Latest linux pull few days ago broke the console.
The error during boot up while switching from init S to M showed
INIT: open(/dev/console): Input/output error
Reverting this commit solves the problem. This is 100% repeatable on a
"real" machine, but does not show up on kvm or vmware guest console.
commit b50989dc444599c8b21edc23536fc305f4e9b7d5
Author: Alan Cox <[email protected]>
Date: Sat Sep 19 13:13:22 2009 -0700
tty: make the kref destructor occur asynchronously
We want to be able to sleep in the destructor for USB at least. It isn't a
hot path so just pushing it to a work queue doesn't really cause any
difficulty.
Signed-off-by: Alan Cox <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>
Thanks,
Jeff
On Thu, 24 Sep 2009 21:24:45 +0800
Jeff Chua <[email protected]> wrote:
> Latest linux pull few days ago broke the console.
>
> The error during boot up while switching from init S to M showed
>
> INIT: open(/dev/console): Input/output error
>
>
> Reverting this commit solves the problem. This is 100% repeatable on a
> "real" machine, but does not show up on kvm or vmware guest console.
It is timing dependant. A discussion and some fixes were posted a while
ago.
On Thu, Sep 24, 2009 at 11:01 PM, Alan Cox <[email protected]> wrote:
> It is timing dependant. A discussion and some fixes were posted a while
> ago.
I must have missed that. Will try to search for it.
Thanks,
Jeff
On Fri, 25 Sep 2009 00:23:27 +0800
"jeff.chua.linux" <[email protected]> wrote:
> On Thu, Sep 24, 2009 at 11:01 PM, Alan Cox <[email protected]>
> wrote:
> > It is timing dependant. A discussion and some fixes were posted a
> > while ago.
>
> I must have missed that. Will try to search for it.
>
> Thanks,
> Jeff
>From Andrew's tree:
tty: fix regression caused by "tty: make the kref destructor occur
asynchronously"
which allows both a sync and an async element to the destruction
cleanup. The console uses the synchronous one which avoids the delay
before it becomes available again.
On Fri, Sep 25, 2009 at 6:56 PM, Alan Cox <[email protected]> wrote:
> From Andrew's tree:
>
> ? ? tty: fix regression caused by "tty: make the kref destructor occur
> ? ? asynchronously"
>
> which allows both a sync and an async element to the destruction
> cleanup. The console uses the synchronous one which avoids the delay
> before it becomes available again.
Alan,
Is this going into Linus's branch soon? Otherwise, please kindly post
the full patch so that I can test it instead of pulling 400MB from
Andrew's tree.
Thanks,
Jeff
> Is this going into Linus's branch soon? Otherwise, please kindly post
> the full patch so that I can test it instead of pulling 400MB from
> Andrew's tree.
I don't have Andrews tty patches here either sorry but I thought he kept
them broken out on kernel.org ?
Alan Cox wrote:
>> Is this going into Linus's branch soon? Otherwise, please kindly post
>> the full patch so that I can test it instead of pulling 400MB from
>> Andrew's tree.
>
> I don't have Andrews tty patches here either sorry but I thought he kept
> them broken out on kernel.org ?
> --
http://userweb.kernel.org/~akpm/mmotm/
see the broken-out/ subdir.
~Randy
On Fri, Sep 25, 2009 at 11:52 PM, Randy Dunlap <[email protected]> wrote:
> Alan Cox wrote:
>>> Is this going into Linus's branch soon? Otherwise, please kindly post
>>> the full patch so that I can test it instead of pulling 400MB from
>>> Andrew's tree.
>>
>> I don't have Andrews tty patches here either sorry but I thought he kept
>> them broken out on kernel.org ?
>> --
>
> http://userweb.kernel.org/~akpm/mmotm/
> see the broken-out/ subdir.
I see. But that patch is not there. Did find it on google.
http://www.spinics.net/lists/mm-commits/msg71772.html
So, I'll test it now.
Thanks,
Jeff
On Sat, Sep 26, 2009 at 12:00 AM, Jeff Chua <[email protected]> wrote:
> On Fri, Sep 25, 2009 at 11:52 PM, Randy Dunlap <[email protected]> wrote:
>> Alan Cox wrote:
>>>> Is this going into Linus's branch soon? Otherwise, please kindly post
>>>> the full patch so that I can test it instead of pulling 400MB from
>>>> Andrew's tree.
>>> I don't have Andrews tty patches here either sorry but I thought he kept
>>> them broken out on kernel.org ?
>> http://userweb.kernel.org/~akpm/mmotm/
>> see the broken-out/ subdir.
>
> I see. But that patch is not there. Did find it on google.
>
> http://www.spinics.net/lists/mm-commits/msg71772.html
>
> So, I'll test it now.
Alan, Randy,
That patch works!
Thanks,
Jeff
On Sat, 26 Sep 2009, Jeff Chua wrote:
> On Sat, Sep 26, 2009 at 12:00 AM, Jeff Chua <[email protected]> wrote:
>> On Fri, Sep 25, 2009 at 11:52 PM, Randy Dunlap <[email protected]> wrote:
>>> Alan Cox wrote:
>>>>> Is this going into Linus's branch soon? Otherwise, please kindly post
>>>>> the full patch so that I can test it instead of pulling 400MB from
>>>>> Andrew's tree.
>>>> I don't have Andrews tty patches here either sorry but I thought he kept
>>>> them broken out on kernel.org ?
>>> http://userweb.kernel.org/~akpm/mmotm/
>>> see the broken-out/ subdir.
>>
>> I see. But that patch is not there. Did find it on google.
>>
>> http://www.spinics.net/lists/mm-commits/msg71772.html
>>
>> So, I'll test it now.
>
> Alan, Randy,
>
> That patch works!
>
> Thanks,
> Jeff
Here, it doesn't fix
http://marc.info/?l=linux-kernel&m=125382651922142&w=2
But I found something in syslog:
Sep 25 16:17:31 pervalidus agetty[1465]: /dev/tty2: No such file or directory
Sep 25 16:17:31 pervalidus agetty[1464]: /dev/tty1: No such file or directory
Sep 25 16:17:31 pervalidus agetty[1466]: /dev/tty3: No such file or directory
...
So, why it works with 2.6.30.7 and not 2.6.30.1 ?
And I also noticed that it takes longer with 2.6.30.7 than
before for the login and password prompts to appear.
In /etc/inittab, I have
c1:1235:respawn:/sbin/agetty 38400 tty1 linux
c2:1235:respawn:/sbin/agetty 38400 tty2 linux
c3:1235:respawn:/sbin/agetty 38400 tty3 linux
c4:1235:respawn:/sbin/agetty 38400 vc/4 linux (yes, nothing
about vc/4 in the logs)
So, with 2.6.30, udev doesn't create the symlinks from vc/*, or
it's something in the kernel that prevents it ?