outstanding tmio patches

outstanding tmio patches

[Guennadi Liakhovetski]

I thought it might help to list all the recent patches in the order they 
should apply to make it a bit easier...

[1/6,v2] mmc: tmio: split core functionality, DMA and MFD glue
[2/6] sh: ecovec: use the CONFIG_MMC_TMIO symbols instead of MFD
[3/6,v2] mmc: tmio: convert the SDHI MMC driver from MFD to a standard platform driver
[4/6] ARM: mach-shmobile: convert boards to use linux/mmc/sh_mobile_sdhi.h
[5/6] sh: convert boards to use linux/mmc/sh_mobile_sdhi.h
[6/6] mfd: remove now redundant sh_mobile_sdhi.h header

[1/4,v3] mmc: tmio: only access registers above 0xff, if available
[2/4,v3] ARM: mach-shmobile: fix SDHI IO address-range
[3/4,v3] sh: fix SDHI IO address-range
[4/4,v3] mmc: tmio: remove work-around for unmasked SDIO interrupts

[PATCH/RFC,1/2,v2] mmc: tmio: fix power-mode interpretation
[PATCH/RFC,2/2,v2] mmc: tmio: support aggressive clock gating

All available in the patchwork 
https://patchwork.kernel.org/project/linux-mmc/list/. The last two 
patches are somewhat risky, but given the timeframe, we could push them 
and fix any unlikely regressions later.

Thanks
Guennadi
---
Guennadi Liakhovetski, Ph.D.
Freelance Open-Source Software Developer
http://www.open-technology.de/

Re: outstanding tmio patches

[Guennadi Liakhovetski]

A small update: two more patches have to be applied first, that are also 
still not in the mmc tree, submitted on the 7th of March:

[1/2] mmc: tmio: Improve DMA stability on sh-mobile
[2/2] mmc: tmio: use PIO for short transfers

On Wed, 23 Mar 2011, Guennadi Liakhovetski wrote:

> I thought it might help to list all the recent patches in the order they 
> should apply to make it a bit easier...
> 
> [1/6,v2] mmc: tmio: split core functionality, DMA and MFD glue
> [2/6] sh: ecovec: use the CONFIG_MMC_TMIO symbols instead of MFD
> [3/6,v2] mmc: tmio: convert the SDHI MMC driver from MFD to a standard platform driver
> [4/6] ARM: mach-shmobile: convert boards to use linux/mmc/sh_mobile_sdhi.h
> [5/6] sh: convert boards to use linux/mmc/sh_mobile_sdhi.h
> [6/6] mfd: remove now redundant sh_mobile_sdhi.h header
> 
> [1/4,v3] mmc: tmio: only access registers above 0xff, if available
> [2/4,v3] ARM: mach-shmobile: fix SDHI IO address-range
> [3/4,v3] sh: fix SDHI IO address-range
> [4/4,v3] mmc: tmio: remove work-around for unmasked SDIO interrupts
> 
> [PATCH/RFC,1/2,v2] mmc: tmio: fix power-mode interpretation
> [PATCH/RFC,2/2,v2] mmc: tmio: support aggressive clock gating
> 
> All available in the patchwork 
> https://patchwork.kernel.org/project/linux-mmc/list/. The last two 
> patches are somewhat risky, but given the timeframe, we could push them 
> and fix any unlikely regressions later.
> 
> Thanks
> Guennadi
> ---
> Guennadi Liakhovetski, Ph.D.
> Freelance Open-Source Software Developer
> http://www.open-technology.de/
> 

---
Guennadi Liakhovetski, Ph.D.
Freelance Open-Source Software Developer
http://www.open-technology.de/

Re: outstanding tmio patches

[Chris Ball]

Hi Guennadi,

On Wed, Mar 23 2011, Guennadi Liakhovetski wrote:
> I thought it might help to list all the recent patches in the order they 
> should apply to make it a bit easier...
>
> [1/2] mmc: tmio: Improve DMA stability on sh-mobile

Some serious merge conflicts with Linus W's dmaengine cleanup here.

> [2/2] mmc: tmio: use PIO for short transfers
> [1/6,v2] mmc: tmio: split core functionality, DMA and MFD glue

And again here.  These flag-day patches that move around thousands of
lines of code are extremely difficult to pull off without announcing
them far ahead of time and getting everyone else to hold off from
merging code first; you're essentially saying that development is closed
and everyone else has to wait for your patch to be merged and then
rebase onto it.

I'm open to suggestions on what to do next -- if you find it easier
to prepare a git tree with everything applied for me to pull from
I'd be happy to do that, although I think we should be preparing this
much churn long before the merge window opens if it's supposed to go
into mainline for the current cycle; the merge window should be for
merging linux-next into mainline.

I think my own suggestion, without knowing too much about tmio, would be
to apply the four patches that provide hardware fixes now, for -rc2:

[1/4,v3] mmc: tmio: only access registers above 0xff, if available
[2/4,v3] ARM: mach-shmobile: fix SDHI IO address-range
[3/4,v3] sh: fix SDHI IO address-range
[4/4,v3] mmc: tmio: remove work-around for unmasked SDIO interrupts

and push the module split and clock gating work out to the start of the
next merge window.

Let me know what you think, thanks,

- Chris.
-- 
Chris Ball   <cjb@laptop.org>   <http://printf.net/>
One Laptop Per Child

Re: outstanding tmio patches

[Guennadi Liakhovetski]

Hi Chris

On Thu, 24 Mar 2011, Chris Ball wrote:

> Hi Guennadi,
> 
> On Wed, Mar 23 2011, Guennadi Liakhovetski wrote:
> > I thought it might help to list all the recent patches in the order they 
> > should apply to make it a bit easier...
> >
> > [1/2] mmc: tmio: Improve DMA stability on sh-mobile
> 
> Some serious merge conflicts with Linus W's dmaengine cleanup here.

Sorry, I must be missing something. I'm basing my patches on the current 
(of yesterday) next tree, which already contains Linus' patches:

mmc: tmio_mmc: use dmaengine helpers, drop submit check
mmc: tmio_mmc: drop dma_sglen state variable
mmc: tmio_mmc: unmap with the proper sglen
mmc: tmio_mmc: map DMA buffers on the DMA engine device

or which his patches do you mean?

Thanks
Guennadi

> > [2/2] mmc: tmio: use PIO for short transfers
> > [1/6,v2] mmc: tmio: split core functionality, DMA and MFD glue
> 
> And again here.  These flag-day patches that move around thousands of
> lines of code are extremely difficult to pull off without announcing
> them far ahead of time and getting everyone else to hold off from
> merging code first; you're essentially saying that development is closed
> and everyone else has to wait for your patch to be merged and then
> rebase onto it.
> 
> I'm open to suggestions on what to do next -- if you find it easier
> to prepare a git tree with everything applied for me to pull from
> I'd be happy to do that, although I think we should be preparing this
> much churn long before the merge window opens if it's supposed to go
> into mainline for the current cycle; the merge window should be for
> merging linux-next into mainline.
> 
> I think my own suggestion, without knowing too much about tmio, would be
> to apply the four patches that provide hardware fixes now, for -rc2:
> 
> [1/4,v3] mmc: tmio: only access registers above 0xff, if available
> [2/4,v3] ARM: mach-shmobile: fix SDHI IO address-range
> [3/4,v3] sh: fix SDHI IO address-range
> [4/4,v3] mmc: tmio: remove work-around for unmasked SDIO interrupts
> 
> and push the module split and clock gating work out to the start of the
> next merge window.
> 
> Let me know what you think, thanks,
> 
> - Chris.
> -- 
> Chris Ball   <cjb@laptop.org>   <http://printf.net/>
> One Laptop Per Child
> 

---
Guennadi Liakhovetski, Ph.D.
Freelance Open-Source Software Developer
http://www.open-technology.de/

Re: outstanding tmio patches

[Chris Ball]

Hi,

On Fri, Mar 25 2011, Guennadi Liakhovetski wrote:
>> > [1/2] mmc: tmio: Improve DMA stability on sh-mobile
>> 
>> Some serious merge conflicts with Linus W's dmaengine cleanup here.
>
> Sorry, I must be missing something. I'm basing my patches on the current 
> (of yesterday) next tree, which already contains Linus' patches:

I agree that linux-next contains these patches, but I don't agree that
"mmc: tmio: Improve DMA stability on sh-mobile" (which was written back
on March 7th before they were merged) avoids merge conflicts with them.

Thanks,

- Chris.
-- 
Chris Ball   <cjb@laptop.org>   <http://printf.net/>
One Laptop Per Child

Re: outstanding tmio patches

[Guennadi Liakhovetski]

On Fri, 25 Mar 2011, Chris Ball wrote:

> Hi,
> 
> On Fri, Mar 25 2011, Guennadi Liakhovetski wrote:
> >> > [1/2] mmc: tmio: Improve DMA stability on sh-mobile
> >> 
> >> Some serious merge conflicts with Linus W's dmaengine cleanup here.
> >
> > Sorry, I must be missing something. I'm basing my patches on the current 
> > (of yesterday) next tree, which already contains Linus' patches:
> 
> I agree that linux-next contains these patches, but I don't agree that
> "mmc: tmio: Improve DMA stability on sh-mobile" (which was written back
> on March 7th before they were merged) avoids merge conflicts with them.

Ok, now I understand: I rebased those patches locally, but forgot to 
resend, sorry. Coming in a minute.

Thanks
Guennadi
---
Guennadi Liakhovetski, Ph.D.
Freelance Open-Source Software Developer
http://www.open-technology.de/

Re: outstanding tmio patches

[Chris Ball]

Hi Guennadi,

On Thu, Mar 24 2011, Chris Ball wrote:
>> [1/2] mmc: tmio: Improve DMA stability on sh-mobile
>> [2/2] mmc: tmio: use PIO for short transfers

Thanks, these two are applied now.

>> [1/6,v2] mmc: tmio: split core functionality, DMA and MFD glue

But now we have a merge conflict here, since we just changed the
codebase that this patch moves around.

- Chris.
-- 
Chris Ball   <cjb@laptop.org>   <http://printf.net/>
One Laptop Per Child

Re: outstanding tmio patches

[Guennadi Liakhovetski]

On Fri, 25 Mar 2011, Chris Ball wrote:

> Hi Guennadi,
> 
> On Thu, Mar 24 2011, Chris Ball wrote:
> >> [1/2] mmc: tmio: Improve DMA stability on sh-mobile
> >> [2/2] mmc: tmio: use PIO for short transfers
> 
> Thanks, these two are applied now.
> 
> >> [1/6,v2] mmc: tmio: split core functionality, DMA and MFD glue
> 
> But now we have a merge conflict here, since we just changed the
> codebase that this patch moves around.

Hm, strange, now all patches should be correct - if the correct versions 
are taken. Let's do it easier: I've uploaded all 15 patches (14 mine and 1 
from Simon) as a quilt series to

http://download.open-technology.de/tmio-for-2.6.39/

Please, just try those.

Thanks
Guennadi
---
Guennadi Liakhovetski, Ph.D.
Freelance Open-Source Software Developer
http://www.open-technology.de/

Re: outstanding tmio patches

[Chris Ball]

Hi,

On Fri, Mar 25 2011, Guennadi Liakhovetski wrote:
> Hm, strange, now all patches should be correct - if the correct versions 
> are taken. Let's do it easier: I've uploaded all 15 patches (14 mine and 1 
> from Simon) as a quilt series to
>
> http://download.open-technology.de/tmio-for-2.6.39/

Thanks, I just needed to pull changes from Linus into mmc-next.

I've pushed everything out to mmc-next now.  If anyone else is able to
grab mmc-next and supply a Tested-by: for their hardware, that'd be
appreciated.

Also, if anyone thinks that some of this should wait until .40, now's a
good time to speak up.  :)

- Chris.
-- 
Chris Ball   <cjb@laptop.org>   <http://printf.net/>
One Laptop Per Child

Re: outstanding tmio patches

[Simon Horman]

On Fri, Mar 25, 2011 at 10:54:48AM -0400, Chris Ball wrote:
> Hi,
> 
> On Fri, Mar 25 2011, Guennadi Liakhovetski wrote:
> > Hm, strange, now all patches should be correct - if the correct versions 
> > are taken. Let's do it easier: I've uploaded all 15 patches (14 mine and 1 
> > from Simon) as a quilt series to
> >
> > http://download.open-technology.de/tmio-for-2.6.39/
> 
> Thanks, I just needed to pull changes from Linus into mmc-next.
> 
> I've pushed everything out to mmc-next now.  If anyone else is able to
> grab mmc-next and supply a Tested-by: for their hardware, that'd be
> appreciated.

Hi,

I have done some tests on a Renesas Mackerel board (Guennadi likely has
one too) both with and without my "mmc, ARM: Add zboot from eSD support for
SuperH Mobile ARM" change - which needs to go through RMK's tree
and is the motivation for the " mmc: tmio_mmc: Move some defines into a
shared header" patch.

I didn't notice any problems.

FWIW

Tested-by: Simon Horman <horms@verge.net.au>

> Also, if anyone thinks that some of this should wait until .40, now's a
> good time to speak up.  :)

I have no particular objections though I do agree that it
is a lot of churn at this stage of the release cycle.

Re: outstanding tmio patches

[Chris Ball]

Hi Guennadi,

On Fri, Mar 25 2011, Guennadi Liakhovetski wrote:
> Let's do it easier: I've uploaded all 15 patches (14 mine and 1 from
> Simon) as a quilt series to
>
> http://download.open-technology.de/tmio-for-2.6.39/
>
> Please, just try those.

These made it in as the last merge before the .39 window closed.  :)
Let's talk a few weeks before the .40 merge window opens about what the
.40 tmio merge should look like; it would have been better for these to
have spent longer in linux-next before merging.

Thanks,

- Chris.
-- 
Chris Ball   <cjb@laptop.org>   <http://printf.net/>
One Laptop Per Child

Re: outstanding tmio patches

[Guennadi Liakhovetski]

Hi Chris

On Tue, 29 Mar 2011, Chris Ball wrote:

> Hi Guennadi,
> 
> On Fri, Mar 25 2011, Guennadi Liakhovetski wrote:
> > Let's do it easier: I've uploaded all 15 patches (14 mine and 1 from
> > Simon) as a quilt series to
> >
> > http://download.open-technology.de/tmio-for-2.6.39/
> >
> > Please, just try those.
> 
> These made it in as the last merge before the .39 window closed.  :)

Great! Thanks a lot!

> Let's talk a few weeks before the .40 merge window opens about what the
> .40 tmio merge should look like; it would have been better for these to
> have spent longer in linux-next before merging.

Sure, we should certainly try to sort things out much earlier in the 
development cycle, than this time.

Thanks
Guennadi
---
Guennadi Liakhovetski, Ph.D.
Freelance Open-Source Software Developer
http://www.open-technology.de/