I confirmed your modification works fine on 2.6.32.19.el6.x86_64. I applied
your modification for kernel/kgdb.c instead of kernel/debug/gdbstub.c
because of 2.6.32.19.el6.x86_64.
Thanks.
-caz
-----Original Message-----
From: Caz Yokoyama [mailto:caz@xxxxxxxxxxx]
Sent: Sunday, August 22, 2010 8:49 PM
To: 'DDD'
Cc: 'Jason Wessel'; 'KGDB Mailing List'; 'kdb@xxxxxxxxxxx'; 'Matt W.
Benjamin'
Subject: RE: [Kgdb-bugreport] kdb working, but kgdb not? 2.6.35-rc3
Hello Dongdong and Jason,
I'll confirm your modification on my target which has 2.6.32.19.el6.x86_64
kernel. Thank you.
-caz
-----Original Message-----
From: DDD [mailto:dongdong.deng@xxxxxxxxxxxxx]
Sent: Sunday, August 22, 2010 8:12 PM
To: Caz Yokoyama
Cc: 'Jason Wessel'; 'KGDB Mailing List'; kdb@xxxxxxxxxxx; 'Matt W. Benjamin'
Subject: Re: [Kgdb-bugreport] kdb working, but kgdb not? 2.6.35-rc3
Hi Caz,
Thanks for your info. Jason have committed a patch to Linux 2.6.35-rc6
for this problem.
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdif
f;h=fb82c0ff27b2c40c6f7a3d1a94cafb154591fa80;hp=1396a21ba0d4ec381db19bc9cd5b
6f25a89cf633
Cheers,
Dongdong
Caz Yokoyama wrote:
> Hello Jason,
>
> Could you return '\0' for "qTStatus"? It seems gdb asks gdbserver trace
> point status on break-in. qTStatus is one of them. I believe kgdb stub
> returns 'E' for "qTStatus". I am asking you to return '\0' instead of 'E'.
> If kgdb stub returns '\0', gdb works fine with Linux kgdb without
> modification on gdb at least break-in, continue and quit.
>
> Thanks.
>
> -caz
>
> _____
>
> From: caz yokoyama [mailto:cazyokoyama@xxxxxxxxx]
> Sent: Monday, June 28, 2010 9:02 AM
> To: Jason Wessel
> Cc: Matt W. Benjamin; KGDB Mailing List; kdb@xxxxxxxxxxx
> Subject: Re: [Kgdb-bugreport] kdb working, but kgdb not? 2.6.35-rc3
>
>
>
> Yes. As you find in the patch, there are 2 commands are required to
support.
> I don't know if they can be implemented on kgdb. I need investigation.
I'll
> let you know when I find.
>
> -caz
>
> On Mon, Jun 28, 2010 at 5:36 PM, Jason Wessel <jason.wessel@xxxxxxxxxxxxx>
> wrote:
>
> On 06/28/2010 10:33 AM, caz yokoyama wrote:
>> I recognized this problem and have a patch. It was caused by introduction
> of
>> new commands between gdb and gdbserver which Linux kernel does not
> support.
>> Here is the patch.
>>
>>
>
> Hi Caz,
>
> Is it the case that the kgdb stub should be responding with some sort of
> message?
>
> Certainly we are always open to improving the kgdb side too.
>
> Cheers,
> Jason.
>
>
>
>
|