From d4f2ec208cb1be22deb096a97515ff3c9e153df3 Mon Sep 17 00:00:00 2001 From: Alexander Graf Date: Fri, 13 Jan 2012 17:05:41 +0100 Subject: [PATCH] linux-user: fix segfault deadlock MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When entering the guest we take a lock to ensure that nobody else messes with our TB chaining while we're doing it. If we get a segfault inside that code, we manage to work on, but will not unlock the lock. This patch forces unlocking of that lock in the segv handler. I'm not sure this is the right approach though. Maybe we should rather make sure we don't segfault in the code? I would greatly appreciate someone more intelligible than me to look at this :). Example code to trigger this is at: http://csgraf.de/tmp/conftest.c Reported-by: Fabio Erculiani Signed-off-by: Alexander Graf [AF: Drop spinlock_safe_unlock() and switch to tb_lock_reset() (bonzini)] Signed-off-by: Andreas Färber --- user-exec.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/user-exec.c b/user-exec.c index a8f95fa1e1..c262653604 100644 --- a/user-exec.c +++ b/user-exec.c @@ -78,6 +78,10 @@ static inline int handle_cpu_signal(uintptr_t pc, unsigned long address, printf("qemu: SIGSEGV pc=0x%08lx address=%08lx w=%d oldset=0x%08lx\n", pc, address, is_write, *(unsigned long *)old_set); #endif + + /* Maybe we're still holding the TB fiddling lock? */ + tb_lock_reset(); + /* XXX: locking issue */ if (is_write && h2g_valid(address)) { switch (page_unprotect(h2g(address), pc)) {