Difference between revisions of "Panda How to kernel 3 1 rcx"
KenMcGuire (talk | contribs) |
KenMcGuire (talk | contribs) |
||
Line 17: | Line 17: | ||
== rc2 == | == rc2 == | ||
− | Yep, -rc2 is out and I haven't had the time to fiddle with it as yet. | + | Yep, -rc2 is out and I haven't had the time to fiddle much with it as yet, on to -rc3. |
− | + | In short, -rc2 fixes some of the wl12xx issues, but still requires a patch to function. | |
+ | There is still an unresolved start up issue with wl12xx, in that sometimes it will work, sometimes, not. Same kernel, MLO, & u-boot and userspace. | ||
== rc3 == | == rc3 == | ||
− | -rc3 is out and there are no OMAP or Panda changes | + | -rc3 is out and there are no OMAP or Panda changes. However there is a boot issue with Panda |
− | + | as described here: https://lkml.org/lkml/2011/8/25/117 | |
− | + | ||
+ | This issue has been resolved, so be sure to get this commit or later from git: | ||
+ | http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=commit;h=69dd3d8e29e294caaf63eb5e8a72d250279f9e5f | ||
+ | |||
Revision as of 10:35, 25 August 2011
Contents
Introduction
Continuing Happy 20th Birthday Greetings to Linux!! The merge window for 3.1 is over and the 3.1-rc1 kernel has been released. Lots of work on the omap platforms and on pandaboard. Let's see what -rc1 has brought.
rc1
There is still an issue with the setting up of the wl12xx:
[ 0.181457] error setting wl12xx data
The extremely simple workaround available for 3.0, moved to twl_common.c in -rc1, doesn't produce the working wlan as in 3.0. But applying that to the -rc2 code does result in a functional wl12xx driver.
More to come as these issue are investigated.
rc2
Yep, -rc2 is out and I haven't had the time to fiddle much with it as yet, on to -rc3. In short, -rc2 fixes some of the wl12xx issues, but still requires a patch to function. There is still an unresolved start up issue with wl12xx, in that sometimes it will work, sometimes, not. Same kernel, MLO, & u-boot and userspace.
rc3
-rc3 is out and there are no OMAP or Panda changes. However there is a boot issue with Panda as described here: https://lkml.org/lkml/2011/8/25/117
This issue has been resolved, so be sure to get this commit or later from git: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=commit;h=69dd3d8e29e294caaf63eb5e8a72d250279f9e5f
DVI Patch
This patch is necessary to make 720p resolution available.
0003-omap4-pandaboard-dvi720p.patch
Apply it like so: (from inside the kernel sources directory)
patch -p1 < 0001-omap4-pandaboard-dvi720p.patch
Then use either of the following configs
make ARCH=arm omap2plus_defconfig or make ARCH=arm panda_dvi_defconfig
Compile as above.
I2C Patch
i2c character device driver
Ever since 2.6.38, the i2c character device driver support from user space has been broken for OMAP44xx. * A fix has been submitted for the linux-omap-2.6 branch, but it may be a while till it gets into an -rcx. See http://git.kernel.org/?p=linux/kernel/git/tmlind/linux-omap-2.6.git;a=commit;h=39fe1a6fafe1e85c183379af9f3ceda7cd24bd65 for the commit. * A quick fix for this issue is available File:Omap44xx-i2c-fix.patch
Apply it like so: (from inside the kernel sources directory)
patch -p1 < Omap44xx-i2c-fix.patch
Compile as above.
Building
Building 3.1-rc1 or -rc2 is basically the same as How to build 2.6.38 kernel, but of course you need the 3.1-rc1/rc2 sources. The omap2plus_defconfig works and illustrates the wl12xx issue, with or without a patch to twl_common.c. config.3.1-rc1.1 works with the dvi patch and i2c patch.
Testing
fbtest on DVI Port
After booting run fbtest to see a nice test pattern from the dvi port.
Switching primary display to the HDMI port
Make sure that a monitor is plugged into the HDMI port before doing the following:
# Enable HDMI echo "1" > /sys/devices/platform/omapdss/display1/enabled # Disable overlay0 (an overlay must be disabled before changing its properties) echo "0" > /sys/devices/platform/omapdss/overlay0/enabled # Set the manager of overlay0 to display1 which is HDMI echo "tv" > /sys/devices/platform/omapdss/overlay0/manager # Enable overlay0 echo "1" > /sys/devices/platform/omapdss/overlay0/enabled
And content on overlay 0 of primary lcd would be transferred to HDMI. It works similarly for all other overlay's.
Switching primary display to the DVI port
See: http://omappedia.org/wiki/Bootargs_for_enabling_display for lots of useful info on the display subsystem. Be aware that the display, manager and overlay numbers don't match the panda configuration.
Make sure that a monitor is plugged into the DVI port before doing the following:
# Disable HDMI echo "0" > /sys/devices/platform/omapdss/display1/enabled # Disable overlay0 (an overlay must be disabled before changing its properties) echo "0" > /sys/devices/platform/omapdss/overlay0/enabled # Set the manager of overlay0 to display0 which is DVI echo "lcd2" > /sys/devices/platform/omapdss/overlay0/manager # Enable overlay0 echo "1" > /sys/devices/platform/omapdss/overlay0/enabled
The above commands should be run from the serial console and the cable should be in the destination port before running the commands.
fbtest on HDMI Port
Run fbtest to see a nice test pattern from the HDMI port.
i2cdetect
You can run i2cdetect and the results should look like this:
# i2cdetect -y -r 1 0 1 2 3 4 5 6 7 8 9 a b c d e f 00: -- -- -- -- -- -- -- -- -- -- -- -- -- 10: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 20: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 30: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 40: -- -- -- -- -- -- -- -- UU UU UU UU -- -- -- -- 50: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 60: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 70: -- -- -- -- -- -- -- --