N8x0 to mainline patches

N8x0 to mainline patches

[Kalle Valo]

Hello,

I started working on adding N8x0 support to mainline, currently serial
console works and I can mount initfs and rootfs on N800. The patches are
available here:

http://www.valot.fi/kalle/n8x0/patches/n8x0-add/

To get all this coordinated, I created a wiki page for this:

http://elinux.org/N800

If there's something wrong or missing, please update it or send me an
email so that I can update it.

-- 
Kalle Valo

Re: N8x0 to mainline patches

[Felipe Balbi]

Hi,

On Mon, Jul 13, 2009 at 07:51:10AM +0200, ext Kalle Valo wrote:
> Hello,
> 
> I started working on adding N8x0 support to mainline, currently serial
> console works and I can mount initfs and rootfs on N800. The patches are
> available here:
> 
> http://www.valot.fi/kalle/n8x0/patches/n8x0-add/

I'd say at least patch #2 should be sent for to l-o right now since it
gets rid of another OMAP_TAG.

All the other patches (besides #1) look fine and I guess you could send
all of them here for people to take a look at.

But patch #2 clearly doesn't belong to that series and should be
integrated

-- 
balbi

Re: N8x0 to mainline patches

[Kalle Valo]

Felipe Balbi <felipe.balbi@nokia.com> writes:

> Hi,
>
> On Mon, Jul 13, 2009 at 07:51:10AM +0200, ext Kalle Valo wrote:
>> Hello,
>> 
>> I started working on adding N8x0 support to mainline, currently serial
>> console works and I can mount initfs and rootfs on N800. The patches are
>> available here:
>> 
>> http://www.valot.fi/kalle/n8x0/patches/n8x0-add/
>
> I'd say at least patch #2 should be sent for to l-o right now since it
> gets rid of another OMAP_TAG.

I didn't implement omap1 part yet. As soon as I have done that, I'll
send it.

> All the other patches (besides #1) 

Yeah, patch #1 is some weird linker magic which I need to look at more.
The patch is in the set just to get things started and I'll remove it
from the final patchset.

> look fine and I guess you could send all of them here for people to
> take a look at.

Ok, I'll do that soon but I'll try to implement a bit more to the board
file first.

> But patch #2 clearly doesn't belong to that series and should be
> integrated

I agree. But isn't Tony on vacation the whole July and there is nobody
to push it to linux-omap?

Thanks for looking at the patches!

-- 
Kalle Valo

Re: N8x0 to mainline patches

[Felipe Balbi]

Hi,

On Mon, Jul 13, 2009 at 08:51:45AM +0200, ext Kalle Valo wrote:
> > I'd say at least patch #2 should be sent for to l-o right now since it
> > gets rid of another OMAP_TAG.
> 
> I didn't implement omap1 part yet. As soon as I have done that, I'll
> send it.

ok, good :-)

> I agree. But isn't Tony on vacation the whole July and there is nobody
> to push it to linux-omap?
> 
> Thanks for looking at the patches!

Yes, he is indeed. Good to see we are really getting rid of omap tags
:-)

-- 
balbi

Re: N8x0 to mainline patches

[Luke-Jr]

I noticed that the patch adding board-n8x0.c seems to be much smaller than the 
older one used in Linux-OMAP. In particular, it lacks the keymaps and such. 
Are those somewhere else now? Was the Linux-OMAP keyboard bug taken into 
consideration for the new keyboard config struct?

Luke

Re: N8x0 to mainline patches

[green]

[-- Attachment #1: Type: text/plain, Size: 462 bytes --]

Kalle Valo wrote at 2009-07-13 00:51 -0500:
> I started working on adding N8x0 support to mainline, currently serial
> console works and I can mount initfs and rootfs on N800. The patches are
> available here:
> 
> http://www.valot.fi/kalle/n8x0/patches/n8x0-add/

Thanks for your work.

> To get all this coordinated, I created a wiki page for this:
> 
> http://elinux.org/N800

I have added a links section and will probably be doing more there.

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 197 bytes --]

Re: N8x0 to mainline patches

[Kalle Valo]

"Luke-Jr" <luke@dashjr.org> writes:

> I noticed that the patch adding board-n8x0.c seems to be much smaller
> than the older one used in Linux-OMAP. In particular, it lacks the
> keymaps and such. Are those somewhere else now?

It's just not implemented yet. If you look at the plan, you see lots of
things we need to implement:

http://www.elinux.org/N800#The_Plan

> Was the Linux-OMAP keyboard bug taken into consideration for the new
> keyboard config struct?

I'm not familiar with the issue. But if it makes sense, sure, we want to
include the fix.

-- 
Kalle Valo

Re: N8x0 to mainline patches

[Arnaud Ebalard]

Hi,

Kalle Valo <kalle.valo@iki.fi> writes:

> "Luke-Jr" <luke@dashjr.org> writes:
>
>> I noticed that the patch adding board-n8x0.c seems to be much smaller
>> than the older one used in Linux-OMAP. In particular, it lacks the
>> keymaps and such. Are those somewhere else now? 
>
> It's just not implemented yet. If you look at the plan, you see lots
> of things we need to implement:
>
> http://www.elinux.org/N800#The_Plan

Just noticed the thread and the link today. I will try and merge some
notes I have on the topic (http://natisbad.org/N810/) tomorrow.


>> Was the Linux-OMAP keyboard bug taken into consideration for the new
>> keyboard config struct?
>
> I'm not familiar with the issue. But if it makes sense, sure, we want
> to include the fix.

I use the patch on a 2.6.30-rc8-omap1 based l-o commit on my N810 and it
fixes the reported issue (missing keys).

BTW, I also use the following patch to be able to use the fn key with
xorg. One would argue this would be better to handle the issue at
another level but it is worth discussing considering what I read on
various ML:

http://lists.freedesktop.org/archives/hal/2009-January/012860.html
http://lists.freedesktop.org/archives/hal/2009-January/012868.html
http://lists.freedesktop.org/archives/hal/2009-January/012794.html
http://www.gossamer-threads.com/lists/maemo/developers/35105,

When the rx44 keymap support is added back in your quilt set, can we
discuss it. IIRC, the same kind of value (i.e. > 255) is also in the
rx-51. Any specific reason? 

Cheers,

a+

Index: linux-omap-2.6/arch/arm/mach-omap2/board-n800.c
===================================================================
--- linux-omap-2.6.orig/arch/arm/mach-omap2/board-n800.c	2009-06-20 12:07:11.000000000 +0200
+++ linux-omap-2.6/arch/arm/mach-omap2/board-n800.c	2009-06-27 20:29:00.000000000 +0200
@@ -85,7 +85,11 @@
 	[0x22] = KEY_SEMICOLON,
 	[0x23] = KEY_MINUS,
 	[0x24] = KEY_EQUAL,
+#ifdef CONFIG_REMAP_N810_FN_KEY
+	[0x2b] = KEY_LEFTALT,
+#else
 	[0x2b] = KEY_FN,
+#endif
 	[0x2c] = KEY_M,
 	[0x2f] = KEY_F8,
 
Index: linux-omap-2.6/drivers/input/keyboard/Kconfig
===================================================================
--- linux-omap-2.6.orig/drivers/input/keyboard/Kconfig	2009-06-20 12:07:11.000000000 +0200
+++ linux-omap-2.6/drivers/input/keyboard/Kconfig	2009-06-20 12:07:35.000000000 +0200
@@ -294,6 +294,20 @@
 	  If you say yes here you get support for the National Semiconductor
 	  LM8323 keypad controller.
 
+config REMAP_N810_FN_KEY
+       boolean "Nokia N810 keyboard Fn key remapping"
+       depends on MACH_NOKIA_N810 && KEYBOARD_LM8323
+       help
+          If you build your kernel for your Nokia N810 (which would
+	  possibly explain you selected LM8323 keypad chip above),
+	  you might want to select this option in order to have the
+	  Fn key (KEY_FN) advertise itself as left alt key
+          (KEY_LEFTALT).
+
+	  This is because KEY_FN value is higher than 255, and Xorg
+	  basically drops those values. Selecting the option allows
+	  you to use the key under X
+
 config KEYBOARD_PXA27x
 	tristate "PXA27x/PXA3xx keypad support"
 	depends on PXA27x || PXA3xx

Re: N8x0 to mainline patches

[Luke-Jr]

> When the rx44 keymap support is added back in your quilt set, can we
> discuss it. IIRC, the same kind of value (i.e. > 255) is also in the
> rx-51. Any specific reason?

I presume the reason is that the "standard" code for the "Fn" key is 464... so 
it is only correct to map "Fn" to KEY_FN. ;)

There is a software Fn key on some Mac laptops as well. Do they work? Why/how?

Note: before I installed KDE 4 on my N810, Fn was working in X with its usual 
keycode...

Re: N8x0 to mainline patches

[Arnaud Ebalard]

Hi,

"Luke-Jr" <luke@dashjr.org> writes:

>> When the rx44 keymap support is added back in your quilt set, can we
>> discuss it. IIRC, the same kind of value (i.e. > 255) is also in the
>> rx-51. Any specific reason?
>
> I presume the reason is that the "standard" code for the "Fn" key is 464... so 
> it is only correct to map "Fn" to KEY_FN. ;)

I know, but from the document I read on the topic, xorg does not support
keycode value higher than 255. I must definitely be missing something.

> There is a software Fn key on some Mac laptops as well. Do they work? Why/how?

My old laptop was a powerbook 12" running Debian and that key never
worked, IIRC.

> Note: before I installed KDE 4 on my N810, Fn was working in X with its usual 
> keycode...

If you manage to find out why/how it worked at some point, I am
interested.

Cheers,

a+

Re: N8x0 to mainline patches

[Tony Lindgren]

* Felipe Balbi <felipe.balbi@nokia.com> [090713 09:54]:
> Hi,
> 
> On Mon, Jul 13, 2009 at 08:51:45AM +0200, ext Kalle Valo wrote:
> > > I'd say at least patch #2 should be sent for to l-o right now since it
> > > gets rid of another OMAP_TAG.
> > 
> > I didn't implement omap1 part yet. As soon as I have done that, I'll
> > send it.
> 
> ok, good :-)
> 
> > I agree. But isn't Tony on vacation the whole July and there is nobody
> > to push it to linux-omap?
> > 
> > Thanks for looking at the patches!
> 
> Yes, he is indeed. Good to see we are really getting rid of omap tags
> :-)

Hey, nice job cleaning up the n8x0 board support! Let's get those into
l-o master and queued up for the next merge window. (Except for the
hacks that is).

Tony

Re: N8x0 to mainline patches

[Kalle Valo]

Tony Lindgren <tony@atomide.com> writes:

> Hey, nice job cleaning up the n8x0 board support! Let's get those into
> l-o master and queued up for the next merge window. (Except for the
> hacks that is).

What did you think of the file layout? I decided to go with one file
approach (against your advice, sorry) because I didn't see any benefit
from having separate files.

My plan is this:

1. first finish OMAP_TAG_UART removal patch and send it for review

2. finins and send tusb6010 patch

+ omap2-add-usb-tusb6010-c

2. fold these n8x0 board file patches into one and send it for review:

+ n8x0-add-board-file
+ n8x0-onenand-support
+ n8x0-add-spi-support
+ n8x0-add-usb-support

Because the merge window is getting close, I'll do this ASAP.

There's a lot of to do still, but it's to get basic support to mainline
already now, makes it easier for us to implement the missing pieces.

-- 
Kalle Valo

Re: N8x0 to mainline patches

[Kalle Valo]

On Thu, Aug 6, 2009 at 7:21 PM, Kalle Valo<kalle.valo@iki.fi> wrote:
> My plan is this:
>
> 1. first finish OMAP_TAG_UART removal patch and send it for review
>
> 2. finins and send tusb6010 patch
>
> + omap2-add-usb-tusb6010-c
>
> 2. fold these n8x0 board file patches into one and send it for review:
>
> + n8x0-add-board-file
> + n8x0-onenand-support
> + n8x0-add-spi-support
> + n8x0-add-usb-support

This was item 3, of course. And I forget n8x0_defconfig, I'll send
that as a separate patch.

Tony, does this look ok to you?

Kalle

Re: N8x0 to mainline patches

[Tony Lindgren]

* Kalle Valo <kalle.valo@iki.fi> [090806 19:29]:
> On Thu, Aug 6, 2009 at 7:21 PM, Kalle Valo<kalle.valo@iki.fi> wrote:
> > My plan is this:
> >
> > 1. first finish OMAP_TAG_UART removal patch and send it for review

This would be great.

> > 2. finins and send tusb6010 patch
> >
> > + omap2-add-usb-tusb6010-c

The tusb6010 could be added later on as a fix too if you want to cut
down dependencies.

> > 2. fold these n8x0 board file patches into one and send it for review:
> >
> > + n8x0-add-board-file
> > + n8x0-onenand-support
> > + n8x0-add-spi-support
> > + n8x0-add-usb-support

Yeh one patch for the board file.
 
> This was item 3, of course. And I forget n8x0_defconfig, I'll send
> that as a separate patch.

Yeah it's easier to read the series that way.
 
> Tony, does this look ok to you?

Looks good to me!

Tony

Re: N8x0 to mainline patches

[Tony Lindgren]

* Kalle Valo <kalle.valo@iki.fi> [090806 19:23]:
> Tony Lindgren <tony@atomide.com> writes:
> 
> > Hey, nice job cleaning up the n8x0 board support! Let's get those into
> > l-o master and queued up for the next merge window. (Except for the
> > hacks that is).
> 
> What did you think of the file layout? I decided to go with one file
> approach (against your advice, sorry) because I didn't see any benefit
> from having separate files.

Your one file approach looks good to me.

Tony

Re: N8x0 to mainline patches

[Kalle Valo]

Tony Lindgren <tony@atomide.com> writes:

>> > 2. finins and send tusb6010 patch
>> >
>> > + omap2-add-usb-tusb6010-c
>
> The tusb6010 could be added later on as a fix too if you want to cut
> down dependencies.

Good point, I'll send that separately.

Current n8x0 patches are here:

http://www.valot.fi/kalle/n8x0/patches/n8x0-add/2009-08-06-19:21/

They are just pending final testing and I'll send them after I have done
that, hopefully tonight.

Tony, on which tree should I base the patches? They are now against
linus's tree.

-- 
Kalle Valo

Re: N8x0 to mainline patches

[Tony Lindgren]

* Kalle Valo <kalle.valo@iki.fi> [090807 10:44]:
> Tony Lindgren <tony@atomide.com> writes:
> 
> >> > 2. finins and send tusb6010 patch
> >> >
> >> > + omap2-add-usb-tusb6010-c
> >
> > The tusb6010 could be added later on as a fix too if you want to cut
> > down dependencies.
> 
> Good point, I'll send that separately.
> 
> Current n8x0 patches are here:
> 
> http://www.valot.fi/kalle/n8x0/patches/n8x0-add/2009-08-06-19:21/
> 
> They are just pending final testing and I'll send them after I have done
> that, hopefully tonight.
> 
> Tony, on which tree should I base the patches? They are now against
> linus's tree.

Great, yeah it's best to base all the patches against Linus' tree where
possible.

I'll start piling up the omap-upstream queue for the merge window and
will add your serial patch there. So we still have few more days to
deal with the board file patch.

Tony