On Fri, 01 Aug 2008, greg@xxxxxxxxxxxx wrote: > On Jul 29, 11:19am, Andrew Vasquez wrote: > } Subject: Re: Poisoning of Linux initiators on SCST reboot. > > Good afternoon to everyone, hope the week is ending well. > > > Ok, here's a patch against Linus' latest git tree which corrects the > > issue locally. Could you try it out on your test configuration? > > We'd like to perform some additional sanity-testing locally before > > we submit the fix formally for upstream consumption. > > The patch worked perfectly on our test configuration. After the SCST > target reboot the client initiator picked up on the presence of the > storage and properly discovered it. > > Evidently the qla2xxx driver has changed significantly from where > Linus' tree was at when you sent the patch. My local git clone is up > to date with Linus' git tree which has > 94ad374a0751f40d25e22e036c37f7263569d24c at its head and the patch > didn't apply. > > The kernel I tested against was from my local git clone of Linus' tree > which has 94ad374a0751f40d25e22e036c37f7263569d24c at its head. The > qla2xxx driver has evidently changed significantly since the patch > didn't apply cleanly against this tree. It was trivial, however, to > update the patch. Thank you for the update. The patch is already in our patch-queue for our next -rcX submission. > > Sidenote: we'll also need to backport some form of this fix to > > stable, but qla2xxx in kernels upto 2.6.26 didn't utilize the > > terminate-rport-io infrastructure: > > That would seem to be the case. It doesn't appear 2.6.26 even has the > necessary infra-structure to support the patch. > > Any idea on how tough a backport to 2.6.26 or earlier would be? > Running 'head-of-git' from Linus' tree would be somewhat of a stretch > for production applications.... :-)( We are still looking to finalize a migration path. As it stands now, the least invasive mechanism looks like backporting: [SCSI] qla2xxx: Add dev_loss_tmo_callbk/terminate_rport_io callback support. http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=5f3a9a207f1fccde476dd31b4c63ead2967d934f [SCSI] qla2xxx: Set an rport's dev_loss_tmo value in a consistent manner. http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=85821c906cf3563a00a3d98fa380a2581a7a5ff1 then adding this patch should suffice. We'll do some more testing locally. Regards, Andrew Vasquez -- To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
