linux 2.6.34 patches

43 views
Skip to first unread message

Martin Guy

unread,
May 14, 2010, 1:25:46 PM5/14/10
to si...@googlegroups.com
Hi
I've added a set of kernel patches to the googlecode repository for
linux-2.6.34.
2.6.34 isn't released yet (it's currently at release candidate 7) but
they should be ok for the final version.
2.6.34 is the first mainline kernel that includes the Sim.One as an
official platform \O/

The most significant item in our 2.6.34 patches at googlecode is a new
SD/MMC card driver by Mika Westerberger (which should be included in
2.6.35). It runs at 377 kB/sec instead of 256 kB/sec and uses 55% CPU
to do that instead of 100%, being interrupt-driven instead of
busy-waiting. This makes putting a rootfs on the SD card reliable and
faster, as well as letting you use a swap partition there to improve
memory use.

What is still missing, compared to the Cirrus 2.6.24.7 kernel is:

- Access to the battery-backed DS1337 real time clock on the I2C bus.
At present it only sees the EP93xx RTC, which resets to 1 Jan 1970
when power is removed. There is an EP93xx I2C bus driver in the 2.6.32
patches, but it doesn't see the DS1337 for some reason.

- A driver for the CS4202 audio chip on the AC97 bus. Ryan Mallon is
working on an audio driver for the I2S bus - our AC97 driver should be
similar to this.

M

Florian Fainelli

unread,
May 14, 2010, 6:13:04 PM5/14/10
to si...@googlegroups.com, Martin Guy
Hi Martin,

Le vendredi 14 mai 2010 19:25:46, Martin Guy a écrit :
> Hi
> I've added a set of kernel patches to the googlecode repository for
> linux-2.6.34.
> 2.6.34 isn't released yet (it's currently at release candidate 7) but
> they should be ok for the final version.
> 2.6.34 is the first mainline kernel that includes the Sim.One as an
> official platform \O/
>
> The most significant item in our 2.6.34 patches at googlecode is a new
> SD/MMC card driver by Mika Westerberger (which should be included in
> 2.6.35). It runs at 377 kB/sec instead of 256 kB/sec and uses 55% CPU
> to do that instead of 100%, being interrupt-driven instead of
> busy-waiting. This makes putting a rootfs on the SD card reliable and
> faster, as well as letting you use a swap partition there to improve
> memory use.

I saw this on linux-arm-kernel, really good work!

>
> What is still missing, compared to the Cirrus 2.6.24.7 kernel is:
>
> - Access to the battery-backed DS1337 real time clock on the I2C bus.
> At present it only sees the EP93xx RTC, which resets to 1 Jan 1970
> when power is removed. There is an EP93xx I2C bus driver in the 2.6.32
> patches, but it doesn't see the DS1337 for some reason.

I believe we can use the ds1337 driver without much troubles, there are no
valid reasons why ep93xx-rtc should be a separate driver as it is just
wrapping the ds1337 driver.

>
> - A driver for the CS4202 audio chip on the AC97 bus. Ryan Mallon is
> working on an audio driver for the I2S bus - our AC97 driver should be
> similar to this.

Oh great, I did not have as much as time as I hoped to work seriously on this.
--
Florian

Martin Guy

unread,
May 14, 2010, 7:30:53 PM5/14/10
to Florian Fainelli, si...@googlegroups.com
On 5/14/10, Florian Fainelli <f.fai...@gmail.com> wrote:
> > - Access to the battery-backed DS1337 real time clock on the I2C bus.
> > At present it only sees the EP93xx RTC, which resets to 1 Jan 1970
> > when power is removed. There is an EP93xx I2C bus driver in the 2.6.32
> > patches, but it doesn't see the DS1337 for some reason.
>
> I believe we can use the ds1337 driver without much troubles, there are no
> valid reasons why ep93xx-rtc should be a separate driver as it is just
> wrapping the ds1337 driver.

Yes. There is a ds1337-on-i2c-bus driver in the kernel already and in
our 2.6.32 patches there is one patch that says it implements an
i2c-bus-on-ep93xx driver, but if you enable them both no magic happens
:(

M

westeri

unread,
May 29, 2010, 2:47:52 AM5/29/10
to sim1


On May 15, 2:30 am, Martin Guy <martinw...@gmail.com> wrote:
For me, RTC works fine on Sim.One (2.6.34 kernel):

[ 1.810000] rtc-ds1307 0-0068: rtc core: registered ds1337 as rtc0
...
[ 1.930000] rtc-ds1307 0-0068: setting system clock to 2010-05-29
09:45:19 UTC (1275126319)

You just need to make sure that CONFIG_I2C_GPIO is set and
CONFIG_RTC_DRV_EP93XX is disabled.

Regards,
MW

Martin Guy

unread,
May 29, 2010, 9:51:49 PM5/29/10
to si...@googlegroups.com
On 5/29/10, westeri <mika.we...@gmail.com> wrote:
> For me, RTC works fine on Sim.One (2.6.34 kernel):
>
> [ 1.810000] rtc-ds1307 0-0068: rtc core: registered ds1337 as rtc0
> ...
> [ 1.930000] rtc-ds1307 0-0068: setting system clock to 2010-05-29
> 09:45:19 UTC (1275126319)
>
> You just need to make sure that CONFIG_I2C_GPIO is set and
> CONFIG_RTC_DRV_EP93XX is disabled.

Thanks! That must be what I was missing!

Now that 2.6.34 has been released, I've uploaded the minimal ("make
simone_defconfig") kernel image as well as the Debian-configured
version and its modules to the google code downloads page. It seems
to work OK with the existing debian-lenny-rootfs.

The only thing it's missing now is the audio driver.

M

Reply all
Reply to author
Forward
0 new messages