Page MenuHomeFreeBSD

[PR 206219] Kernel panic from lagg_ioctl and lagg_port_ioctl
ClosedPublic

Authored by rpokala on Jan 14 2016, 1:46 AM.
Tags
None
Referenced Files
F82196110: D4929.id12279.diff
Fri, Apr 26, 9:00 AM
F82196010: D4929.id12278.diff
Fri, Apr 26, 8:59 AM
F82195997: D4929.id12334.diff
Fri, Apr 26, 8:59 AM
F82195993: D4929.id.diff
Fri, Apr 26, 8:59 AM
F82195988: D4929.id12340.diff
Fri, Apr 26, 8:59 AM
F82161138: D4929.diff
Fri, Apr 26, 1:56 AM
F82124696: D4929.id12278.diff
Thu, Apr 25, 6:47 PM
Unknown Object (File)
Thu, Apr 25, 5:16 AM
Subscribers

Details

Summary

r287723 removed some cleanup from lagg(4), which leads to panics when
changing configuration. Restore the spirit of the code which was removed.
This issue has been refactored out of existence in -HEAD, so this patch is
directly against stable/10.

Test Plan

Create and destroy lagg, add and remove ports, switch laggprotos; used to panic, doesn't anymore.

Diff Detail

Repository
rS FreeBSD src repository - subversion
Lint
Lint Not Applicable
Unit
Tests Not Applicable

Event Timeline

rpokala retitled this revision from to [PR 206219] Kernel panic from lagg_ioctl and lagg_port_ioctl r287723 removed some cleanup from lagg(4), which leads to panics when changing configuration. Restore the spirit of the code which was removed. This issue has been refactored out of....
rpokala updated this object.
rpokala edited the test plan for this revision. (Show Details)
rpokala retitled this revision from [PR 206219] Kernel panic from lagg_ioctl and lagg_port_ioctl r287723 removed some cleanup from lagg(4), which leads to panics when changing configuration. Restore the spirit of the code which was removed. This issue has been refactored out of... to [PR 206219] Kernel panic from lagg_ioctl and lagg_port_ioctl.Jan 14 2016, 1:47 AM
rpokala updated this object.
rpokala edited edge metadata.

Cleanup suggested by @darius-dons.net.au : no need to unlock, when it is immediately locked again.

This revision is now accepted and ready to land.Jan 14 2016, 2:45 AM
rpokala edited edge metadata.

Upon further examination, we discovered that some detach routines do
LAGG_WUNLOCK() and some don't. To minimize changes, leave the locking as-is,
and only do the NULLing.

This revision now requires review to proceed.Jan 14 2016, 11:38 PM
rpokala edited edge metadata.
hiren added a reviewer: hiren.
hiren added subscribers: hrs, hiren.

looks okay to me. I am not sure what was the rationale for removing NULL initializations in the first place. May be @hrs could answer.

This revision is now accepted and ready to land.Jan 15 2016, 12:19 AM
This revision was automatically updated to reflect the committed changes.