Replace all setjmp()/longjmp() with sigsetjmp()/siglongjmp()
The setjmp() function doesn't specify whether signal masks are saved and
restored; on Linux they are not, but on BSD (including MacOSX) they are.
We want to have consistent behaviour across platforms, so we should
always use "don't save/restore signal mask" (this is also generally
going to be faster). This also works around a bug in MacOSX where the
signal-restoration on longjmp() affects the signal mask for a completely
different thread, not just the mask for the thread which did the longjmp.
The most visible effect of this was that ctrl-C was ignored on MacOSX
because the CPU thread did a longjmp which resulted in its signal mask
being applied to every thread, so that all threads had SIGINT and SIGTERM
blocked.
The POSIX-sanctioned portable way to do a jump without affecting signal
masks is to siglongjmp() to a sigjmp_buf which was created by calling
sigsetjmp() with a zero savemask parameter, so change all uses of
setjmp()/longjmp() accordingly. [Technically POSIX allows sigsetjmp(buf, 0)
to save the signal mask; however the following siglongjmp() must not
restore the signal mask, so the pair can be effectively considered as
"sigjmp/longjmp which don't touch the mask".]
For Windows we provide a trivial sigsetjmp/siglongjmp in terms of
setjmp/longjmp -- this is OK because no user will ever pass a non-zero
savemask.
The setjmp() uses in tests/tcg/test-i386.c and tests/tcg/linux-test.c
are left untouched because these are self-contained singlethreaded
test programs intended to be run under QEMU's Linux emulation, so they
have neither the portability nor the multithreading issues to deal with.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <rth@twiddle.net>
Tested-by: Stefan Weil <sw@weilnetz.de>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
diff --git a/coroutine-sigaltstack.c b/coroutine-sigaltstack.c
index e37ebac..1fb41c9 100644
--- a/coroutine-sigaltstack.c
+++ b/coroutine-sigaltstack.c
@@ -45,7 +45,7 @@
typedef struct {
Coroutine base;
void *stack;
- jmp_buf env;
+ sigjmp_buf env;
} CoroutineUContext;
/**
@@ -59,7 +59,7 @@
CoroutineUContext leader;
/** Information for the signal handler (trampoline) */
- jmp_buf tr_reenter;
+ sigjmp_buf tr_reenter;
volatile sig_atomic_t tr_called;
void *tr_handler;
} CoroutineThreadState;
@@ -115,8 +115,8 @@
static void coroutine_bootstrap(CoroutineUContext *self, Coroutine *co)
{
/* Initialize longjmp environment and switch back the caller */
- if (!setjmp(self->env)) {
- longjmp(*(jmp_buf *)co->entry_arg, 1);
+ if (!sigsetjmp(self->env, 0)) {
+ siglongjmp(*(sigjmp_buf *)co->entry_arg, 1);
}
while (true) {
@@ -145,14 +145,14 @@
/*
* Here we have to do a bit of a ping pong between the caller, given that
* this is a signal handler and we have to do a return "soon". Then the
- * caller can reestablish everything and do a longjmp here again.
+ * caller can reestablish everything and do a siglongjmp here again.
*/
- if (!setjmp(coTS->tr_reenter)) {
+ if (!sigsetjmp(coTS->tr_reenter, 0)) {
return;
}
/*
- * Ok, the caller has longjmp'ed back to us, so now prepare
+ * Ok, the caller has siglongjmp'ed back to us, so now prepare
* us for the real machine state switching. We have to jump
* into another function here to get a new stack context for
* the auto variables (which have to be auto-variables
@@ -179,7 +179,7 @@
/* The way to manipulate stack is with the sigaltstack function. We
* prepare a stack, with it delivering a signal to ourselves and then
- * put setjmp/longjmp where needed.
+ * put sigsetjmp/siglongjmp where needed.
* This has been done keeping coroutine-ucontext as a model and with the
* pth ideas (GNU Portable Threads). See coroutine-ucontext for the basics
* of the coroutines and see pth_mctx.c (from the pth project) for the
@@ -220,7 +220,7 @@
/*
* Now transfer control onto the signal stack and set it up.
- * It will return immediately via "return" after the setjmp()
+ * It will return immediately via "return" after the sigsetjmp()
* was performed. Be careful here with race conditions. The
* signal can be delivered the first time sigsuspend() is
* called.
@@ -261,8 +261,8 @@
* type-conversion warnings related to the `volatile' qualifier and
* the fact that `jmp_buf' usually is an array type.
*/
- if (!setjmp(old_env)) {
- longjmp(coTS->tr_reenter, 1);
+ if (!sigsetjmp(old_env, 0)) {
+ siglongjmp(coTS->tr_reenter, 1);
}
/*
@@ -311,9 +311,9 @@
s->current = to_;
- ret = setjmp(from->env);
+ ret = sigsetjmp(from->env, 0);
if (ret == 0) {
- longjmp(to->env, action);
+ siglongjmp(to->env, action);
}
return ret;
}