In testing some new code I'm working on that uses shm_mq to do Cool
Stuff (TM), I discovered that my code was mysteriously failing
assertions when I hit ^C in the middle of the test. This led me on a
lengthy, mostly-misguided hunt for the culprit. I eventually
discovered that the problem wasn't in my new code at all, but was
rather an oversight in the shm_mq stuff I previously committed. So, I
intend to commit and back-patch the attached fix shortly.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company
commit 42900c00c831607ec435228e346ccbb8617be521
Author: Robert Haas <rh...@postgresql.org>
Date: Wed Jul 23 16:29:40 2014 -0400
Prevent shm_mq_send from reading uninitialized memory.
shm_mq_send_bytes didn't invariably initialize *bytes_written before
returning, which would cause shm_mq_send to read from uninitialized
memory and add the value it found there to mqh->mqh_partial_bytes.
This could cause the next attempt to send a message via the queue to
fail an assertion (if the queue was detached) or copy data from a
garbage pointer value into the queue (if non-blocking mode was in use).
diff --git a/src/backend/storage/ipc/shm_mq.c b/src/backend/storage/ipc/shm_mq.c
index 6f9c3a3..d96627a 100644
--- a/src/backend/storage/ipc/shm_mq.c
+++ b/src/backend/storage/ipc/shm_mq.c
@@ -676,7 +676,10 @@ shm_mq_send_bytes(shm_mq_handle *mqh, Size nbytes, void *data, bool nowait,
/* Bail out if the queue has been detached. */
if (detached)
+ {
+ *bytes_written = sent;
return SHM_MQ_DETACHED;
+ }
if (available == 0)
{
@@ -691,12 +694,16 @@ shm_mq_send_bytes(shm_mq_handle *mqh, Size nbytes, void *data, bool nowait,
if (nowait)
{
if (shm_mq_get_receiver(mq) == NULL)
+ {
+ *bytes_written = sent;
return SHM_MQ_WOULD_BLOCK;
+ }
}
else if (!shm_mq_wait_internal(mq, &mq->mq_receiver,
mqh->mqh_handle))
{
mq->mq_detached = true;
+ *bytes_written = sent;
return SHM_MQ_DETACHED;
}
mqh->mqh_counterparty_attached = true;
--
Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers
