[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20180420123852.ox5xcmwgj4krzapf@wfg-t540p.sh.intel.com>
Date: Fri, 20 Apr 2018 20:38:52 +0800
From: Fengguang Wu <fengguang.wu@...el.com>
To: Peter Oberparleiter <oberpar@...ux.ibm.com>
Cc: Arnd Bergmann <arnd@...db.de>,
Riku Voipio <riku.voipio@...aro.org>,
Frantisek Hrbata <fhrbata@...hat.com>,
Masahiro Yamada <yamada.m@...panasonic.com>,
linux-kernel@...r.kernel.org, lkp@...org
Subject: Re: [gcov_module_notifier] WARNING: CPU: 0 PID: 155 at
mm/slab_common.c:996 kmalloc_slab+0x1f/0x79
Hi Peter,
On Fri, Apr 20, 2018 at 01:08:07PM +0200, Peter Oberparleiter wrote:
>On 19.04.2018 04:58, Fengguang Wu wrote:
>> FYI this happens in mainline kernel 4.17.0-rc1.
>> It at least dates back to v4.5 .
>
>This is likely the result of compiling the kernel with GCC 7 while
>specifying that gcov-kernel should expect GCC <= 3.4 format data:
>
>dmesg:
>> (gcc version 7.3.0 (Debian 7.3.0-1)) #31 Mon Apr 16 23:17:50 CST 2018
>
>.config:
>> # CONFIG_GCOV_FORMAT_AUTODETECT is not set
>> CONFIG_GCOV_FORMAT_3_4=y
>> # CONFIG_GCOV_FORMAT_4_7 is not set
>
>As such, it is an expected test failure. Assuming that this is a
>randconfig test, my recommendation would be to blacklist
>"CONFIG_GCOV_FORMAT_3_4=y" when using GCC >3.4.
Ah OK, thanks! It looks easy to fix for us. Interestingly there's
the default GCOV_FORMAT_AUTODETECT option, however the randconfig
test obviously (blindly) will try other options if ever they are
selectable.
Thanks,
Fengguang
Powered by blists - more mailing lists