Re: [PATCH] block: enable per-cpu bio cache by default

[fengnan chang <fengnanchang@gmail.com>]

On Mon, Oct 13, 2025 at 1:19 PM Ming Lei <ming.lei@redhat.com> wrote:
>
> On Sat, Oct 11, 2025 at 09:33:12AM +0800, Fengnan Chang wrote:
> > Per cpu bio cache was only used in the io_uring + raw block device,
> > after commit 12e4e8c7ab59 ("io_uring/rw: enable bio caches for IRQ
> > rw"),  bio_put is safe for task and irq context, bio_alloc_bioset is
> > safe for task context and no one calls in irq context, so we can enable
> > per cpu bio cache by default.
> >
> > Benchmarked with t/io_uring and ext4+nvme:
> > taskset -c 6 /root/fio/t/io_uring  -p0 -d128 -b4096 -s1 -c1 -F1 -B1 -R1
> > -X1 -n1 -P1  /mnt/testfile
> > base IOPS is 562K, patch IOPS is 574K. The CPU usage of bio_alloc_bioset
> > decrease from 1.42% to 1.22%.
> >
> > The worst case is allocate bio in CPU A but free in CPU B, still use
> > t/io_uring and ext4+nvme:
> > base IOPS is 648K, patch IOPS is 647K.
>
> Just be curious, how do you run the remote bio free test? If the nvme is 1:1
> mapping, you may not trigger it.

I modified the nvme driver, reduce the number of queues.

>
> BTW, ublk has this kind of remote bio free trouble, but not see IOPS drop
> with this patch.
>
> The patch itself looks fine for me.
>
>
> Thanks,
> Ming
>
>