Patchwork [meta-fsl-arm-extra,2/2] wandboard: Change kernel to 3.10.17

login
register
mail settings
Submitter John Weber
Date March 24, 2014, 7:58 p.m.
Message ID <1395691102-14206-2-git-send-email-rjohnweber@gmail.com>
Download mbox | patch
Permalink /patch/69117/
State Superseded
Delegated to: Otavio Salvador
Headers show

Comments

John Weber - March 24, 2014, 7:58 p.m.
Move default kernel selection to 3.10.17 in advance of the 1.6 release,
because currently 3.0.35_4.0.0 is broken in master.  There are known issues
with the kernel that will be resolved with community assistance.  These
include:

- HDMI:  Using X11 causes HDMI output to default to 1024x768 resolution
         regardless of the 'video=mxcfb...' bootarg selection.
- HDMI, EDID:  Does not work (but this did not work on 3.0.35 either).
- Sound, SGTL5000: After POR, the codec chip works fine.  However, after
         playing sound through the coded, a board soft reset or pushbutton
         reset results in a hung I2C2 bus, causing devices on i2c2 to fail
         to initialize.
- Sound, HDMI audio output: Using HDMI audio output seems to cause very slow
         playback of both audio and video, and can be seen using aplay or
         gst-launch.

Signed-off-by: John Weber <rjohnweber@gmail.com>
---
 conf/machine/include/wandboard.inc |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Fabio Estevam - March 24, 2014, 8:15 p.m.
Hi John,

On Mon, Mar 24, 2014 at 4:58 PM, John Weber <rjohnweber@gmail.com> wrote:
> Move default kernel selection to 3.10.17 in advance of the 1.6 release,
> because currently 3.0.35_4.0.0 is broken in master.  There are known issues
> with the kernel that will be resolved with community assistance.  These
> include:
>
> - HDMI:  Using X11 causes HDMI output to default to 1024x768 resolution
>          regardless of the 'video=mxcfb...' bootarg selection.
> - HDMI, EDID:  Does not work (but this did not work on 3.0.35 either).
> - Sound, SGTL5000: After POR, the codec chip works fine.  However, after
>          playing sound through the coded, a board soft reset or pushbutton
>          reset results in a hung I2C2 bus, causing devices on i2c2 to fail
>          to initialize.

I have fixed this issue already:
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/sound/soc/codecs/sgtl5000.c?id=af8ee11209e749c75eabf32b2a4ca631f396acf8

Regards,

Fabio Estevam
John Weber - March 24, 2014, 9:33 p.m.
Hi Fabio,

On 3/24/14, 3:15 PM, Fabio Estevam wrote:
> Hi John,
>
> On Mon, Mar 24, 2014 at 4:58 PM, John Weber <rjohnweber@gmail.com> wrote:
>> Move default kernel selection to 3.10.17 in advance of the 1.6 release,
>> because currently 3.0.35_4.0.0 is broken in master.  There are known issues
>> with the kernel that will be resolved with community assistance.  These
>> include:
>>
>> - HDMI:  Using X11 causes HDMI output to default to 1024x768 resolution
>>           regardless of the 'video=mxcfb...' bootarg selection.
>> - HDMI, EDID:  Does not work (but this did not work on 3.0.35 either).
>> - Sound, SGTL5000: After POR, the codec chip works fine.  However, after
>>           playing sound through the coded, a board soft reset or pushbutton
>>           reset results in a hung I2C2 bus, causing devices on i2c2 to fail
>>           to initialize.
> I have fixed this issue already:
> https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/sound/soc/codecs/sgtl5000.c?id=af8ee11209e749c75eabf32b2a4ca631f396acf8
Thanks for noticing this. I did see your patches.  Unfortunately, I don't think 
that the bug fixes you pushed resolved this particular problem.  I have a local 
branch where I've merged a lot of the ASoC patches (starting with yours) and 
they didn't help.  I suspect that something else is causing the codec to be naughty.
> Regards,
>
> Fabio Estevam
Thanks,
John
Fabio Estevam - March 24, 2014, 10:47 p.m.
Hi John,

On Mon, Mar 24, 2014 at 6:33 PM, John Weber <rjohnweber@gmail.com> wrote:

> Thanks for noticing this. I did see your patches.  Unfortunately, I don't
> think that the bug fixes you pushed resolved this particular problem.  I
> have a local branch where I've merged a lot of the ASoC patches (starting
> with yours) and they didn't help.  I suspect that something else is causing
> the codec to be naughty.

I noticed this same bug on several boards: mx28evk, sabrelite,
wandboard and it fixes the issue for me on all of them.

I have just tried a mainline kernel and I am able to play audio just
fine after a reboot on a mx6qsabresd.

Regards,

Fabio Estevam
Fabio Estevam - March 24, 2014, 10:50 p.m.
On Mon, Mar 24, 2014 at 7:47 PM, Fabio Estevam <festevam@gmail.com> wrote:
> Hi John,
>
> On Mon, Mar 24, 2014 at 6:33 PM, John Weber <rjohnweber@gmail.com> wrote:
>
>> Thanks for noticing this. I did see your patches.  Unfortunately, I don't
>> think that the bug fixes you pushed resolved this particular problem.  I
>> have a local branch where I've merged a lot of the ASoC patches (starting
>> with yours) and they didn't help.  I suspect that something else is causing
>> the codec to be naughty.
>
> I noticed this same bug on several boards: mx28evk, sabrelite,
> wandboard and it fixes the issue for me on all of them.
>
> I have just tried a mainline kernel and I am able to play audio just
> fine after a reboot on a mx6qsabresd.

I meant "on a mx6q-wandboard".
John Weber - March 25, 2014, 5:47 p.m.
Hi Fabio -
On 3/24/14, 3:15 PM, Fabio Estevam wrote:
> Hi John,
>
> On Mon, Mar 24, 2014 at 4:58 PM, John Weber <rjohnweber@gmail.com> wrote:
>> Move default kernel selection to 3.10.17 in advance of the 1.6 release,
>> because currently 3.0.35_4.0.0 is broken in master.  There are known issues
>> with the kernel that will be resolved with community assistance.  These
>> include:
>>
>> - HDMI:  Using X11 causes HDMI output to default to 1024x768 resolution
>>           regardless of the 'video=mxcfb...' bootarg selection.
>> - HDMI, EDID:  Does not work (but this did not work on 3.0.35 either).
>> - Sound, SGTL5000: After POR, the codec chip works fine.  However, after
>>           playing sound through the coded, a board soft reset or pushbutton
>>           reset results in a hung I2C2 bus, causing devices on i2c2 to fail
>>           to initialize.
> I have fixed this issue already:
> https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/sound/soc/codecs/sgtl5000.c?id=af8ee11209e749c75eabf32b2a4ca631f396acf8
>
> Regards,
>
> Fabio Estevam
Thanks for your help with this.  I was able to make the correct changes this 
time and fixed the problem.  Just sent a v2.

John

Patch

diff --git a/conf/machine/include/wandboard.inc b/conf/machine/include/wandboard.inc
index ff6fb52..d8a0348 100644
--- a/conf/machine/include/wandboard.inc
+++ b/conf/machine/include/wandboard.inc
@@ -4,7 +4,7 @@  include conf/machine/include/imx-base.inc
 include conf/machine/include/tune-cortexa9.inc
 
 PREFERRED_PROVIDER_virtual/kernel ?= "linux-wandboard"
-PREFERRED_VERSION_linux-wandboard ?= "3.0.35"
+PREFERRED_VERSION_linux-wandboard ?= "3.10.17"
 
 SERIAL_CONSOLE = "115200 ttymxc0"