[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAGXu5jKMUrqOo4gUPLvWg3mLxYjqYE-8wUx_FJmTR_JxDuwGWQ@mail.gmail.com>
Date: Tue, 8 Sep 2015 16:42:07 -0700
From: Kees Cook <keescook@...omium.org>
To: Hiraku Toyooka <hiraku.toyooka.gu@...achi.com>
Cc: LKML <linux-kernel@...r.kernel.org>,
Tony Luck <tony.luck@...el.com>,
Linux API <linux-api@...r.kernel.org>,
Anton Vorontsov <anton@...msg.org>,
Shuah Khan <shuahkh@....samsung.com>,
Mark Salyzyn <salyzyn@...roid.com>,
Colin Cross <ccross@...roid.com>,
Seiji Aguchi <seiji.aguchi@....com>
Subject: Re: [PATCH 0/2] selftests/pstore: add pstore test script
On Tue, Sep 8, 2015 at 4:06 AM, Hiraku Toyooka
<hiraku.toyooka.gu@...achi.com> wrote:
> These scripts include test cases which check pstore behavior. This
> is useful to avoid regressions of pstore.
>
> Pstore is used across kernel crash, so these test cases are split
> into three parts.
>
> - pstore_tests: check pstore behavior before crash
> - pstore_post_reboot_tests: check pstore behavior after crash and reboot
> - pstore_crash_test: cause kernel crash and reboot
>
> The pstore_test and the pstore_post_reboot_tests are the actual scripts
> for testing pstore and are executed in usual selftest's "run_test" target.
> On the other hand, the pstore_crash_test is to cause kernel panic and reboot,
> so it is executed in new "run_pstore_crash" target which is specified ad-hoc
> by users. In addition, there is a "common_tests" script which includes
> utilities and test cases used commonly in these scripts.
>
> When the pstore_crash_test is executed, it creates a file as a reboot flag.
> The pstore_post_reboot_tests detects whether the file exists or not. If the
> file doesn't exists, the test cases are skipped.
>
> These scripts expect that one pstore backend is registered before the
> scripts are executed.
> Assumed use case is following.
>
> # cd linux/tools/testing/selftests
> # make run_tests -C pstore
> make: Entering directory '/home/root/selftests/pstore'
> === Pstore unit tests (pstore_tests)===
> Checking pstore backend is registered ... ok
> Checking pstore console is registered ... ok
> Checking /dev/pmsg0 exists ... ok
> Writing TEST_STRING to /dev/pmsg0 ... ok
> selftests: pstore_tests [PASS]
> === Pstore unit tests (pstore_post_reboot_tests)===
> Checking pstore backend is registered ... ok
> pstore_crash_test has not been executed yet. we skip further tests.
> selftests: pstore_post_reboot_tests [PASS]
> make: Leaving directory '/home/root/selftests/pstore'
> # make run_pstore_crash
> ...
> (kernel crash and reboot)
> ...
> # make run_tests -C pstore
> make: Entering directory '/home/root/selftests/pstore'
> === Pstore unit tests (pstore_tests)===
> Checking pstore backend is registered ... ok
> Checking pstore console is registered ... ok
> Checking /dev/pmsg0 exists ... ok
> Writing TEST_STRING to /dev/pmsg0 ... ok
> selftests: pstore_tests [PASS]
> === Pstore unit tests (pstore_post_reboot_tests)===
> Checking pstore backend is registered ... ok
> Mounting pstore filesystem ... ok
> Checking dmesg files exist in pstore filesystem ... ok
> dmesg-ramoops-0
> dmesg-ramoops-1
> Checking console files exist in pstore filesystem ... ok
> console-ramoops-0
> Checking pmsg files exist in pstore filesystem ... ok
> pmsg-ramoops-0
> Checking dmesg files contains oops end marker
> dmesg-ramoops-0 ... ok
> dmesg-ramoops-1 ... ok
> Checking console file contains oops end marker ... ok
> Checking pmsg file contains TEST_STRING ... ok
> Removing all files in pstore filesystem
> console-ramoops-0 ... ok
> dmesg-ramoops-0 ... ok
> dmesg-ramoops-1 ... ok
> pmsg-ramoops-0 ... ok
> selftests: pstore_post_reboot_tests [PASS]
> make: Leaving directory '/home/root/selftests/pstore'
>
>
> We can also see test logs later.
>
> # cat pstore/logs/20150903-111158/pstore_tests.log
> ...
Thanks for working on this! Pstore does need some selftests, so this
is greatly appreciated. :) I sent some improvement ideas in separate
emails.
-Kees
--
Kees Cook
Chrome OS Security
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists