linux-kernel - Very low responsiveness for pstore ftrace

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>] [day] [month] [year] [list]

Date:	Tue, 9 Feb 2016 14:21:12 -0800
From:	"Shi, Yang" <yang.shi@...aro.org>
To:	anton@...msg.org, ccross@...roid.com, keescook@...omium.org,
	tony.luck@...el.com
Cc:	LKML <linux-kernel@...r.kernel.org>, linux-fsdevel@...r.kernel.org
Subject: Very low responsiveness for pstore ftrace

Hi folks,

I'm newbie to pstore. I just did a quick try for pstore ftrace. But, I 
experience very serious responsiveness problem. When running "echo 1 > 
/sys/kernel/debug/pstore/record_ftrace", it took a couple of minutes to 
return (here return means to get prompt back) on my arm64 machine (8 A57 
cores). time command shows:

# time echo 1 > /sys/kernel/debug/pstore/record_ftrace
hrtimer: interrupt took 17532560 ns

real    7m35.717s
user    0m0.004s
sys     2m25.244s

I had a quick look at the code, it looks there is a long critical 
section with irq disabled.

If with CONFIG_PSTORE_CONSOLE enabled, the machine even can't response 
interrupts. The pstore console code also need acquire the same lock with 
irq disabled.

The low responsiveness sounds not reasonable to me.

Any idea?

Thanks,
Yang