8:14 a.m.
This update fixes a bug and improves register restoration for active
tasks. See the ChangeLog below.
Note that there's no support range change on kernel versions and
architectures.
Thanks Dave for your reviewing.
ChangeLog:
- Fix a bug that for a multi-thread tasks, register values for all the
tasks are equal to the register values of the task specified by user
in command-line.
(d.hatayama(a)jp.fujitsu.com)
- Retrieve register values in exception frame on kernel stack that is
leaved by KDUMP, NETDUMP or DISKDUMP mechanism. Also, use register
values included in note information of KVMDUMP format.
(d.hatayama(a)jp.fujitsu.com)
Thanks.
HATAYAMA Daisuke
6:09 a.m.
On Thu, Feb 10, 2011 at 1:44 AM, HATAYAMA Daisuke <d.hatayama(a)jp.fujitsu.com
wrote:
This update fixes a bug and improves register restoration for active
tasks. See the ChangeLog below.
Note that there's no support range change on kernel versions and
architectures.
[snip]
Let me just send a shout out, and say thanks for gcore! For reasons I may
never know (human error? communication?) a developer (vendor) asked me for a
crash dump, when he really wanted a kill -6 of a user process. If you
haven't kept it in mind already, I am looking forward to writing out a new
core file of just the user process. Not a high priority, but something that
would be grand for shipping off the user core when the crash dump is not
needed in whole.
Thanks again!
Scott Edwards
-- Daxal Communications - http://www.daxal.com/?from=crash+mailinglist
8:04 a.m.
----- Original Message -----
Let me just send a shout out, and say thanks for gcore! For reasons
I
may never know (human error? communication?) a developer (vendor)
asked me for a crash dump, when he really wanted a kill -6 of a user
process. If you haven't kept it in mind already, I am looking forward
to writing out a new core file of just the user process. Not a high
priority, but something that would be grand for shipping off the user
core when the crash dump is not needed in whole.
Thanks again!
Scott Edwards
-- Daxal Communications - http://www.daxal.com/?from=crash+mailinglist
Hi Scott,
Help me out -- doesn't gcore do exactly what you're asking for?
Dave
8:33 a.m.
Hi Scott,
Help me out -- doesn't gcore do exactly what you're asking for?
Dave
Why it appears so. I'm pretty sure I did not have access to gcore before. I
still have a ~6mo old 16GB crash dump tucked away somewhere, and want to
retry analysis after I find it. It's probably a short circuit. I've been
fuzzy ever since I left work yesterday. Car was stuck twice on icy+snowy
hills, took three hours for a normal 40 minute commute, and I'm sore all
over. :) I'm doing fine now. But still a big thank you for someone reaching
for this fruit that's a little too high up for me.
Regards,
Scott Edwards
-- Daxal Communications - http://www.daxal.com/?from=crash+mailinglist
7:36 a.m.
The gcore files on http://people.redhat.com/anderson/extensions.html
have been updated.
Thanks,
Dave
----- Original Message -----
This update fixes a bug and improves register restoration for active
tasks. See the ChangeLog below.
Note that there's no support range change on kernel versions and
architectures.
Thanks Dave for your reviewing.
ChangeLog:
- Fix a bug that for a multi-thread tasks, register values for all the
tasks are equal to the register values of the task specified by user
in command-line.
(d.hatayama(a)jp.fujitsu.com)
- Retrieve register values in exception frame on kernel stack that is
leaved by KDUMP, NETDUMP or DISKDUMP mechanism. Also, use register
values included in note information of KVMDUMP format.
(d.hatayama(a)jp.fujitsu.com)
Thanks.
HATAYAMA Daisuke
5145
days inactive
486
days old
devel@lists.crash-utility.osci.io
4 comments
3 participants
participants (3)
-
Dave Anderson -
HATAYAMA Daisuke -
Scott Edwards