* RE: A clash happened in kernel by glibc-2.33
[not found] ` <87669f23-c5f0-e742-5dc1-638f5bc89bec@redhat.com>
@ 2021-07-01 1:49 ` Zeng, MingwangX
2021-07-01 6:06 ` Florian Weimer
0 siblings, 1 reply; 2+ messages in thread
From: Zeng, MingwangX @ 2021-07-01 1:49 UTC (permalink / raw)
To: libc-alpha
Hi, alpha
I tried kernel versions 4.9.257, 4.9.258, and they all crashed.
I don't use the option --enable-kernel, as we update kernel more frequently than glibc .
Is this option --enable-kernel=4.9.255 compatible with versions after 4.9.255?
BR,
Mingwang
-----Original Message-----
From: Carlos O'Donell <carlos@redhat.com>
Sent: Wednesday, June 30, 2021 10:29 PM
To: Zeng, MingwangX <mingwangx.zeng@intel.com>; libc-maintainers@gnu.org
Subject: Re: A clash happened in kernel by glibc-2.33
On 6/30/21 2:29 AM, Zeng, MingwangX wrote:
> Hi, Glibc maintainers
Zeng,
Please send this to libc-alpha@sourceware.org, and we'll have a development discussion there?
Please provide the exact value of --enable-kernel that was used to trigger this defect and why you selected that given kernel version.
No default selection will trigger this because we set the Z component of the version to 0 in the defaults.
> A clash happened in kernel by glibc-2.33.
>
> Scene:
> when kernel mini version lager then 255, glibc could cause the clash.
>
> FATAL: kernel too old
> Kernel panic - not syncing: Attempted to kill init!
> exitcode=0x00007f00
>
> Reason:
> sysdeps/unix/sysv/linux/dl-sysdep.c
> line:
> 110 while ((*cp >= '0') && (*cp <= '9'))
> 111 {
> 112 unsigned int here = *cp++ - '0';
> 113
> 114 while ((*cp >= '0') && (*cp <= '9'))
> 115 {
> 116 here *= 10;
> 117 here += *cp++ - '0';
> 118 }
> 119
> 120 ++parts;
> 121 version <<= 8;
> 122 version | = here; // root case.
>
> Such as Kernel version is 4.9.256 , it will be : version |= 256 ;
> Version == 4.9.0
>
> Would you please provide me with a temporary patch, think you very much!
>
> BR,
> Mingwang.
>
>
>
--
Cheers,
Carlos.
^ permalink raw reply [flat|nested] 2+ messages in thread