONFIG_DEBUG_LOCK_ALLOC=n
+CONFIG_RCU_CPU_STALL_INFO=n
+CONFIG_RCU_CPU_STALL_VERBOSE=n
+CONFIG_RCU_BOOST=n
+CONFIG_DEBUG_OBJECTS_RCU_HEAD=n
diff --git a/rcutorture/configs/rcu/TREE08.boot b/rcutorture/configs/rcu/TREE08.boot
new file mode 100644
index 0000000..3b42b8b
--- /dev/null
+++ b/rcutorture/configs/rcu/TREE08.boot
@@ -0,0 +1 @@
+rcutorture.torture_type=sched
diff --git a/rcutorture/configs/rcu/TREE09 b/rcutorture/configs/rcu/TREE09
new file mode 100644
index 0000000..b7a62a5
--- /dev/null
+++ b/rcutorture/configs/rcu/TREE09
@@ -0,0 +1,19 @@
+CONFIG_SMP=n
+CONFIG_NR_CPUS=1
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_HZ_PERIODIC=n
+CONFIG_NO_HZ_IDLE=y
+CONFIG_NO_HZ_FULL=n
+CONFIG_RCU_TRACE=n
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_RCU_NOCB_CPU=n
+CONFIG_DEBUG_LOCK_ALLOC=n
+CONFIG_RCU_CPU_STALL_INFO=n
+CONFIG_RCU_CPU_STALL_VERBOSE=n
+CONFIG_RCU_BOOST=n
+CONFIG_DEBUG_OBJECTS_RCU_HEAD=n
diff --git a/rcutorture/configs/rcu/v0.0/CFLIST b/rcutorture/configs/rcu/v0.0/CFLIST
new file mode 100644
index 0000000..1822394
--- /dev/null
+++ b/rcutorture/configs/rcu/v0.0/CFLIST
@@ -0,0 +1,14 @@
+P1-S-T-NH-SD-SMP-HP
+P2-2-t-nh-sd-SMP-hp
+P3-3-T-nh-SD-SMP-hp
+P4-A-t-NH-sd-SMP-HP
+P5-U-T-NH-sd-SMP-hp
+N1-S-T-NH-SD-SMP-HP
+N2-2-t-nh-sd-SMP-hp
+N3-3-T-nh-SD-SMP-hp
+N4-A-t-NH-sd-SMP-HP
+N5-U-T-NH-sd-SMP-hp
+PT1-nh
+PT2-NH
+NT1-nh
+NT3-NH
diff --git a/rcutorture/configs/rcu/v0.0/N1-S-T-NH-SD-SMP-HP b/rcutorture/configs/rcu/v0.0/N1-S-T-NH-SD-SMP-HP
new file mode 100644
index 0000000..d3ef873
--- /dev/null
+++ b/rcutorture/configs/rcu/v0.0/N1-S-T-NH-SD-SMP-HP
@@ -0,0 +1,18 @@
+CONFIG_RCU_TRACE=y
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=8
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=y
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v0.0/N2-2-t-nh-sd-SMP-hp b/rcutorture/configs/rcu/v0.0/N2-2-t-nh-sd-SMP-hp
new file mode 100644
index 0000000..02e4185
--- /dev/null
+++ b/rcutorture/configs/rcu/v0.0/N2-2-t-nh-sd-SMP-hp
@@ -0,0 +1,20 @@
+CONFIG_RCU_TRACE=n
+CONFIG_NO_HZ=n
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=4
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v0.0/N3-3-T-nh-SD-SMP-hp b/rcutorture/configs/rcu/v0.0/N3-3-T-nh-SD-SMP-hp
new file mode 100644
index 0000000..b3100f6
--- /dev/null
+++ b/rcutorture/configs/rcu/v0.0/N3-3-T-nh-SD-SMP-hp
@@ -0,0 +1,22 @@
+CONFIG_RCU_TRACE=y
+CONFIG_NO_HZ=n
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=2
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_PROVE_LOCKING=y
+CONFIG_PROVE_RCU=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v0.0/N4-A-t-NH-sd-SMP-HP b/rcutorture/configs/rcu/v0.0/N4-A-t-NH-sd-SMP-HP
new file mode 100644
index 0000000..c56b445
--- /dev/null
+++ b/rcutorture/configs/rcu/v0.0/N4-A-t-NH-sd-SMP-HP
@@ -0,0 +1,18 @@
+CONFIG_RCU_TRACE=n
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=6
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=y
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v0.0/N5-U-T-NH-sd-SMP-hp b/rcutorture/configs/rcu/v0.0/N5-U-T-NH-sd-SMP-hp
new file mode 100644
index 0000000..90d924f
--- /dev/null
+++ b/rcutorture/configs/rcu/v0.0/N5-U-T-NH-sd-SMP-hp
@@ -0,0 +1,22 @@
+CONFIG_RCU_TRACE=y
+CONFIG_DEBUG_KERNEL=y
+CONFIG_RCU_CPU_STALL_INFO=y
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=6
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=y
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v0.0/NT1-nh b/rcutorture/configs/rcu/v0.0/NT1-nh
new file mode 100644
index 0000000..023f312
--- /dev/null
+++ b/rcutorture/configs/rcu/v0.0/NT1-nh
@@ -0,0 +1,23 @@
+#CHECK#CONFIG_TINY_RCU=y
+CONFIG_RCU_TRACE=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+#
+CONFIG_SMP=n
+#
+CONFIG_HOTPLUG_CPU=n
+#
+CONFIG_NO_HZ=n
+#
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+CONFIG_PROVE_LOCKING=y
+CONFIG_PROVE_RCU=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v0.0/NT3-NH b/rcutorture/configs/rcu/v0.0/NT3-NH
new file mode 100644
index 0000000..6fd0235
--- /dev/null
+++ b/rcutorture/configs/rcu/v0.0/NT3-NH
@@ -0,0 +1,20 @@
+#CHECK#CONFIG_TINY_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+#
+CONFIG_SMP=n
+#
+CONFIG_HOTPLUG_CPU=n
+#
+CONFIG_NO_HZ=y
+#
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v0.0/P1-S-T-NH-SD-SMP-HP b/rcutorture/configs/rcu/v0.0/P1-S-T-NH-SD-SMP-HP
new file mode 100644
index 0000000..f72402d
--- /dev/null
+++ b/rcutorture/configs/rcu/v0.0/P1-S-T-NH-SD-SMP-HP
@@ -0,0 +1,19 @@
+CONFIG_RCU_TRACE=y
+CONFIG_RCU_CPU_STALL_INFO=y
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=8
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=y
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v0.0/P2-2-t-nh-sd-SMP-hp b/rcutorture/configs/rcu/v0.0/P2-2-t-nh-sd-SMP-hp
new file mode 100644
index 0000000..0f3b667
--- /dev/null
+++ b/rcutorture/configs/rcu/v0.0/P2-2-t-nh-sd-SMP-hp
@@ -0,0 +1,20 @@
+CONFIG_RCU_TRACE=n
+CONFIG_NO_HZ=n
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=4
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v0.0/P3-3-T-nh-SD-SMP-hp b/rcutorture/configs/rcu/v0.0/P3-3-T-nh-SD-SMP-hp
new file mode 100644
index 0000000..b035e14
--- /dev/null
+++ b/rcutorture/configs/rcu/v0.0/P3-3-T-nh-SD-SMP-hp
@@ -0,0 +1,20 @@
+CONFIG_RCU_TRACE=y
+CONFIG_NO_HZ=n
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=2
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v0.0/P4-A-t-NH-sd-SMP-HP b/rcutorture/configs/rcu/v0.0/P4-A-t-NH-sd-SMP-HP
new file mode 100644
index 0000000..3ccf6a9
--- /dev/null
+++ b/rcutorture/configs/rcu/v0.0/P4-A-t-NH-sd-SMP-HP
@@ -0,0 +1,22 @@
+CONFIG_RCU_TRACE=n
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=6
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=y
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_RT_MUTEXES=y
+CONFIG_RCU_BOOST=y
+CONFIG_RCU_BOOST_PRIO=2
+CONFIG_PROVE_LOCKING=y
+CONFIG_PROVE_RCU=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v0.0/P5-U-T-NH-sd-SMP-hp b/rcutorture/configs/rcu/v0.0/P5-U-T-NH-sd-SMP-hp
new file mode 100644
index 0000000..a55c008
--- /dev/null
+++ b/rcutorture/configs/rcu/v0.0/P5-U-T-NH-sd-SMP-hp
@@ -0,0 +1,27 @@
+CONFIG_RCU_TRACE=y
+CONFIG_RCU_CPU_STALL_INFO=y
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=6
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=y
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_DEBUG_KERNEL=y
+CONFIG_DEBUG_OBJECTS=y
+CONFIG_DEBUG_OBJECTS_RCU_HEAD=y
+CONFIG_RT_MUTEXES=y
+CONFIG_RCU_BOOST=y
+CONFIG_RCU_BOOST_PRIO=2
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v0.0/PT1-nh b/rcutorture/configs/rcu/v0.0/PT1-nh
new file mode 100644
index 0000000..e3361c3
--- /dev/null
+++ b/rcutorture/configs/rcu/v0.0/PT1-nh
@@ -0,0 +1,23 @@
+CONFIG_TINY_PREEMPT_RCU=y
+CONFIG_RCU_BOOST=y
+CONFIG_RCU_BOOST_PRIO=2
+CONFIG_RCU_TRACE=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+#
+CONFIG_SMP=n
+#
+CONFIG_HOTPLUG_CPU=n
+#
+CONFIG_NO_HZ=n
+#
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v0.0/PT2-NH b/rcutorture/configs/rcu/v0.0/PT2-NH
new file mode 100644
index 0000000..64abfc3
--- /dev/null
+++ b/rcutorture/configs/rcu/v0.0/PT2-NH
@@ -0,0 +1,22 @@
+CONFIG_TINY_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+#
+CONFIG_SMP=n
+#
+CONFIG_HOTPLUG_CPU=n
+#
+CONFIG_NO_HZ=y
+#
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+CONFIG_PROVE_LOCKING=y
+CONFIG_PROVE_RCU=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v0.0/ver_functions.sh b/rcutorture/configs/rcu/v0.0/ver_functions.sh
new file mode 100644
index 0000000..5ace37a
--- /dev/null
+++ b/rcutorture/configs/rcu/v0.0/ver_functions.sh
@@ -0,0 +1,33 @@
+#!/bin/bash
+#
+# Kernel-version-dependent shell functions for the rest of the scripts.
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, you can access it online at
+# http://www.gnu.org/licenses/gpl-2.0.html.
+#
+# Copyright (C) IBM Corporation, 2013
+#
+# Authors: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
+
+# per_version_boot_params bootparam-string config-file seconds
+#
+# Adds per-version torture-module parameters to kernels supporting them.
+# Which old kernels do not.
+per_version_boot_params () {
+ echo rcutorture.stat_interval=15 \
+ rcutorture.shutdown_secs=$3 \
+ rcutorture.rcutorture_runnable=1 \
+ rcutorture.test_no_idle_hz=1 \
+ rcutorture.verbose=1
+}
diff --git a/rcutorture/configs/rcu/v3.12/CFLIST b/rcutorture/configs/rcu/v3.12/CFLIST
new file mode 100644
index 0000000..da4cbc6
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.12/CFLIST
@@ -0,0 +1,17 @@
+sysidleY.2013.06.19a
+sysidleN.2013.06.19a
+P1-S-T-NH-SD-SMP-HP
+P2-2-t-nh-sd-SMP-hp
+P3-3-T-nh-SD-SMP-hp
+P4-A-t-NH-sd-SMP-HP
+P5-U-T-NH-sd-SMP-hp
+P6---t-nh-SD-smp-hp
+N1-S-T-NH-SD-SMP-HP
+N2-2-t-nh-sd-SMP-hp
+N3-3-T-nh-SD-SMP-hp
+N4-A-t-NH-sd-SMP-HP
+N5-U-T-NH-sd-SMP-hp
+PT1-nh
+PT2-NH
+NT1-nh
+NT3-NH
diff --git a/rcutorture/configs/rcu/v3.12/N1-S-T-NH-SD-SMP-HP b/rcutorture/configs/rcu/v3.12/N1-S-T-NH-SD-SMP-HP
new file mode 100644
index 0000000..d81e11d
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.12/N1-S-T-NH-SD-SMP-HP
@@ -0,0 +1,19 @@
+CONFIG_RCU_TRACE=y
+CONFIG_RCU_FAST_NO_HZ=y
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=8
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=y
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.12/N2-2-t-nh-sd-SMP-hp b/rcutorture/configs/rcu/v3.12/N2-2-t-nh-sd-SMP-hp
new file mode 100644
index 0000000..02e4185
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.12/N2-2-t-nh-sd-SMP-hp
@@ -0,0 +1,20 @@
+CONFIG_RCU_TRACE=n
+CONFIG_NO_HZ=n
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=4
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.12/N3-3-T-nh-SD-SMP-hp b/rcutorture/configs/rcu/v3.12/N3-3-T-nh-SD-SMP-hp
new file mode 100644
index 0000000..b3100f6
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.12/N3-3-T-nh-SD-SMP-hp
@@ -0,0 +1,22 @@
+CONFIG_RCU_TRACE=y
+CONFIG_NO_HZ=n
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=2
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_PROVE_LOCKING=y
+CONFIG_PROVE_RCU=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.12/N4-A-t-NH-sd-SMP-HP b/rcutorture/configs/rcu/v3.12/N4-A-t-NH-sd-SMP-HP
new file mode 100644
index 0000000..c56b445
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.12/N4-A-t-NH-sd-SMP-HP
@@ -0,0 +1,18 @@
+CONFIG_RCU_TRACE=n
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=6
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=y
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.12/N5-U-T-NH-sd-SMP-hp b/rcutorture/configs/rcu/v3.12/N5-U-T-NH-sd-SMP-hp
new file mode 100644
index 0000000..90d924f
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.12/N5-U-T-NH-sd-SMP-hp
@@ -0,0 +1,22 @@
+CONFIG_RCU_TRACE=y
+CONFIG_DEBUG_KERNEL=y
+CONFIG_RCU_CPU_STALL_INFO=y
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=6
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=y
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.12/N6---t-nh-SD-smp-hp b/rcutorture/configs/rcu/v3.12/N6---t-nh-SD-smp-hp
new file mode 100644
index 0000000..0ccc36d
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.12/N6---t-nh-SD-smp-hp
@@ -0,0 +1,19 @@
+CONFIG_RCU_TRACE=n
+CONFIG_NO_HZ=n
+CONFIG_SMP=y
+CONFIG_NR_CPUS=1
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.12/N7-4-T-NH-SD-SMP-HP b/rcutorture/configs/rcu/v3.12/N7-4-T-NH-SD-SMP-HP
new file mode 100644
index 0000000..3f640cf
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.12/N7-4-T-NH-SD-SMP-HP
@@ -0,0 +1,26 @@
+CONFIG_RCU_TRACE=y
+CONFIG_DEBUG_KERNEL=y
+CONFIG_RCU_CPU_STALL_INFO=y
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=2
+CONFIG_NR_CPUS=16
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=y
+CONFIG_RCU_NOCB_CPU=y
+CONFIG_RCU_NOCB_CPU_NONE=y
+CONFIG_RCU_NOCB_CPU_ZERO=n
+CONFIG_RCU_NOCB_CPU_ALL=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.12/N8-2-T-NH-SD-SMP-HP b/rcutorture/configs/rcu/v3.12/N8-2-T-NH-SD-SMP-HP
new file mode 100644
index 0000000..285da2d
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.12/N8-2-T-NH-SD-SMP-HP
@@ -0,0 +1,22 @@
+CONFIG_RCU_TRACE=y
+CONFIG_DEBUG_KERNEL=y
+CONFIG_RCU_CPU_STALL_INFO=y
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=14
+CONFIG_NR_CPUS=16
+CONFIG_RCU_FANOUT_EXACT=y
+CONFIG_HOTPLUG_CPU=y
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.12/NT1-nh b/rcutorture/configs/rcu/v3.12/NT1-nh
new file mode 100644
index 0000000..023f312
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.12/NT1-nh
@@ -0,0 +1,23 @@
+#CHECK#CONFIG_TINY_RCU=y
+CONFIG_RCU_TRACE=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+#
+CONFIG_SMP=n
+#
+CONFIG_HOTPLUG_CPU=n
+#
+CONFIG_NO_HZ=n
+#
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+CONFIG_PROVE_LOCKING=y
+CONFIG_PROVE_RCU=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.12/NT3-NH b/rcutorture/configs/rcu/v3.12/NT3-NH
new file mode 100644
index 0000000..6fd0235
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.12/NT3-NH
@@ -0,0 +1,20 @@
+#CHECK#CONFIG_TINY_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+#
+CONFIG_SMP=n
+#
+CONFIG_HOTPLUG_CPU=n
+#
+CONFIG_NO_HZ=y
+#
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.12/P1-S-T-NH-SD-SMP-HP b/rcutorture/configs/rcu/v3.12/P1-S-T-NH-SD-SMP-HP
new file mode 100644
index 0000000..9647c44
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.12/P1-S-T-NH-SD-SMP-HP
@@ -0,0 +1,20 @@
+CONFIG_RCU_TRACE=y
+CONFIG_RCU_CPU_STALL_INFO=y
+CONFIG_NO_HZ=y
+CONFIG_RCU_FAST_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=8
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=y
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.12/P2-2-t-nh-sd-SMP-hp b/rcutorture/configs/rcu/v3.12/P2-2-t-nh-sd-SMP-hp
new file mode 100644
index 0000000..0f3b667
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.12/P2-2-t-nh-sd-SMP-hp
@@ -0,0 +1,20 @@
+CONFIG_RCU_TRACE=n
+CONFIG_NO_HZ=n
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=4
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.12/P3-3-T-nh-SD-SMP-hp b/rcutorture/configs/rcu/v3.12/P3-3-T-nh-SD-SMP-hp
new file mode 100644
index 0000000..b035e14
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.12/P3-3-T-nh-SD-SMP-hp
@@ -0,0 +1,20 @@
+CONFIG_RCU_TRACE=y
+CONFIG_NO_HZ=n
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=2
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.12/P4-A-t-NH-sd-SMP-HP b/rcutorture/configs/rcu/v3.12/P4-A-t-NH-sd-SMP-HP
new file mode 100644
index 0000000..3ccf6a9
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.12/P4-A-t-NH-sd-SMP-HP
@@ -0,0 +1,22 @@
+CONFIG_RCU_TRACE=n
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=6
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=y
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_RT_MUTEXES=y
+CONFIG_RCU_BOOST=y
+CONFIG_RCU_BOOST_PRIO=2
+CONFIG_PROVE_LOCKING=y
+CONFIG_PROVE_RCU=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.12/P5-U-T-NH-sd-SMP-hp b/rcutorture/configs/rcu/v3.12/P5-U-T-NH-sd-SMP-hp
new file mode 100644
index 0000000..a55c008
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.12/P5-U-T-NH-sd-SMP-hp
@@ -0,0 +1,27 @@
+CONFIG_RCU_TRACE=y
+CONFIG_RCU_CPU_STALL_INFO=y
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=6
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=y
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_DEBUG_KERNEL=y
+CONFIG_DEBUG_OBJECTS=y
+CONFIG_DEBUG_OBJECTS_RCU_HEAD=y
+CONFIG_RT_MUTEXES=y
+CONFIG_RCU_BOOST=y
+CONFIG_RCU_BOOST_PRIO=2
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.12/P6---t-nh-SD-smp-hp b/rcutorture/configs/rcu/v3.12/P6---t-nh-SD-smp-hp
new file mode 100644
index 0000000..f4c9175
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.12/P6---t-nh-SD-smp-hp
@@ -0,0 +1,18 @@
+CONFIG_RCU_TRACE=n
+CONFIG_NO_HZ=n
+CONFIG_SMP=n
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.12/P7-4-T-NH-SD-SMP-HP b/rcutorture/configs/rcu/v3.12/P7-4-T-NH-SD-SMP-HP
new file mode 100644
index 0000000..77a8c5b
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.12/P7-4-T-NH-SD-SMP-HP
@@ -0,0 +1,30 @@
+CONFIG_RCU_TRACE=y
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=2
+CONFIG_NR_CPUS=16
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=y
+CONFIG_RCU_NOCB_CPU=y
+CONFIG_RCU_NOCB_CPU_NONE=n
+CONFIG_RCU_NOCB_CPU_ZERO=n
+CONFIG_RCU_NOCB_CPU_ALL=y
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_PROVE_LOCKING=y
+CONFIG_PROVE_RCU=y
+CONFIG_DEBUG_KERNEL=y
+CONFIG_DEBUG_OBJECTS=y
+CONFIG_DEBUG_OBJECTS_RCU_HEAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_SLUB=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.12/P7-4-T-NH-SD-SMP-HP-all b/rcutorture/configs/rcu/v3.12/P7-4-T-NH-SD-SMP-HP-all
new file mode 100644
index 0000000..0eecebc
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.12/P7-4-T-NH-SD-SMP-HP-all
@@ -0,0 +1,30 @@
+CONFIG_RCU_TRACE=y
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=2
+CONFIG_NR_CPUS=16
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=y
+CONFIG_RCU_NOCB_CPU=y
+CONFIG_RCU_NOCB_CPU_NONE=y
+CONFIG_RCU_NOCB_CPU_ZERO=n
+CONFIG_RCU_NOCB_CPU_ALL=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_PROVE_LOCKING=y
+CONFIG_PROVE_RCU=y
+CONFIG_DEBUG_KERNEL=y
+CONFIG_DEBUG_OBJECTS=y
+CONFIG_DEBUG_OBJECTS_RCU_HEAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_SLUB=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.12/P7-4-T-NH-SD-SMP-HP-none b/rcutorture/configs/rcu/v3.12/P7-4-T-NH-SD-SMP-HP-none
new file mode 100644
index 0000000..0eecebc
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.12/P7-4-T-NH-SD-SMP-HP-none
@@ -0,0 +1,30 @@
+CONFIG_RCU_TRACE=y
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=2
+CONFIG_NR_CPUS=16
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=y
+CONFIG_RCU_NOCB_CPU=y
+CONFIG_RCU_NOCB_CPU_NONE=y
+CONFIG_RCU_NOCB_CPU_ZERO=n
+CONFIG_RCU_NOCB_CPU_ALL=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_PROVE_LOCKING=y
+CONFIG_PROVE_RCU=y
+CONFIG_DEBUG_KERNEL=y
+CONFIG_DEBUG_OBJECTS=y
+CONFIG_DEBUG_OBJECTS_RCU_HEAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_SLUB=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.12/P7-4-T-NH-SD-SMP-hp b/rcutorture/configs/rcu/v3.12/P7-4-T-NH-SD-SMP-hp
new file mode 100644
index 0000000..588bc70
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.12/P7-4-T-NH-SD-SMP-hp
@@ -0,0 +1,30 @@
+CONFIG_RCU_TRACE=y
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=2
+CONFIG_NR_CPUS=16
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=n
+CONFIG_RCU_NOCB_CPU=y
+CONFIG_RCU_NOCB_CPU_NONE=n
+CONFIG_RCU_NOCB_CPU_ZERO=y
+CONFIG_RCU_NOCB_CPU_ALL=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_PROVE_LOCKING=y
+CONFIG_PROVE_RCU=y
+CONFIG_DEBUG_KERNEL=y
+CONFIG_DEBUG_OBJECTS=y
+CONFIG_DEBUG_OBJECTS_RCU_HEAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_SLUB=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.12/PT1-nh b/rcutorture/configs/rcu/v3.12/PT1-nh
new file mode 100644
index 0000000..e3361c3
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.12/PT1-nh
@@ -0,0 +1,23 @@
+CONFIG_TINY_PREEMPT_RCU=y
+CONFIG_RCU_BOOST=y
+CONFIG_RCU_BOOST_PRIO=2
+CONFIG_RCU_TRACE=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+#
+CONFIG_SMP=n
+#
+CONFIG_HOTPLUG_CPU=n
+#
+CONFIG_NO_HZ=n
+#
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.12/PT2-NH b/rcutorture/configs/rcu/v3.12/PT2-NH
new file mode 100644
index 0000000..64abfc3
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.12/PT2-NH
@@ -0,0 +1,22 @@
+CONFIG_TINY_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+#
+CONFIG_SMP=n
+#
+CONFIG_HOTPLUG_CPU=n
+#
+CONFIG_NO_HZ=y
+#
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+CONFIG_PROVE_LOCKING=y
+CONFIG_PROVE_RCU=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.3/CFLIST b/rcutorture/configs/rcu/v3.3/CFLIST
new file mode 100644
index 0000000..1822394
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.3/CFLIST
@@ -0,0 +1,14 @@
+P1-S-T-NH-SD-SMP-HP
+P2-2-t-nh-sd-SMP-hp
+P3-3-T-nh-SD-SMP-hp
+P4-A-t-NH-sd-SMP-HP
+P5-U-T-NH-sd-SMP-hp
+N1-S-T-NH-SD-SMP-HP
+N2-2-t-nh-sd-SMP-hp
+N3-3-T-nh-SD-SMP-hp
+N4-A-t-NH-sd-SMP-HP
+N5-U-T-NH-sd-SMP-hp
+PT1-nh
+PT2-NH
+NT1-nh
+NT3-NH
diff --git a/rcutorture/configs/rcu/v3.3/N1-S-T-NH-SD-SMP-HP b/rcutorture/configs/rcu/v3.3/N1-S-T-NH-SD-SMP-HP
new file mode 100644
index 0000000..d81e11d
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.3/N1-S-T-NH-SD-SMP-HP
@@ -0,0 +1,19 @@
+CONFIG_RCU_TRACE=y
+CONFIG_RCU_FAST_NO_HZ=y
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=8
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=y
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.3/N2-2-t-nh-sd-SMP-hp b/rcutorture/configs/rcu/v3.3/N2-2-t-nh-sd-SMP-hp
new file mode 100644
index 0000000..02e4185
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.3/N2-2-t-nh-sd-SMP-hp
@@ -0,0 +1,20 @@
+CONFIG_RCU_TRACE=n
+CONFIG_NO_HZ=n
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=4
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.3/N3-3-T-nh-SD-SMP-hp b/rcutorture/configs/rcu/v3.3/N3-3-T-nh-SD-SMP-hp
new file mode 100644
index 0000000..b3100f6
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.3/N3-3-T-nh-SD-SMP-hp
@@ -0,0 +1,22 @@
+CONFIG_RCU_TRACE=y
+CONFIG_NO_HZ=n
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=2
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_PROVE_LOCKING=y
+CONFIG_PROVE_RCU=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.3/N4-A-t-NH-sd-SMP-HP b/rcutorture/configs/rcu/v3.3/N4-A-t-NH-sd-SMP-HP
new file mode 100644
index 0000000..c56b445
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.3/N4-A-t-NH-sd-SMP-HP
@@ -0,0 +1,18 @@
+CONFIG_RCU_TRACE=n
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=6
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=y
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.3/N5-U-T-NH-sd-SMP-hp b/rcutorture/configs/rcu/v3.3/N5-U-T-NH-sd-SMP-hp
new file mode 100644
index 0000000..90d924f
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.3/N5-U-T-NH-sd-SMP-hp
@@ -0,0 +1,22 @@
+CONFIG_RCU_TRACE=y
+CONFIG_DEBUG_KERNEL=y
+CONFIG_RCU_CPU_STALL_INFO=y
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=6
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=y
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.3/NT1-nh b/rcutorture/configs/rcu/v3.3/NT1-nh
new file mode 100644
index 0000000..023f312
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.3/NT1-nh
@@ -0,0 +1,23 @@
+#CHECK#CONFIG_TINY_RCU=y
+CONFIG_RCU_TRACE=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+#
+CONFIG_SMP=n
+#
+CONFIG_HOTPLUG_CPU=n
+#
+CONFIG_NO_HZ=n
+#
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+CONFIG_PROVE_LOCKING=y
+CONFIG_PROVE_RCU=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.3/NT3-NH b/rcutorture/configs/rcu/v3.3/NT3-NH
new file mode 100644
index 0000000..6fd0235
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.3/NT3-NH
@@ -0,0 +1,20 @@
+#CHECK#CONFIG_TINY_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+#
+CONFIG_SMP=n
+#
+CONFIG_HOTPLUG_CPU=n
+#
+CONFIG_NO_HZ=y
+#
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.3/P1-S-T-NH-SD-SMP-HP b/rcutorture/configs/rcu/v3.3/P1-S-T-NH-SD-SMP-HP
new file mode 100644
index 0000000..9647c44
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.3/P1-S-T-NH-SD-SMP-HP
@@ -0,0 +1,20 @@
+CONFIG_RCU_TRACE=y
+CONFIG_RCU_CPU_STALL_INFO=y
+CONFIG_NO_HZ=y
+CONFIG_RCU_FAST_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=8
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=y
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.3/P2-2-t-nh-sd-SMP-hp b/rcutorture/configs/rcu/v3.3/P2-2-t-nh-sd-SMP-hp
new file mode 100644
index 0000000..0f3b667
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.3/P2-2-t-nh-sd-SMP-hp
@@ -0,0 +1,20 @@
+CONFIG_RCU_TRACE=n
+CONFIG_NO_HZ=n
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=4
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.3/P3-3-T-nh-SD-SMP-hp b/rcutorture/configs/rcu/v3.3/P3-3-T-nh-SD-SMP-hp
new file mode 100644
index 0000000..b035e14
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.3/P3-3-T-nh-SD-SMP-hp
@@ -0,0 +1,20 @@
+CONFIG_RCU_TRACE=y
+CONFIG_NO_HZ=n
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=2
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.3/P4-A-t-NH-sd-SMP-HP b/rcutorture/configs/rcu/v3.3/P4-A-t-NH-sd-SMP-HP
new file mode 100644
index 0000000..3ccf6a9
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.3/P4-A-t-NH-sd-SMP-HP
@@ -0,0 +1,22 @@
+CONFIG_RCU_TRACE=n
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=6
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=y
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_RT_MUTEXES=y
+CONFIG_RCU_BOOST=y
+CONFIG_RCU_BOOST_PRIO=2
+CONFIG_PROVE_LOCKING=y
+CONFIG_PROVE_RCU=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.3/P5-U-T-NH-sd-SMP-hp b/rcutorture/configs/rcu/v3.3/P5-U-T-NH-sd-SMP-hp
new file mode 100644
index 0000000..a55c008
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.3/P5-U-T-NH-sd-SMP-hp
@@ -0,0 +1,27 @@
+CONFIG_RCU_TRACE=y
+CONFIG_RCU_CPU_STALL_INFO=y
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=6
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=y
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_DEBUG_KERNEL=y
+CONFIG_DEBUG_OBJECTS=y
+CONFIG_DEBUG_OBJECTS_RCU_HEAD=y
+CONFIG_RT_MUTEXES=y
+CONFIG_RCU_BOOST=y
+CONFIG_RCU_BOOST_PRIO=2
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.3/PT1-nh b/rcutorture/configs/rcu/v3.3/PT1-nh
new file mode 100644
index 0000000..e3361c3
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.3/PT1-nh
@@ -0,0 +1,23 @@
+CONFIG_TINY_PREEMPT_RCU=y
+CONFIG_RCU_BOOST=y
+CONFIG_RCU_BOOST_PRIO=2
+CONFIG_RCU_TRACE=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+#
+CONFIG_SMP=n
+#
+CONFIG_HOTPLUG_CPU=n
+#
+CONFIG_NO_HZ=n
+#
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.3/PT2-NH b/rcutorture/configs/rcu/v3.3/PT2-NH
new file mode 100644
index 0000000..64abfc3
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.3/PT2-NH
@@ -0,0 +1,22 @@
+CONFIG_TINY_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+#
+CONFIG_SMP=n
+#
+CONFIG_HOTPLUG_CPU=n
+#
+CONFIG_NO_HZ=y
+#
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+CONFIG_PROVE_LOCKING=y
+CONFIG_PROVE_RCU=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.3/ver_functions.sh b/rcutorture/configs/rcu/v3.3/ver_functions.sh
new file mode 100644
index 0000000..bae5569
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.3/ver_functions.sh
@@ -0,0 +1,44 @@
+#!/bin/bash
+#
+# Kernel-version-dependent shell functions for the rest of the scripts.
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, you can access it online at
+# http://www.gnu.org/licenses/gpl-2.0.html.
+#
+# Copyright (C) IBM Corporation, 2013
+#
+# Authors: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
+
+# rcutorture_param_onoff bootparam-string config-file
+#
+# Adds onoff rcutorture module parameters to kernels having it.
+rcutorture_param_onoff () {
+ if ! bootparam_hotplug_cpu "$1" && configfrag_hotplug_cpu "$2"
+ then
+ echo CPU-hotplug kernel, adding rcutorture onoff. 1>&2
+ echo rcutorture.onoff_interval=3 rcutorture.onoff_holdoff=30
+ fi
+}
+
+# per_version_boot_params bootparam-string config-file seconds
+#
+# Adds per-version torture-module parameters to kernels supporting them.
+per_version_boot_params () {
+ echo $1 `rcutorture_param_onoff "$1" "$2"` \
+ rcutorture.stat_interval=15 \
+ rcutorture.shutdown_secs=$3 \
+ rcutorture.rcutorture_runnable=1 \
+ rcutorture.test_no_idle_hz=1 \
+ rcutorture.verbose=1
+}
diff --git a/rcutorture/configs/rcu/v3.5/CFLIST b/rcutorture/configs/rcu/v3.5/CFLIST
new file mode 100644
index 0000000..1822394
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.5/CFLIST
@@ -0,0 +1,14 @@
+P1-S-T-NH-SD-SMP-HP
+P2-2-t-nh-sd-SMP-hp
+P3-3-T-nh-SD-SMP-hp
+P4-A-t-NH-sd-SMP-HP
+P5-U-T-NH-sd-SMP-hp
+N1-S-T-NH-SD-SMP-HP
+N2-2-t-nh-sd-SMP-hp
+N3-3-T-nh-SD-SMP-hp
+N4-A-t-NH-sd-SMP-HP
+N5-U-T-NH-sd-SMP-hp
+PT1-nh
+PT2-NH
+NT1-nh
+NT3-NH
diff --git a/rcutorture/configs/rcu/v3.5/N1-S-T-NH-SD-SMP-HP b/rcutorture/configs/rcu/v3.5/N1-S-T-NH-SD-SMP-HP
new file mode 100644
index 0000000..d81e11d
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.5/N1-S-T-NH-SD-SMP-HP
@@ -0,0 +1,19 @@
+CONFIG_RCU_TRACE=y
+CONFIG_RCU_FAST_NO_HZ=y
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=8
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=y
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.5/N2-2-t-nh-sd-SMP-hp b/rcutorture/configs/rcu/v3.5/N2-2-t-nh-sd-SMP-hp
new file mode 100644
index 0000000..02e4185
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.5/N2-2-t-nh-sd-SMP-hp
@@ -0,0 +1,20 @@
+CONFIG_RCU_TRACE=n
+CONFIG_NO_HZ=n
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=4
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.5/N3-3-T-nh-SD-SMP-hp b/rcutorture/configs/rcu/v3.5/N3-3-T-nh-SD-SMP-hp
new file mode 100644
index 0000000..b3100f6
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.5/N3-3-T-nh-SD-SMP-hp
@@ -0,0 +1,22 @@
+CONFIG_RCU_TRACE=y
+CONFIG_NO_HZ=n
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=2
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_PROVE_LOCKING=y
+CONFIG_PROVE_RCU=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.5/N4-A-t-NH-sd-SMP-HP b/rcutorture/configs/rcu/v3.5/N4-A-t-NH-sd-SMP-HP
new file mode 100644
index 0000000..c56b445
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.5/N4-A-t-NH-sd-SMP-HP
@@ -0,0 +1,18 @@
+CONFIG_RCU_TRACE=n
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=6
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=y
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.5/N5-U-T-NH-sd-SMP-hp b/rcutorture/configs/rcu/v3.5/N5-U-T-NH-sd-SMP-hp
new file mode 100644
index 0000000..90d924f
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.5/N5-U-T-NH-sd-SMP-hp
@@ -0,0 +1,22 @@
+CONFIG_RCU_TRACE=y
+CONFIG_DEBUG_KERNEL=y
+CONFIG_RCU_CPU_STALL_INFO=y
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=6
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=y
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+#CHECK#CONFIG_TREE_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.5/NT1-nh b/rcutorture/configs/rcu/v3.5/NT1-nh
new file mode 100644
index 0000000..023f312
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.5/NT1-nh
@@ -0,0 +1,23 @@
+#CHECK#CONFIG_TINY_RCU=y
+CONFIG_RCU_TRACE=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+#
+CONFIG_SMP=n
+#
+CONFIG_HOTPLUG_CPU=n
+#
+CONFIG_NO_HZ=n
+#
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+CONFIG_PROVE_LOCKING=y
+CONFIG_PROVE_RCU=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.5/NT3-NH b/rcutorture/configs/rcu/v3.5/NT3-NH
new file mode 100644
index 0000000..6fd0235
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.5/NT3-NH
@@ -0,0 +1,20 @@
+#CHECK#CONFIG_TINY_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+#
+CONFIG_SMP=n
+#
+CONFIG_HOTPLUG_CPU=n
+#
+CONFIG_NO_HZ=y
+#
+CONFIG_PREEMPT_NONE=y
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=n
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.5/P1-S-T-NH-SD-SMP-HP b/rcutorture/configs/rcu/v3.5/P1-S-T-NH-SD-SMP-HP
new file mode 100644
index 0000000..9647c44
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.5/P1-S-T-NH-SD-SMP-HP
@@ -0,0 +1,20 @@
+CONFIG_RCU_TRACE=y
+CONFIG_RCU_CPU_STALL_INFO=y
+CONFIG_NO_HZ=y
+CONFIG_RCU_FAST_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=8
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=y
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.5/P2-2-t-nh-sd-SMP-hp b/rcutorture/configs/rcu/v3.5/P2-2-t-nh-sd-SMP-hp
new file mode 100644
index 0000000..0f3b667
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.5/P2-2-t-nh-sd-SMP-hp
@@ -0,0 +1,20 @@
+CONFIG_RCU_TRACE=n
+CONFIG_NO_HZ=n
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=4
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.5/P3-3-T-nh-SD-SMP-hp b/rcutorture/configs/rcu/v3.5/P3-3-T-nh-SD-SMP-hp
new file mode 100644
index 0000000..b035e14
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.5/P3-3-T-nh-SD-SMP-hp
@@ -0,0 +1,20 @@
+CONFIG_RCU_TRACE=y
+CONFIG_NO_HZ=n
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=2
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.5/P4-A-t-NH-sd-SMP-HP b/rcutorture/configs/rcu/v3.5/P4-A-t-NH-sd-SMP-HP
new file mode 100644
index 0000000..3ccf6a9
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.5/P4-A-t-NH-sd-SMP-HP
@@ -0,0 +1,22 @@
+CONFIG_RCU_TRACE=n
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=6
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=n
+CONFIG_HOTPLUG_CPU=y
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_RT_MUTEXES=y
+CONFIG_RCU_BOOST=y
+CONFIG_RCU_BOOST_PRIO=2
+CONFIG_PROVE_LOCKING=y
+CONFIG_PROVE_RCU=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.5/P5-U-T-NH-sd-SMP-hp b/rcutorture/configs/rcu/v3.5/P5-U-T-NH-sd-SMP-hp
new file mode 100644
index 0000000..a55c008
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.5/P5-U-T-NH-sd-SMP-hp
@@ -0,0 +1,27 @@
+CONFIG_RCU_TRACE=y
+CONFIG_RCU_CPU_STALL_INFO=y
+CONFIG_NO_HZ=y
+CONFIG_SMP=y
+CONFIG_RCU_FANOUT=6
+CONFIG_NR_CPUS=8
+CONFIG_RCU_FANOUT_EXACT=y
+CONFIG_HOTPLUG_CPU=n
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+#CHECK#CONFIG_TREE_PREEMPT_RCU=y
+CONFIG_DEBUG_KERNEL=y
+CONFIG_DEBUG_OBJECTS=y
+CONFIG_DEBUG_OBJECTS_RCU_HEAD=y
+CONFIG_RT_MUTEXES=y
+CONFIG_RCU_BOOST=y
+CONFIG_RCU_BOOST_PRIO=2
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.5/PT1-nh b/rcutorture/configs/rcu/v3.5/PT1-nh
new file mode 100644
index 0000000..e3361c3
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.5/PT1-nh
@@ -0,0 +1,23 @@
+CONFIG_TINY_PREEMPT_RCU=y
+CONFIG_RCU_BOOST=y
+CONFIG_RCU_BOOST_PRIO=2
+CONFIG_RCU_TRACE=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+#
+CONFIG_SMP=n
+#
+CONFIG_HOTPLUG_CPU=n
+#
+CONFIG_NO_HZ=n
+#
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.5/PT2-NH b/rcutorture/configs/rcu/v3.5/PT2-NH
new file mode 100644
index 0000000..64abfc3
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.5/PT2-NH
@@ -0,0 +1,22 @@
+CONFIG_TINY_PREEMPT_RCU=y
+CONFIG_RCU_TORTURE_TEST=m
+CONFIG_MODULE_UNLOAD=y
+CONFIG_SUSPEND=n
+CONFIG_HIBERNATION=n
+#
+CONFIG_SMP=n
+#
+CONFIG_HOTPLUG_CPU=n
+#
+CONFIG_NO_HZ=y
+#
+CONFIG_PREEMPT_NONE=n
+CONFIG_PREEMPT_VOLUNTARY=n
+CONFIG_PREEMPT=y
+CONFIG_PROVE_LOCKING=y
+CONFIG_PROVE_RCU=y
+CONFIG_SYSFS_DEPRECATED_V2=y
+CONFIG_IKCONFIG=y
+CONFIG_IKCONFIG_PROC=y
+CONFIG_PRINTK_TIME=y
+
diff --git a/rcutorture/configs/rcu/v3.5/ver_functions.sh b/rcutorture/configs/rcu/v3.5/ver_functions.sh
new file mode 100644
index 0000000..8977d8d
--- /dev/null
+++ b/rcutorture/configs/rcu/v3.5/ver_functions.sh
@@ -0,0 +1,57 @@
+#!/bin/bash
+#
+# Kernel-version-dependent shell functions for the rest of the scripts.
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, you can access it online at
+# http://www.gnu.org/licenses/gpl-2.0.html.
+#
+# Copyright (C) IBM Corporation, 2013
+#
+# Authors: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
+
+# rcutorture_param_n_barrier_cbs bootparam-string
+#
+# Adds n_barrier_cbs rcutorture module parameter to kernels having it.
+rcutorture_param_n_barrier_cbs () {
+ if echo $1 | grep -q "rcutorture\.n_barrier_cbs"
+ then
+ :
+ else
+ echo rcutorture.n_barrier_cbs=4
+ fi
+}
+
+# rcutorture_param_onoff bootparam-string config-file
+#
+# Adds onoff rcutorture module parameters to kernels having it.
+rcutorture_param_onoff () {
+ if ! bootparam_hotplug_cpu "$1" && configfrag_hotplug_cpu "$2"
+ then
+ echo CPU-hotplug kernel, adding rcutorture onoff. 1>&2
+ echo rcutorture.onoff_interval=3 rcutorture.onoff_holdoff=30
+ fi
+}
+
+# per_version_boot_params bootparam-string config-file seconds
+#
+# Adds per-version torture-module parameters to kernels supporting them.
+per_version_boot_params () {
+ echo $1 `rcutorture_param_onoff "$1" "$2"` \
+ `rcutorture_param_n_barrier_cbs "$1"` \
+ rcutorture.stat_interval=15 \
+ rcutorture.shutdown_secs=$3 \
+ rcutorture.rcutorture_runnable=1 \
+ rcutorture.test_no_idle_hz=1 \
+ rcutorture.verbose=1
+}
diff --git a/rcutorture/configs/rcu/ver_functions.sh b/rcutorture/configs/rcu/ver_functions.sh
new file mode 100644
index 0000000..8977d8d
--- /dev/null
+++ b/rcutorture/configs/rcu/ver_functions.sh
@@ -0,0 +1,57 @@
+#!/bin/bash
+#
+# Kernel-version-dependent shell functions for the rest of the scripts.
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, you can access it online at
+# http://www.gnu.org/licenses/gpl-2.0.html.
+#
+# Copyright (C) IBM Corporation, 2013
+#
+# Authors: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
+
+# rcutorture_param_n_barrier_cbs bootparam-string
+#
+# Adds n_barrier_cbs rcutorture module parameter to kernels having it.
+rcutorture_param_n_barrier_cbs () {
+ if echo $1 | grep -q "rcutorture\.n_barrier_cbs"
+ then
+ :
+ else
+ echo rcutorture.n_barrier_cbs=4
+ fi
+}
+
+# rcutorture_param_onoff bootparam-string config-file
+#
+# Adds onoff rcutorture module parameters to kernels having it.
+rcutorture_param_onoff () {
+ if ! bootparam_hotplug_cpu "$1" && configfrag_hotplug_cpu "$2"
+ then
+ echo CPU-hotplug kernel, adding rcutorture onoff. 1>&2
+ echo rcutorture.onoff_interval=3 rcutorture.onoff_holdoff=30
+ fi
+}
+
+# per_version_boot_params bootparam-string config-file seconds
+#
+# Adds per-version torture-module parameters to kernels supporting them.
+per_version_boot_params () {
+ echo $1 `rcutorture_param_onoff "$1" "$2"` \
+ `rcutorture_param_n_barrier_cbs "$1"` \
+ rcutorture.stat_interval=15 \
+ rcutorture.shutdown_secs=$3 \
+ rcutorture.rcutorture_runnable=1 \
+ rcutorture.test_no_idle_hz=1 \
+ rcutorture.verbose=1
+}
diff --git a/rcutorture/doc/TINY_RCU.txt b/rcutorture/doc/TINY_RCU.txt
new file mode 100644
index 0000000..28db67b
--- /dev/null
+++ b/rcutorture/doc/TINY_RCU.txt
@@ -0,0 +1,40 @@
+This document gives a brief rationale for the TINY_RCU test cases.
+
+
+Kconfig Parameters:
+
+CONFIG_DEBUG_LOCK_ALLOC -- Do all three and none of the three.
+CONFIG_PREEMPT_COUNT
+CONFIG_RCU_TRACE
+
+The theory here is that randconfig testing will hit the other six possible
+combinations of these parameters.
+
+
+Kconfig Parameters Ignored:
+
+CONFIG_DEBUG_OBJECTS_RCU_HEAD
+CONFIG_PROVE_RCU
+
+ In common code tested by TREE_RCU test cases.
+
+CONFIG_NO_HZ_FULL_SYSIDLE
+CONFIG_RCU_NOCB_CPU
+CONFIG_RCU_USER_QS
+
+ Meaningless for TINY_RCU.
+
+CONFIG_RCU_STALL_COMMON
+CONFIG_RCU_TORTURE_TEST
+
+ Redundant with CONFIG_RCU_TRACE.
+
+CONFIG_HOTPLUG_CPU
+CONFIG_PREEMPT
+CONFIG_PREEMPT_RCU
+CONFIG_SMP
+CONFIG_TINY_RCU
+CONFIG_TREE_PREEMPT_RCU
+CONFIG_TREE_RCU
+
+ All forced by CONFIG_TINY_RCU.
diff --git a/rcutorture/doc/TREE_RCU-kconfig.txt b/rcutorture/doc/TREE_RCU-kconfig.txt
new file mode 100644
index 0000000..3e588db
--- /dev/null
+++ b/rcutorture/doc/TREE_RCU-kconfig.txt
@@ -0,0 +1,94 @@
+This document gives a brief rationale for the TREE_RCU-related test
+cases, a group that includes TREE_PREEMPT_RCU.
+
+
+Kconfig Parameters:
+
+CONFIG_DEBUG_LOCK_ALLOC -- Do three, covering CONFIG_PROVE_LOCKING & not.
+CONFIG_DEBUG_OBJECTS_RCU_HEAD -- Do one.
+CONFIG_HOTPLUG_CPU -- Do half. (Every second.)
+CONFIG_HZ_PERIODIC -- Do one.
+CONFIG_NO_HZ_IDLE -- Do those not otherwise specified. (Groups of two.)
+CONFIG_NO_HZ_FULL -- Do two, one with CONFIG_NO_HZ_FULL_SYSIDLE.
+CONFIG_NO_HZ_FULL_SYSIDLE -- Do one.
+CONFIG_PREEMPT -- Do half. (First three and #8.)
+CONFIG_PROVE_LOCKING -- Do all but two, covering CONFIG_PROVE_RCU and not.
+CONFIG_PROVE_RCU -- Do all but one under CONFIG_PROVE_LOCKING.
+CONFIG_RCU_BOOST -- one of TREE_PREEMPT_RCU.
+CONFIG_RCU_BOOST_PRIO -- set to 2 for _BOOST testing.
+CONFIG_RCU_CPU_STALL_INFO -- do one with and without _VERBOSE.
+CONFIG_RCU_CPU_STALL_VERBOSE -- do one with and without _INFO.
+CONFIG_RCU_FANOUT -- Cover hierarchy as currently, but overlap with others.
+CONFIG_RCU_FANOUT_EXACT -- Do one.
+CONFIG_RCU_FANOUT_LEAF -- Do one non-default.
+CONFIG_RCU_FAST_NO_HZ -- Do one, but not with CONFIG_RCU_NOCB_CPU_ALL.
+CONFIG_RCU_NOCB_CPU -- Do three, see below.
+CONFIG_RCU_NOCB_CPU_ALL -- Do one.
+CONFIG_RCU_NOCB_CPU_NONE -- Do one.
+CONFIG_RCU_NOCB_CPU_ZERO -- Do one.
+CONFIG_RCU_TRACE -- Do half.
+CONFIG_SMP -- Need one !SMP for TREE_PREEMPT_RCU.
+RCU-bh: Do one with PREEMPT and one with !PREEMPT.
+RCU-sched: Do one with PREEMPT but not BOOST.
+
+
+Hierarchy:
+
+TREE01. CONFIG_NR_CPUS=8, CONFIG_RCU_FANOUT=8, CONFIG_RCU_FANOUT_EXACT=n.
+TREE02. CONFIG_NR_CPUS=8, CONFIG_RCU_FANOUT=3, CONFIG_RCU_FANOUT_EXACT=n,
+ CONFIG_RCU_FANOUT_LEAF=3.
+TREE03. CONFIG_NR_CPUS=8, CONFIG_RCU_FANOUT=4, CONFIG_RCU_FANOUT_EXACT=n,
+ CONFIG_RCU_FANOUT_LEAF=4.
+TREE04. CONFIG_NR_CPUS=8, CONFIG_RCU_FANOUT=2, CONFIG_RCU_FANOUT_EXACT=n,
+ CONFIG_RCU_FANOUT_LEAF=2.
+TREE05. CONFIG_NR_CPUS=8, CONFIG_RCU_FANOUT=6, CONFIG_RCU_FANOUT_EXACT=n
+ CONFIG_RCU_FANOUT_LEAF=6.
+TREE06. CONFIG_NR_CPUS=8, CONFIG_RCU_FANOUT=6, CONFIG_RCU_FANOUT_EXACT=y
+ CONFIG_RCU_FANOUT_LEAF=6.
+TREE07. CONFIG_NR_CPUS=16, CONFIG_RCU_FANOUT=2, CONFIG_RCU_FANOUT_EXACT=n,
+ CONFIG_RCU_FANOUT_LEAF=2.
+TREE08. CONFIG_NR_CPUS=16, CONFIG_RCU_FANOUT=3, CONFIG_RCU_FANOUT_EXACT=y,
+ CONFIG_RCU_FANOUT_LEAF=2.
+TREE09. CONFIG_NR_CPUS=1.
+
+
+Kconfig Parameters Ignored:
+
+CONFIG_64BIT
+
+ Used only to check CONFIG_RCU_FANOUT value, inspection suffices.
+
+CONFIG_NO_HZ_FULL_SYSIDLE_SMALL
+
+ Defer until Frederic uses this.
+
+CONFIG_PREEMPT_COUNT
+CONFIG_PREEMPT_RCU
+
+ Redundant with CONFIG_PREEMPT, ignore.
+
+CONFIG_RCU_BOOST_DELAY
+
+ Inspection suffices, ignore.
+
+CONFIG_RCU_CPU_STALL_TIMEOUT
+
+ Inspection suffices, ignore.
+
+CONFIG_RCU_STALL_COMMON
+
+ Implied by TREE_RCU and TREE_PREEMPT_RCU.
+
+CONFIG_RCU_TORTURE_TEST
+CONFIG_RCU_TORTURE_TEST_RUNNABLE
+
+ Always used in KVM testing.
+
+CONFIG_RCU_USER_QS
+
+ Redundant with CONFIG_NO_HZ_FULL.
+
+CONFIG_TREE_PREEMPT_RCU
+CONFIG_TREE_RCU
+
+ These are controlled by CONFIG_PREEMPT.
diff --git a/rcutorture/doc/initrd.txt b/rcutorture/doc/initrd.txt
new file mode 100644
index 0000000..49d134c
--- /dev/null
+++ b/rcutorture/doc/initrd.txt
@@ -0,0 +1,90 @@
+This document describes one way to create the initrd directory hierarchy
+in order to allow an initrd to be built into your kernel. The trick
+here is to steal the initrd file used on your Linux laptop, Ubuntu in
+this case. There are probably much better ways of doing this.
+
+That said, here are the commands:
+
+------------------------------------------------------------------------
+zcat /initrd.img > /tmp/initrd.img.zcat
+mkdir initrd
+cd initrd
+cpio -id < /tmp/initrd.img.zcat
+------------------------------------------------------------------------
+
+Interestingly enough, if you are running rcutorture, you don't really
+need userspace in many cases. Running without userspace has the
+advantage of allowing you to test your kernel independently of the
+distro in place, the root-filesystem layout, and so on. To make this
+happen, put the following script in the initrd's tree's "/init" file,
+with 0755 mode.
+
+------------------------------------------------------------------------
+#!/bin/sh
+
+[ -d /dev ] || mkdir -m 0755 /dev
+[ -d /root ] || mkdir -m 0700 /root
+[ -d /sys ] || mkdir /sys
+[ -d /proc ] || mkdir /proc
+[ -d /tmp ] || mkdir /tmp
+mkdir -p /var/lock
+mount -t sysfs -o nodev,noexec,nosuid sysfs /sys
+mount -t proc -o nodev,noexec,nosuid proc /proc
+# Some things don't work properly without /etc/mtab.
+ln -sf /proc/mounts /etc/mtab
+
+# Note that this only becomes /dev on the real filesystem if udev's scripts
+# are used; which they will be, but it's worth pointing out
+if ! mount -t devtmpfs -o mode=0755 udev /dev; then
+ echo "W: devtmpfs not available, falling back to tmpfs for /dev"
+ mount -t tmpfs -o mode=0755 udev /dev
+ [ -e /dev/console ] || mknod --mode=600 /dev/console c 5 1
+ [ -e /dev/kmsg ] || mknod --mode=644 /dev/kmsg c 1 11
+ [ -e /dev/null ] || mknod --mode=666 /dev/null c 1 3
+fi
+
+mkdir /dev/pts
+mount -t devpts -o noexec,nosuid,gid=5,mode=0620 devpts /dev/pts || true
+mount -t tmpfs -o "nosuid,size=20%,mode=0755" tmpfs /run
+mkdir /run/initramfs
+# compatibility symlink for the pre-oneiric locations
+ln -s /run/initramfs /dev/.initramfs
+
+# Export relevant variables
+export ROOT=
+export ROOTDELAY=
+export ROOTFLAGS=
+export ROOTFSTYPE=
+export IP=
+export BOOT=
+export BOOTIF=
+export UBIMTD=
+export break=
+export init=/sbin/init
+export quiet=n
+export readonly=y
+export rootmnt=/root
+export debug=
+export panic=
+export blacklist=
+export resume=
+export resume_offset=
+export recovery=
+
+for i in /sys/devices/system/cpu/cpu*/online
+do
+ case $i in
+ '/sys/devices/system/cpu/cpu0/online')
+ ;;
+ '/sys/devices/system/cpu/cpu*/online')
+ ;;
+ *)
+ echo 1 > $i
+ ;;
+ esac
+done
+
+while :
+do
+ sleep 10
+done
diff --git a/rcutorture/doc/rcu-test-image.txt b/rcutorture/doc/rcu-test-image.txt
new file mode 100644
index 0000000..66efb59
--- /dev/null
+++ b/rcutorture/doc/rcu-test-image.txt
@@ -0,0 +1,42 @@
+This document describes one way to created the rcu-test-image file
+that contains the filesystem used by the guest-OS kernel. There are
+probably much better ways of doing this, and this filesystem could no
+doubt be smaller. It is probably also possible to simply download
+an appropriate image from any number of places.
+
+That said, here are the commands:
+
+------------------------------------------------------------------------
+dd if=/dev/zero of=rcu-test-image bs=400M count=1
+mkfs.ext3 ./rcu-test-image
+sudo mount -o loop ./rcu-test-image /mnt
+
+# Replace "precise" below with your favorite Ubuntu release.
+# Empirical evidence says this image will work for 64-bit, but...
+# Note that debootstrap does take a few minutes to run. Or longer.
+sudo debootstrap --verbose --arch i386 precise /mnt http://archive.ubuntu.com/ubuntu
+cat << '___EOF___' | sudo dd of=/mnt/etc/fstab
+# UNCONFIGURED FSTAB FOR BASE SYSTEM
+#
+/dev/vda / ext3 defaults 1 1
+dev /dev tmpfs rw 0 0
+tmpfs /dev/shm tmpfs defaults 0 0
+devpts /dev/pts devpts gid=5,mode=620 0 0
+sysfs /sys sysfs defaults 0 0
+proc /proc proc defaults 0 0
+___EOF___
+sudo umount /mnt
+------------------------------------------------------------------------
+
+
+References:
+
+ http://sripathikodi.blogspot.com/2010/02/creating-kvm-bootable-fedora-system.html
+ https://help.ubuntu.com/community/KVM/CreateGuests
+ https://help.ubuntu.com/community/JeOSVMBuilder
+ http://wiki.libvirt.org/page/UbuntuKVMWalkthrough
+ http://www.moe.co.uk/2011/01/07/pci_add_option_rom-failed-to-find-romfile-pxe-rtl8139-bin/ -- "apt-get install kvm-pxe"
+ http://www.landley.net/writing/rootfs-howto.html
+ http://en.wikipedia.org/wiki/Initrd
+ http://en.wikipedia.org/wiki/Cpio
+ http://wiki.libvirt.org/page/UbuntuKVMWalkthrough
diff --git a/sysctl/Makefile b/sysctl/Makefile
new file mode 100644
index 0000000..0a92ada
--- /dev/null
+++ b/sysctl/Makefile
@@ -0,0 +1,19 @@
+# Makefile for sysctl selftests.
+# Expects kernel.sysctl_writes_strict=1.
+
+# No binaries, but make sure arg-less "make" doesn't trigger "run_tests".
+all:
+
+# Allow specific tests to be selected.
+test_num:
+ @/bin/sh ./run_numerictests
+
+test_string:
+ @/bin/sh ./run_stringtests
+
+run_tests: all test_num test_string
+
+# Nothing to clean up.
+clean:
+
+.PHONY: all run_tests clean test_num test_string
diff --git a/sysctl/common_tests b/sysctl/common_tests
new file mode 100644
index 0000000..17d534b
--- /dev/null
+++ b/sysctl/common_tests
@@ -0,0 +1,109 @@
+#!/bin/sh
+
+TEST_FILE=$(mktemp)
+
+echo "== Testing sysctl behavior against ${TARGET} =="
+
+set_orig()
+{
+ echo "${ORIG}" > "${TARGET}"
+}
+
+set_test()
+{
+ echo "${TEST_STR}" > "${TARGET}"
+}
+
+verify()
+{
+ local seen
+ seen=$(cat "$1")
+ if [ "${seen}" != "${TEST_STR}" ]; then
+ return 1
+ fi
+ return 0
+}
+
+trap 'set_orig; rm -f "${TEST_FILE}"' EXIT
+
+rc=0
+
+echo -n "Writing test file ... "
+echo "${TEST_STR}" > "${TEST_FILE}"
+if ! verify "${TEST_FILE}"; then
+ echo "FAIL" >&2
+ exit 1
+else
+ echo "ok"
+fi
+
+echo -n "Checking sysctl is not set to test value ... "
+if verify "${TARGET}"; then
+ echo "FAIL" >&2
+ exit 1
+else
+ echo "ok"
+fi
+
+echo -n "Writing sysctl from shell ... "
+set_test
+if ! verify "${TARGET}"; then
+ echo "FAIL" >&2
+ exit 1
+else
+ echo "ok"
+fi
+
+echo -n "Resetting sysctl to original value ... "
+set_orig
+if verify "${TARGET}"; then
+ echo "FAIL" >&2
+ exit 1
+else
+ echo "ok"
+fi
+
+# Now that we've validated the sanity of "set_test" and "set_orig",
+# we can use those functions to set starting states before running
+# specific behavioral tests.
+
+echo -n "Writing entire sysctl in single write ... "
+set_orig
+dd if="${TEST_FILE}" of="${TARGET}" bs=4096 2>/dev/null
+if ! verify "${TARGET}"; then
+ echo "FAIL" >&2
+ rc=1
+else
+ echo "ok"
+fi
+
+echo -n "Writing middle of sysctl after synchronized seek ... "
+set_test
+dd if="${TEST_FILE}" of="${TARGET}" bs=1 seek=1 skip=1 2>/dev/null
+if ! verify "${TARGET}"; then
+ echo "FAIL" >&2
+ rc=1
+else
+ echo "ok"
+fi
+
+echo -n "Writing beyond end of sysctl ... "
+set_orig
+dd if="${TEST_FILE}" of="${TARGET}" bs=20 seek=2 2>/dev/null
+if verify "${TARGET}"; then
+ echo "FAIL" >&2
+ rc=1
+else
+ echo "ok"
+fi
+
+echo -n "Writing sysctl with multiple long writes ... "
+set_orig
+(perl -e 'print "A" x 50;'; echo "${TEST_STR}") | \
+ dd of="${TARGET}" bs=50 2>/dev/null
+if verify "${TARGET}"; then
+ echo "FAIL" >&2
+ rc=1
+else
+ echo "ok"
+fi
diff --git a/sysctl/run_numerictests b/sysctl/run_numerictests
new file mode 100644
index 0000000..8510f93
--- /dev/null
+++ b/sysctl/run_numerictests
@@ -0,0 +1,10 @@
+#!/bin/sh
+
+SYSCTL="/proc/sys"
+TARGET="${SYSCTL}/vm/swappiness"
+ORIG=$(cat "${TARGET}")
+TEST_STR=$(( $ORIG + 1 ))
+
+. ./common_tests
+
+exit $rc
diff --git a/sysctl/run_stringtests b/sysctl/run_stringtests
new file mode 100644
index 0000000..90a9293
--- /dev/null
+++ b/sysctl/run_stringtests
@@ -0,0 +1,77 @@
+#!/bin/sh
+
+SYSCTL="/proc/sys"
+TARGET="${SYSCTL}/kernel/domainname"
+ORIG=$(cat "${TARGET}")
+TEST_STR="Testing sysctl"
+
+. ./common_tests
+
+# Only string sysctls support seeking/appending.
+MAXLEN=65
+
+echo -n "Writing entire sysctl in short writes ... "
+set_orig
+dd if="${TEST_FILE}" of="${TARGET}" bs=1 2>/dev/null
+if ! verify "${TARGET}"; then
+ echo "FAIL" >&2
+ rc=1
+else
+ echo "ok"
+fi
+
+echo -n "Writing middle of sysctl after unsynchronized seek ... "
+set_test
+dd if="${TEST_FILE}" of="${TARGET}" bs=1 seek=1 2>/dev/null
+if verify "${TARGET}"; then
+ echo "FAIL" >&2
+ rc=1
+else
+ echo "ok"
+fi
+
+echo -n "Checking sysctl maxlen is at least $MAXLEN ... "
+set_orig
+perl -e 'print "A" x ('"${MAXLEN}"'-2), "B";' | \
+ dd of="${TARGET}" bs="${MAXLEN}" 2>/dev/null
+if ! grep -q B "${TARGET}"; then
+ echo "FAIL" >&2
+ rc=1
+else
+ echo "ok"
+fi
+
+echo -n "Checking sysctl keeps original string on overflow append ... "
+set_orig
+perl -e 'print "A" x ('"${MAXLEN}"'-1), "B";' | \
+ dd of="${TARGET}" bs=$(( MAXLEN - 1 )) 2>/dev/null
+if grep -q B "${TARGET}"; then
+ echo "FAIL" >&2
+ rc=1
+else
+ echo "ok"
+fi
+
+echo -n "Checking sysctl stays NULL terminated on write ... "
+set_orig
+perl -e 'print "A" x ('"${MAXLEN}"'-1), "B";' | \
+ dd of="${TARGET}" bs="${MAXLEN}" 2>/dev/null
+if grep -q B "${TARGET}"; then
+ echo "FAIL" >&2
+ rc=1
+else
+ echo "ok"
+fi
+
+echo -n "Checking sysctl stays NULL terminated on overwrite ... "
+set_orig
+perl -e 'print "A" x ('"${MAXLEN}"'-1), "BB";' | \
+ dd of="${TARGET}" bs=$(( $MAXLEN + 1 )) 2>/dev/null
+if grep -q B "${TARGET}"; then
+ echo "FAIL" >&2
+ rc=1
+else
+ echo "ok"
+fi
+
+exit $rc
diff --git a/timers/Makefile b/timers/Makefile
new file mode 100644
index 0000000..eb2859f
--- /dev/null
+++ b/timers/Makefile
@@ -0,0 +1,8 @@
+all:
+ gcc posix_timers.c -o posix_timers -lrt
+
+run_tests: all
+ ./posix_timers
+
+clean:
+ rm -f ./posix_timers
diff --git a/timers/posix_timers.c b/timers/posix_timers.c
new file mode 100644
index 0000000..41bd855
--- /dev/null
+++ b/timers/posix_timers.c
@@ -0,0 +1,221 @@
+/*
+ * Copyright (C) 2013 Red Hat, Inc., Frederic Weisbecker <fweisbec@redhat.com>
+ *
+ * Licensed under the terms of the GNU GPL License version 2
+ *
+ * Selftests for a few posix timers interface.
+ *
+ * Kernel loop code stolen from Steven Rostedt <srostedt@redhat.com>
+ */
+
+#include <sys/time.h>
+#include <stdio.h>
+#include <signal.h>
+#include <unistd.h>
+#include <time.h>
+#include <pthread.h>
+
+#define DELAY 2
+#define USECS_PER_SEC 1000000
+
+static volatile int done;
+
+/* Busy loop in userspace to elapse ITIMER_VIRTUAL */
+static void user_loop(void)
+{
+ while (!done);
+}
+
+/*
+ * Try to spend as much time as possible in kernelspace
+ * to elapse ITIMER_PROF.
+ */
+static void kernel_loop(void)
+{
+ void *addr = sbrk(0);
+
+ while (!done) {
+ brk(addr + 4096);
+ brk(addr);
+ }
+}
+
+/*
+ * Sleep until ITIMER_REAL expiration.
+ */
+static void idle_loop(void)
+{
+ pause();
+}
+
+static void sig_handler(int nr)
+{
+ done = 1;
+}
+
+/*
+ * Check the expected timer expiration matches the GTOD elapsed delta since
+ * we armed the timer. Keep a 0.5 sec error margin due to various jitter.
+ */
+static int check_diff(struct timeval start, struct timeval end)
+{
+ long long diff;
+
+ diff = end.tv_usec - start.tv_usec;
+ diff += (end.tv_sec - start.tv_sec) * USECS_PER_SEC;
+
+ if (abs(diff - DELAY * USECS_PER_SEC) > USECS_PER_SEC / 2) {
+ printf("Diff too high: %lld..", diff);
+ return -1;
+ }
+
+ return 0;
+}
+
+static int check_itimer(int which)
+{
+ int err;
+ struct timeval start, end;
+ struct itimerval val = {
+ .it_value.tv_sec = DELAY,
+ };
+
+ printf("Check itimer ");
+
+ if (which == ITIMER_VIRTUAL)
+ printf("virtual... ");
+ else if (which == ITIMER_PROF)
+ printf("prof... ");
+ else if (which == ITIMER_REAL)
+ printf("real... ");
+
+ fflush(stdout);
+
+ done = 0;
+
+ if (which == ITIMER_VIRTUAL)
+ signal(SIGVTALRM, sig_handler);
+ else if (which == ITIMER_PROF)
+ signal(SIGPROF, sig_handler);
+ else if (which == ITIMER_REAL)
+ signal(SIGALRM, sig_handler);
+
+ err = gettimeofday(&start, NULL);
+ if (err < 0) {
+ perror("Can't call gettimeofday()\n");
+ return -1;
+ }
+
+ err = setitimer(which, &val, NULL);
+ if (err < 0) {
+ perror("Can't set timer\n");
+ return -1;
+ }
+
+ if (which == ITIMER_VIRTUAL)
+ user_loop();
+ else if (which == ITIMER_PROF)
+ kernel_loop();
+ else if (which == ITIMER_REAL)
+ idle_loop();
+
+ gettimeofday(&end, NULL);
+ if (err < 0) {
+ perror("Can't call gettimeofday()\n");
+ return -1;
+ }
+
+ if (!check_diff(start, end))
+ printf("[OK]\n");
+ else
+ printf("[FAIL]\n");
+
+ return 0;
+}
+
+static int check_timer_create(int which)
+{
+ int err;
+ timer_t id;
+ struct timeval start, end;
+ struct itimerspec val = {
+ .it_value.tv_sec = DELAY,
+ };
+
+ printf("Check timer_create() ");
+ if (which == CLOCK_THREAD_CPUTIME_ID) {
+ printf("per thread... ");
+ } else if (which == CLOCK_PROCESS_CPUTIME_ID) {
+ printf("per process... ");
+ }
+ fflush(stdout);
+
+ done = 0;
+ err = timer_create(which, NULL, &id);
+ if (err < 0) {
+ perror("Can't create timer\n");
+ return -1;
+ }
+ signal(SIGALRM, sig_handler);
+
+ err = gettimeofday(&start, NULL);
+ if (err < 0) {
+ perror("Can't call gettimeofday()\n");
+ return -1;
+ }
+
+ err = timer_settime(id, 0, &val, NULL);
+ if (err < 0) {
+ perror("Can't set timer\n");
+ return -1;
+ }
+
+ user_loop();
+
+ gettimeofday(&end, NULL);
+ if (err < 0) {
+ perror("Can't call gettimeofday()\n");
+ return -1;
+ }
+
+ if (!check_diff(start, end))
+ printf("[OK]\n");
+ else
+ printf("[FAIL]\n");
+
+ return 0;
+}
+
+int main(int argc, char **argv)
+{
+ int err;
+
+ printf("Testing posix timers. False negative may happen on CPU execution \n");
+ printf("based timers if other threads run on the CPU...\n");
+
+ if (check_itimer(ITIMER_VIRTUAL) < 0)
+ return -1;
+
+ if (check_itimer(ITIMER_PROF) < 0)
+ return -1;
+
+ if (check_itimer(ITIMER_REAL) < 0)
+ return -1;
+
+ if (check_timer_create(CLOCK_THREAD_CPUTIME_ID) < 0)
+ return -1;
+
+ /*
+ * It's unfortunately hard to reliably test a timer expiration
+ * on parallel multithread cputime. We could arm it to expire
+ * on DELAY * nr_threads, with nr_threads busy looping, then wait
+ * the normal DELAY since the time is elapsing nr_threads faster.
+ * But for that we need to ensure we have real physical free CPUs
+ * to ensure true parallelism. So test only one thread until we
+ * find a better solution.
+ */
+ if (check_timer_create(CLOCK_PROCESS_CPUTIME_ID) < 0)
+ return -1;
+
+ return 0;
+}
diff --git a/user/Makefile b/user/Makefile
new file mode 100644
index 0000000..396255b
--- /dev/null
+++ b/user/Makefile
@@ -0,0 +1,13 @@
+# Makefile for user memory selftests
+
+# No binaries, but make sure arg-less "make" doesn't trigger "run_tests"
+all:
+
+run_tests: all
+ @if /sbin/modprobe test_user_copy ; then \
+ rmmod test_user_copy; \
+ echo "user_copy: ok"; \
+ else \
+ echo "user_copy: [FAIL]"; \
+ exit 1; \
+ fi
diff --git a/vm/.gitignore b/vm/.gitignore
new file mode 100644
index 0000000..ff1bb16
--- /dev/null
+++ b/vm/.gitignore
@@ -0,0 +1,4 @@
+hugepage-mmap
+hugepage-shm
+map_hugetlb
+thuge-gen
diff --git a/vm/Makefile b/vm/Makefile
new file mode 100644
index 0000000..3f94e1a
--- /dev/null
+++ b/vm/Makefile
@@ -0,0 +1,15 @@
+# Makefile for vm selftests
+
+CC = $(CROSS_COMPILE)gcc
+CFLAGS = -Wall
+BINARIES = hugepage-mmap hugepage-shm map_hugetlb thuge-gen hugetlbfstest
+
+all: $(BINARIES)
+%: %.c
+ $(CC) $(CFLAGS) -o $@ $^
+
+run_tests: all
+ @/bin/sh ./run_vmtests || (echo "vmtests: [FAIL]"; exit 1)
+
+clean:
+ $(RM) $(BINARIES)
diff --git a/vm/hugepage-mmap.c b/vm/hugepage-mmap.c
new file mode 100644
index 0000000..a10f310
--- /dev/null
+++ b/vm/hugepage-mmap.c
@@ -0,0 +1,92 @@
+/*
+ * hugepage-mmap:
+ *
+ * Example of using huge page memory in a user application using the mmap
+ * system call. Before running this application, make sure that the
+ * administrator has mounted the hugetlbfs filesystem (on some directory
+ * like /mnt) using the command mount -t hugetlbfs nodev /mnt. In this
+ * example, the app is requesting memory of size 256MB that is backed by
+ * huge pages.
+ *
+ * For the ia64 architecture, the Linux kernel reserves Region number 4 for
+ * huge pages. That means that if one requires a fixed address, a huge page
+ * aligned address starting with 0x800000... will be required. If a fixed
+ * address is not required, the kernel will select an address in the proper
+ * range.
+ * Other architectures, such as ppc64, i386 or x86_64 are not so constrained.
+ */
+
+#include <stdlib.h>
+#include <stdio.h>
+#include <unistd.h>
+#include <sys/mman.h>
+#include <fcntl.h>
+
+#define FILE_NAME "huge/hugepagefile"
+#define LENGTH (256UL*1024*1024)
+#define PROTECTION (PROT_READ | PROT_WRITE)
+
+/* Only ia64 requires this */
+#ifdef __ia64__
+#define ADDR (void *)(0x8000000000000000UL)
+#define FLAGS (MAP_SHARED | MAP_FIXED)
+#else
+#define ADDR (void *)(0x0UL)
+#define FLAGS (MAP_SHARED)
+#endif
+
+static void check_bytes(char *addr)
+{
+ printf("First hex is %x\n", *((unsigned int *)addr));
+}
+
+static void write_bytes(char *addr)
+{
+ unsigned long i;
+
+ for (i = 0; i < LENGTH; i++)
+ *(addr + i) = (char)i;
+}
+
+static int read_bytes(char *addr)
+{
+ unsigned long i;
+
+ check_bytes(addr);
+ for (i = 0; i < LENGTH; i++)
+ if (*(addr + i) != (char)i) {
+ printf("Mismatch at %lu\n", i);
+ return 1;
+ }
+ return 0;
+}
+
+int main(void)
+{
+ void *addr;
+ int fd, ret;
+
+ fd = open(FILE_NAME, O_CREAT | O_RDWR, 0755);
+ if (fd < 0) {
+ perror("Open failed");
+ exit(1);
+ }
+
+ addr = mmap(ADDR, LENGTH, PROTECTION, FLAGS, fd, 0);
+ if (addr == MAP_FAILED) {
+ perror("mmap");
+ unlink(FILE_NAME);
+ exit(1);
+ }
+
+ printf("Returned address is %p\n", addr);
+ check_bytes(addr);
+ write_bytes(addr);
+ ret = read_bytes(addr);
+
+ munmap(addr, LENGTH);
+ close(fd);
+ unlink(FILE_NAME);
+
+ return ret;
+}
diff --git a/vm/hugepage-shm.c b/vm/hugepage-shm.c
new file mode 100644
index 0000000..0d0ef4f
--- /dev/null
+++ b/vm/hugepage-shm.c
@@ -0,0 +1,100 @@
+/*
+ * hugepage-shm:
+ *
+ * Example of using huge page memory in a user application using Sys V shared
+ * memory system calls. In this example the app is requesting 256MB of
+ * memory that is backed by huge pages. The application uses the flag
+ * SHM_HUGETLB in the shmget system call to inform the kernel that it is
+ * requesting huge pages.
+ *
+ * For the ia64 architecture, the Linux kernel reserves Region number 4 for
+ * huge pages. That means that if one requires a fixed address, a huge page
+ * aligned address starting with 0x800000... will be required. If a fixed
+ * address is not required, the kernel will select an address in the proper
+ * range.
+ * Other architectures, such as ppc64, i386 or x86_64 are not so constrained.
+ *
+ * Note: The default shared memory limit is quite low on many kernels,
+ * you may need to increase it via:
+ *
+ * echo 268435456 > /proc/sys/kernel/shmmax
+ *
+ * This will increase the maximum size per shared memory segment to 256MB.
+ * The other limit that you will hit eventually is shmall which is the
+ * total amount of shared memory in pages. To set it to 16GB on a system
+ * with a 4kB pagesize do:
+ *
+ * echo 4194304 > /proc/sys/kernel/shmall
+ */
+
+#include <stdlib.h>
+#include <stdio.h>
+#include <sys/types.h>
+#include <sys/ipc.h>
+#include <sys/shm.h>
+#include <sys/mman.h>
+
+#ifndef SHM_HUGETLB
+#define SHM_HUGETLB 04000
+#endif
+
+#define LENGTH (256UL*1024*1024)
+
+#define dprintf(x) printf(x)
+
+/* Only ia64 requires this */
+#ifdef __ia64__
+#define ADDR (void *)(0x8000000000000000UL)
+#define SHMAT_FLAGS (SHM_RND)
+#else
+#define ADDR (void *)(0x0UL)
+#define SHMAT_FLAGS (0)
+#endif
+
+int main(void)
+{
+ int shmid;
+ unsigned long i;
+ char *shmaddr;
+
+ shmid = shmget(2, LENGTH, SHM_HUGETLB | IPC_CREAT | SHM_R | SHM_W);
+ if (shmid < 0) {
+ perror("shmget");
+ exit(1);
+ }
+ printf("shmid: 0x%x\n", shmid);
+
+ shmaddr = shmat(shmid, ADDR, SHMAT_FLAGS);
+ if (shmaddr == (char *)-1) {
+ perror("Shared memory attach failure");
+ shmctl(shmid, IPC_RMID, NULL);
+ exit(2);
+ }
+ printf("shmaddr: %p\n", shmaddr);
+
+ dprintf("Starting the writes:\n");
+ for (i = 0; i < LENGTH; i++) {
+ shmaddr[i] = (char)(i);
+ if (!(i % (1024 * 1024)))
+ dprintf(".");
+ }
+ dprintf("\n");
+
+ dprintf("Starting the Check...");
+ for (i = 0; i < LENGTH; i++)
+ if (shmaddr[i] != (char)i) {
+ printf("\nIndex %lu mismatched\n", i);
+ exit(3);
+ }
+ dprintf("Done.\n");
+
+ if (shmdt((const void *)shmaddr) != 0) {
+ perror("Detach failure");
+ shmctl(shmid, IPC_RMID, NULL);
+ exit(4);
+ }
+
+ shmctl(shmid, IPC_RMID, NULL);
+
+ return 0;
+}
diff --git a/vm/hugetlbfstest.c b/vm/hugetlbfstest.c
new file mode 100644
index 0000000..ea40ff8
--- /dev/null
+++ b/vm/hugetlbfstest.c
@@ -0,0 +1,84 @@
+#define _GNU_SOURCE
+#include <assert.h>
+#include <fcntl.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/mman.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <unistd.h>
+
+typedef unsigned long long u64;
+
+static size_t length = 1 << 24;
+
+static u64 read_rss(void)
+{
+ char buf[4096], *s = buf;
+ int i, fd;
+ u64 rss;
+
+ fd = open("/proc/self/statm", O_RDONLY);
+ assert(fd > 2);
+ memset(buf, 0, sizeof(buf));
+ read(fd, buf, sizeof(buf) - 1);
+ for (i = 0; i < 1; i++)
+ s = strchr(s, ' ') + 1;
+ rss = strtoull(s, NULL, 10);
+ return rss << 12; /* assumes 4k pagesize */
+}
+
+static void do_mmap(int fd, int extra_flags, int unmap)
+{
+ int *p;
+ int flags = MAP_PRIVATE | MAP_POPULATE | extra_flags;
+ u64 before, after;
+
+ before = read_rss();
+ p = mmap(NULL, length, PROT_READ | PROT_WRITE, flags, fd, 0);
+ assert(p != MAP_FAILED ||
+ !"mmap returned an unexpected error");
+ after = read_rss();
+ assert(llabs(after - before - length) < 0x40000 ||
+ !"rss didn't grow as expected");
+ if (!unmap)
+ return;
+ munmap(p, length);
+ after = read_rss();
+ assert(llabs(after - before) < 0x40000 ||
+ !"rss didn't shrink as expected");
+}
+
+static int open_file(const char *path)
+{
+ int fd, err;
+
+ unlink(path);
+ fd = open(path, O_CREAT | O_RDWR | O_TRUNC | O_EXCL
+ | O_LARGEFILE | O_CLOEXEC, 0600);
+ assert(fd > 2);
+ unlink(path);
+ err = ftruncate(fd, length);
+ assert(!err);
+ return fd;
+}
+
+int main(void)
+{
+ int hugefd, fd;
+
+ fd = open_file("/dev/shm/hugetlbhog");
+ hugefd = open_file("/hugepages/hugetlbhog");
+
+ system("echo 100 > /proc/sys/vm/nr_hugepages");
+ do_mmap(-1, MAP_ANONYMOUS, 1);
+ do_mmap(fd, 0, 1);
+ do_mmap(-1, MAP_ANONYMOUS | MAP_HUGETLB, 1);
+ do_mmap(hugefd, 0, 1);
+ do_mmap(hugefd, MAP_HUGETLB, 1);
+ /* Leak the last one to test do_exit() */
+ do_mmap(-1, MAP_ANONYMOUS | MAP_HUGETLB, 0);
+ printf("oll korrekt.\n");
+ return 0;
+}
diff --git a/vm/map_hugetlb.c b/vm/map_hugetlb.c
new file mode 100644
index 0000000..ac56639
--- /dev/null
+++ b/vm/map_hugetlb.c
@@ -0,0 +1,79 @@
+/*
+ * Example of using hugepage memory in a user application using the mmap
+ * system call with MAP_HUGETLB flag. Before running this program make
+ * sure the administrator has allocated enough default sized huge pages
+ * to cover the 256 MB allocation.
+ *
+ * For ia64 architecture, Linux kernel reserves Region number 4 for hugepages.
+ * That means the addresses starting with 0x800000... will need to be
+ * specified. Specifying a fixed address is not required on ppc64, i386
+ * or x86_64.
+ */
+#include <stdlib.h>
+#include <stdio.h>
+#include <unistd.h>
+#include <sys/mman.h>
+#include <fcntl.h>
+
+#define LENGTH (256UL*1024*1024)
+#define PROTECTION (PROT_READ | PROT_WRITE)
+
+#ifndef MAP_HUGETLB
+#define MAP_HUGETLB 0x40000 /* arch specific */
+#endif
+
+/* Only ia64 requires this */
+#ifdef __ia64__
+#define ADDR (void *)(0x8000000000000000UL)
+#define FLAGS (MAP_PRIVATE | MAP_ANONYMOUS | MAP_HUGETLB | MAP_FIXED)
+#else
+#define ADDR (void *)(0x0UL)
+#define FLAGS (MAP_PRIVATE | MAP_ANONYMOUS | MAP_HUGETLB)
+#endif
+
+static void check_bytes(char *addr)
+{
+ printf("First hex is %x\n", *((unsigned int *)addr));
+}
+
+static void write_bytes(char *addr)
+{
+ unsigned long i;
+
+ for (i = 0; i < LENGTH; i++)
+ *(addr + i) = (char)i;
+}
+
+static int read_bytes(char *addr)
+{
+ unsigned long i;
+
+ check_bytes(addr);
+ for (i = 0; i < LENGTH; i++)
+ if (*(addr + i) != (char)i) {
+ printf("Mismatch at %lu\n", i);
+ return 1;
+ }
+ return 0;
+}
+
+int main(void)
+{
+ void *addr;
+ int ret;
+
+ addr = mmap(ADDR, LENGTH, PROTECTION, FLAGS, 0, 0);
+ if (addr == MAP_FAILED) {
+ perror("mmap");
+ exit(1);
+ }
+
+ printf("Returned address is %p\n", addr);
+ check_bytes(addr);
+ write_bytes(addr);
+ ret = read_bytes(addr);
+
+ munmap(addr, LENGTH);
+
+ return ret;
+}
diff --git a/vm/run_vmtests b/vm/run_vmtests
new file mode 100644
index 0000000..c87b681
--- /dev/null
+++ b/vm/run_vmtests
@@ -0,0 +1,93 @@
+#!/bin/bash
+#please run as root
+
+#we need 256M, below is the size in kB
+needmem=262144
+mnt=./huge
+exitcode=0
+
+#get pagesize and freepages from /proc/meminfo
+while read name size unit; do
+ if [ "$name" = "HugePages_Free:" ]; then
+ freepgs=$size
+ fi
+ if [ "$name" = "Hugepagesize:" ]; then
+ pgsize=$size
+ fi
+done < /proc/meminfo
+
+#set proper nr_hugepages
+if [ -n "$freepgs" ] && [ -n "$pgsize" ]; then
+ nr_hugepgs=`cat /proc/sys/vm/nr_hugepages`
+ needpgs=`expr $needmem / $pgsize`
+ if [ $freepgs -lt $needpgs ]; then
+ lackpgs=$(( $needpgs - $freepgs ))
+ echo $(( $lackpgs + $nr_hugepgs )) > /proc/sys/vm/nr_hugepages
+ if [ $? -ne 0 ]; then
+ echo "Please run this test as root"
+ exit 1
+ fi
+ fi
+else
+ echo "no hugetlbfs support in kernel?"
+ exit 1
+fi
+
+mkdir $mnt
+mount -t hugetlbfs none $mnt
+
+echo "--------------------"
+echo "running hugepage-mmap"
+echo "--------------------"
+./hugepage-mmap
+if [ $? -ne 0 ]; then
+ echo "[FAIL]"
+ exitcode=1
+else
+ echo "[PASS]"
+fi
+
+shmmax=`cat /proc/sys/kernel/shmmax`
+shmall=`cat /proc/sys/kernel/shmall`
+echo 268435456 > /proc/sys/kernel/shmmax
+echo 4194304 > /proc/sys/kernel/shmall
+echo "--------------------"
+echo "running hugepage-shm"
+echo "--------------------"
+./hugepage-shm
+if [ $? -ne 0 ]; then
+ echo "[FAIL]"
+ exitcode=1
+else
+ echo "[PASS]"
+fi
+echo $shmmax > /proc/sys/kernel/shmmax
+echo $shmall > /proc/sys/kernel/shmall
+
+echo "--------------------"
+echo "running map_hugetlb"
+echo "--------------------"
+./map_hugetlb
+if [ $? -ne 0 ]; then
+ echo "[FAIL]"
+ exitcode=1
+else
+ echo "[PASS]"
+fi
+
+echo "--------------------"
+echo "running hugetlbfstest"
+echo "--------------------"
+./hugetlbfstest
+if [ $? -ne 0 ]; then
+ echo "[FAIL]"
+ exitcode=1
+else
+ echo "[PASS]"
+fi
+
+#cleanup
+umount $mnt
+rm -rf $mnt
+echo $nr_hugepgs > /proc/sys/vm/nr_hugepages
+exit $exitcode
diff --git a/vm/thuge-gen.c b/vm/thuge-gen.c
new file mode 100644
index 0000000..c879572
--- /dev/null
+++ b/vm/thuge-gen.c
@@ -0,0 +1,254 @@
+/* Test selecting other page sizes for mmap/shmget.
+
+ Before running this huge pages for each huge page size must have been
+ reserved.
+ For large pages beyond MAX_ORDER (like 1GB on x86) boot options must be used.
+ Also shmmax must be increased.
+ And you need to run as root to work around some weird permissions in shm.
+ And nothing using huge pages should run in parallel.
+ When the program aborts you may need to clean up the shm segments with
+ ipcrm -m by hand, like this
+ sudo ipcs | awk '$1 == "0x00000000" {print $2}' | xargs -n1 sudo ipcrm -m
+ (warning this will remove all if someone else uses them) */
+
+#define _GNU_SOURCE 1
+#include <sys/mman.h>
+#include <stdlib.h>
+#include <stdio.h>
+#include <sys/ipc.h>
+#include <sys/shm.h>
+#include <sys/stat.h>
+#include <glob.h>
+#include <assert.h>
+#include <unistd.h>
+#include <stdarg.h>
+#include <string.h>
+
+#define err(x) perror(x), exit(1)
+
+#define MAP_HUGE_2MB (21 << MAP_HUGE_SHIFT)
+#define MAP_HUGE_1GB (30 << MAP_HUGE_SHIFT)
+#define MAP_HUGE_SHIFT 26
+#define MAP_HUGE_MASK 0x3f
+#define MAP_HUGETLB 0x40000
+
+#define SHM_HUGETLB 04000 /* segment will use huge TLB pages */
+#define SHM_HUGE_SHIFT 26
+#define SHM_HUGE_MASK 0x3f
+#define SHM_HUGE_2MB (21 << SHM_HUGE_SHIFT)
+#define SHM_HUGE_1GB (30 << SHM_HUGE_SHIFT)
+
+#define NUM_PAGESIZES 5
+
+#define NUM_PAGES 4
+
+#define Dprintf(fmt...) // printf(fmt)
+
+unsigned long page_sizes[NUM_PAGESIZES];
+int num_page_sizes;
+
+int ilog2(unsigned long v)
+{
+ int l = 0;
+ while ((1UL << l) < v)
+ l++;
+ return l;
+}
+
+void find_pagesizes(void)
+{
+ glob_t g;
+ int i;
+ glob("/sys/kernel/mm/hugepages/hugepages-*kB", 0, NULL, &g);
+ assert(g.gl_pathc <= NUM_PAGESIZES);
+ for (i = 0; i < g.gl_pathc; i++) {
+ sscanf(g.gl_pathv[i], "/sys/kernel/mm/hugepages/hugepages-%lukB",
+ &page_sizes[i]);
+ page_sizes[i] <<= 10;
+ printf("Found %luMB\n", page_sizes[i] >> 20);
+ }
+ num_page_sizes = g.gl_pathc;
+ globfree(&g);
+}
+
+unsigned long default_huge_page_size(void)
+{
+ unsigned long hps = 0;
+ char *line = NULL;
+ size_t linelen = 0;
+ FILE *f = fopen("/proc/meminfo", "r");
+ if (!f)
+ return 0;
+ while (getline(&line, &linelen, f) > 0) {
+ if (sscanf(line, "Hugepagesize: %lu kB", &hps) == 1) {
+ hps <<= 10;
+ break;
+ }
+ }
+ free(line);
+ return hps;
+}
+
+void show(unsigned long ps)
+{
+ char buf[100];
+ if (ps == getpagesize())
+ return;
+ printf("%luMB: ", ps >> 20);
+ fflush(stdout);
+ snprintf(buf, sizeof buf,
+ "cat /sys/kernel/mm/hugepages/hugepages-%lukB/free_hugepages",
+ ps >> 10);
+ system(buf);
+}
+
+unsigned long read_sysfs(int warn, char *fmt, ...)
+{
+ char *line = NULL;
+ size_t linelen = 0;
+ char buf[100];
+ FILE *f;
+ va_list ap;
+ unsigned long val = 0;
+
+ va_start(ap, fmt);
+ vsnprintf(buf, sizeof buf, fmt, ap);
+ va_end(ap);
+
+ f = fopen(buf, "r");
+ if (!f) {
+ if (warn)
+ printf("missing %s\n", buf);
+ return 0;
+ }
+ if (getline(&line, &linelen, f) > 0) {
+ sscanf(line, "%lu", &val);
+ }
+ fclose(f);
+ free(line);
+ return val;
+}
+
+unsigned long read_free(unsigned long ps)
+{
+ return read_sysfs(ps != getpagesize(),
+ "/sys/kernel/mm/hugepages/hugepages-%lukB/free_hugepages",
+ ps >> 10);
+}
+
+void test_mmap(unsigned long size, unsigned flags)
+{
+ char *map;
+ unsigned long before, after;
+ int err;
+
+ before = read_free(size);
+ map = mmap(NULL, size*NUM_PAGES, PROT_READ|PROT_WRITE,
+ MAP_PRIVATE|MAP_ANONYMOUS|MAP_HUGETLB|flags, 0, 0);
+
+ if (map == (char *)-1) err("mmap");
+ memset(map, 0xff, size*NUM_PAGES);
+ after = read_free(size);
+ Dprintf("before %lu after %lu diff %ld size %lu\n",
+ before, after, before - after, size);
+ assert(size == getpagesize() || (before - after) == NUM_PAGES);
+ show(size);
+ err = munmap(map, size);
+ assert(!err);
+}
+
+void test_shmget(unsigned long size, unsigned flags)
+{
+ int id;
+ unsigned long before, after;
+ int err;
+
+ before = read_free(size);
+ id = shmget(IPC_PRIVATE, size * NUM_PAGES, IPC_CREAT|0600|flags);
+ if (id < 0) err("shmget");
+
+ struct shm_info i;
+ if (shmctl(id, SHM_INFO, (void *)&i) < 0) err("shmctl");
+ Dprintf("alloc %lu res %lu\n", i.shm_tot, i.shm_rss);
+
+
+ Dprintf("id %d\n", id);
+ char *map = shmat(id, NULL, 0600);
+ if (map == (char*)-1) err("shmat");
+
+ shmctl(id, IPC_RMID, NULL);
+
+ memset(map, 0xff, size*NUM_PAGES);
+ after = read_free(size);
+
+ Dprintf("before %lu after %lu diff %ld size %lu\n",
+ before, after, before - after, size);
+ assert(size == getpagesize() || (before - after) == NUM_PAGES);
+ show(size);
+ err = shmdt(map);
+ assert(!err);
+}
+
+void sanity_checks(void)
+{
+ int i;
+ unsigned long largest = getpagesize();
+
+ for (i = 0; i < num_page_sizes; i++) {
+ if (page_sizes[i] > largest)
+ largest = page_sizes[i];
+
+ if (read_free(page_sizes[i]) < NUM_PAGES) {
+ printf("Not enough huge pages for page size %lu MB, need %u\n",
+ page_sizes[i] >> 20,
+ NUM_PAGES);
+ exit(0);
+ }
+ }
+
+ if (read_sysfs(0, "/proc/sys/kernel/shmmax") < NUM_PAGES * largest) {
+ printf("Please do echo %lu > /proc/sys/kernel/shmmax", largest * NUM_PAGES);
+ exit(0);
+ }
+
+#if defined(__x86_64__)
+ if (largest != 1U<<30) {
+ printf("No GB pages available on x86-64\n"
+ "Please boot with hugepagesz=1G hugepages=%d\n", NUM_PAGES);
+ exit(0);
+ }
+#endif
+}
+
+int main(void)
+{
+ int i;
+ unsigned default_hps = default_huge_page_size();
+
+ find_pagesizes();
+
+ sanity_checks();
+
+ for (i = 0; i < num_page_sizes; i++) {
+ unsigned long ps = page_sizes[i];
+ int arg = ilog2(ps) << MAP_HUGE_SHIFT;
+ printf("Testing %luMB mmap with shift %x\n", ps >> 20, arg);
+ test_mmap(ps, MAP_HUGETLB | arg);
+ }
+ printf("Testing default huge mmap\n");
+ test_mmap(default_hps, SHM_HUGETLB);
+
+ puts("Testing non-huge shmget");
+ test_shmget(getpagesize(), 0);
+
+ for (i = 0; i < num_page_sizes; i++) {
+ unsigned long ps = page_sizes[i];
+ int arg = ilog2(ps) << SHM_HUGE_SHIFT;
+ printf("Testing %luMB shmget with shift %x\n", ps >> 20, arg);
+ test_shmget(ps, SHM_HUGETLB | arg);
+ }
+ puts("default huge shmget");
+ test_shmget(default_hps, SHM_HUGETLB);
+
+ return 0;
+}