Skip to contentSkip to conversation

N810: latest linux-omap-2.6.git master

DISCUSSION Luke-Jr 6 participants 14 messages May 9, 2009 – May 23, 2009

lore
Share & tools
lore

Luke-Jr

author
#0 lore raw

Having trouble getting "2.6.30-rc4-omap1" to boot my US N810 (running Gentoo). The initial problem appears to be related to /dev/mmcblk* missing, but I can't debug it very far due to the second problem I encountered:
The keys "fghjvbnm" do not appear to work. I added debug info to print when the keyboard FIFO is empty, and also when it has a valid mapping, but these keys do not appear to be triggering the 'process_keys' function at all.

Any suggestions?

Thanks,

Luke

#1 lore raw
Quoted 3 lines · On Sat, 9 May 2009 17:45:00 -0500 "Luke-Jr" <[email protected]> wrote:

Having trouble getting "2.6.30-rc4-omap1" to boot my US N810 (running Gentoo). The initial problem appears to be related to /dev/mmcblk* missing, but I can't debug it very far due to the second problem I

For me mmc is working. I'm booting through the modified initfs with official major number 179 for /dev/mmcblk* device nodes.

Quoted 5 lines

encountered: The keys "fghjvbnm" do not appear to work. I added debug info to print when the keyboard FIFO is empty, and also when it has a valid mapping, but these keys do not appear to be triggering the 'process_keys' function at all.

I see this problem as well. I think this wasn't here just couple weeks ago after Tony's code reset to mainline patches but I'm not sure was it still 2.6.29 when I last time used the keyboard.

Signature · 1 line
Jarkko
#2 lore raw
Quoted 9 lines · Jarkko Nikula

On Sat, 9 May 2009 17:45:00 -0500

Quoted 3 lines · Luke-Jr

Having trouble getting "2.6.30-rc4-omap1" to boot my US N810 (running Gentoo). The initial problem appears to be related to /dev/mmcblk* missing, but I can't debug it very far due to the second problem I

For me mmc is working. I'm booting through the modified initfs with official major number 179 for /dev/mmcblk* device nodes.

Modified initfs? Is there something needed beyond copying the filesystem's modules?

Luke

#3 lore raw
Quoted 6 lines · On Tue, 12 May 2009 18:18:13 -0500 Luke Dashjr <[email protected]> wrote:
Quoted 2 lines

For me mmc is working. I'm booting through the modified initfs with official major number 179 for /dev/mmcblk* device nodes.

Modified initfs? Is there something needed beyond copying the filesystem's modules?

Actually I'm not using modules from initfs at all nor updating them but I have compiled the ext2 in so that initfs can mount the ext2 formatted root from mmc (flasher --set-root-device=mmc).

This device node update to initfs is required since Maemo 2.6.21 kernel & initfs are using major number 254 for mmcblk whereas newer kernels are using 179.

I basically copied initfs from running system or fiasco image, I don't remember, modified it on desktop and then created new jffs2 image out of it. That can be then flashed with "flasher -f -n initfs.jffs2".

I think it would be possible to boot modified initfs also from another mmc partition so that initfs on NAND can be intact.

Signature · 1 line
Jarkko
author
#4 lore raw
Quoted 3 lines · Jarkko Nikula

This device node update to initfs is required since Maemo 2.6.21 kernel & initfs are using major number 254 for mmcblk whereas newer kernels are using 179.

Yes, that change seemed to get me booting... however, with "rc6", the keyboard no longer works *at all*, and in neither rc4 nor rc6 does p54spi actually work.

Tried setting up g_ether, but for some reason I couldn't get it to auto- configure (via Gentoo's boot scripts) for rc6.

Is anyone actually trying to work on or maintain N8x0 support?

#5 lore raw
Quoted 4 lines · On Tue, 19 May 2009 00:09:20 -0500 "Luke-Jr" <[email protected]> wrote:

Yes, that change seemed to get me booting... however, with "rc6", the keyboard no longer works *at all*, and in neither rc4 nor rc6 does p54spi actually work.

For me keyboard still works as partially (commit 0b2ce840c9fb58cf489bfdfc2d77f99a09a5dca3). I wonder what have caused it to break during 2.6.30-rc since at quick look I don't point any special from Tony's "Extra omap code in linux-omap tree" patches and patches before those are from February?

git whatchanged arch/arm/mach-omap2/board-n800.c
git whatchanged drivers/input/keyboard/lm8323.c

Quoted 3 lines

Tried setting up g_ether, but for some reason I couldn't get it to auto- configure (via Gentoo's boot scripts) for rc6.

For me it works but usually only if cable has been plugged in while booting. Same with Beagle as well. Any idea from MUSB guys?

Quoted 1 line

Is anyone actually trying to work on or maintain N8x0 support?

I think N8x0 support is a little bit behind from other boards. Lot of active development is going around on newer and more open platforms like Beagle and not all patches are necessary tested on N8x0 (or other boards too!). Of course active hacker base helps to point out breakages.

Signature · 1 line
Jarkko
#6 lore raw
Quoted 7 lines · Jarkko Nikula
Quoted 4 lines · On Tue, 19 May 2009 00:09:20 -0500 "Luke-Jr" <[email protected]> wrote:

Yes, that change seemed to get me booting... however, with "rc6", the keyboard no longer works *at all*, and in neither rc4 nor rc6 does p54spi actually work.

Quoted 8 lines

For me keyboard still works as partially (commit 0b2ce840c9fb58cf489bfdfc2d77f99a09a5dca3). I wonder what have caused it to break during 2.6.30-rc since at quick look I don't point any special from Tony's "Extra omap code in linux-omap tree" patches and patches before those are from February?

git whatchanged arch/arm/mach-omap2/board-n800.c
git whatchanged drivers/input/keyboard/lm8323.c

While waiting to get things in sync with mainline, something like this might be handy too:

$ git diff omap-2.6.28..master -- arch/arm/mach-omap2/board-n800.c
Quoted 6 lines
Quoted 3 lines

Tried setting up g_ether, but for some reason I couldn't get it to auto- configure (via Gentoo's boot scripts) for rc6.

For me it works but usually only if cable has been plugged in while booting. Same with Beagle as well. Any idea from MUSB guys?

Quoted 6 lines
Quoted 1 line

Is anyone actually trying to work on or maintain N8x0 support?

I think N8x0 support is a little bit behind from other boards. Lot of active development is going around on newer and more open platforms like Beagle and not all patches are necessary tested on N8x0 (or other boards too!). Of course active hacker base helps to point out breakages.

We should try to get the board-n8x0 stuff in shape for mainline maybe the next merge window after this, anybody willing to spend some cycles on preparing the patches against the mainline tree?

Tony

#7 lore raw
Quoted 10 lines · Tony Lindgren
Quoted 6 lines
Quoted 1 line

Is anyone actually trying to work on or maintain N8x0 support?

I think N8x0 support is a little bit behind from other boards. Lot of active development is going around on newer and more open platforms like Beagle and not all patches are necessary tested on N8x0 (or other boards too!). Of course active hacker base helps to point out breakages.

We should try to get the board-n8x0 stuff in shape for mainline maybe the next merge window after this, anybody willing to spend some cycles on preparing the patches against the mainline tree?

So that's something like three months from now? Hmm. If someone else doesn't volunteer I can try to spend some time on this. I would love to see n800 and n810 supported in mainline.

But I cannot make any promises, though. It would be great if someone else can step up.

Signature · 1 line
Kalle Valo
#8 lore raw
Quoted 13 lines · Kalle Valo
Quoted 10 lines · Tony Lindgren
Quoted 6 lines
Quoted 1 line

Is anyone actually trying to work on or maintain N8x0 support?

I think N8x0 support is a little bit behind from other boards. Lot of active development is going around on newer and more open platforms like Beagle and not all patches are necessary tested on N8x0 (or other boards too!). Of course active hacker base helps to point out breakages.

We should try to get the board-n8x0 stuff in shape for mainline maybe the next merge window after this, anybody willing to spend some cycles on preparing the patches against the mainline tree?

Quoted 6 lines

So that's something like three months from now? Hmm. If someone else doesn't volunteer I can try to spend some time on this. I would love to see n800 and n810 supported in mainline.

But I cannot make any promises, though. It would be great if someone else can step up.

OK, I can help a bit too. Here's what I have in mind we should do:

- Rename the board-*c files to board-n8x0-*.c. I can do this assuming people are OK with this.

- Get the gpio-switch replacement completed and into mainline. Jani, any news on that? There's a threaad about that at:
http://thread.gmane.org/gmane.linux.kernel/809866

- Stop using omap specific ATAGs and pass all the gpio etc info from platform_data. This should be easy with CONFIG_ATAGS_PROC that shows the passed ATAGs in /proc. I can work on this too.

- Get drivers/cbus into mainline. Kalle, do you want to do this?

I guess there's some other clean-up needed too that we can all work once we have the initial patches working against the mainline.

Regards,

Tony

#9 lore raw
Quoted 2 lines · On Wed, 20 May 2009 08:13:02 -0700 Tony Lindgren <[email protected]> wrote:

OK, I can help a bit too. Here's what I have in mind we should do:

I can participate at least to testing if nothing else.

Quoted 2 lines

- Get drivers/cbus into mainline. Kalle, do you want to do this?

I can be part of process how to remove drivers/cbus/retu-headset.c and integrate that functionality with sound/soc/omap/n810.c.

Signature · 1 line
Jarkko
#10 lore raw
Quoted 4 lines
Quoted 1 line · Tony Lindgren

OK, I can help a bit too. Here's what I have in mind we should do:

I can participate at least to testing if nothing else.

So am I. Testing and bughunting are what I can do best. Or development of small comprehensible features.

Thanks.
-- Max

#11 lore raw
In reply to Tony Lindgren
Quoted 9 lines · Tony Lindgren

OK, I can help a bit too. Here's what I have in mind we should do:

- Rename the board-*c files to board-n8x0-*.c. I can do this assuming people are OK with this.

- Get the gpio-switch replacement completed and into mainline. Jani, any news on that? There's a threaad about that at:
http://thread.gmane.org/gmane.linux.kernel/809866

Quoted 5 lines

- Stop using omap specific ATAGs and pass all the gpio etc info from platform_data. This should be easy with CONFIG_ATAGS_PROC that shows the passed ATAGs in /proc. I can work on this too.

- Get drivers/cbus into mainline. Kalle, do you want to do this?

Sure, I can work with cbus. I can't give any promises how long it will take, so don't have your hopes up :)

Quoted 2 lines

I guess there's some other clean-up needed too that we can all work once we have the initial patches working against the mainline.

Yes, I agree.

For reference, we discussed a bit about this last September:

http://www.spinics.net/lists/linux-omap/msg05505.html

Signature · 1 line
Kalle Valo
#12 lore raw
In reply to Tony Lindgren

Tony Lindgren wrote at 2009-05-20 09:13 -0600:

Quoted 15 lines · Kalle Valo
Quoted 14 lines
Quoted 8 lines · Tony Lindgren
Quoted 4 lines

I think N8x0 support is a little bit behind from other boards. Lot of active development is going around on newer and more open platforms like Beagle and not all patches are necessary tested on N8x0 (or other boards too!). Of course active hacker base helps to point out breakages.

We should try to get the board-n8x0 stuff in shape for mainline maybe the next merge window after this, anybody willing to spend some cycles on preparing the patches against the mainline tree?

So that's something like three months from now? Hmm. If someone else doesn't volunteer I can try to spend some time on this. I would love to see n800 and n810 supported in mainline.

Quoted 7 lines
Quoted 2 lines

But I cannot make any promises, though. It would be great if someone else can step up.

OK, I can help a bit too. Here's what I have in mind we should do:

I guess there's some other clean-up needed too that we can all work once we have the initial patches working against the mainline.

It will be great to see some movement forward for the n8x0! I can't do much with the skills and time I have but I will be testing when I can, especially if I can somehow get booting from the internal mmc to work again. I have been working with that for a while (with basically no knowledge about C) but have not really accomplised much, except learn more about git.

#13 lore raw
In reply to Luke-Jr
Quoted 3 lines

Yes, that change seemed to get me booting... however, with "rc6", the keyboard no longer works *at all*, and in neither rc4 nor rc6 does p54spi actually work.

For p54spi to minimally work you need at least the following patches from linux-wireless:

320abdf21a27e9f662f70d7726f56cd458343d3b
b4668d4eb59e0bc89059ba258c4aa278dcc3f872

For it to work good, all patches to drivers/net/wireless/p54 from linux- wireless are required.

Thanks.
-- Max