Re: linux-next: Tree for Sept 11 (drm related: boot problems on amd64)

From: Sedat Dilek
Date: Wed Sep 12 2012 - 23:01:27 EST


On Wed, Sep 12, 2012 at 9:52 PM, Bjorn Helgaas <bhelgaas@xxxxxxxxxx> wrote:
> On Tue, Sep 11, 2012 at 4:04 PM, Sedat Dilek <sedat.dilek@xxxxxxxxx> wrote:
>> On Tue, Sep 11, 2012 at 8:29 PM, Sedat Dilek <sedat.dilek@xxxxxxxxx> wrote:
>>> On Tue, Sep 11, 2012 at 8:12 PM, Sedat Dilek <sedat.dilek@xxxxxxxxx> wrote:
>>>> On Tue, Sep 11, 2012 at 8:31 AM, Stephen Rothwell <sfr@xxxxxxxxxxxxxxxx> wrote:
>>>>> Hi all,
>>>>>
>>>>> Changes since 201209010:
>>>>>
>>>>> New tree: ixp4xx
>>>>>
>>>>> The pci tree gained a build failure so I used the version from
>>>>> next-20120910.
>>>>>
>>>>> The regulator tree lost its build failure.
>>>>>
>>>>> The staging tree lost its build failure.
>>>>>
>>>>> The akpm tree lost a few patches that turned up elsewhere.
>>>>>
>>>>> ----------------------------------------------------------------------------
>>>>>
>>>>
>>>> Hi,
>>>>
>>>> today's and yesterday's Linux-Next is broken for me again.
>>>> I tried with systemd and upstart on Ubuntu/precise, with nomodeset and
>>>> rescue boot-option.
>>>>
>>>> With rescue boot-option I see this in my logs (Intel sandy-bridge
>>>> ultrabook here):
>>>>
>>>> Sep 11 18:43:36 fambox kernel: [ 9.654972] [drm:drm_pci_agp_init]
>>>> *ERROR* Cannot initialize the agpgart module.
>>>> Sep 11 18:43:36 fambox kernel: [ 9.654980] DRM: Fill_in_dev failed.
>>>>
>>>> I have not checked any MLs... coming from hospital right now.
>>>>
>>>
>>> More ERRORs:
>>>
>>> # grep -A1 "ERROR" /var/log/kern.log
>>> Sep 10 17:51:29 fambox kernel: [ 10.205818]
>>> [drm:i915_get_bridge_dev] *ERROR* bridge device not found
>>> Sep 10 17:51:29 fambox kernel: [ 10.206055] i915: probe of
>>> 0000:00:02.0 failed with error -5
>>> --
>>> Sep 10 15:53:00 fambox kernel: [ 10.500387]
>>> [drm:i915_get_bridge_dev] *ERROR* bridge device not found
>>> Sep 10 15:53:00 fambox kernel: [ 10.500602] i915: probe of
>>> 0000:00:02.0 failed with error -5
>>> --
>>> Sep 11 20:41:01 fambox kernel: [ 9.636010]
>>> [drm:i915_get_bridge_dev] *ERROR* bridge device not found
>>> Sep 11 20:41:01 fambox kernel: [ 9.636202] i915: probe of
>>> 0000:00:02.0 failed with error -5
>>> --
>>> Sep 11 18:42:18 fambox kernel: [ 10.132229]
>>> [drm:i915_get_bridge_dev] *ERROR* bridge device not found
>>> Sep 11 18:42:18 fambox kernel: [ 10.132433] i915: probe of
>>> 0000:00:02.0 failed with error -5
>>> --
>>> Sep 11 18:43:36 fambox kernel: [ 9.654972] [drm:drm_pci_agp_init]
>>> *ERROR* Cannot initialize the agpgart module.
>>> Sep 11 18:43:36 fambox kernel: [ 9.654980] DRM: Fill_in_dev failed.
>>> --
>>> Sep 11 19:52:10 fambox kernel: [ 9.545562]
>>> [drm:i915_get_bridge_dev] *ERROR* bridge device not found
>>> Sep 11 19:52:10 fambox kernel: [ 9.545798] i915: probe of
>>> 0000:00:02.0 failed with error -5
>>> --
>>> Sep 11 20:04:09 fambox kernel: [ 9.798233]
>>> [drm:i915_get_bridge_dev] *ERROR* bridge device not found
>>> Sep 11 20:04:09 fambox kernel: [ 9.798557] i915: probe of
>>> 0000:00:02.0 failed with error -5
>>>
>>
>> [ CC Bjorn (pci maintainer) ]
>>
>> I pulled in pci.git#next up to commit
>> 9c33332178e6ba49fe48c468edc08ad94b53e1b1 ("Merge branch
>> 'pci/gavin-window-alignment' into next") on top of next-20120911.
>> This lets my machine boot, but freezes somewhere else.
>
> We had a PCI bug in an earlier version of this patch:
> http://git.kernel.org/?p=linux/kernel/git/helgaas/pci.git;a=commitdiff;h=b9443f401bb20ae6414e3e68bca0413bad28b689
> that caused lspci to fail.
>

I retested with next-20120912 and my machine boots, so no more
drm-related errors causing boot-failures.

> I suspect this also caused the issue you saw. We've since fixed it,
> but let me know if you see a PCI-related issue again.
>

Immediately after pressing any single key at X-login (Ubuntu/precise
IIRC uses lightdm + unity-greeter), my machine produces a
kernel-panic.
I have taken a photo and will send a separate email on this.
Can't say ATM what is the root-cause for it.
But that is what I have seen after my merging of the "clean"
pci.git/next into yesterday's "nine-eleven" release.

Further questions (not issue related):

While doing a simple grepping for new "pcie-cap" patterns (a list at
[1]) came with commit 8c0d3a02c1309eb6112d2e7c8172e8ceb26ecfca ("PCI:
Add accessors for PCI Express Capability") on drivers/gpu/
directory...more simplifications possible (I had only a quick view)?

- Sedat -

[1] http://git.kernel.org/?p=linux/kernel/git/helgaas/pci.git;a=commitdiff;h=8c0d3a02c1309eb6112d2e7c8172e8ceb26ecfca#patch2

> Bjorn
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/