Please note that User Registration has been temporarily disabled due to a recent increase in automated registrations. If anyone needs an account, please request one here: RequestAccount. Thanks for your patience!--Wmat (talk)
Please email User:Wmat if you experience any issues with the Request Account form.

Difference between revisions of "Panda How to kernel 3 1 rcx"

From eLinux.org
Jump to: navigation, search
 
(9 intermediate revisions by the same user not shown)
Line 3: Line 3:
 
The merge window for 3.1 is over and the 3.1-rc1 kernel has been released.
 
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
 
Lots of work on the omap platforms and on pandaboard. Let's see
what -rc1 has brought. (This is a still a WIP though, so beware of taking any of this on face value)
+
what -rc1 has brought.
 
+
  
 
== rc1 ==
 
== rc1 ==
Line 12: Line 11:
 
  [    0.181457] error setting wl12xx data
 
  [    0.181457] error setting wl12xx data
  
The extremely simple workaround available for 3.0, moved to twl_common.c, doesn't produce the working wlan as in 3.0.
+
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.
 
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
 +
 +
== rc4 ==
 +
-rc4 has been released and yet again there are no OMAP or Panda changes. The issue introduced in -rc3 noted above, has been resolved.
 +
 +
If you build the wl12xx driver as a module, and modprobe it, the wlan seems to start up fine every time, must be some sort of race when it is compiled into the kernel.
 +
 +
== rc5 ==
 +
-rc5 has been released, and kernel.org is still sort of sick. Not to worry, Linus has a github repository that you can pull from (if you dare ;>),
 +
see https://lkml.org/lkml/2011/9/4/92 for details. From the changelog, it doesn't look like any omap4 or PandaBoard related changes, so we'll wait a bit to see if kernel.org recovers after the holiday.
 +
 +
== rc6 ==
 +
-rc6 has been released, and kernel.org is still down. I cloned Linus's github repo so as not to disturb my kernel.org repo (yes I know I could have pulled from the github repo, and with a few other magic git incantations resulted in the same thing...maybe...). There have been some OMAP4 related changes, but applying the same patches as -rc4, a working kernel as described below is the result.
 +
 +
== rc7 ==
 +
-rc7 has been released and Linus says that 3.1 won't happen till kernel.org is back, so we may have a few more -rcx's ahead of us.
 +
No OMAP or Panda related changes in -rc7, the same patches that worked for -rc4 still work (some with a bit of an offset).
 +
 +
== git.kerenl.org ==
 +
It's back, mostly, as of 10/3/2011. Stay tuned.
 +
 +
 +
== rc8 ==
 +
Skipped
 +
 +
== rc9 ==
 +
With git.kerne.org back, Linus has resumed the 3.1-rcx series on http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git
 +
 +
3.1 should be released soon, -rc9 is out which should be pretty close. The same patches that worked for -rc4 still work (some with a bit of an offset).
 +
 +
 +
== USB Performance improvement ==
 +
This patch fixs one performance bug on ARM Cortex A9 dual core
 +
platform, which has been reported on quite a few ARM machines
 +
(OMAP4, Tegra 2, snowball...), see details from link of
 +
https://bugs.launchpad.net/bugs/709245.
 +
 +
[[Media:0006-omap4-usb-improvement.patch|0006-omap4-usb-improvement.patch]]
 +
 +
Apply it like so: (from inside the kernel sources directory)
 +
 +
patch -p1 < 0006-omap4-usb-improvement.patch
 +
 +
== wlan12xx patch ==
 +
This patch is necessary to resolve the issue noted in 3.0 and 3.1-rc1 above, however the code has moved to twl_common.c and so an new patch is presented here.
 +
 +
[[Media:0001a-omap4-pandaboard-wlan-fix.patch|0001a-omap4-pandaboard-wlan-fix.patch]]
 +
 +
Apply it like so: (from inside the kernel sources directory)
 +
 +
patch -p1 < 0001a-omap4-pandaboard-wlan-fix.patch
  
 
== DVI Patch ==
 
== DVI Patch ==
Line 24: Line 85:
 
Apply it like so: (from inside the kernel sources directory)
 
Apply it like so: (from inside the kernel sources directory)
  
  patch -p1 < 0001-omap4-pandaboard-dvi720p.patch
+
  patch -p1 < 0003-omap4-pandaboard-dvi720p.patch
  
 
Then use either of the following configs
 
Then use either of the following configs
Line 40: Line 101:
 
  * A fix has been submitted for the linux-omap-2.6 branch, but it may be a while till it gets into an -rcx.
 
  * 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.
 
  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|Available]]
+
  * A quick fix for this issue is available [[Media:omap44xx-i2c-fix.patch|omap44xx-i2c-fix.patch]]
 +
 
 +
 
  
 
Apply it like so: (from inside the kernel sources directory)
 
Apply it like so: (from inside the kernel sources directory)
Line 49: Line 112:
  
 
== Building ==
 
== Building ==
Building 3.1-rc1 is basically the same as [[Panda_How_to_kernel_2_6_38|How to build 2.6.38 kernel]], but of course you need the 3.1-rc1 sources. The omap2plus_defconfig works and illustrates the wl12xx issue, with or without a patch to twl_common.c.  [[Media:config.3.1-rc1.1|config.3.1-rc1.1]] works with the dvi patch and i2c patch.
+
Building 3.1-rc1 or -rc2 is basically the same as [[Panda_How_to_kernel_2_6_38|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.  [[Media:config.3.1-rc1.1|config.3.1-rc1.1]] works with the dvi patch and i2c patch.
 +
 
 +
Building 3.1-rc4 -rc6/7 or -rc9 is a bit different if you want to have a working wlan.
 +
 
 +
In order to ensure that the wlan starts up consistently, it is recommended that the wl12xx driver be built as a module and started after the PandaBoard has booted.
 +
 
 +
Grab the 3.1-rc4 or -rc6/7/9 sources and use [[Media:config.3.1-rc4.1|config.3.1-rc4.1]], [[Media:config.3.1-rc6.1|config.3.1-rc6.1]], [[Media:config.3.1-rc7.1|config.3.1-rc7.1]] or [[Media:config.3.1-rc9.1|config.3.1-rc9.1]]as the .config (this requires you patch the dvi & i2c as above)
 +
 
 +
Then compile like so:
 +
 
 +
make ARCH=arm CROSS_COMPILE=Path_to_your/arm-2010q1/bin/arm-none-linux-gnueabi- uImage
 +
 
 +
Then compile the modules like so:
 +
 
 +
make ARCH=arm CROSS_COMPILE=/home/kenm/Panda/arm-2010q1/bin/arm-none-linux-gnueabi- modules
 +
 
 +
"Install" the modules to somewhere convenient:
 +
 
 +
make ARCH=arm CROSS_COMPILE=/home/kenm/Panda/arm-2010q1/bin/arm-none-linux-gnueabi- INSTALL_MOD_PATH=../testlib_omap modules_install
 +
 
 +
Copy lib/modules/3.1.0-rc4-dirty/, lib/modules/3.1.0-rc6-dirty/, lib/modules/3.1.0-rc7-dirty/ or lib/modules/3.1.0-rc9-dirty/ to your SD card (as root), boot up the Pandaboard.
 +
 
  
 
== Testing ==
 
== Testing ==
Line 116: Line 200:
 
  60: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --  
 
  60: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --  
 
  70: -- -- -- -- -- -- -- --
 
  70: -- -- -- -- -- -- -- --
 +
 +
=== wlan ===
 +
 +
Run the following commands after the PandaBoard is booted:
 +
 +
modprobe wl12xx_sdio
 +
ifconfig wlan0 up
 +
iwconfig wlan0 essid "Your AccessPoint Name"
 +
udhcpc -i wlan0
 +
 +
If your network is set up to provide DHCP services, the PandaBoard will get all the "right stuff(tm)" and you will be able to access the Internet.
 +
 +
# ping www.google.com
 +
PING www.google.com (74.125.73.99): 56 data bytes
 +
64 bytes from 74.125.73.99: seq=0 ttl=43 time=62.683 ms
 +
64 bytes from 74.125.73.99: seq=1 ttl=43 time=54.077 ms
 +
64 bytes from 74.125.73.99: seq=2 ttl=43 time=51.484 ms
 +
64 bytes from 74.125.73.99: seq=3 ttl=43 time=54.108 ms
 +
 +
=== USB Performance ===
 +
 +
Insert a USB memory stick into one of the usb ports
 +
 +
Run dmesg to see what sdx the stick was recognised as, then:
 +
 +
hdparam -tT /dev/sdx
 +
 +
If you did this on a kernel without the patch you will see some improvement, if you run the same command on a desktop Linux system, with the same USB memory stick, the PandaBoard's speed should roughly be the same.
 +
  
  

Latest revision as of 17:14, 31 March 2012

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

rc4

-rc4 has been released and yet again there are no OMAP or Panda changes. The issue introduced in -rc3 noted above, has been resolved.

If you build the wl12xx driver as a module, and modprobe it, the wlan seems to start up fine every time, must be some sort of race when it is compiled into the kernel.

rc5

-rc5 has been released, and kernel.org is still sort of sick. Not to worry, Linus has a github repository that you can pull from (if you dare ;>), see https://lkml.org/lkml/2011/9/4/92 for details. From the changelog, it doesn't look like any omap4 or PandaBoard related changes, so we'll wait a bit to see if kernel.org recovers after the holiday.

rc6

-rc6 has been released, and kernel.org is still down. I cloned Linus's github repo so as not to disturb my kernel.org repo (yes I know I could have pulled from the github repo, and with a few other magic git incantations resulted in the same thing...maybe...). There have been some OMAP4 related changes, but applying the same patches as -rc4, a working kernel as described below is the result.

rc7

-rc7 has been released and Linus says that 3.1 won't happen till kernel.org is back, so we may have a few more -rcx's ahead of us. No OMAP or Panda related changes in -rc7, the same patches that worked for -rc4 still work (some with a bit of an offset).

git.kerenl.org

It's back, mostly, as of 10/3/2011. Stay tuned.


rc8

Skipped

rc9

With git.kerne.org back, Linus has resumed the 3.1-rcx series on http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git

3.1 should be released soon, -rc9 is out which should be pretty close. The same patches that worked for -rc4 still work (some with a bit of an offset).


USB Performance improvement

This patch fixs one performance bug on ARM Cortex A9 dual core platform, which has been reported on quite a few ARM machines (OMAP4, Tegra 2, snowball...), see details from link of https://bugs.launchpad.net/bugs/709245.

0006-omap4-usb-improvement.patch

Apply it like so: (from inside the kernel sources directory)

patch -p1 < 0006-omap4-usb-improvement.patch

wlan12xx patch

This patch is necessary to resolve the issue noted in 3.0 and 3.1-rc1 above, however the code has moved to twl_common.c and so an new patch is presented here.

0001a-omap4-pandaboard-wlan-fix.patch

Apply it like so: (from inside the kernel sources directory)

patch -p1 < 0001a-omap4-pandaboard-wlan-fix.patch

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 < 0003-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 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.

Building 3.1-rc4 -rc6/7 or -rc9 is a bit different if you want to have a working wlan.

In order to ensure that the wlan starts up consistently, it is recommended that the wl12xx driver be built as a module and started after the PandaBoard has booted.

Grab the 3.1-rc4 or -rc6/7/9 sources and use config.3.1-rc4.1, config.3.1-rc6.1, config.3.1-rc7.1 or config.3.1-rc9.1as the .config (this requires you patch the dvi & i2c as above)

Then compile like so:

make ARCH=arm CROSS_COMPILE=Path_to_your/arm-2010q1/bin/arm-none-linux-gnueabi- uImage

Then compile the modules like so:

make ARCH=arm CROSS_COMPILE=/home/kenm/Panda/arm-2010q1/bin/arm-none-linux-gnueabi- modules

"Install" the modules to somewhere convenient:

make ARCH=arm CROSS_COMPILE=/home/kenm/Panda/arm-2010q1/bin/arm-none-linux-gnueabi- INSTALL_MOD_PATH=../testlib_omap modules_install 

Copy lib/modules/3.1.0-rc4-dirty/, lib/modules/3.1.0-rc6-dirty/, lib/modules/3.1.0-rc7-dirty/ or lib/modules/3.1.0-rc9-dirty/ to your SD card (as root), boot up the Pandaboard.


Testing

fbtest on DVI Port

After booting run fbtest to see a nice test pattern from the dvi port.

Fbtest.jpg

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.

Fbtest2.jpg

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: -- -- -- -- -- -- -- --

wlan

Run the following commands after the PandaBoard is booted:

modprobe wl12xx_sdio
ifconfig wlan0 up
iwconfig wlan0 essid "Your AccessPoint Name"
udhcpc -i wlan0

If your network is set up to provide DHCP services, the PandaBoard will get all the "right stuff(tm)" and you will be able to access the Internet.

# ping www.google.com
PING www.google.com (74.125.73.99): 56 data bytes
64 bytes from 74.125.73.99: seq=0 ttl=43 time=62.683 ms
64 bytes from 74.125.73.99: seq=1 ttl=43 time=54.077 ms
64 bytes from 74.125.73.99: seq=2 ttl=43 time=51.484 ms
64 bytes from 74.125.73.99: seq=3 ttl=43 time=54.108 ms

USB Performance

Insert a USB memory stick into one of the usb ports

Run dmesg to see what sdx the stick was recognised as, then:

hdparam -tT /dev/sdx

If you did this on a kernel without the patch you will see some improvement, if you run the same command on a desktop Linux system, with the same USB memory stick, the PandaBoard's speed should roughly be the same.