Re: [PATCH] block-mq:Fix the null memory access while setting tags cpumask

Raghavendra K T [off-list ref] writes:

In nr_hw_queues >1 cases when certain number of cpus are onlined/or
offlined, that results change in request_queue map in block-mq layer,
we see the kernel dumping like:

What version is that patch against?  This problem should be fixed by the
patch set from Akinobu Mita.  I tested cpu hot-plugging on my system
with an nvme device, and it is fixed there on mainline.

Cheers,
Jeff

On 10/13/2015 10:17 PM, Jeff Moyer wrote:

Raghavendra K T [off-list ref] writes:

quoted

In nr_hw_queues >1 cases when certain number of cpus are onlined/or
offlined, that results change in request_queue map in block-mq layer,
we see the kernel dumping like:

What version is that patch against?  This problem should be fixed by the
patch set from Akinobu Mita.  I tested cpu hot-plugging on my system
with an nvme device, and it is fixed there on mainline.

Hi Jeff,

My patch was against (4.3.0-rc2) bcee19f424a.

I do see now that  commit 1356aae08338 (blk-mq: avoid setting
hctx->tags->cpumask before allocation) has equivalent fix.

Thanks for the information.