* Regression since 4.1
@ 2015-12-09 17:01 Tobias Geiger
2015-12-09 19:07 ` Alex Williamson
0 siblings, 1 reply; 5+ messages in thread
From: Tobias Geiger @ 2015-12-09 17:01 UTC (permalink / raw)
To: iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA
Hello!
in short:
No Kernel since 4.1 boots on my machine.
I can make newer kernels boot, but ONLY when i remove "iommu=on
intel_iommu=on" from the cmdline.
If i leave it in, as with all the kernels before and including 4.1, newer
kernels dont boot, and in a really not very informative way:
They dont even print a single message to the console, i only see the grub-
messages (like "loading kernel" and "loading initial ramdisk") - after these
messages NO more output of any kind - also no SysRQ keys working - nothing.
Because of this complete lack of information, i guess its hard to debug for
you the cause of this.
Perhaps i can help - but how?
i'm writing from a pure IOMMU-User perspective - i can compile kernels, but
never git-bisected, so maybe a hint to a tutorial for that would be great!
Greetings and thanks for your insight!
Tobias
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: Regression since 4.1
2015-12-09 17:01 Regression since 4.1 Tobias Geiger
@ 2015-12-09 19:07 ` Alex Williamson
[not found] ` <1449688068.15753.461.camel-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
0 siblings, 1 reply; 5+ messages in thread
From: Alex Williamson @ 2015-12-09 19:07 UTC (permalink / raw)
To: Tobias Geiger; +Cc: iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA
On Wed, 2015-12-09 at 18:01 +0100, Tobias Geiger wrote:
> Hello!
>
> in short:
> No Kernel since 4.1 boots on my machine.
>
> I can make newer kernels boot, but ONLY when i remove "iommu=on
> intel_iommu=on" from the cmdline.
> If i leave it in, as with all the kernels before and including 4.1, newer
> kernels dont boot, and in a really not very informative way:
> They dont even print a single message to the console, i only see the grub-
> messages (like "loading kernel" and "loading initial ramdisk") - after these
> messages NO more output of any kind - also no SysRQ keys working - nothing.
>
> Because of this complete lack of information, i guess its hard to debug for
> you the cause of this.
> Perhaps i can help - but how?
>
> i'm writing from a pure IOMMU-User perspective - i can compile kernels, but
> never git-bisected, so maybe a hint to a tutorial for that would be great!
Try simply removing the "quiet" and "splash" or "rhgb" options if there
are any from the kernel command line in grub first. You can temporarily
just editing the commandline manually as you boot. If you don't have a
serial console available, at least have a camera handy to take a picture
of the screen. You can also add "ignore_loglevel" to the kernel
commandline to get more logging. Thanks,
Alex
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: Regression since 4.1
[not found] ` <1449688068.15753.461.camel-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
@ 2015-12-09 23:07 ` Tobias Geiger
2015-12-28 16:49 ` Joerg Roedel
0 siblings, 1 reply; 5+ messages in thread
From: Tobias Geiger @ 2015-12-09 23:07 UTC (permalink / raw)
To: Alex Williamson; +Cc: iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA
Hello Alex,
thanks for looking into this.
Nothing helps - not adding "ignore_loglevel" , not removing everything but the
most necessary (="BOOT_IMAGE=/boot/vmlinuz-4.3.0-trunk-amd64 ro
root=LABEL=dom0root panic=15 iommu=on intel_iommu=on ignore_loglevel") - grub
loads the kernel, loads initrd, execs the kernel - and then nothing. no output
of any kind from the kernel. no sysrq, no panic-reboot despite panic=15 ...
nothing. not even a blinking cursor, just a steady cursor which is still from
grub i guess...
It all comes down to kernel > 4.1 && iommu=on = not working here
Hardware is:
Intel DX58SO, Intel CPU i7 920... all pretty solid running for years now,
always with iommo=on.
I'd love to debug this further, just dont know how...
Greetings!
Tobias
Am Mittwoch, 9. Dezember 2015, 12:07:48 schrieb Alex Williamson:
> Try simply removing the "quiet" and "splash" or "rhgb" options if there
> are any from the kernel command line in grub first. You can temporarily
> just editing the commandline manually as you boot. If you don't have a
> serial console available, at least have a camera handy to take a picture
> of the screen. You can also add "ignore_loglevel" to the kernel
> commandline to get more logging. Thanks,
>
> Alex
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: Regression since 4.1
2015-12-09 23:07 ` Tobias Geiger
@ 2015-12-28 16:49 ` Joerg Roedel
2016-04-04 17:06 ` Tom Wimmenhove
0 siblings, 1 reply; 5+ messages in thread
From: Joerg Roedel @ 2015-12-28 16:49 UTC (permalink / raw)
To: Tobias Geiger; +Cc: iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA
Hi Tobias,
On Thu, Dec 10, 2015 at 12:07:57AM +0100, Tobias Geiger wrote:
> It all comes down to kernel > 4.1 && iommu=on = not working here
>
> Hardware is:
> Intel DX58SO, Intel CPU i7 920... all pretty solid running for years now,
> always with iommo=on.
>
> I'd love to debug this further, just dont know how...
Sounds like only bisecting the issue will help here. You can find an
introduction on bisecting here:
https://git-scm.com/book/en/v2/Git-Tools-Debugging-with-Git
Would be great if you can find the time to bisect the issue.
Joerg
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: Regression since 4.1
2015-12-28 16:49 ` Joerg Roedel
@ 2016-04-04 17:06 ` Tom Wimmenhove
0 siblings, 0 replies; 5+ messages in thread
From: Tom Wimmenhove @ 2016-04-04 17:06 UTC (permalink / raw)
To: iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA
Joerg Roedel <joro@...> writes:
>
> Hi Tobias,
>
> On Thu, Dec 10, 2015 at 12:07:57AM +0100, Tobias Geiger wrote:
> > It all comes down to kernel > 4.1 && iommu=on = not working here
> >
> > Hardware is:
> > Intel DX58SO, Intel CPU i7 920... all pretty solid running for years
now,
> > always with iommo=on.
> >
> > I'd love to debug this further, just dont know how...
>
> Sounds like only bisecting the issue will help here. You can find an
> introduction on bisecting here:
>
> https://git-scm.com/book/en/v2/Git-Tools-Debugging-with-Git
>
> Would be great if you can find the time to bisect the issue.
>
> Joerg
>
>
I can confirm this bug with kernel "4.2.0-16-generic #19-Ubuntu" on an
Intel DX58SO board with a W3520 Xeon CPU.
Before I was running 3.16.0 which did not lock up at boot but did have
occasional DMAR issues with GPU passthrough while running a Windows 7 gues
in Qemu. I don't know if the two are related.
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2016-04-04 17:06 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2015-12-09 17:01 Regression since 4.1 Tobias Geiger
2015-12-09 19:07 ` Alex Williamson
[not found] ` <1449688068.15753.461.camel-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
2015-12-09 23:07 ` Tobias Geiger
2015-12-28 16:49 ` Joerg Roedel
2016-04-04 17:06 ` Tom Wimmenhove
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).