geli panics my system after suspend-resume: g_eli_orphan_spoil_assert() called for cd0.eli

[Pawel Jakub Dawidek]

On Sun, Jun 28, 2015 at 08:30:11PM -0500, Matthew D. Fuller wrote:
> On Mon, Jun 15, 2015 at 01:01:28AM -0500 I heard the voice of
> Matthew D. Fuller, and lo! it spake thus:
> > 
> > But presumably geli should treat that spoil event the same as an
> > orpah, which at least avoids the panic and does the least wrong
> > thing in the circumstances.  Since the events DO occur (the code
> > suggests the original thought was that they couldn't), we might as
> > well deal with them as gracefully as possible.
> 
> Patch stuck in bugzilla at
> <https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=201185> to not get
> lost.

Hi Matthew.

The spoil event happens when provider is open for writting for the first
time (its open-for-write counter goes from 0 to >0). Since GELI opens
its providers always exclusively there can be no write opens, thus the
spoil event is unexpected.

To see what exactly is happening I'd need to see the console output from
before that happens with kern.geom.debugflags set to 5.

-- 
Pawel Jakub Dawidek                       http://www.wheelsystems.com
FreeBSD committer                         http://www.FreeBSD.org
Am I Evil? Yes, I Am!                     http://mobter.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <http://lists.freebsd.org/pipermail/freebsd-geom/attachments/20150701/7de96e59/attachment.bin>