From: tixy@linaro.org (Jon Medhurst (Tixy))
To: linux-arm-kernel@lists.infradead.org
Subject: [RESEND][PATCH v15 7/7] ARM: kprobes: enable OPTPROBES for ARM 32
Date: Tue, 09 Dec 2014 09:38:35 +0000 [thread overview]
Message-ID: <1418117915.3641.16.camel@linaro.org> (raw)
In-Reply-To: <54865AF2.5040707@huawei.com>
On Tue, 2014-12-09 at 10:14 +0800, Wang Nan wrote:
> Hi all,
>
> Looks like v15 patch solves the problems found in patch v13 and v14. I run Tixy's test
> cases in a loop for a whole night on two real unpublished Hisilicon ARM A15 SoCs (I
> backported kprobes to 3.10 kernel), no problem arise, they are still running now.
I get a reproducible failure (below), but I believe it is a flaw in the
test code. Now that we have optimised probes there is a window of
opportunity for and interrupt to happen between the the kprobe which
sets up the test preconditions and the tested probe executing. That will
change the contents of stack memory below SP and make it inconsistent
between test runs.
This recent flurry of kprobes work has been eating up my time somewhat
and I have other urgent matters I should attend to, so I won't be
looking at this problem for a few days.
That test failure...
strh r3, [r13, #-64]! @ e16d34b0
FAIL: test memory differs
FAIL: Test strh r3, [r13, #-64]!
FAIL: Scenario 1
initial_regs:
r0 21522152 | r1 21522052 | r2 21522352 | r3 12345678
r4 21522552 | r5 21522452 | r6 21522752 | r7 21522652
r8 21522952 | r9 21522852 | r10 21522b52 | r11 21522a52
r12 21522d52 | sp bf053ea0 | lr 21522f52 | pc 80028cf0
cpsr 18010013
expected_regs:
r0 21522152 | r1 21522052 | r2 21522352 | r3 12345678
r4 21522552 | r5 21522452 | r6 21522752 | r7 21522652
r8 21522952 | r9 21522852 | r10 21522b52 | r11 21522a52
r12 21522d52 | sp bf053e60 | lr 21522f52 | pc 80028cf4
cpsr 18010013
result_regs:
r0 21522152 | r1 21522052 | r2 21522352 | r3 12345678
r4 21522552 | r5 21522452 | r6 21522752 | r7 21522652
r8 21522952 | r9 21522852 | r10 21522b52 | r11 21522a52
r12 21522d52 | sp bf053e60 | lr 21522f52 | pc 80028cf4
cpsr 18010013
current_stack=bf053da0
expected_memory:
21525678 12345678 21522552 21522452
21522752 21522652 21522952 21522852
21522b52 21522a52 21522d52 bf053ea0
21522f52 80028cf0 18010113 ffffffff
result_memory:
45675678 4567baab 4567bbab 4567bcab
4567bdab 4567beab 4567bfab 4567c0ab
4567c1ab 4567c2ab 4567c3ab 4567c4ab
4567c5ab 4567c6ab 4567c7ab 4567c8ab
strh r3, [r13, #-64-8]! @ e16d34b8
strh r4, [r14, #-64-8]! @ e16e44b8
--
Tixy
next prev parent reply other threads:[~2014-12-09 9:38 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-12-09 2:14 [RESEND][PATCH v15 7/7] ARM: kprobes: enable OPTPROBES for ARM 32 Wang Nan
2014-12-09 9:38 ` Jon Medhurst (Tixy) [this message]
-- strict thread matches above, loose matches on Subject: below --
2014-12-08 14:09 [RESEND][PATCH v15 0/7] ARM: kprobes: OPTPROBES and other improvements Wang Nan
2014-12-08 14:09 ` [RESEND][PATCH v15 7/7] ARM: kprobes: enable OPTPROBES for ARM 32 Wang Nan
2014-12-09 6:47 ` Masami Hiramatsu
2014-12-09 9:14 ` Jon Medhurst (Tixy)
2014-12-09 10:25 ` Masami Hiramatsu
2014-12-09 10:12 ` Wang Nan
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1418117915.3641.16.camel@linaro.org \
--to=tixy@linaro.org \
--cc=linux-arm-kernel@lists.infradead.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).