Bug 104 - Setuid problem in Cygwin TORQUE
: Setuid problem in Cygwin TORQUE
Status: RESOLVED FIXED
Product: TORQUE
pbs_mom
: 2.5.x
: PC Linux
: P5 enhancement
Assigned To: David Beer
:
:
:
  Show dependency treegraph
 
Reported: 2010-12-07 09:13 MST by David Beer
Modified: 2010-12-08 11:43 MST (History)
5 users (show)

See Also:


Attachments
This is Vikentsi's proposed patch (2.05 KB, patch)
2010-12-07 09:14 MST, David Beer
Details | Diff


Note

You need to log in before you can comment on or make changes to this bug.


Description David Beer 2010-12-07 09:13:35 MST
Reported by Vikentsi Lapa:

> Hi, all.
> 
> I send my mail once again, because previous stay without answer.
> 
> My job file is
> 
> #!/bin/sh
> #PBS -l nodes=4:ppn=2,walltime=00:00:20
> sleep 40
> 
> 
> After job run I have stderr file with
> 
> PBSTest.?1642
> =>> PBS: job killed: walltime 51 exceeded limit 20
> 
> Error is cygwin related. They appear because we don't have root
> (UID=0) user in
> cygwin enviroment and seteuid(0) not work. Error appear only if one of
> limit
> exceeded. Attached patch fix this differences. Please, examine it and
> add to repository.
>
Comment 1 David Beer 2010-12-07 09:14:20 MST
Created an attachment (id=67) [details]
This is Vikentsi's proposed patch
Comment 2 David Beer 2010-12-07 09:15:06 MST
I have looked at the patch and propose that we apply it, except that the
changes are all encased in 
#ifdef CYGWIN
statements.

Any objections?
Comment 3 Garrick Staples 2010-12-07 10:56:08 MST
(In reply to comment #2)
> I have looked at the patch and propose that we apply it, except that the
> changes are all encased in 
> #ifdef CYGWIN
> statements.
> 
> Any objections?

I don't see anything cygwin-specific in that patch. Why would you want to ifdef
it?

I worked really hard to remove most of the platfom ifdef's! Please don't add
them again!  If code should be conditionally compiled, then add the necessary
feature check in configure.ac
Comment 4 Glen 2010-12-07 11:14:11 MST
(In reply to comment #2)
> I have looked at the patch and propose that we apply it, except that the
> changes are all encased in 
> #ifdef CYGWIN
> statements.
> 
> Any objections?


I don't see a reason for the #ifdef
Comment 5 David Beer 2010-12-08 11:29:37 MST
> 
> I don't see a reason for the #ifdef

My mistake - it doesn't need it. I see where he sets it.
Comment 6 Ken Nielson 2010-12-08 11:43:54 MST
This patch has been committed to 2.5-fixes, 3.0-fixes and trunk as is. No
#ifdef CYGWIN.