flock user defined function

[Chris Goughnour]

I'm trying to write two C language user defined functions, lockfile() and
unlockfile(), that call flock using LOCK_EX and LOCK_UN respectively. If I
call lockfile from a first psql process it returns successfully. Calling
lockfile from a second psql process blocks. However, when I call unlockfile
from the first psql process, the second process still blocks. The lockfile
call from the second psql proccess doesn't return until I kill the first
psql process.
Any suggestions? Thanks in advance.
Chris Goughnour

PG_FUNCTION_INF O_V1(lockFile);
Datum lockFile(PG_FUN CTION_ARGS){
text *t=PG_GETARG_TE XT_P(0);
char *path=palloc(VA RSIZE(t)-VARHDRSZ+1);
int fileHandle,stat us;
memcpy((void *)path,(void *)VARDATA(t),VA RSIZE(t)-VARHDRSZ);
path[VARSIZE(t)-VARHDRSZ]=0;
fileHandle=open ((const char *)path,O_RDONLY );
if(fileHandle==-1){
PG_RETURN_INT32 (-1);
}
if(flock(fileHa ndle,LOCK_EX)==-1){
PG_RETURN_INT32 (-1);
}
PG_RETURN_INT32 (0);
}

PG_FUNCTION_INF O_V1(unlockFile );
Datum unlockFile(PG_F UNCTION_ARGS){
text *t=PG_GETARG_TE XT_P(0);
char *path=palloc(VA RSIZE(t)-VARHDRSZ+1);
int fileHandle;
memcpy((void *)path,(void *)VARDATA(t),VA RSIZE(t)-VARHDRSZ);
path[VARSIZE(t)-VARHDRSZ]=0;
fileHandle=open ((const char *)path,O_RDONLY );
if(fileHandle==-1){
PG_RETURN_INT32 (-1);
}
if(flock(fileHa ndle,LOCK_UN)==-1){
PG_RETURN_INT32 (-1);
}
PG_RETURN_INT32 (0);
}
---------------------------(end of broadcast)---------------------------
TIP 3: if posting/reading through Usenet, please send an appropriate
subscribe-nomail command to ma*******@postg resql.org so that your
message can get through to the mailing list cleanly

[Martijn van Oosterhout]

On Tue, Jun 22, 2004 at 02:49:27PM -0700, Chris Goughnour wrote:

I'm trying to write two C language user defined functions, lockfile() and
unlockfile(), that call flock using LOCK_EX and LOCK_UN respectively. IfI
call lockfile from a first psql process it returns successfully. Calling
lockfile from a second psql process blocks. However, when I call unlockfile
from the first psql process, the second process still blocks. The lockfile
call from the second psql proccess doesn't return until I kill the first
psql process.
Any suggestions? Thanks in advance.
Chris Goughnour
<snip>

Where do you close the file? That might cause some issues.
--
Martijn van Oosterhout <kl*****@svana. org> http://svana.org/kleptog/ Patent. n. Genius is 5% inspiration and 95% perspiration. A patent is a
tool for doing 5% of the work and then sitting around waiting for someone
else to do the other 95% so you can sue them.

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQFA2K7GY5T wig3Ge+YRAv3wAJ 0UVrdFN+OlT8jh0 UrFKDS1FQG5PwCf Quz3
kMpd0/j8qyAql8yfqMozF To=
=fVSM
-----END PGP SIGNATURE-----

[Doug McNaught]

Martijn van Oosterhout <kl*****@svana. org> writes:

On Tue, Jun 22, 2004 at 02:49:27PM -0700, Chris Goughnour wrote:

I'm trying to write two C language user defined functions, lockfile() and
unlockfile(), that call flock using LOCK_EX and LOCK_UN respectively. If I
call lockfile from a first psql process it returns successfully. Calling
lockfile from a second psql process blocks. However, when I call unlockfile
from the first psql process, the second process still blocks. The lockfile
call from the second psql proccess doesn't return until I kill the first
psql process.
Any suggestions? Thanks in advance.
Chris Goughnour

<snip>

Where do you close the file? That might cause some issues.

Yeah, it's generally best not to call LOCK_UN at all, but just to
close the file (which will release the locks). Otherwise, if you are
using stdio, you can get a situation where the file is unlocked but
its stdio buffer has not been flushed, leading to the corruption you
were trying to avoid by locking the file...

-Doug

---------------------------(end of broadcast)---------------------------
TIP 2: you can get off all lists at once with the unregister command
(send "unregister YourEmailAddres sHere" to ma*******@postg resql.org)

[Tom Lane]

Chris Goughnour <cg********@hot p.com> writes:

Any suggestions? Thanks in advance.

I believe locks are associated with file descriptors (what you're
miscalling a handle). The unlock function cannot release a lock
that is held via a different descriptor. What it needs to be doing
is closing the descriptor that lockFile opened. This would also
solve the rather serious descriptor-leak problem you've got.

regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 7: don't forget to increase your free space map settings

[Chris Goughnour]

Thanks. Yeah, I figured that out after Martijn's response. I'm just
returning the file descriptor from lockFile, passing it to unlockFile and
closing the descriptor there. It works fine now. Thanks for edifying a
clueless novice such as myself. :-)

Chris Goughnour <cg********@hot p.com> writes:

Any suggestions? Thanks in advance.

I believe locks are associated with file descriptors (what you're
miscalling a handle). The unlock function cannot release a lock
that is held via a different descriptor. What it needs to be doing
is closing the descriptor that lockFile opened. This would also
solve the rather serious descriptor-leak problem you've got.

regards, tom lane

---------------------------(end of broadcast)---------------------------
TIP 5: Have you checked our extensive FAQ?

http://www.postgresql.org/docs/faqs/FAQ.html