[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <191b6d46-cb8d-4e5d-8c4c-9dae91e32a45@intel.com>
Date: Wed, 2 Mar 2022 16:32:02 +0800
From: Yin Fengwei <fengwei.yin@...el.com>
To: kernel test robot <oliver.sang@...el.com>,
Brian Foster <bfoster@...hat.com>
CC: "Darrick J. Wong" <djwong@...nel.org>,
Dave Chinner <dchinner@...hat.com>,
LKML <linux-kernel@...r.kernel.org>, <lkp@...ts.01.org>,
<lkp@...el.com>, <ying.huang@...el.com>, <feng.tang@...el.com>,
<zhengjun.xing@...ux.intel.com>
Subject: Re: [xfs] 6191cf3ad5: stress-ng.rename.ops_per_sec -73.5% regression
Hi,
On 2022/1/31 23:49, kernel test robot wrote:
> 0.00 +59.8 59.78 perf-profile.calltrace.cycles-pp.osq_lock.__mutex_lock.flush_workqueue.xfs_fs_statfs.statfs_by_dentry
> 23.44 ± 2% +61.0 84.49 perf-profile.calltrace.cycles-pp.vfs_statfs.user_statfs.__do_sys_statfs.do_syscall_64.entry_SYSCALL_64_after_hwframe
> 23.40 ± 2% +61.1 84.47 perf-profile.calltrace.cycles-pp.statfs_by_dentry.vfs_statfs.user_statfs.__do_sys_statfs.do_syscall_64
We did further check of the perf data. From above three lines,
the patch brought higher lock content (+59.8) in this extreme cases.
Regarding the patch could fix race issue, we will not tracking this
report. Let us know if you have any comments. Thanks.
Regards
Yin, Fengwei
Powered by blists - more mailing lists