Skip to contentSkip to conversation

Fix use after stack unwind in fzsync lib

PATCH Superseded Martin Doucha 3 participants 4 messages Mar 26, 2020 – Mar 27, 2020

lore
Share & tools
lore
#0 lore raw

PATCH v3 1/1 Fix use after stack unwind in fzsync lib

tst_fzsync_pair_reset() passes a local variable to thread B which may be already unwinded by the time the thread wrapper function executes. If new variables get allocated and initialized on stack between pthread_create() and thread wrapper execution, thread B will segfault.

---

Changes since v1:
- Use tst_alloc() instead of SAFE_MALLOC() to prevent memory leak even if SAFE_PTHREAD_CREATE() fails
- Revert changes in tst_fzsync_thread_wrapper() as they're no longer needed

Changes since v2:
- Use static local variable instead to prevent memory leaks

 include/tst_fuzzy_sync.h | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
include/tst_fuzzy_sync.h +4 -1
index c1d0b00f9..9ff3a78ae 100644
@@ -297,7 +297,10 @@ static void tst_fzsync_pair_reset(struct tst_fzsync_pair *pair,
 	pair->b_cntr = 0;
 	pair->exit = 0;
 	if (run_b) {
-		struct tst_fzsync_run_thread wrap_run_b = {.func = run_b, .arg = NULL};
+		static struct tst_fzsync_run_thread wrap_run_b;
+
+		wrap_run_b.func = run_b;
+		wrap_run_b.arg = NULL;
 		SAFE_PTHREAD_CREATE(&pair->thread_b, 0, tst_fzsync_thread_wrapper, &wrap_run_b);
 	}
 
Signature · 1 line
2.25.1
#2 lore raw

PATCH v3 1/1 Fix use after stack unwind in fzsync lib

In reply to Martin Doucha

Hi Martin,

Quoted 4 lines

tst_fzsync_pair_reset() passes a local variable to thread B which may be already unwinded by the time the thread wrapper function executes. If new variables get allocated and initialized on stack between pthread_create() and thread wrapper execution, thread B will segfault.

I naively assumed this would fix SIGKILL signal for CVE 2018-1000199 test, but it didn't.

Kind regards,
Petr

#3 lore raw

PATCH v3 1/1 Fix use after stack unwind in fzsync lib

Quoted 9 lines · Petr Vorel

Hi Martin,

Quoted 4 lines

tst_fzsync_pair_reset() passes a local variable to thread B which may be already unwinded by the time the thread wrapper function executes. If new variables get allocated and initialized on stack between pthread_create() and thread wrapper execution, thread B will segfault.

I naively assumed this would fix SIGKILL signal for CVE 2018-1000199 test, but it didn't.

No, the fuzzysync fix is unrelated to the CVE-2018-1000199 test. I ran into mysterious segfaults while writing a test for CVE-2018-18559 which I haven't finished yet.

Signature · 7 lines
Martin Doucha   [email protected]
QA Engineer for Software Maintenance
SUSE LINUX, s.r.o.
CORSO IIa
Krizikova 148/34
186 00 Prague 8
Czech Republic