[Gta04-owner] wifi patches in letux-base?

[H. Nikolaus Schaller]

Hi Andreas,

Am 24.01.2016 um 22:27 schrieb H. Nikolaus Schaller <hns at goldelico.com>:

> Hi Andreas,
> 
> Am 24.01.2016 um 21:22 schrieb Andreas Kemnade <andreas at kemnade.info>:
> 
>> Hi,
>> 
>> I have done some further analysis:
>> 
>> On Sat, 23 Jan 2016 23:08:01 +0100
>> "H. Nikolaus Schaller" <hns at goldelico.com> wrote:
>> 
>>> 
>>> Am 23.01.2016 um 19:03 schrieb Andreas Kemnade <andreas at kemnade.info>:
>>> 
>>>> Hi,
>>>> 
>>>> On Sat, 23 Jan 2016 13:13:03 +0100
>>>> "H. Nikolaus Schaller" <hns at goldelico.com> wrote:
>>>> 
>>>>> Hi,
>>>>> 
>>>>> Am 23.01.2016 um 11:36 schrieb Andreas Kemnade <andreas at kemnade.info>:
>>>>> 
>>>>>> Hi,
>>>>>> 
>>>>>> I found out that there is stuff in letux-base which is not upstreamed.
>>>>>> Is everything there considered not upstreamable? What is considered a
>>>>>> good fix and what not?
>>>>> 
>>>>> Well, some things in letux-base are purely historical and inherited from 3.7-neil...
>>>>> So they might be wrong today.
>>>>> 
>>>> That is all 2013/2014. So it is not that historcal
>>>> 
>>>> $ git shortlog --no-merges v4.4...goldelico2/letux-base -- drivers/net/wireless/libertas 
>>>> Dan Williams (1):
>>>>    libertas/sdio: fix releasing memory twice.
>>>> 
>> superseeded by 
>> libertas: move firmware lifetime handling to firmware.c
>> 1dfba3060fe7ee03ccec25a91d35085142dfc295
>> probably merged with that to the proper thing.
> 
> I think merging something twice is not harmful.
> 
> We probably cherry-picked a preliminary version which was changed on its way
> through linux-next to linus/master.
> 
>> 
>> 
>>>> NeilBrown (4):
>>>>    libertas/sdio: make sure card enters power-off when firmware is loaded.
>> 
>> falso alert
>> https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/drivers/net/wireless/libertas/if_sdio.c?id=d94248ac2aef78e39a76321458a78ba7805dba2e
>>>>    libertas - hack to fix suspend.
>>>>    libertas/sdio: make power management dependant on MMC_CAP_POWER_OFF_CARD
>>>>    libertas/sdio: simplify suspend/resume.
>>>> 
>> These things are not upstreamed and require more analysis.
>> I will check these.
>> Still waiting to be ported from my 3.7-neil-plus-plus:
>> 8b9da484f3f4e7129debf887f373dd2d95f96068
>> wireless:libertas: fix race conditions on resume
>> That is definitively needed in recent kernels.
> 
> I have cherry-picked and fixed 3 merge conflicts. Two appeared to be easy but one (if_sdio_power_restore)
> wasn't. Please look into it when I push the merge results in the next days.

Ok, here it is:

http://git.goldelico.com/?p=gta04-kernel.git;a=shortlog;h=refs/heads/work/kemnade/wifi-sdio-fixes

git rebase was clever enough to detect the rename to a marvell/ subdirectory. So it was easy to rebase this patch set.
The most difficult thing was that omapdss is now duplicated. One version is in drivers/video/fbdev/omap2/omapdss
and the other in drivers/gpu/drm/omapdss. Both have their own versions of panel drivers.

Since we have switched to DRM (because omapfb didn't work well any more), I have (re)moved our specific panel drivers
to drivers/gpu/drm as well.

But for the GTA04 it isn't a problem since our td028ttec1 panel driver is already upstream.

What I don't know is how this affects the PVR/SGX stuff (which didn't work anyways since 3.8 or so). I think there
was an omaplfb bridge for fbdev based systems and a DRM based variant. I can only hope that TI is publishing
something new.

BR,
Nikolaus