From 1d48fdd9d84aab1bd32c1f70947932f5d90f92aa Mon Sep 17 00:00:00 2001 From: Peter Maydell Date: Tue, 14 Jun 2016 12:49:18 +0100 Subject: linux-user: Don't use sigfillset() on uc->uc_sigmask The kernel and libc have different ideas about what a sigset_t is -- for the kernel it is only _NSIG / 8 bytes in size (usually 8 bytes), but for libc it is much larger, 128 bytes. In most situations the difference doesn't matter, because if you pass a pointer to a libc sigset_t to the kernel it just acts on the first 8 bytes of it, but for the ucontext_t* argument to a signal handler it trips us up. The kernel allocates this ucontext_t on the stack according to its idea of the sigset_t type, but the type of the ucontext_t defined by the libc headers uses the libc type, and so do the manipulator functions like sigfillset(). This means that (1) sizeof(uc->uc_sigmask) is much larger than the actual space used on the stack (2) sigfillset(&uc->uc_sigmask) will write garbage 0xff bytes off the end of the structure, which can trash data that was on the stack before the signal handler was invoked, and may result in a crash after the handler returns To avoid this, we use a memset() of the correct size to fill the signal mask rather than using the libc function. This fixes a problem where we would crash at least some of the time on an i386 host when a signal was taken. Signed-off-by: Peter Maydell Reviewed-by: Laurent Vivier Signed-off-by: Riku Voipio --- linux-user/signal.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) (limited to 'linux-user/signal.c') diff --git a/linux-user/signal.c b/linux-user/signal.c index e2d55fff97..9d980456ec 100644 --- a/linux-user/signal.c +++ b/linux-user/signal.c @@ -636,8 +636,16 @@ static void host_signal_handler(int host_signum, siginfo_t *info, * code in case the guest code provokes one in the window between * now and it getting out to the main loop. Signals will be * unblocked again in process_pending_signals(). + * + * WARNING: we cannot use sigfillset() here because the uc_sigmask + * field is a kernel sigset_t, which is much smaller than the + * libc sigset_t which sigfillset() operates on. Using sigfillset() + * would write 0xff bytes off the end of the structure and trash + * data on the struct. + * We can't use sizeof(uc->uc_sigmask) either, because the libc + * headers define the struct field with the wrong (too large) type. */ - sigfillset(&uc->uc_sigmask); + memset(&uc->uc_sigmask, 0xff, SIGSET_T_SIZE); sigdelset(&uc->uc_sigmask, SIGSEGV); sigdelset(&uc->uc_sigmask, SIGBUS); -- cgit v1.2.3-55-g7522