Difference between revisions of "Panda How to kernel 3 0 rcx"
KenMcGuire (talk | contribs) |
KenMcGuire (talk | contribs) |
||
Line 29: | Line 29: | ||
The oops occurs when the EHCI driver tries to access the hardware which has no power. | The oops occurs when the EHCI driver tries to access the hardware which has no power. | ||
+ | |||
+ | There is also an issue with the setting up of the wl12xx: | ||
+ | |||
+ | [ 0.181457] error setting wl12xx data | ||
+ | |||
+ | Which is a problem new to 3.0-rc1, still present in linux-omap-2.6.git. | ||
Revision as of 08:34, 1 June 2011
Contents
Introduction
The 3.0-rc1 kernel has been released.
rc1
There have been regressions since 2.6.39.
Building
Building 3.0-rc1 is basically the same as How to build 2.6.38 kernel, but of course you need the -rc1 sources. The omap2plus_defconfig works once the mmc ptch is applied. Enabling USB results in very ugly kernel oopses.
MMC Patch
A patch is available which addresses the mmc issue noted above:
0002-omap4-pandaboard-fix-mmc-support.patch
Apply it like so: (from inside the kernel sources directory)
patch -p1 < 0002-omap4-pandaboard-fix-mmc-support.patch
This issue has been fixed in linux-omap-2.6.git, but the oopses are still present when EHCI is enabled in .config.
It looks like they are related to VUSB power not being set up correctly as evidenced by:
[ 0.243896] machine_constraints_voltage: VUSB: failed to apply 3300000uV constraint [ 0.245208] twl_reg twl_reg.46: can't register VUSB, -22 [ 0.245239] twl_reg: probe of twl_reg.46 failed with error -22
which occurs after the board-omap4panda.c file init's the various regulators, and happens whether the EHCI driver is selected in the .config or not.
The oops occurs when the EHCI driver tries to access the hardware which has no power.
There is also an issue with the setting up of the wl12xx:
[ 0.181457] error setting wl12xx data
Which is a problem new to 3.0-rc1, still present in linux-omap-2.6.git.
DVI Patch
Apply it like so: (from inside the kernel sources directory)
patch -p1 < 0001-omap4-pandaboard-fix-dvi-support.patch
Then use either of the following configs
make ARCH=arm omap2plus_defconfig or make ARCH=arm panda_dvi_defconfig
Then compile with
make ARCH=arm CROSS_COMPILE=Path_to_your/arm-2009q3/bin/arm-none-linux-gnueabi- uImage or make ARCH=arm CROSS_COMPILE=Path_to_your/arm-2010q1/bin/arm-none-linux-gnueabi- uImage
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
I2C Patch
Apply it like so: (from inside the kernel sources directory)
patch -p1 < Omap44xx-i2c-fix.patch
It applies with some fuzzes and offsets, but does apply successfuly. Then compile as above.
Testing: No testing possible yet till the booting problems are resolved
After booting run fbtest to see a nice test pattern from the dvi port.
Then 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: -- -- -- -- -- -- -- --
Stay tuned as the 3.0-rcx series evolves, where there needs to be some serious bug squashing.