From: David Daney <ddaney@caviumnetworks.com>
To: Aaro Koskinen <aaro.koskinen@iki.fi>
Cc: Ralf Baechle <ralf@linux-mips.org>,
Aaro Koskinen <aaro.koskinen@nokia.com>,
Joshua Kinard <kumba@gentoo.org>, <linux-mips@linux-mips.org>,
"Hill, Steven" <Steven.Hill@cavium.com>
Subject: Re: THP broken on OCTEON?
Date: Mon, 23 May 2016 12:03:34 -0700 [thread overview]
Message-ID: <57435406.1060104@caviumnetworks.com> (raw)
In-Reply-To: <20160523185226.GA1253@raspberrypi.musicnaut.iki.fi>
On 05/23/2016 11:52 AM, Aaro Koskinen wrote:
> On Mon, May 23, 2016 at 09:21:22AM -0700, David Daney wrote:
>> On 05/23/2016 08:20 AM, Ralf Baechle wrote:
>>> On Mon, May 23, 2016 at 06:13:46PM +0300, Aaro Koskinen wrote:
>>>> I'm getting kernel crashes (see below) reliably when building Perl in
>>>> parallel (make -j16) on OCTEON EBH5600 board (8 cores, 4 GB RAM) with
>>>> Linux 4.6.
>>>>
>>>> It seems that CONFIG_TRANSPARENT_HUGEPAGE has something to do with the
>>>> issue - disabling it makes build go through fine.
>>>>
>>>> Any ideas?
>>>
>>> I thought it was working except on SGI Origin 200/2000 aka IP27 where
>>> Joshua Kinard (added to cc) was hitting issues as well.
>>>
>>> Joshua, does that similar to the issues you were hitting?
>>
>> There is nothing OCTEON specific in the THP code, or huge pages in general.
>>
>> That said, we have seen other THP related failures, and have never been able
>> to find the cause.
>>
>> If someone can come up with a reproducible test case that triggers quickly,
>> we can run it in our simulator and easily find the problem.
>
> Trying to build Perl is a reliable reproducer. Is that too heavyweight
> for your simulator?
>
> I was able to reproduce this also on EdgeRouter Pro, but there the kernel
> does not fail, only compiler dies with SIGBUS:
>
> [ 315.095264] Data bus error, epc == 0000000000a801c4, ra == 0000000000a80624
>
> And without THP the build is fine.
>
> I also tried CN68XX board with 16 GB RAM and also there I get SIGBUS failure
> instead of Machine Check.
>
Yes. I think the problem is some sort of corruption of the page tables.
This may show up as MachineCheck Errors, or bus errors, or SIGSEGV.
David.
WARNING: multiple messages have this Message-ID (diff)
From: David Daney <ddaney@caviumnetworks.com>
To: Aaro Koskinen <aaro.koskinen@iki.fi>
Cc: Ralf Baechle <ralf@linux-mips.org>,
Aaro Koskinen <aaro.koskinen@nokia.com>,
Joshua Kinard <kumba@gentoo.org>,
linux-mips@linux-mips.org, "Hill,
Steven" <Steven.Hill@cavium.com>
Subject: Re: THP broken on OCTEON?
Date: Mon, 23 May 2016 12:03:34 -0700 [thread overview]
Message-ID: <57435406.1060104@caviumnetworks.com> (raw)
Message-ID: <20160523190334.szE3C_KTfTarrgX7-T0GgCZxtAmFOHSiBbDREvgS0Go@z> (raw)
In-Reply-To: <20160523185226.GA1253@raspberrypi.musicnaut.iki.fi>
On 05/23/2016 11:52 AM, Aaro Koskinen wrote:
> On Mon, May 23, 2016 at 09:21:22AM -0700, David Daney wrote:
>> On 05/23/2016 08:20 AM, Ralf Baechle wrote:
>>> On Mon, May 23, 2016 at 06:13:46PM +0300, Aaro Koskinen wrote:
>>>> I'm getting kernel crashes (see below) reliably when building Perl in
>>>> parallel (make -j16) on OCTEON EBH5600 board (8 cores, 4 GB RAM) with
>>>> Linux 4.6.
>>>>
>>>> It seems that CONFIG_TRANSPARENT_HUGEPAGE has something to do with the
>>>> issue - disabling it makes build go through fine.
>>>>
>>>> Any ideas?
>>>
>>> I thought it was working except on SGI Origin 200/2000 aka IP27 where
>>> Joshua Kinard (added to cc) was hitting issues as well.
>>>
>>> Joshua, does that similar to the issues you were hitting?
>>
>> There is nothing OCTEON specific in the THP code, or huge pages in general.
>>
>> That said, we have seen other THP related failures, and have never been able
>> to find the cause.
>>
>> If someone can come up with a reproducible test case that triggers quickly,
>> we can run it in our simulator and easily find the problem.
>
> Trying to build Perl is a reliable reproducer. Is that too heavyweight
> for your simulator?
>
> I was able to reproduce this also on EdgeRouter Pro, but there the kernel
> does not fail, only compiler dies with SIGBUS:
>
> [ 315.095264] Data bus error, epc == 0000000000a801c4, ra == 0000000000a80624
>
> And without THP the build is fine.
>
> I also tried CN68XX board with 16 GB RAM and also there I get SIGBUS failure
> instead of Machine Check.
>
Yes. I think the problem is some sort of corruption of the page tables.
This may show up as MachineCheck Errors, or bus errors, or SIGSEGV.
David.
next prev parent reply other threads:[~2016-05-23 19:03 UTC|newest]
Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-05-23 15:13 THP broken on OCTEON? Aaro Koskinen
2016-05-23 15:20 ` Ralf Baechle
2016-05-23 16:21 ` David Daney
2016-05-23 18:52 ` Aaro Koskinen
2016-05-23 19:03 ` David Daney [this message]
2016-05-23 19:03 ` David Daney
2016-05-23 19:08 ` Joshua Kinard
2016-05-23 20:02 ` Alastair Bridgewater
2016-05-23 18:57 ` Joshua Kinard
2016-05-23 19:22 ` Ralf Baechle
2016-05-23 19:40 ` Joshua Kinard
2016-05-23 20:01 ` Ralf Baechle
2016-05-24 21:21 ` Aaro Koskinen
2016-05-24 22:39 ` David Daney
2016-05-25 13:41 ` Aaro Koskinen
2016-05-26 9:33 ` Joshua Kinard
2016-05-26 13:36 ` Aaro Koskinen
2016-05-26 17:59 ` David Daney
2016-05-26 19:23 ` Aaro Koskinen
2016-05-26 22:13 ` David Daney
2016-05-27 17:14 ` Aaro Koskinen
2016-05-27 21:03 ` Joshua Kinard
2016-05-27 22:05 ` Aaro Koskinen
2016-05-27 22:22 ` Joshua Kinard
2016-06-22 22:05 ` David Daney
2016-06-23 12:08 ` Aaro Koskinen
2016-06-23 12:08 ` Aaro Koskinen
2016-06-24 11:38 ` Joshua Kinard
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=57435406.1060104@caviumnetworks.com \
--to=ddaney@caviumnetworks.com \
--cc=Steven.Hill@cavium.com \
--cc=aaro.koskinen@iki.fi \
--cc=aaro.koskinen@nokia.com \
--cc=kumba@gentoo.org \
--cc=linux-mips@linux-mips.org \
--cc=ralf@linux-mips.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