linux-kernel - Re: [PATCH RFT 0/3] blk-mq: Optimise blk_mq_queue_tag_busy

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <9859e133-e3b8-4e53-dfad-cbf75ed3102f@huawei.com>
Date:   Thu, 25 Nov 2021 14:09:33 +0000
From:   John Garry <john.garry@...wei.com>
To:     Kashyap Desai <kashyap.desai@...adcom.com>, <axboe@...nel.dk>
CC:     <linux-block@...r.kernel.org>, <linux-kernel@...r.kernel.org>,
        <ming.lei@...hat.com>, <hare@...e.de>
Subject: Re: [PATCH RFT 0/3] blk-mq: Optimise blk_mq_queue_tag_busy_iter() for
 shared tags

On 25/11/2021 13:46, Kashyap Desai wrote:
>> John -
>>
>> I will try something this week and let you know.
> John - I tested patchset and looks good.  Issue reported at below thread is
> fixed.
> https://lore.kernel.org/linux-block/e4e92abbe9d52bcba6b8cc6c91c442cc@mail.gmail.com/
> 
> Here is perf top data.
> 
> 5.70%  [megaraid_sas]                [k] complete_cmd_fusion
>     3.80%  [megaraid_sas]                [k]
> megasas_build_and_issue_cmd_fusion
>     2.75%  [kernel]                      [k] syscall_return_via_sysret
>     2.68%  [kernel]                      [k] entry_SYSCALL_64
>     2.22%  [kernel]                      [k] io_submit_one
>     2.19%  [megaraid_sas]                [k] megasas_build_ldio_fusion
>     1.95%  [kernel]                      [k] llist_add_batch
>     1.80%  [kernel]                      [k] llist_reverse_order
>     1.79%  [kernel]                      [k] scsi_complete
>     1.73%  [kernel]                      [k] scsi_queue_rq
>     1.66%  [kernel]                      [k] check_preemption_disabled
>     1.37%  [megaraid_sas]                [k] megasas_queue_command
>     1.16%  [kernel]                      [k] native_irq_return_iret
>     1.11%  [kernel]                      [k] aio_complete_rw
>     1.07%  [kernel]                      [k] read_tsc
>     1.06%  fio                           [.] __fio_gettime
>     1.05%  [kernel]                      [k] flush_smp_call_function_queue
>     1.03%  [kernel]                      [k] blk_complete_reqs
>     1.00%  [kernel]                      [k] blk_mq_free_request
>     0.98%  [kernel]                      [k] sbitmap_get
> 
> 
> I will continue testing and let you know how it goes.

ok, good to know, thanks. But I would still like to know what is 
triggering blk_mq_queue_tag_busy_iter() so often. Indeed, as mentioned 
in this cover letter, this function was hardly optimised before for 
shared sbitmap.

And any opinion whether we would want this as a fix? Information 
requested above would help explain why we would need it as a fix.

Cheers,
John