Project

General

Profile

Actions

Bug #521

closed

syscall number clash with OpenAFS

Added by Garrett D'Amore over 12 years ago. Updated 9 months ago.

Status:
Closed
Priority:
High
Category:
kernel
Start date:
2010-12-15
Due date:
% Done:

0%

Estimated time:
Difficulty:
Medium
Tags:
needs-triage
Gerrit CR:
External Bug:

Description

From Frank B. @ Oracle:

Hey Garrett,

you may want to do the OpenAFS folks the same favour as we did via
although I can not claim I know about openAFS on illumos, you may want to
ask them here:

http://lists.openafs.org/mailman/listinfo/port-solaris

I filed this bug for them and Roger fixed it:

http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=7003018

see:

http://lists.openafs.org/pipermail/port-solaris/2010-November/thread.html
http://lists.openafs.org/pipermail/port-solaris/2010-December/thread.html

we "broke" their syscall number abuse in build 135 with:

http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6906485

this affects ON and the solaris 10 brand btw.

cheers
frankB

Actions #1

Updated by Albert Lee over 12 years ago

Just in case it matters in the future, it would be good to know whether all of the *at vectors were relocated for the fix, or just unlinkat.

Actions #2

Updated by Andrew Deason about 11 years ago

  • Difficulty set to Medium
  • Tags set to needs-triage

Albert Lee wrote:

Just in case it matters in the future, it would be good to know whether all of the *at vectors were relocated for the fix, or just unlinkat.

From the release of Solaris 11, it looks like just unlinkat was moved (it was moved from 65 to 76). There are several other differences in the syscall numbers that appear to be unrelated to this... but that appears to be the only difference for the *at calls.

Any chance this could be moved in illumos?

Actions #3

Updated by Garrett D'Amore about 11 years ago

  • Assignee set to Garrett D'Amore

I'd like to go ahead and do this. I'm going to go ahead and assign it to myself.

Actions #4

Updated by Garrett D'Amore 9 months ago

I guess this is probably defunct now? If it still needs to be addressed please reply here, and I can try to put together a change to do it. This feels through the cracks a decade ago. Sorry.

If it's irrelevant now, let me know and I'll just close it.

Actions #5

Updated by Garrett D'Amore 9 months ago

  • Status changed from New to Closed

I'm closing this... I believe OpenAFS has probably made the change to use ioctl instead of custom sys calls, and nobody is using problematic versions of OpenAFS anymore. The fact that nobody else has complained in 10 years makes me convinced we can just move on.

Actions

Also available in: Atom PDF