[Letux-kernel] Long-Term kernels

H. Nikolaus Schaller hns at goldelico.com
Mon Dec 21 20:12:36 CET 2020 

> Am 21.12.2020 um 19:59 schrieb David Shah <dave at ds0.me>:
> 
> I am just repeating
> 
> xset dpms force off
> xset dpms force on
> 
> until it fails - once it's failed you can get it to come back by
> running the sequence of commands again a few times.

Ah, ok. Haven't tried that ever on the Pyra :)

Maybe because I had trouble waking up my HDMI-Monitor when running Letux
kernel on the RasPi 3B+. So I thought that the command is somehow broken.

> 
> Note that this is with the OMAP5 X11 driver - 
> https://dev.pyra-handheld.com/packages/xserver-xorg-video-armsoc-omap5
> 
> I can't see a particularly concerning difference between our disable
> function (which is where the problems seem to start) and other panel
> drivers. I half wonder if the problem might be in the OMAP DSI stuff
> rather than the panel driver.

Yes, it is unlikely a problem of the panel driver (which follows structure
of the otm8009a driver which is used by some STM mipi dsi host). As said
Tomi was not sure about the OMAPDRM driver doing all sequencing
correctly.

So let's wait until the new OMAPDRM subsystem arrives in linux-next and
then submit fixes we can also backport to our letux-5.10.y series.

BR,
Nikolaus

> 
> Best
> 
> David
> 
> On Mon, 2020-12-21 at 19:56 +0100, H. Nikolaus Schaller wrote:
>> 
>>> Am 21.12.2020 um 19:51 schrieb David Shah <dave at ds0.me>:
>>> 
>>> And... seems I spoke to soon about DPMS working ;) - but at least
>>> it
>>> does now work sometimes.
>>> 
>>> There's now a new, intermittent DSI related failure going on that I
>>> think is unrelated to the last problem:
>>> 
>>> [   53.141659] panel-btl507212-w677l 58004000.encoder.0:
>>> w677l_disable
>>> [   53.451626] DSI: omapdss DSI error: DSI error, irqstatus 4000
>>> [   53.460384] panel-btl507212-w677l 58004000.encoder.0:
>>> w677l_unprepare
>>> [   53.468136] panel-btl507212-w677l 58004000.encoder.0:
>>> power_off()
>>> [   53.486342] panel-btl507212-w677l 58004000.encoder.0:
>>> w677l_reset(active)
>>> [   53.503567] panel-btl507212-w677l 58004000.encoder.0:
>>> w677l_regulator(off)
>>> [   53.531918] panel-btl507212-w677l 58004000.encoder.0:
>>> w677l_unprepare finished
>>> [   53.541620] panel-btl507212-w677l 58004000.encoder.0:
>>> w677l_prepare
>>> [   53.549286] panel-btl507212-w677l 58004000.encoder.0: power_on()
>>> [   53.555619] panel-btl507212-w677l 58004000.encoder.0:
>>> w677l_reset(active)
>>> [   53.563112] panel-btl507212-w677l 58004000.encoder.0:
>>> w677l_regulator(on)
>>> [   53.636501] panel-btl507212-w677l 58004000.encoder.0:
>>> w677l_reset(inactive)
>>> [   53.666897] panel-btl507212-w677l 58004000.encoder.0: dsi:
>>> w677l_write(dcs 11) [0]
>>> [   53.684628] DSI: omapdss DSI error: Error while sending BTA:
>>> 4000
>>> [   53.691461] DSI: omapdss DSI error: bta sync failed
>>> [   53.697731] panel-btl507212-w677l 58004000.encoder.0: write
>>> cmd/reg(11) failed: -5
>>> [   53.708303] panel-btl507212-w677l 58004000.encoder.0: sequence
>>> failed: 0
>>> [   53.732510] panel-btl507212-w677l 58004000.encoder.0:
>>> w677l_enable
>>> [   53.739193] panel-btl507212-w677l 58004000.encoder.0:
>>> w677l_enable()
>>> powered on()
>>> 
>>> 
>>> it's possible that the panel is being disabled at the wrong point
>>> and
>>> messing up the DSI state.
>> 
>> Indeed. Tomi was not very sure about properly sequencing these
>> things.
>> (Note: he has a new e-mail address).
>> 
>> Is there a method or specific command sequence to trigger or
>> reproduce this problem?
>> 
>> BR and thanks,
>> Nikolaus
>> 
>> 
>>> 
>>> 
>>> On Mon, 2020-12-21 at 18:40 +0000, David Shah wrote:
>>>> It was just another pyra_defconfig problem in the end, sent a
>>>> patch.
>>>> 
>>>> And the DPMS issue now seems to be resolved :)
>>>> 
>>>> Next problem I've noticed is that pvr init is failing, I think I
>>>> saw
>>>> some conversation between you and Tony about that back in
>>>> November
>>>> that
>>>> I'll look into.
>>>> 
>>>> Best
>>>> 
>>>> David
>>>> 
>>>> On Mon, 2020-12-21 at 19:32 +0100, H. Nikolaus Schaller wrote:
>>>>> 
>>>>>> Am 21.12.2020 um 19:21 schrieb David Shah <dave at ds0.me>:
>>>>>> 
>>>>>> Trying letux-5.10.y and the annoying display-not-being-loaded
>>>>>> issue
>>>>>> has
>>>>>> come back to bite again, even though we are now baking
>>>>>> omapdrm
>>>>>> into
>>>>>> the
>>>>>> kernel in the Pyra config.
>>>>> 
>>>>> Strangely I haven't seen that recently. And it is questionable
>>>>> if
>>>>> it
>>>>> is solved by the new panel driver architecture.
>>>>> 
>>>>> But you could do a test by merging letux/boe-w677-dsi-panel-v4.
>>>>> This
>>>>> should include the new omapdrm and the new panel driver.
>>>>> 
>>>>> If not someone has to dig into it and search for the
>>>>> initialization
>>>>> race.
>>>>> 
>>>>> BR,
>>>>> Nikolaus
>>>>> 
>>>>>> 
>>>>>> On Mon, 2020-12-21 at 18:14 +0100, H. Nikolaus Schaller
>>>>>> wrote:
>>>>>>> 
>>>>>>>> Am 21.12.2020 um 18:00 schrieb David Shah <dave at ds0.me>:
>>>>>>>> 
>>>>>>>> It would be nice to get 5.10 up and running on the Pyra,
>>>>>>> 
>>>>>>> Well, for me it is running and I am not aware of bigger
>>>>>>> issues.
>>>>>>> But I
>>>>>>> am not using it daily so that I may miss some problems.
>>>>>>> 
>>>>>>>> despite not containing
>>>>>>>> the new panel driver yet (afaik).
>>>>>>> 
>>>>>>> Yes, the new panel driver will not come before v5.12. But
>>>>>>> IMHO
>>>>>>> it
>>>>>>> has
>>>>>>> no functional difference. Is just much cleaner and
>>>>>>> upstreamable
>>>>>>> code.
>>>>>>> So no need to wait for it.
>>>>>>> 
>>>>>>>> 
>>>>>>>> Though I haven't tested, I suspect the biggest issue is
>>>>>>>> going
>>>>>>>> to be
>>>>>>>> the
>>>>>>>> DPMS/stuck atomic update issue that is somewhere in the
>>>>>>>> display/DRM
>>>>>>>> code
>>>>>>>> discussed previously and has likely been present since
>>>>>>>> 5.7
>>>>>>>> ish.
>>>>>>> 
>>>>>>> I think there was some fix a while ago which may also be in
>>>>>>> 5.9.
>>>>>>> But I currently don't have the Pyra in operation to check.
>>>>>>> 
>>>>>>> Generally I think the best strategy for PyraOS would be to
>>>>>>> use
>>>>>>> and
>>>>>>> maintain
>>>>>>> 5.10.y as the next long-term kernel.
>>>>>>> 
>>>>>>> BR,
>>>>>>> Nikolaus
>>>>>>> 
>>>>>>>> 
>>>>>>>> David
>>>>>>>> 
>>>>>>>> On Mon, 2020-12-21 at 16:12 +0100, H. Nikolaus Schaller
>>>>>>>> wrote:
>>>>>>>>> v5.9.16 was just tagged [EOL].
>>>>>>>>> 
>>>>>>>>> According to
>>>>>>>>> https://www.kernel.org/category/releases.html the
>>>>>>>>> long-term
>>>>>>>>> kernels will be:
>>>>>>>>> 
>>>>>>>>> - 5.10
>>>>>>>>> - 5.4
>>>>>>>>> - 4.19
>>>>>>>>> - 4.14
>>>>>>>>> - 4.9
>>>>>>>>> - 4.4   (maintenance ends in 2 months - but we anyways
>>>>>>>>> have
>>>>>>>>> no
>>>>>>>>> letux-4.4)
>>>>>>>>> 
>>>>>>>>> For Letux-Kernels we will follow this scheme.
>>>>>>>>> 
>>>>>>>>> BR,
>>>>>>>>> Nikolaus
>>>>>>>>> 
>>>>>>>>> _______________________________________________
>>>>>>>>> https://projects.goldelico.com/p/gta04-kernel/
>>>>>>>>> Letux-kernel mailing list
>>>>>>>>> Letux-kernel at openphoenux.org
>>>>>>>>> http://lists.goldelico.com/mailman/listinfo.cgi/letux-kernel
>>>>>>>> 
>>>>>>>> 
>>>>>>>> _______________________________________________
>>>>>>>> https://projects.goldelico.com/p/gta04-kernel/
>>>>>>>> Letux-kernel mailing list
>>>>>>>> Letux-kernel at openphoenux.org
>>>>>>>> http://lists.goldelico.com/mailman/listinfo.cgi/letux-kernel
>>>>>>> 
>>>>>>> _______________________________________________
>>>>>>> Kernel mailing list
>>>>>>> Kernel at pyra-handheld.com
>>>>>>> http://pyra-handheld.com/cgi-bin/mailman/listinfo/kernel
>>>>>> 
>>>>>> 
>>>>> 
>>>> 
>>>> 
>>>> _______________________________________________
>>>> https://projects.goldelico.com/p/gta04-kernel/
>>>> Letux-kernel mailing list
>>>> Letux-kernel at openphoenux.org
>>>> http://lists.goldelico.com/mailman/listinfo.cgi/letux-kernel
>>> 
>>> 
>> 
> 
>

More information about the Letux-kernel mailing list