On 03/08/2013 11:14 AM, richard -rw- weinberger wrote:
> On Thu, Mar 7, 2013 at 6:28 PM, Toralf Förster wrote:
>> On 03/07/2013 05:34 PM, richard -rw- weinberger wrote:
>>> So, xterm is broken in 3.9-rc1?
>>> Please bisect it.
>>>
>> Cannot be reproduced under host kernel 3.7.10, and host kernel 3
On Thu, Mar 7, 2013 at 6:28 PM, Toralf Förster wrote:
> 2013-03-07T18:23:59.185+01:00 n22stab4 kernel: WARNING: at
> drivers/tty/tty_buffer.c:428 flush_to_ldisc+0x51/0x190()
> 2013-03-07T18:23:59.185+01:00 n22stab4 kernel: tty is NULL
> 2013-03-07T18:23:59.185+01:00 n22stab4 kernel: 3979fea0: [<
On 03/07/2013 12:00 AM, richard -rw- weinberger wrote:
> Found the issue. A very nasty use-after-free tty issue which happened
> only sometimes.
> Please test the attached patch.
> /me tested x86_64 and x86 successfully.
>
Today I tested a patched 3.9 kernel for the UML guest several times,
host k
On 03/08/2013 11:14 AM, richard -rw- weinberger wrote:
> On Thu, Mar 7, 2013 at 6:28 PM, Toralf Förster wrote:
>> On 03/07/2013 05:34 PM, richard -rw- weinberger wrote:
>>> So, xterm is broken in 3.9-rc1?
>>> Please bisect it.
>>>
>> Cannot be reproduced under host kernel 3.7.10, and host kernel 3
On Thu, Mar 7, 2013 at 6:28 PM, Toralf Förster wrote:
> On 03/07/2013 05:34 PM, richard -rw- weinberger wrote:
>> So, xterm is broken in 3.9-rc1?
>> Please bisect it.
>>
> Cannot be reproduced under host kernel 3.7.10, and host kernel 3.8.2 suffers
> from a complete hang if an (unprivileged) user
On 03/07/2013 05:34 PM, richard -rw- weinberger wrote:
> So, xterm is broken in 3.9-rc1?
> Please bisect it.
>
Cannot be reproduced under host kernel 3.7.10, and host kernel 3.8.2 suffers
from a complete hang if an (unprivileged) user starts UML and just wait for
a while (already reported to this
On Thu, Mar 7, 2013 at 5:11 PM, Toralf Förster wrote:
>> issue is fixed here too - thx.
>
> FWIW something like "con=pts con0=fd:0,fd:1" is now mandatory,
> otherwise an xterm is open and user root can login, but I experienced
> strange things with my Gentoo :
So, xterm is broken in 3.9-rc1?
Plea
On 03/07/2013 12:00 AM, richard -rw- weinberger wrote:
> On Wed, Mar 6, 2013 at 7:24 PM, richard -rw- weinberger
> wrote:
>> On Wed, Mar 6, 2013 at 7:10 PM, Toralf Förster
>> wrote:
>>> On 03/06/2013 12:32 AM, richard -rw- weinberger wrote:
What is your console? I suspect it is not stdio_co
On Wed, Mar 6, 2013 at 7:24 PM, richard -rw- weinberger
wrote:
> On Wed, Mar 6, 2013 at 7:10 PM, Toralf Förster wrote:
>> On 03/06/2013 12:32 AM, richard -rw- weinberger wrote:
>>> What is your console? I suspect it is not stdio_console.
>>
>> I tried both with "con=pts con0=fd:0,fd:1" as well as
On Wed, Mar 6, 2013 at 7:10 PM, Toralf Förster wrote:
> On 03/06/2013 12:32 AM, richard -rw- weinberger wrote:
>> What is your console? I suspect it is not stdio_console.
>
> I tried both with "con=pts con0=fd:0,fd:1" as well as without specifying
> console/tty at the command line - no difference.
On 03/06/2013 12:32 AM, richard -rw- weinberger wrote:
> What is your console? I suspect it is not stdio_console.
I tried both with "con=pts con0=fd:0,fd:1" as well as without specifying
console/tty at the command line - no difference.
But for completeness I attached the .config I used
--
MfG/S
On Tue, Mar 5, 2013 at 6:57 PM, Toralf Förster wrote:
> On 03/05/2013 06:30 PM, richard -rw- weinberger wrote:
>> On Tue, Mar 5, 2013 at 4:33 PM, Toralf Förster
>> wrote:
>>> On 03/05/2013 01:51 AM, Pierluigi Rolando wrote:
I think what's going on
here is that UML gets a signal it can
On 03/05/2013 06:30 PM, richard -rw- weinberger wrote:
> On Tue, Mar 5, 2013 at 4:33 PM, Toralf Förster wrote:
>> On 03/05/2013 01:51 AM, Pierluigi Rolando wrote:
>>> I think what's going on
>>> here is that UML gets a signal it can't handle when you resize.
>>
>> Yes - something _is_ wrong with
On Tue, Mar 5, 2013 at 4:33 PM, Toralf Förster wrote:
> On 03/05/2013 01:51 AM, Pierluigi Rolando wrote:
>> I think what's going on
>> here is that UML gets a signal it can't handle when you resize.
>
> Yes - something _is_ wrong with the signal handling, but my command line
> looks not suspicou
On 03/05/2013 01:51 AM, Pierluigi Rolando wrote:
> I think what's going on
> here is that UML gets a signal it can't handle when you resize.
Yes - something _is_ wrong with the signal handling, but my command line looks
not suspicous, or ? :
/usr/local/bin/linux-v3.8.2 earlyprintk
ubda=/home/tf
Hi,
UML user here. I'm able to reproduce a similar crash 100% of the time
with a 3.7 UML kernel if I configure my UML consoles some specific
way. I don't remember all the details of it. I think what's going on
here is that UML gets a signal it can't handle when you resize. Not
attaching to a pts (o
16 matches
Mail list logo
