linux-ext4 - [Bug 217965] ext4(?) regression since 6.5.0 on sata hdd

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: Sun, 07 Jan 2024 02:56:44 +0000
From: bugzilla-daemon@...nel.org
To: linux-ext4@...r.kernel.org
Subject: [Bug 217965] ext4(?) regression since 6.5.0 on sata hdd

https://bugzilla.kernel.org/show_bug.cgi?id=217965

--- Comment #65 from Matthew Stapleton (matthew4196@...il.com) ---
I'm not sure if this will be worth looking into, but with the cpu usage
problem, I've done some more testing and "perf report --stdio -i perf.data" is
showing around 2 to 3 times as much cpu time since the commit that added the
CR_BEST_AVAIL_LEN feature: 7e170922f06bf .  With
/sys/fs/ext4/<dev>/mb_best_avail_max_trim_order set to 0 it goes a bit lower,
but is still at least 2 times cpu.  With perf record -ag sleep 25 which is the
time it takes the extract to run, it's goes from about 2% to 5% so fairly
small, but when measuring with 5 second intervals this can go from around 5% to
15%.  This cpu usage is concentrated in ext4_get_group_desc and
ext4_get_group_info and within these functions, perf is picking up more cpu
time in __rcu_read_lock and __rcu_read_unlock.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.