On Fri, 2014-08-01 at 07:56 -0700, Roland Dreier wrote: > On Thu, Jul 31, 2014 at 5:22 PM, Nicholas A. Bellinger > <nab@xxxxxxxxxxxxxxx> wrote: > > Roland + Joern, I'm specifically talking about the following patches: > > > > https://github.com/rolandd/qla-target/commit/77bd729e5e9f858495dfded7c08e548bc9ecc9be > > https://github.com/rolandd/qla-target/commit/79ae07a141950c0008480a46f91d72318a1d799e > > > > Can you please confirm if these are related to what deeepdish has > > reported here..? > > Our latest tree doesn't use the btree -> radix tree conversion -- we > decided there is nothing wrong with the btree code itself. So from > that POV upstream is fine. > > However the first (77bd...) commit may be related. Certainly the > comment is true: the qla2xxx target stuff is racy, especially around > fabric changes (initiators/sessions coming and going). It looks like > the original trace has to do with session update, so yeah, it probably > is hitting one of those races. > Just a heads up that commit 77bd729e5 from above has been included into target-pending/for-next: qla_target: don't delete changed nacls https://git.kernel.org/cgit/linux/kernel/git/nab/target-pending.git/commit/?h=for-next&id=f4c24db1b7ad0ce84409e15744d26c6f86a96840 Deeepdish, it would be really helpful if you can verify the following patch on your tcm_qla2xxx setup, and see if you can reproduce the loss of FC connectivity with a Windows 2012 R2 initiator. Note that Joern's patch still has a WARN_ON() in place when the scenario occurs (eg: you'll still see a stack backtrace in dmesg when the issue occurs), but now tcm_qla2xxx should be able to successfully recover. Please let us know if/when your able to verify. --nab -- To unsubscribe from this list: send the line "unsubscribe target-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
