On 2014-11-05 09:42, Takashi Iwai wrote:
At Wed, 05 Nov 2014 09:16:01 +0100,
David Henningsson wrote:
On 2014-11-05 09:00, Jiri Slaby wrote:
On 11/04/2014, 12:02 PM, David Henningsson wrote:
On 2014-10-31 18:28, Takashi Iwai wrote:
At Fri, 31 Oct 2014 12:07:13 +0100,
Jiri Slaby wrote:
On 10/08/2014, 10:09 AM, David Henningsson wrote:
Hi stable maintainers,
These two patches: c149dcb5c60bfea8871f16dfcc0690255eeb825f
drm/i915: provide interface for audio driver to query cdclk
and
e4d9e513dedb5ac4e166c1053314fa935ddecc8c
ALSA: hda - restore BCLK M/N value as per CDCLK for HSW/BDW display HDA
controller
...never seem to have made it into 3.12+ stable kernels. They fix an
issue where playback rate over Haswell HDMI audio can be faster than
normal.
Now, the latter of these two patches requires a restructure of moving
things from hda_intel.c to hda_priv.h, which looked tedious and
potentially error-prone to backport in itself, so instead I rewrote the
function slightly, see attachment.
Do you think this is a reasonable approach?
I seem not to find out a way to apply those to 3.12 cleanly :/. Any
ideas/help appreciated.
FWIW, SLE12 has already three backported patches relevant with this:
patches.drivers/drm-i915-HD-audio-Don-t-continue-probing-when-nomode
for 74b0c2d75fb4cc89173944e6d8f9eb47aca0c343
patches.drivers/drm-i915-provide-interface-for-audio-driver-to-query
for c149dcb5c60bfea8871f16dfcc0690255eeb825f
patches.drivers/hda-0038-restore-BCLK-M-N-value-as-per-CDCLK-for-HSW
for e4d9e513dedb5ac4e166c1053314fa935ddecc8c
The first one changes the i915 function type, so this may conflict
with David's backport.
Hmm, interesting as a reference - is there a public repository where
these patches exist, or is it for Suse's eyes only?
Hi,
we have all kernel branches at kernel.opensuse.org:
http://kernel.opensuse.org/cgit/kernel/tree/?h=SLE12
Also is this understanding correct: Jiri makes a stable 3.12 kernel
backport, posts that on kernel.org, and then on top of that, SUSE adds
its own patches (e g with new hardware support as demanded from
customers), and that in turn becomes the SLE12 kernel? And in this case,
the patches made it into the SLE12 patch set, rather than the stable
3.12 patch set?
Yes, exactly.
thanks,
Thanks. So Takashi's backport of this:
http://kernel.opensuse.org/cgit/kernel/commit/?h=SLE12&id=c35bc9cf53fd397b3d55c754bc274928b46cc6f0
...would work just as well as my version of the patch. It's mostly a
matter of in which file you put the function.
So how about this: Let's first assume you have *not* applied the patch
that was cc:ed to stable called "ALSA: hda - restore BCLK M/N values
when resuming HSW/BDW display controller", if you did, revert it.
Then you can apply the version referenced above, by Takashi. You will
still need "drm/i915: provide interface for audio driver to query cdclk"
but that one shouldn't cause any trouble to apply.
Does this make sense to everyone?
Also the commit 74b0c2d75fb4cc89173944e6d8f9eb47aca0c343 needs
backports before others. It makes other backports easier, not only
fixing the bug itself.
If David is fine with them, Jiri, could you queue the three patches to
3.12 stable?
patches.drivers/drm-i915-HD-audio-Don-t-continue-probing-when-nomode
(upstream 74b0c2d75fb4cc89173944e6d8f9eb47aca0c343)
patches.drivers/drm-i915-provide-interface-for-audio-driver-to-query
(upstream c149dcb5c60bfea8871f16dfcc0690255eeb825f)
patches.drivers/hda-0038-restore-BCLK-M-N-value-as-per-CDCLK-for-HSW
(upstream e4d9e513dedb5ac4e166c1053314fa935ddecc8c)
Yes, I'm fine with the "don't continue probing when nomodeset..." patch
too. It looks reasonable to me.
thanks,
Takashi
--
David Henningsson, Canonical Ltd.
https://launchpad.net/~diwic
--
To unsubscribe from this list: send the line "unsubscribe stable" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html