linux-kernel - Re: [PATCH v2 00/12] Improve Raid5 Lock Contention

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]

Date:   Mon, 25 Apr 2022 09:39:30 -0600
From:   Logan Gunthorpe <logang@...tatee.com>
To:     Guoqing Jiang <guoqing.jiang@...ux.dev>, Xiao Ni <xni@...hat.com>
Cc:     open list <linux-kernel@...r.kernel.org>,
        linux-raid <linux-raid@...r.kernel.org>,
        Song Liu <song@...nel.org>,
        Christoph Hellwig <hch@...radead.org>,
        Stephen Bates <sbates@...thlin.com>,
        Martin Oliveira <Martin.Oliveira@...eticom.com>,
        David Sloan <David.Sloan@...eticom.com>
Subject: Re: [PATCH v2 00/12] Improve Raid5 Lock Contention



On 2022-04-24 02:00, Guoqing Jiang wrote:
> 
> 
> On 4/22/22 12:02 AM, Logan Gunthorpe wrote:
>>
>> On 2022-04-21 02:45, Xiao Ni wrote:
>>> Could you share the commands to get the test result (lock contention
>>> and performance)?
>> Sure. The performance we were focused on was large block writes. So we
>> setup raid5 instances with varying number of disks and ran the following
>> fio script directly on the drive.
>>
>> [simple]
>> filename=/dev/md0
>> ioengine=libaio
>> rw=write
>> direct=1
>> size=8G
>> blocksize=2m
>> iodepth=16
>> runtime=30s
>> time_based=1
>> offset_increment=8G
>> numjobs=12
>> 
>> (We also played around with tuning this but didn't find substantial
>> changes once the bottleneck was hit)
> 
> Nice, I suppose other IO patterns keep the same performance as before.
> 
>> We tuned md with parameters like:
>>
>> echo 4 > /sys/block/md0/md/group_thread_cnt
>> echo 8192 > /sys/block/md0/md/stripe_cache_size
>>
>> For lock contention stats, we just used lockstat[1]; roughly like:
>>
>> echo 1 > /proc/sys/kernel/lock_stat
>> fio test.fio
>> echo 0 > /proc/sys/kernel/lock_stat
>> cat /proc/lock_stat
>>
>> And compared the before and after.
> 
> Thanks for your effort, besides the performance test, please try to run
> mdadm test suites to avoid regression.

Yeah, is there any documentation for that? I tried to look into it but
couldn't figure out how it's run.

I do know that lkp-tests has run it on this series as I did get an error
from it. But while I'm pretty sure that error has been resolved, I was
never able to figure out how to run them locally.

Thanks,

Logan