https://bugzilla.kernel.org/show_bug.cgi?id=199435
--- Comment #9 from Don ([email protected]) --- When you applied the 4.16 hpsa driver patches, was this patch also applied? commit 84676c1f21e8ff54befe985f4f14dc1edc10046b Author: Christoph Hellwig <[email protected]> Date: Fri Jan 12 10:53:05 2018 +0800 genirq/affinity: assign vectors to all possible CPUs Currently we assign managed interrupt vectors to all present CPUs. This works fine for systems were we only online/offline CPUs. But in case of systems that support physical CPU hotplug (or the virtualized version of it) this means the additional CPUs covered for in the ACPI tables or on the command line are not catered for. To fix this we'd either need to introduce new hotplug CPU states just for this case, or we can start assining vectors to possible but not present CPUs. Reported-by: Christian Borntraeger <[email protected]> Tested-by: Christian Borntraeger <[email protected]> Tested-by: Stefan Haberland <[email protected]> Fixes: 4b855ad37194 ("blk-mq: Create hctx for each present CPU") Cc: [email protected] Cc: Thomas Gleixner <[email protected]> Signed-off-by: Christoph Hellwig <[email protected]> Signed-off-by: Jens Axboe <[email protected]> The above patch is why the hpsa-fix-selection-of-reply-queue patch was needed. If not, I would redact that patch because it may be causing your issues. There was another patch required for the hpsa-fix-selection-of-reply-queue patch: scsi-introduce-force-blk-mq. The errors shown in your logs indicate issues with DMA transfers of your data. Unaligned partial completion errors are usually issues with the scatter/gather buffers that represent your data buffers. I would like to eliminate using the 4.16 hpsa driver in a 4.11 kernel. Can you try our out-of-box driver? I'll attach this to the BZ. You compile it with make -f Makefile.alt The name is hpsa-3.4.20-136.tar.bz2 -------- commit 8b834bff1b73dce46f4e9f5e84af6f73fed8b0ef Author: Ming Lei <[email protected]> Date: Tue Mar 13 17:42:39 2018 +0800 scsi: hpsa: fix selection of reply queue Since commit 84676c1f21e8 ("genirq/affinity: assign vectors to all possible CPUs") we could end up with an MSI-X vector that did not have any online CPUs mapped. This would lead to I/O hangs since there was no CPU to receive the completion. Retrieve IRQ affinity information using pci_irq_get_affinity() and use this mapping to choose a reply queue. [mkp: tweaked commit desc] Cc: Hannes Reinecke <[email protected]> Cc: "Martin K. Petersen" <[email protected]>, Cc: James Bottomley <[email protected]>, Cc: Christoph Hellwig <[email protected]>, Cc: Don Brace <[email protected]> Cc: Kashyap Desai <[email protected]> Cc: Laurence Oberman <[email protected]> Cc: Meelis Roos <[email protected]> Cc: Artem Bityutskiy <[email protected]> Cc: Mike Snitzer <[email protected]> Fixes: 84676c1f21e8 ("genirq/affinity: assign vectors to all possible CPUs") Signed-off-by: Ming Lei <[email protected]> Tested-by: Laurence Oberman <[email protected]> Tested-by: Don Brace <[email protected]> Tested-by: Artem Bityutskiy <[email protected]> Acked-by: Don Brace <[email protected]> Reviewed-by: Christoph Hellwig <[email protected]> Signed-off-by: Martin K. Petersen <[email protected]> I believe this patch is also required. commit cf2a0ce8d1c25c8cc4509874d270be8fc6026cc3 Author: Ming Lei <[email protected]> Date: Tue Mar 13 17:42:41 2018 +0800 scsi: introduce force_blk_mq From scsi driver view, it is a bit troublesome to support both blk-mq and non-blk-mq at the same time, especially when drivers need to support multi hw-queue. This patch introduces 'force_blk_mq' to scsi_host_template so that drivers can provide blk-mq only support, so driver code can avoid the trouble for supporting both. Cc: Omar Sandoval <[email protected]>, Cc: "Martin K. Petersen" <[email protected]>, Cc: James Bottomley <[email protected]>, Cc: Christoph Hellwig <[email protected]>, Cc: Don Brace <[email protected]> Cc: Kashyap Desai <[email protected]> Cc: Mike Snitzer <[email protected]> Cc: Laurence Oberman <[email protected]> Reviewed-by: Hannes Reinecke <[email protected]> Reviewed-by: Christoph Hellwig <[email protected]> Signed-off-by: Ming Lei <[email protected]> -- You are receiving this mail because: You are the assignee for the bug.
