Drop last two remaining omap3 legacy board files

Let's drop the last two remaining omap3 legacy boot board files. Let's only use the device tree based booting known to work for these two boards.

We still have two omap3 boards booting in legacy mode in addition to device tree based booting. All the other ten or so omap2+ SoCs have been booting using device tree mode only for years now. This has allowed us to get rid of quite a bit of arch/arm/mach-omap2 related platform init code in favor of dts and driver changes.

Pretty much the only remaining known users for omap3 legacy boot board files are Kevin's and Russell's boot test systems, and N900 kernel tree maintained by Pali and Ivaylo. And all of them are also supporting the device tree based booting. The legacy booting mode has been kept around mostly to verify against regressions.

As there is still a slim chance of possible other uses of the mainline kernel for these boards, let's just drop the board files for v4.8, and let's not touch the related platform init code until around v4.9 time if no issues are found. This makes it trivial for us to add back the board files with a simple revert.

Signed-off-by Tony Lindgren <[email protected]>

---
 arch/arm/mach-omap2/Kconfig                  |    6 -
 arch/arm/mach-omap2/Makefile                 |    3 -
 arch/arm/mach-omap2/board-rx51-peripherals.c | 1317 --------------------------
 arch/arm/mach-omap2/board-rx51-video.c       |   67 --
 arch/arm/mach-omap2/board-rx51.c             |  141 ---
 arch/arm/mach-omap2/board-rx51.h             |   11 -
 6 files changed, 1545 deletions(-)
 delete mode 100644 arch/arm/mach-omap2/board-rx51-peripherals.c
 delete mode 100644 arch/arm/mach-omap2/board-rx51-video.c
 delete mode 100644 arch/arm/mach-omap2/board-rx51.c
 delete mode 100644 arch/arm/mach-omap2/board-rx51.h

Hi! I just quickly looked at N900 status matrix https://elinux.org/N900
and there is still one thing not finished in DT: It is slot-name for omap_hsmmc driver. So switch from legacy board code to DT is regression.

Hi,

Quoted 3 lines · Pali Roh?r

Hi! I just quickly looked at N900 status matrix https://elinux.org/N900
and there is still one thing not finished in DT: It is slot-name for omap_hsmmc driver. So switch from legacy board code to DT is regression.

That's fixed with the following in v4.7-rc series:

db863d896681 ("mmc: omap_hsmmc: Check if MMC slot name is passed in pdata") 10c1f7d32bf4 ("ARM: OMAP2+: n900 needs MMC slot names for legacy user space")

Quoted 7 lines · On Wednesday 15 June 2016 04:02:02 Tony Lindgren wrote:

Hi,

Quoted 3 lines · Pali Roh?r

Hi! I just quickly looked at N900 status matrix https://elinux.org/N900
and there is still one thing not finished in DT: It is slot-name for omap_hsmmc driver. So switch from legacy board code to DT is regression.

Quoted 4 lines

That's fixed with the following in v4.7-rc series:

db863d896681 ("mmc: omap_hsmmc: Check if MMC slot name is passed in pdata") 10c1f7d32bf4 ("ARM: OMAP2+: n900 needs MMC slot names for legacy user space")

Great! Sorry, but now wiki is not fully updated with last changes and I do not have in my mind what was finished/merged and what not. It is going hard for me to track everything.

Quoted 10 lines · Pali Roh?r

Quoted 9 lines · On Wednesday 15 June 2016 04:02:02 Tony Lindgren wrote:

Hi,

Quoted 3 lines · Pali Roh?r

Hi! I just quickly looked at N900 status matrix https://elinux.org/N900
and there is still one thing not finished in DT: It is slot-name for omap_hsmmc driver. So switch from legacy board code to DT is regression.

That's fixed with the following in v4.7-rc series:

Quoted 6 lines

Quoted 2 lines

db863d896681 ("mmc: omap_hsmmc: Check if MMC slot name is passed in pdata") 10c1f7d32bf4 ("ARM: OMAP2+: n900 needs MMC slot names for legacy user space")

Great! Sorry, but now wiki is not fully updated with last changes and I do not have in my mind what was finished/merged and what not. It is going hard for me to track everything.

OK well maybe take one more look and then ack if no more issues.

Quoted 14 lines · Tony Lindgren

Quoted 13 lines · Pali Roh?r

Quoted 11 lines · On Wednesday 15 June 2016 04:02:02 Tony Lindgren wrote:

Hi,

Quoted 3 lines · Pali Roh?r

Hi! I just quickly looked at N900 status matrix https://elinux.org/N900
and there is still one thing not finished in DT: It is slot-name for omap_hsmmc driver. So switch from legacy board code to DT is regression.

That's fixed with the following in v4.7-rc series:

db863d896681 ("mmc: omap_hsmmc: Check if MMC slot name is passed in pdata") 10c1f7d32bf4 ("ARM: OMAP2+: n900 needs MMC slot names for legacy user space")

Quoted 5 lines

Quoted 3 lines

Great! Sorry, but now wiki is not fully updated with last changes and I do not have in my mind what was finished/merged and what not. It is going hard for me to track everything.

OK well maybe take one more look and then ack if no more issues.

Pali, any news on this?

Quoted 19 lines · On Tuesday 21 June 2016 04:25:36 Tony Lindgren wrote:

Quoted 18 lines · Tony Lindgren

Quoted 16 lines · Pali Roh?r

Quoted 11 lines · On Wednesday 15 June 2016 04:02:02 Tony Lindgren wrote:

Hi,

Quoted 3 lines · Pali Roh?r

Hi! I just quickly looked at N900 status matrix https://elinux.org/N900
and there is still one thing not finished in DT: It is slot-name for omap_hsmmc driver. So switch from legacy board code to DT is regression.

That's fixed with the following in v4.7-rc series:

db863d896681 ("mmc: omap_hsmmc: Check if MMC slot name is passed in pdata") 10c1f7d32bf4 ("ARM: OMAP2+: n900 needs MMC slot names for legacy user space")

Great! Sorry, but now wiki is not fully updated with last changes and I do not have in my mind what was finished/merged and what not. It is going hard for me to track everything.

Quoted 3 lines

Quoted 1 line

OK well maybe take one more look and then ack if no more issues.

Pali, any news on this?

I have not looked at this deeply yet. Please give me some more time.

Quoted 23 lines · Pali Roh?r

Quoted 22 lines · On Tuesday 21 June 2016 04:25:36 Tony Lindgren wrote:

Quoted 21 lines · Tony Lindgren

Quoted 19 lines · Pali Roh?r

Quoted 8 lines · On Wednesday 15 June 2016 04:02:02 Tony Lindgren wrote:

Hi,

Quoted 4 lines · Pali Roh?r

Hi! I just quickly looked at N900 status matrix
https://elinux.org/N900 and there is still one thing not finished in DT: It is slot-name for omap_hsmmc driver. So switch from legacy board code to DT is regression.

Quoted 5 lines

That's fixed with the following in v4.7-rc series:

db863d896681 ("mmc: omap_hsmmc: Check if MMC slot name is passed in pdata") 10c1f7d32bf4 ("ARM: OMAP2+: n900 needs MMC slot names for legacy user space")

Great! Sorry, but now wiki is not fully updated with last changes and I do not have in my mind what was finished/merged and what not. It is going hard for me to track everything.

Quoted 7 lines

Quoted 4 lines

Quoted 2 lines

OK well maybe take one more look and then ack if no more issues.

Pali, any news on this?

I have not looked at this deeply yet. Please give me some more time.

OK I'd like to apply this series by Monday if no issues to have enough time for it in Linux next. Does that sound OK to you?

Regards,

Tony

Quoted 6 lines

OK I'd like to apply this series by Monday if no issues to have enough time for it in Linux next. Does that sound OK to you?

Regards,

Tony

Ok, now I looked at it and tested current version from torvalds tree.

There is still missing fixed support for IR, but that is waiting in media tree queue for next merge window.

So after IR patches land in torvalds tree, I'm fine with dropping legacy board code for n900 and you can add my Acked-by.

Sorry that it look so long time, but Nokia N900 is omap3 device with most features & drivers.

I need to say that Nokia N900 support has good support in mainline kernel, probably the best from all mobile phones. It is almost feature complete, missing parts are: DSP, both cameras, bluetooth and RM radio.

...and graphics acceleration and power management. But who needs accelerated graphics, and power management will be solved. So yes, N900 is the best phone :-).

Quoted 5 lines · Pavel Machek

...and graphics acceleration and power management. But who needs accelerated graphics, and power management will be solved. So yes, N900 is the best phone :-).

Well, actually we have SGX out-of-tree driver working (with some glitches) with 4.7-rc, Maemo 5 needs it :)

Quoted 32 lines · On Tuesday 21 June 2016 13:32:16 Tony Lindgren wrote:

Quoted 23 lines · Pali Roh?r

Quoted 22 lines · On Tuesday 21 June 2016 04:25:36 Tony Lindgren wrote:

Quoted 21 lines · Tony Lindgren

Quoted 19 lines · Pali Roh?r

Quoted 8 lines · On Wednesday 15 June 2016 04:02:02 Tony Lindgren wrote:

Hi,

Quoted 4 lines · Pali Roh?r

Hi! I just quickly looked at N900 status matrix
https://elinux.org/N900 and there is still one thing not finished in DT: It is slot-name for omap_hsmmc driver. So switch from legacy board code to DT is regression.

Quoted 5 lines

That's fixed with the following in v4.7-rc series:

db863d896681 ("mmc: omap_hsmmc: Check if MMC slot name is passed in pdata") 10c1f7d32bf4 ("ARM: OMAP2+: n900 needs MMC slot names for legacy user space")

Great! Sorry, but now wiki is not fully updated with last changes and I do not have in my mind what was finished/merged and what not. It is going hard for me to track everything.

Quoted 7 lines

Quoted 4 lines

Quoted 2 lines

OK well maybe take one more look and then ack if no more issues.

Pali, any news on this?

I have not looked at this deeply yet. Please give me some more time.

Quoted 6 lines

OK I'd like to apply this series by Monday if no issues to have enough time for it in Linux next. Does that sound OK to you?

Regards,

Tony

Ok, now I looked at it and tested current version from torvalds tree.

There is still missing fixed support for IR, but that is waiting in media tree queue for next merge window.

So after IR patches land in torvalds tree, I'm fine with dropping legacy board code for n900 and you can add my Acked-by.

Quoted 4 lines

So after IR patches land in torvalds tree, I'm fine with dropping legacy board code for n900 and you can add my Acked-by.

Tony asked me to repost the driver so to be included in 4.7 (or 4.8), but I am too busy lately to do it, will try to find time the next week. In the meanwhile, if somebody has time and will, feel free to repost the series for me.

Ok, now I looked at it and tested current version from torvalds tree.

There is still missing fixed support for IR, but that is waiting in media tree queue for next merge window.

So after IR patches land in torvalds tree, I'm fine with dropping legacy board code for n900 and you can add my Acked-by.

Sorry that it look so long time, but Nokia N900 is omap3 device with most features & drivers.

I need to say that Nokia N900 support has good support in mainline kernel, probably the best from all mobile phones. It is almost feature complete, missing parts are: DSP, both cameras, bluetooth and RM radio.

So thanks everybody for this big improvement!

Let's drop the last two remaining omap3 legacy boot board files. Let's only use the device tree based booting known to work for these two boards.

We still have two omap3 boards booting in legacy mode in addition to device tree based booting. All the other ten or so omap2+ SoCs have been booting using device tree mode only for years now. This has allowed us to get rid of quite a bit of arch/arm/mach-omap2 related platform init code in favor of dts and driver changes.

Pretty much the only remaining known users for omap3 legacy boot board files are Kevin's and Russell's boot test systems, and N900 kernel tree maintained by Pali and Ivaylo. And all of them are also supporting the device tree based booting. The legacy booting mode has been kept around mostly to verify against regressions.

As there is still a slim chance of possible other uses of the mainline kernel for these boards, let's just drop the board files for v4.8, and let's not touch the related platform init code until around v4.9 time if no issues are found. This makes it trivial for us to add back the board files with a simple revert.

Signed-off-by Tony Lindgren <[email protected]>

Let's drop the last two remaining omap3 legacy boot board files. Let's only use the device tree based booting known to work for these two boards.

We still have two omap3 boards booting in legacy mode in addition to device tree based booting. All the other ten or so omap2+ SoCs have been booting using device tree mode only for years now. This has allowed us to get rid of quite a bit of arch/arm/mach-omap2 related platform init code in favor of dts and driver changes.

Pretty much the only remaining known users for omap3 legacy boot board files are Kevin's and Russell's boot test systems, and N900 kernel tree maintained by Pali and Ivaylo. And all of them are also supporting the device tree based booting. The legacy booting mode has been kept around mostly to verify against regressions.

As there is still a slim chance of possible other uses of the mainline kernel for these boards, let's just drop the board files for v4.8, and let's not touch the related platform init code until around v4.9 time if no issues are found. This makes it trivial for us to add back the board files with a simple revert.

Signed-off-by Tony Lindgren <[email protected]>

---
 arch/arm/mach-omap2/Kconfig                  |    6 -
 arch/arm/mach-omap2/Makefile                 |    3 -
 arch/arm/mach-omap2/board-rx51-peripherals.c | 1317 --------------------------
 arch/arm/mach-omap2/board-rx51-video.c       |   67 --
 arch/arm/mach-omap2/board-rx51.c             |  141 ---
 arch/arm/mach-omap2/board-rx51.h             |   11 -
 6 files changed, 1545 deletions(-)
 delete mode 100644 arch/arm/mach-omap2/board-rx51-peripherals.c
 delete mode 100644 arch/arm/mach-omap2/board-rx51-video.c
 delete mode 100644 arch/arm/mach-omap2/board-rx51.c
 delete mode 100644 arch/arm/mach-omap2/board-rx51.h

Here are two patches to drop the last remaining omap3 legacy boot files. We have device tree based booting working for these boards, and the legacy booting has been mostly kept around to validate against regressions.

Let's not touch the rest of the platform init code until around v4.9 so we can revert these two patches if needed.

Cheers,

Tony

Tony Lindgren (2):
ARM: OMAP2+: Drop legacy board file for n900
ARM: OMAP2+: Drop legacy board file for LDP

Hi Tony,

Quoted 12 lines · Tony Lindgren

Here are two patches to drop the last remaining omap3 legacy boot files. We have device tree based booting working for these boards, and the legacy booting has been mostly kept around to validate against regressions.

Let's not touch the rest of the platform init code until around v4.9 so we can revert these two patches if needed.

Cheers,

Tony

Quoted 3 lines

Tony Lindgren (2):
ARM: OMAP2+: Drop legacy board file for n900
ARM: OMAP2+: Drop legacy board file for LDP

I wonder what happened to these patches. I see, that you have a omap-for-v4.8/legacy branch, which contains them. But as far as I can see its not in linux-next and it was not included in the ARM pull request.