Le Sun, 23 Sep 2018 20:22:56 +0200
Stefan Priebe - Profihost AG <s.priebe@xxxxxxxxxxxx> écrivait:
> Am 22.09.2018 um 23:40 schrieb Bart Van Assche:
> > On 9/18/18 11:10 PM, Stefan Priebe - Profihost AG wrote:
> >> after upgrading the aacraid driver / kernel from aacraid 50792 to
> >> aacraid 50877.
> >
> > The aacraid driver version was updated to 50792 in commit
> > 0662cc968ace ("scsi: aacraid: Update driver version") and to 50877
> > in commit 1cdb74b80f93 ("scsi: aacraid: Update driver version to
> > 50877"). That means that the regression you encountered got
> > introduced after commit 0662cc968ace. 114 changes got checked in
> > after that commit. That's too much to find the root cause by
> > rereading all these changes. Is there any way to trigger the
> > problem faster such that it becomes feasible to run a bisect?
>
> Sadly i'm not able. May be also something else in the kernel has
> changed.
>
> I'm now trying the original out of tree driver from microsemi /
> adaptec: Adaptec aacraid driver 1.2.1.56008src
>
> No idea how those driver versions corespond to the kernel ones.
I can tell you that the last working version for me is 50834. You can
find the version number in aacraid.h, that's the value of
the AAC_DRIVER_BUILD constant.
the 50877 driver has been broken since it was out several months ago!
--
------------------------------------------------------------------------
Emmanuel Florac | Direction technique
| Intellique
| <eflorac@xxxxxxxxxxxxxx>
| +33 1 78 94 84 02
------------------------------------------------------------------------
Attachment:
pgp1Mq21gpItg.pgp
Description: Signature digitale OpenPGP
