On Thu, Aug 6, 2015 at 9:04 AM, Antonin Houska <a...@cybertec.at> wrote: > Can anyone please explain why the following patch shouldn't be applied? > > diff --git a/src/backend/storage/ipc/shm_mq.c > b/src/backend/storage/ipc/shm_mq.c > index 126cb07..4cd52ac 100644 > --- a/src/backend/storage/ipc/shm_mq.c > +++ b/src/backend/storage/ipc/shm_mq.c > @@ -584,7 +584,7 @@ shm_mq_receive(shm_mq_handle *mqh, Size *nbytesp, void > **datap, bool nowait) > if (mqh->mqh_partial_bytes + rb > sizeof(Size)) > lengthbytes = sizeof(Size) - > mqh->mqh_partial_bytes; > else > - lengthbytes = rb - mqh->mqh_partial_bytes; > + lengthbytes = rb; > memcpy(&mqh->mqh_buffer[mqh->mqh_partial_bytes], > rawdata, > lengthbytes); > mqh->mqh_partial_bytes += lengthbytes; > > > I'm failing to understand why anything should be subtracted. Note that the > previous iteration must have called shm_mq_inc_bytes_read(), so "rb" should > not include anything of mqh->mqh_partial_bytes. Thanks.
Hmm, I think you are correct. This would matter in the case where the message length word was read in more than two chunks. I don't *think* that's possible right now because I believe the only systems where MAXIMUM_ALIGNOF < sizeof(Size) are those with MAXIMUM_ALIGNOF == 4 and sizeof(Size) == 8. However, if we had systems where MAXIMUM_ALIGNOF == 4 and sizeof(Size) == 16, or systems where MAXIMUM_ALIGNOF == 2 and sizeof(Size) == 8, this would be a live bug. Thanks for reviewing; I'll go push this change. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company -- Sent via pgsql-hackers mailing list (pgsql-hackers@postgresql.org) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-hackers
