On Mon, Feb 14, 2011 at 1:24 PM, Sedat Dilek [off-list ref] wrote:

On Mon, Feb 14, 2011 at 1:13 PM, Sedat Dilek [off-list ref] wrote:

quoted

Hi,

I have pulled Dave's drm-fixes-staging into today's linux-next and did
not see a KMS-modesetted resolution change in VT (runlevel-3).
After a quick check of my dmesg, I see also my kernel-module for sound
has "unknown parameter".
( Thus, personally I don't think it is a bug from any drm-2.6 tree. )

# dmesg | egrep -i 'unknown|param'
[   12.431216] radeon: Unknown parameter `modeset'
[   12.665911] snd_intel8x0m: Unknown parameter `index'
[   37.311333] radeon: Unknown parameter `modeset'

Any idea what's going on? Stephen, Randy noticed the same like me?

Regards,
- Sedat -

P.S.: Attached are dmesg and kernel dot-config.

[ CC Dmitry Torokhov ]

Could that be a possible fix for me [1]?

- Sedat -

[1] https://patchwork.kernel.org/patch/548891/

[ CC Rusty Russell + Removed drm-2.6/dri-devel folks ]

While digging into the problem, I had a closer look into
linux-next/Trees file [1], which list the included GIT/QUILT
trees/patch-queues, line #121 says:

rr              quilt   http://ozlabs.org/~rusty/kernel/rr-latest/

( Personally, I thought all stuff are pulled from GIT trees, but you
learn more by CSI-ing. )

As a clever guy, I was silently assuming the problem could occur from
any module-related tree (so I checked linux-next GIT repo via "git log
-3 kernel/params.c" which gave me a hint to Dmitry's commit in
quilt/rr, Bingo?).

With this "evidence material" I "tig-ged" into the commits in my local
linux-next GIT repo, which revealed more:

2011-02-14 14:17 Stephen Rothwell   Merge branch 'quilt/rr'
2011-02-11 10:37 Rusty Russell      module: deal with alignment issues
in built-in module parameters FIX
2011-02-07 16:02 Dmitry Torokhov    module: deal with alignment issues
in built-in module parameters
2011-02-01 21:43 Christoph Hellwig  virtio_blk: allow re-reading
config space at runtime
2011-02-07 16:02 Dmitry Torokhov    module: do not hide
__modver_version_show declaration behind ifdef
2011-02-07 16:02 Dmitry Torokhov    module: deal with alignment issues
in built-in module versions

My damn brain can't understand - looking at the series file in [2] -
why the hell the "possible" patch [3] is not included in linux-next?
I did not compare but [3] looks similiar to [4].
I have applied [3] on top my own patch-series and will report later,
if this worked or not.

BUT please, what are the factors for choosing the relevant patches
from quilt/rr?
( For example, the cpumask stuff is awaited by systemd folks. Is this
stuff included? I saw by a very quick look @ git.k.o there is a GIT
repo for that. )

And yeah, slow-ass UserModeSetting suckz here!

/me confused,
- Sedat -

[1] http://git.kernel.org/?p=linux/kernel/git/next/linux-next.git;a=blob;f=Next/Trees
[2] http://ozlabs.org/~rusty/kernel/rr-latest/series
[3] http://ozlabs.org/~rusty/kernel/rr-latest/module:deal_with_alignment_issues_in_built_in_parameters.patch
[4] https://patchwork.kernel.org/patch/548891/