path: root/arch/x86
diff options
authorTony Luck <tony.luck@intel.com>2020-05-20 09:35:46 -0700
committerThomas Gleixner <tglx@linutronix.de>2020-06-11 15:19:17 +0200
commit17fae1294ad9d711b2c3dd0edef479d40c76a5e8 (patch)
tree13d6d176ac7bf835ab28f7b5bfe7e08996c9decb /arch/x86
parentf77d26a9fc525286bcef3d4f98b52e17482cf49c (diff)
x86/{mce,mm}: Unmap the entire page if the whole page is affected and poisoned
An interesting thing happened when a guest Linux instance took a machine check. The VMM unmapped the bad page from guest physical space and passed the machine check to the guest. Linux took all the normal actions to offline the page from the process that was using it. But then guest Linux crashed because it said there was a second machine check inside the kernel with this stack trace: do_memory_failure set_mce_nospec set_memory_uc _set_memory_uc change_page_attr_set_clr cpa_flush clflush_cache_range_opt This was odd, because a CLFLUSH instruction shouldn't raise a machine check (it isn't consuming the data). Further investigation showed that the VMM had passed in another machine check because is appeared that the guest was accessing the bad page. Fix is to check the scope of the poison by checking the MCi_MISC register. If the entire page is affected, then unmap the page. If only part of the page is affected, then mark the page as uncacheable. This assumes that VMMs will do the logical thing and pass in the "whole page scope" via the MCi_MISC register (since they unmapped the entire page). [ bp: Adjust to x86/entry changes. ] Fixes: 284ce4011ba6 ("x86/memory_failure: Introduce {set, clear}_mce_nospec()") Reported-by: Jue Wang <juew@google.com> Signed-off-by: Tony Luck <tony.luck@intel.com> Signed-off-by: Borislav Petkov <bp@suse.de> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Tested-by: Jue Wang <juew@google.com> Cc: <stable@vger.kernel.org> Link: https://lkml.kernel.org/r/20200520163546.GA7977@agluck-desk2.amr.corp.intel.com
Diffstat (limited to 'arch/x86')
2 files changed, 27 insertions, 10 deletions
diff --git a/arch/x86/include/asm/set_memory.h b/arch/x86/include/asm/set_memory.h
index ec2c0a094b5d..5948218f35c5 100644
--- a/arch/x86/include/asm/set_memory.h
+++ b/arch/x86/include/asm/set_memory.h
@@ -86,28 +86,35 @@ int set_direct_map_default_noflush(struct page *page);
extern int kernel_set_to_readonly;
#ifdef CONFIG_X86_64
-static inline int set_mce_nospec(unsigned long pfn)
+ * Prevent speculative access to the page by either unmapping
+ * it (if we do not require access to any part of the page) or
+ * marking it uncacheable (if we want to try to retrieve data
+ * from non-poisoned lines in the page).
+ */
+static inline int set_mce_nospec(unsigned long pfn, bool unmap)
unsigned long decoy_addr;
int rc;
- * Mark the linear address as UC to make sure we don't log more
- * errors because of speculative access to the page.
* We would like to just call:
- * set_memory_uc((unsigned long)pfn_to_kaddr(pfn), 1);
+ * set_memory_XX((unsigned long)pfn_to_kaddr(pfn), 1);
* but doing that would radically increase the odds of a
* speculative access to the poison page because we'd have
* the virtual address of the kernel 1:1 mapping sitting
* around in registers.
* Instead we get tricky. We create a non-canonical address
* that looks just like the one we want, but has bit 63 flipped.
- * This relies on set_memory_uc() properly sanitizing any __pa()
+ * This relies on set_memory_XX() properly sanitizing any __pa()
* results with __PHYSICAL_MASK or PTE_PFN_MASK.
decoy_addr = (pfn << PAGE_SHIFT) + (PAGE_OFFSET ^ BIT(63));
- rc = set_memory_uc(decoy_addr, 1);
+ if (unmap)
+ rc = set_memory_np(decoy_addr, 1);
+ else
+ rc = set_memory_uc(decoy_addr, 1);
if (rc)
pr_warn("Could not invalidate pfn=0x%lx from 1:1 map\n", pfn);
return rc;
diff --git a/arch/x86/kernel/cpu/mce/core.c b/arch/x86/kernel/cpu/mce/core.c
index 30413325de22..ce9120c4f740 100644
--- a/arch/x86/kernel/cpu/mce/core.c
+++ b/arch/x86/kernel/cpu/mce/core.c
@@ -520,6 +520,14 @@ bool mce_is_memory_error(struct mce *m)
+static bool whole_page(struct mce *m)
+ if (!mca_cfg.ser || !(m->status & MCI_STATUS_MISCV))
+ return true;
+ return MCI_MISC_ADDR_LSB(m->misc) >= PAGE_SHIFT;
bool mce_is_correctable(struct mce *m)
if (m->cpuvendor == X86_VENDOR_AMD && m->status & MCI_STATUS_DEFERRED)
@@ -573,7 +581,7 @@ static int uc_decode_notifier(struct notifier_block *nb, unsigned long val,
pfn = mce->addr >> PAGE_SHIFT;
if (!memory_failure(pfn, 0)) {
- set_mce_nospec(pfn);
+ set_mce_nospec(pfn, whole_page(mce));
mce->kflags |= MCE_HANDLED_UC;
@@ -1173,11 +1181,12 @@ static void kill_me_maybe(struct callback_head *cb)
pr_err("Uncorrected hardware memory error in user-access at %llx", p->mce_addr);
- if (!(p->mce_status & MCG_STATUS_RIPV))
+ if (!p->mce_ripv)
flags |= MF_MUST_KILL;
if (!memory_failure(p->mce_addr >> PAGE_SHIFT, flags)) {
- set_mce_nospec(p->mce_addr >> PAGE_SHIFT);
+ set_mce_nospec(p->mce_addr >> PAGE_SHIFT, p->mce_whole_page);
@@ -1331,7 +1340,8 @@ void noinstr do_machine_check(struct pt_regs *regs)
BUG_ON(!on_thread_stack() || !user_mode(regs));
current->mce_addr = m.addr;
- current->mce_status = m.mcgstatus;
+ current->mce_ripv = !!(m.mcgstatus & MCG_STATUS_RIPV);
+ current->mce_whole_page = whole_page(&m);
current->mce_kill_me.func = kill_me_maybe;
if (kill_it)
current->mce_kill_me.func = kill_me_now;