All of lore.kernel.org
 help / color / mirror / Atom feed
* crashdump on PVM Dom0
@ 2017-08-28  8:01 Minjun Hong
  2017-08-28  8:13 ` Jan Beulich
                   ` (2 more replies)
  0 siblings, 3 replies; 7+ messages in thread
From: Minjun Hong @ 2017-08-28  8:01 UTC (permalink / raw)
  To: xen-devel


[-- Attachment #1.1: Type: text/plain, Size: 1461 bytes --]

Hello~
I'm new to develop Xen and my first work is about its scheduler.
After I modified Xen 4.5 source code, my custom Xen has had a problem,
which it causes automatic reboot of Dom0 when I execute a micro benchmark
program.
To debug it, I have tried to use crashdump on Dom0 with the Xen but, I
failed to enable it.
My configurations are following:

1. Base Xen version: 4.5 release
2. OS: Ubuntu 14.04 LTS
3. VM type: PVM

4. Crashdump install:
- install linux-crashdump
- install kexec from source code with '--with-xen' option
- modified '/etc/default/kdump-tools':

> USE_KDUMP=1
>
- kernel boot param setting(/etc/defualt/grub):

>
> GRUB_CMDLINE_LINUX="crashkernel=2G-4G:320M,4G-32G:512M,32G-64G:1024M,64G-128G:2048M,128G-:4096M@32M
> "
> # update-grub2
> # reboot
>

5. Problem:
- current state is "not ready to kdump" when I type "kdump-config show":

> DUMP_MODE:        kdump
> USE_KDUMP:        1
> KDUMP_SYSCTL:     kernel.panic_on_oops=1
> KDUMP_COREDIR:    /var/crash
> crashkernel addr: 0x30000000
> current state:    Not ready to kdump
>

I have done my best to resolve this problem, but I don't have any other way
that I could try anymore.
If you have a hint which helps me use crashdump on PVM Dom0 or any other
way to debug customized Xen that it causes automatic reboot of Dom0 due to
wrong code modification(without any log),
please let me know.
Even if it is very minor thing, I'll really appreciate it.

Thanks to read this.
Sincerely,

[-- Attachment #1.2: Type: text/html, Size: 2667 bytes --]

[-- Attachment #2: Type: text/plain, Size: 127 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: crashdump on PVM Dom0
  2017-08-28  8:01 crashdump on PVM Dom0 Minjun Hong
@ 2017-08-28  8:13 ` Jan Beulich
       [not found]   ` <CAFX_q-HHBCb6KOknVXhu5CirfEckkOS0SW2ZLjoaEcpswwy5Og@mail.gmail.com>
  2017-08-29  9:19 ` Wei Liu
  2017-08-29 10:05 ` Roger Pau Monné
  2 siblings, 1 reply; 7+ messages in thread
From: Jan Beulich @ 2017-08-28  8:13 UTC (permalink / raw)
  To: Minjun Hong; +Cc: xen-devel

>>> On 28.08.17 at 10:01, <nickeysgo@gmail.com> wrote:
> I'm new to develop Xen and my first work is about its scheduler.
> After I modified Xen 4.5 source code, my custom Xen has had a problem,
> which it causes automatic reboot of Dom0 when I execute a micro benchmark
> program.
> To debug it, I have tried to use crashdump on Dom0 with the Xen but, I
> failed to enable it.
> My configurations are following:
> 
> 1. Base Xen version: 4.5 release
> 2. OS: Ubuntu 14.04 LTS
> 3. VM type: PVM
> 
> 4. Crashdump install:
> - install linux-crashdump
> - install kexec from source code with '--with-xen' option
> - modified '/etc/default/kdump-tools':
> 
>> USE_KDUMP=1
>>
> - kernel boot param setting(/etc/defualt/grub):
> 
>>
>> 
> GRUB_CMDLINE_LINUX="crashkernel=2G-4G:320M,4G-32G:512M,32G-64G:1024M,64G-128G
> :2048M,128G-:4096M@32M

It is Xen to control the reserved memory, so "crashkernel=" needs
to be put on the Xen command line (suitably modified to that Xen
actually understands it - see the command line option documentation).

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: crashdump on PVM Dom0
       [not found]     ` <59A3FAF202000078001746B4@prv-mh.provo.novell.com>
@ 2017-08-28 10:45       ` Minjun Hong
  2017-08-28 11:13         ` Jan Beulich
  0 siblings, 1 reply; 7+ messages in thread
From: Minjun Hong @ 2017-08-28 10:45 UTC (permalink / raw)
  To: Jan Beulich; +Cc: xen-devel


[-- Attachment #1.1: Type: text/plain, Size: 1350 bytes --]

Oh!! I'm so sorry Jan Beulich.
It is my first time to use this mailing list.
Please understand my fault, this time only.

Anyway, as you mentioned, I added below line in '/etc/default/grub' file
and reboot:
> GRUB_CMDLINE_XEN="crashkernel=384M-:256M@64M"
Then,
1. I cannot see anything when I type "grep -i crash /proc/iomem"
2. I can see kernel log of Xen hypervisor when I type "sudo xl dmesg | grep
-i crash":
(XEN) Command line: placeholder crashkernel=384M-:256M@64M

I suppose a 'crashkernel' cmdline option of either Xen hypervisor or Dom0
kernel can be enabled.
I know it is quite difficult to advise some problems online but, could you
give me a piece of hint?

Sincerely,

On Mon, Aug 28, 2017 at 6:13 PM, Jan Beulich <JBeulich@suse.com> wrote:

> >>> On 28.08.17 at 11:04, <nickeysgo@gmail.com> wrote:
> > Thank you for your reply.
> > I checked what you mentioned(Xen command line).
> > But I could not find where the Xen command line configuration.
> > As I think, it would be in /etc/default/grub file because Dom0 is used
> for
> > booting like native system.
> > Is it wrong? I didn't get your point?
>
> First of all - please avoid private mails. Keep the list Cc-ed.
>
> And then - I can't tell you how to configure your variant of grub.
> All I can tell you is that the option you're after needs to be
> passed to Xen.
>
> Jan
>
>

[-- Attachment #1.2: Type: text/html, Size: 2068 bytes --]

[-- Attachment #2: Type: text/plain, Size: 127 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: crashdump on PVM Dom0
  2017-08-28 10:45       ` Minjun Hong
@ 2017-08-28 11:13         ` Jan Beulich
  0 siblings, 0 replies; 7+ messages in thread
From: Jan Beulich @ 2017-08-28 11:13 UTC (permalink / raw)
  To: Minjun Hong; +Cc: xen-devel

>>> On 28.08.17 at 12:45, <nickeysgo@gmail.com> wrote:
> Oh!! I'm so sorry Jan Beulich.
> It is my first time to use this mailing list.
> Please understand my fault, this time only.
> 
> Anyway, as you mentioned, I added below line in '/etc/default/grub' file
> and reboot:
>> GRUB_CMDLINE_XEN="crashkernel=384M-:256M@64M"
> Then,
> 1. I cannot see anything when I type "grep -i crash /proc/iomem"
> 2. I can see kernel log of Xen hypervisor when I type "sudo xl dmesg | grep
> -i crash":
> (XEN) Command line: placeholder crashkernel=384M-:256M@64M
> 
> I suppose a 'crashkernel' cmdline option of either Xen hypervisor or Dom0
> kernel can be enabled.
> I know it is quite difficult to advise some problems online but, could you
> give me a piece of hint?

I'm not sure what else hint you're after. You can check the log
so see whether the option was accepted. I'm not sure /proc/iomem
is supposed to have any indication of the crash area.

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: crashdump on PVM Dom0
  2017-08-28  8:01 crashdump on PVM Dom0 Minjun Hong
  2017-08-28  8:13 ` Jan Beulich
@ 2017-08-29  9:19 ` Wei Liu
  2017-08-29 10:05 ` Roger Pau Monné
  2 siblings, 0 replies; 7+ messages in thread
From: Wei Liu @ 2017-08-29  9:19 UTC (permalink / raw)
  To: Minjun Hong; +Cc: Wei Liu, xen-devel


On Mon, Aug 28, 2017 at 05:01:48PM +0900, Minjun Hong wrote:
> Hello~
> I'm new to develop Xen and my first work is about its scheduler.
> After I modified Xen 4.5 source code, my custom Xen has had a problem,
> which it causes automatic reboot of Dom0 when I execute a micro benchmark
> program.
> To debug it, I have tried to use crashdump on Dom0 with the Xen but, I
> failed to enable it.
> My configurations are following:
> 
> 1. Base Xen version: 4.5 release

Off-topic:

If you want to develop Xen I suggest you use xen-unstable branch or the
latest stable release, unless you have (good) reasons to stick with 4.5,
which is a bit old.

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: crashdump on PVM Dom0
  2017-08-28  8:01 crashdump on PVM Dom0 Minjun Hong
  2017-08-28  8:13 ` Jan Beulich
  2017-08-29  9:19 ` Wei Liu
@ 2017-08-29 10:05 ` Roger Pau Monné
  2017-08-29 10:24   ` Minjun Hong
  2 siblings, 1 reply; 7+ messages in thread
From: Roger Pau Monné @ 2017-08-29 10:05 UTC (permalink / raw)
  To: Minjun Hong; +Cc: xen-devel

On Mon, Aug 28, 2017 at 05:01:48PM +0900, Minjun Hong wrote:
> Hello~
> I'm new to develop Xen and my first work is about its scheduler.
> After I modified Xen 4.5 source code, my custom Xen has had a problem,
> which it causes automatic reboot of Dom0 when I execute a micro benchmark
> program.
> To debug it, I have tried to use crashdump on Dom0 with the Xen but, I
> failed to enable it.

My suggestion would be to get the output from the serial port
instead, specially if you plan to do Xen development. It's going to be
much more reliable.

Roger.

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel

^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: crashdump on PVM Dom0
  2017-08-29 10:05 ` Roger Pau Monné
@ 2017-08-29 10:24   ` Minjun Hong
  0 siblings, 0 replies; 7+ messages in thread
From: Minjun Hong @ 2017-08-29 10:24 UTC (permalink / raw)
  To: Roger Pau Monné; +Cc: xen-devel


[-- Attachment #1.1: Type: text/plain, Size: 1169 bytes --]

@Wei Liu
Thanks for your reply.
I should use this version(4.5) because I received a solution developed in
the version.

@Roger Pau Monné
Firstly, Thank you for your reply.
I also considered the option using serial cable but, I think if I can get
debugging information with only SW method(crashdump), that method will be
more flexible way to debug on Xen.
Even if I have struggled to find the way to debug Xen using crashdump, I
failed. So, finally I ordered serial cable.

On Tue, Aug 29, 2017 at 7:05 PM, Roger Pau Monné <roger.pau@citrix.com>
wrote:

> On Mon, Aug 28, 2017 at 05:01:48PM +0900, Minjun Hong wrote:
> > Hello~
> > I'm new to develop Xen and my first work is about its scheduler.
> > After I modified Xen 4.5 source code, my custom Xen has had a problem,
> > which it causes automatic reboot of Dom0 when I execute a micro benchmark
> > program.
> > To debug it, I have tried to use crashdump on Dom0 with the Xen but, I
> > failed to enable it.
>
> My suggestion would be to get the output from the serial port
> instead, specially if you plan to do Xen development. It's going to be
> much more reliable.
>
> Roger.
>

[-- Attachment #1.2: Type: text/html, Size: 2021 bytes --]

[-- Attachment #2: Type: text/plain, Size: 127 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel

^ permalink raw reply	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2017-08-29 10:24 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-08-28  8:01 crashdump on PVM Dom0 Minjun Hong
2017-08-28  8:13 ` Jan Beulich
     [not found]   ` <CAFX_q-HHBCb6KOknVXhu5CirfEckkOS0SW2ZLjoaEcpswwy5Og@mail.gmail.com>
     [not found]     ` <59A3FAF202000078001746B4@prv-mh.provo.novell.com>
2017-08-28 10:45       ` Minjun Hong
2017-08-28 11:13         ` Jan Beulich
2017-08-29  9:19 ` Wei Liu
2017-08-29 10:05 ` Roger Pau Monné
2017-08-29 10:24   ` Minjun Hong

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.