On 4/25/18 2:38 AM, Stefan Haberland wrote: > On 25.04.2018 09:41, Christian Borntraeger wrote: >> On 04/24/2018 10:01 PM, Ming Lei wrote: >>> This reverts commit 37c7c6c76d431dd7ef9c29d95f6052bd425f004c. >>> >>> Turns out some drivers(most are FC drivers) may not use managed >>> IRQ affinity, and has their customized .map_queues meantime, so >>> still keep this code for avoiding regression. >>> >>> Reported-by: Laurence Oberman <loberman@xxxxxxxxxx> >>> Cc: Ewan Milne <emilne@xxxxxxxxxx> >>> Cc: Stefan Haberland <sth@xxxxxxxxxxxxxxxxxx> >>> Cc: Christian Borntraeger <borntraeger@xxxxxxxxxx> >>> Cc: Christoph Hellwig <hch@xxxxxx> >>> Cc: Sagi Grimberg <sagi@xxxxxxxxxxx> >>> Signed-off-by: Ming Lei <ming.lei@xxxxxxxxxx> >> Seems to work ok with 4.17-rc2 + this patch with s390 and dasd devices. >> So one of the other fixes seems to have taken care of the original issue. >> Would be good if Stefan Haberland could also verify that. >> >> > > Looks OK to me, too. > Did some testing with 4.17-rc2 + this patch on LPAR and z/VM with 1-65 > CPUs as > well as dynamic CPU definitionon z/VM and a small DASD regression test. For both you and Christian - it'd be really nice if those positive test results could be included in the forms of a Tested-by or similar in the patch itself. -- Jens Axboe