Lustre devel - Jan 2007 - [Bug 10866] proc_file_read() causes MDS crash on stop

Please don''t reply to lustre-devel. Instead, comment in Bugzilla by
using the following link:
https://bugzilla.lustre.org/show_bug.cgi?id=10866



Created an attachment (id=9376)
Please don''t reply to lustre-devel. Instead, comment in Bugzilla by
using the following link:
 --> (https://bugzilla.lustre.org/attachment.cgi?id=9376&action=view)
additional fixes for b1_5

additional proc files for the MGS, plus fixed up the regression test.
I EXCEPTed the test, but that doesn''t make the slab corruption go away.

I think both the b1_4 and the b1_5 patches for 10866 should be removed from CVS

until this issue is resolved.

Please don''t reply to lustre-devel. Instead, comment in Bugzilla by
using the following link:
https://bugzilla.lustre.org/show_bug.cgi?id=10866

           What    |Removed                     |Added
----------------------------------------------------------------------------
Attachment #9339 is|0                           |1
           obsolete|                            |
Attachment #9376 is|0                           |1
           obsolete|                            |


Created an attachment (id=9377)
Please don''t reply to lustre-devel. Instead, comment in Bugzilla by
using the following link:
 --> (https://bugzilla.lustre.org/attachment.cgi?id=9377&action=view)
net b1_5 patch (reversed)

reverted in cvs, this is the net b1_5 patch (in reversed form)

Please don''t reply to lustre-devel. Instead, comment in Bugzilla by
using the following link:
https://bugzilla.lustre.org/show_bug.cgi?id=10866

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|jxiong@clusterfs.com        |adilger@clusterfs.com


Created an attachment (id=9379)
Please don''t reply to lustre-devel. Instead, comment in Bugzilla by
using the following link:
 --> (https://bugzilla.lustre.org/attachment.cgi?id=9379&action=view)
fixups I was working on before reverting the patch

Please don''t reply to lustre-devel. Instead, comment in Bugzilla by
using the following link:
https://bugzilla.lustre.org/show_bug.cgi?id=10866



(In reply to comment #47)
> Created an attachment (id=9379)
Please don''t reply to lustre-devel. Instead, comment in Bugzilla by
using the following link:
 -->
(https://bugzilla.lustre.org/attachment.cgi?id=9379&action=view)
> fixups I was working on before reverting the patch
>       
Could someone please explain the background of the
"slab corruption" comment and the patch gyrations 
going on with this bug.    We are currently testing 
with an earlier patch for this bug, and we are concerned 
that we missed something, and will need to redo that.

Please don''t reply to lustre-devel. Instead, comment in Bugzilla by
using the following link:
https://bugzilla.lustre.org/show_bug.cgi?id=10866



(In reply to comment #81)
> Could someone please explain the background of the
> "slab corruption" comment and the patch gyrations 
> going on with this bug.    We are currently testing 
> with an earlier patch for this bug, and we are concerned 
> that we missed something, and will need to redo that.      
Sorry Terry; hopefully I''ve marked the comments "public" now.
But I would not use any patches posted here until we resolve the problem.

Please don''t reply to lustre-devel. Instead, comment in Bugzilla by
using the following link:
https://bugzilla.lustre.org/show_bug.cgi?id=10866

           What    |Removed                     |Added
----------------------------------------------------------------------------
Attachment #9320 is|0                           |1
           obsolete|                            |


Created an attachment (id=9470)
Please don''t reply to lustre-devel. Instead, comment in Bugzilla by
using the following link:
 --> (https://bugzilla.lustre.org/attachment.cgi?id=9470&action=view)
Patch for b1_4

Please don''t reply to lustre-devel. Instead, comment in Bugzilla by
using the following link:
https://bugzilla.lustre.org/show_bug.cgi?id=10866



Created an attachment (id=9495)
Please don''t reply to lustre-devel. Instead, comment in Bugzilla by
using the following link:
 --> (https://bugzilla.lustre.org/attachment.cgi?id=9495&action=view)
revised macros and added missing proc

some fixups.  testing now.

Please don''t reply to lustre-devel. Instead, comment in Bugzilla by
using the following link:
https://bugzilla.lustre.org/show_bug.cgi?id=10866

           What    |Removed                     |Added
----------------------------------------------------------------------------
Attachment #9495 is|0                           |1
           obsolete|                            |


Created an attachment (id=9496)
Please don''t reply to lustre-devel. Instead, comment in Bugzilla by
using the following link:
 --> (https://bugzilla.lustre.org/attachment.cgi?id=9496&action=view)
fix for client import procs

resolve deadlock from ldlm_namespace_free from previous version of this patch.

Please don''t reply to lustre-devel. Instead, comment in Bugzilla by
using the following link:
https://bugzilla.lustre.org/show_bug.cgi?id=10866

           What    |Removed                     |Added
----------------------------------------------------------------------------
Attachment #9133 is|0                           |1
           obsolete|                            |
Attachment #9377 is|0                           |1
           obsolete|                            |
Attachment #9379 is|0                           |1
           obsolete|                            |


Created an attachment (id=9512)
Please don''t reply to lustre-devel. Instead, comment in Bugzilla by
using the following link:
 --> (https://bugzilla.lustre.org/attachment.cgi?id=9512&action=view)
fix for client import procs b1_4

1.4 version of attachment 9496.  This bug can re-close as soon as this patch is

approved and landed. 
(I''m also obsoleting the out of date attachments.)

Please don''t reply to lustre-devel. Instead, comment in Bugzilla by
using the following link:
https://bugzilla.lustre.org/show_bug.cgi?id=10866

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Attachment #9512|                            |review+
               Flag|                            |


(From update of attachment 9512)
This looks a lot cleaner to me.
>@@ -419,7 +419,7 @@ void ptlrpc_lprocfs_rpc_sent(struct ptlr
> {
>         struct lprocfs_stats *svc_stats;
>         int opc =  opcode_offset(req->rq_reqmsg->opc);
>-
>+ 
>         svc_stats = req->rq_import->imp_obd->obd_svc_stats;
>         if (svc_stats == NULL || opc <= 0)
>                 return;
Please don''t add whitespace at EOL.

Please don''t reply to lustre-devel. Instead, comment in Bugzilla by
using the following link:
https://bugzilla.lustre.org/show_bug.cgi?id=10866

           What    |Removed                     |Added
----------------------------------------------------------------------------
Attachment #9512 is|0                           |1
           obsolete|                            |


Created an attachment (id=9513)
Please don''t reply to lustre-devel. Instead, comment in Bugzilla by
using the following link:
 --> (https://bugzilla.lustre.org/attachment.cgi?id=9513&action=view)
fix for client import procs b1_4

aargh, ported the wrong version of the patch.  Sorry andreas.