From: Gregory CLEMENT <gregory.clement@bootlin.com>
To: linux-mips@vger.kernel.org
Cc: Alexandre Belloni <alexandre.belloni@bootlin.com>
Subject: [BUG] Crash during futex initialization with gcc 10
Date: Wed, 28 Oct 2020 11:34:50 +0100 [thread overview]
Message-ID: <878sbqd4fp.fsf@BL-laptop> (raw)
Hello,
I saw a regression when building a MIPS kernel with gcc 10:
futex hash table entries: 256 (order: 3, 32768 bytes, linear)
CPU 0 Unable to handle kernel paging request at virtual address 00000000, epc == 801be4f0, ra == 80993778
Oops[#1]:
CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.9.0-rc8-00006-g6185221cb805 #3
$ 0 : 00000000 00000001 00000000 00000000
$ 4 : 87831e28 00000000 00000000 00000000
$ 8 : 8782c000 87888000 00000100 00000051
$12 : 09dcd4c5 87831c69 00000000 00000010
$16 : 80960000 00000001 80960000 8095bc90
$20 : 809d0000 809b3fc4 80987094 00000002
$24 : 00000000 00000000
$28 : 87830000 87831df8 809b3fe4 80993778
Hi : 00000010
Lo : 8a3d70e6
epc : 801be4f0 cmpxchg_futex_value_locked+0x2c/0x78
ra : 80993778 futex_init+0xb4/0x124
Status: 11008403 KERNEL EXL IE
Cause : 00800008 (ExcCode 02)
BadVA : 00000000
PrId : 02019654 (MIPS 24KEc)
Modules linked in:
Process swapper/0 (pid: 1, threadinfo=(ptrval), task=(ptrval), tls=00000000)
Stack : ffe4d6e0 001b291f 80963e60 80964fe0 00000000 87831e24 00000000 00000100
00000100 80a00000 00000000 00000008 809b3fe4 0a27934f 87878880 809936c4
87878880 80960000 00000000 80100120 00000000 00000000 00000000 87831e54
87831e54 87878895 8094cecc 80900000 00000000 00000001 80960000 00000001
809629c4 80908380 808ff7ec 80900000 0000007f 87878895 00000000 0a27934f
...
Call Trace:
[<801be4f0>] cmpxchg_futex_value_locked+0x2c/0x78
Code: 1440000f 00000000 0000000f <c0a30000> 14660005 00000000 00e00825 e0a10000 1020fff9
---[ end trace e34236d22d7ecc0e ]---
Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
---[ end Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b ]---
If I use gcc 9.3 or if I disable the futex in the kernel configuration
there is no crash anymore.
The crash occurs after that the pagefault was disabled and yet we see a
page fault. The exact part where it occurs is here:
https://elixir.bootlin.com/linux/latest/source/arch/mips/include/asm/futex.h#L167
But it is too much MIPS assembly for me !
Does someone has an idea to fix it ?
It is alos possible that the issue was in gcc itself, but we need to
figure out what in order to report it to the gcc guys.
Gregory
--
Gregory Clement, Bootlin
Embedded Linux and Kernel engineering
http://bootlin.com
next reply other threads:[~2020-10-29 1:21 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-10-28 10:34 Gregory CLEMENT [this message]
2020-11-01 11:12 ` [BUG] Crash during futex initialization with gcc 10 Maciej W. Rozycki
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=878sbqd4fp.fsf@BL-laptop \
--to=gregory.clement@bootlin.com \
--cc=alexandre.belloni@bootlin.com \
--cc=linux-mips@vger.kernel.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