2016-09-27 Multimedia group chat report
We had a multimedia group meeting on 2016-09-27. Here's a summary of the
discussions. Please correct any mistake you would notice.
Attendees:
- Kieran
- Laurent
- Morimoto-san
- Niklas
- Ulrich
Magnus was excused.
Topic 1. Status check for the multimedia tasks
----------------------------------------------
* Kieran
Since last meeting:
- Patch generated to apply restrictions for UDS/SRU pipeline configuration
with Partition Algorithm
- Still needs testing/review.
- Stress-testing - segfault investigation
- Identified race in vsp1_video_start_streaming() for multiple inputs
- Prototype writeback support
- vsp1_video v4l2 device node connected to a single WPF entity
- ‘Empty’ frames flowing to userspace.
For the next two weeks:
- Attend Kernel Recipes
- Continue on VSP1-WPF Writeback prototype.
Issues and Blockers:
- WPF is locking up when writeback enabled.
- Debug in progress
* Laurent
Since last meeting:
- Completed M3-W DU support prototype, still needs a bit of polishing
- Progressing in LVDS mode selection for DU.
The hard part is finding the relevant standards to create proper DT bindings.
For the next two weeks:
- Attending kernel recipes this week
- Working on the V4L2 request API next week to post the next version before
ELCE
- Reviewing Gen3 VIN patches
Issues and Blockers:
None.
* Morimoto-san
Since last meeting:
- Modified ALSA sound driver DMA channel get/release timing for MSTP power
save.
Now the driver acquires a channel when starting playback, and releases it when
stopping. Patches should be posted to the mailing list, but they depend on
previous series that still need to be posted.
- Added IPMMU support on ALSA sound driver with Magnus.
This requires a workaround patch (= for sound register), but it works well.
Patches should be posted, with same dependency as above.
- Fixed ALSA sound suspend/resume.
Jinso reported that ALSA sound is broken when resuming from suspend on v4.8-
rc2. The issue was that the sound framework called status SUSPEND when
suspending, but called START when resuming. The sound driver has START/STOP
status support, but doesn't have SUSPEND/RESUME. Adding SUSPEND/RESUME to the
driver fixed this. Patches should be posted, with same dependency as above.
For the next two weeks:
- Continuing work on ALSA
- Attending LinuxCon EU next week
Issues and Blockers:
None
* Niklas
Since last meeting:
- Update HGT testcase to work with all hue area configurations
For the next two weeks:
- Refresh the current VIN Gen3 work on top of media_tree
- Start M3-W VIN support work
Issues and Blockers:
- Still no review on VIN Gen3 patches
At this point reviewers should however wait a few days for the refreshed
version of the patch series.
* Ulrich
Since last meeting:
- Finished Gen2 VIN integration
A trivial fix is needed, next version will be sent soon.
For the next two weeks:
- Update Gen3 HDMI output to use soc_device_match()
Issues and Blockers:
None
Topic 2. Test reports from Jinzai
---------------------------------
We have received testing reports from Jinzai for the 9/M tasks. They noticed
quite a few failures, due to two issues:
- Missing patches
Jinzai tested the latest renesas-drivers release, and not all patches were
merged in the tree. Please make sure you get your patches merged in renesas-
drivers when you send the task report.
- Missing kernel config options
Jinzai used the ARM64 defconfig, which is missing support for several of the
our drivers. To avoid such test failures, a proposal to include the kernel
config options required for testing of each task in the report has been
approved.
A minimal diff against defconfig can be generated with
make ARCH=arm64 savedefconfig
and can then be manually filtered to remove unrelated options.
Topic 3. Agenda for the multimedia face to face meeting in Berlin
-----------------------------------------------------------------
The meeting is scheduled for Sunday morning before ELCE. A draft agenda will
be sent next weekend, based on the Miniperi-2016-10 wiki page available at
https://osdr.renesas.com/projects/linux-kernel-development/wiki/Miniperi-2016-10.
Please propose additional topics in reply to this e-mail.
We will also discuss additional 12/M tasks in Berlin, as well as plans for the
next quarter. Again, if you have identified areas that you think need some
love, please speak out by e-mail.
Topic 4. Additional tasks for 11/M
----------------------------------
Detailed task descriptions have been submitted to Magnus. They should follow
the usual process and the SoWs should hopefully come soon.
</pre>
h1. Multimedia-chat-meeting-2016-09-27
<pre>
<pinchartl> hyvää huomenta for the Europeans, and hyvää iltaa for the Japanese
[15:51]
<neg> morning all [15:52]
<kbingham-> Morning :)
<uli___> good morning
*** horms (~horms@217.111.208.18) has quit: Ping timeout: 244 seconds
<morimoto> hi
<morimoto> Ohayo-gozaimasu [15:53]
<pinchartl> let's wait a minute for Magnus to join
* kbingham- gets tea while waiting :D
<pinchartl> and then get started, as we have a bit less than one hour before
the core meeting
<morimoto> Magnus is preparing Berlin trip now ? He go tomorrow [15:54]
<morimoto> I'm not sure
<pinchartl> he just told me he won't be able to join, so let's get started
[15:55]
<pinchartl> first topic, tasks status update
<pinchartl> let's use the real alphabetical order today
<pinchartl> Kieran, you're first
*** horms (~horms@217.111.208.18) has joined channel #periperi
<pinchartl> (as usual, Since last meeting:
<pinchartl> For the next two weeks:
<pinchartl> Issues and Blockers:)
<kbingham-> As sent in a mail for ease, but pasted here [15:56]
<kbingham-> * Since last meeting
<kbingham-> - Patch generated to apply restrictions for UDS/SRU pipeline
configuration with Partition Algorithm Needs testing
<kbingham-> - Stress-testing - segfault investigation
<kbingham-> = Identified race in vsp1_video_start_streaming() for
multiple inputs
<kbingham-> - Prototype writeback support
<kbingham-> = vsp1_video v4l2 device node connected to a single WPF
entity
<kbingham-> = Empty frames flowing to userspace.
<kbingham-> * For the next two weeks
<kbingham-> - Attend Kernel Recipes
<kbingham-> - Continue on VSP1-WPF Writeback prototype.
<kbingham-> * Blockers / Issues
<kbingham-> - WPF is locking up when writeback enabled. Debug in progress
[15:57]
<kbingham-> Need to work through getting the WPF to writeback to memory
successfully, but that will be next week now.
<pinchartl> thanks
<pinchartl> do you need help with WPF writeback ? [15:58]
<kbingham-> I've only had a few hours of playing / debugging so far - so not
'yet'
<pinchartl> ok [15:59]
<kbingham-> s/few/couple/
<pinchartl> I believe I'm next
<pinchartl> since last meeting I've started work on the 11/M tasks [16:00]
<pinchartl> M3-W DU support is in good shape and just needs a bit of polishing
<pinchartl> and I'm progressing in LVDS mode selection
<pinchartl> the hard part is finding the relevant standards to create proper
DT bindings
*** horms (~horms@217.111.208.18) has quit: Ping timeout: 252 seconds [16:01]
<pinchartl> in the next two weeks I'll attend kernel recipes this week
<pinchartl> and then work on the V4L2 request API next week to post the next
version before ELCE
<pinchartl> no issue or blocker so far
<pinchartl> next, Morimoto-san [16:02]
<morimoto> Since Last Week
<morimoto> 1st, I modified ALSA sound driver DMA channel get/release timing
for MSTP power save. Now It get channel when it start playback, and
release it when it stopped. I would like to post this patch to ML,
but I need to post previous patches first.
<morimoto> 2nd, I added IPMMU support on ALSA sound driver with Magnus. It
needs workaround patch (= for sound register), but it works well. I
would like to post this patch too, but ...
<morimoto> 3rd, Jinso reported that ALSA sound has damage when resume on
v4.8-rc2. The issue was sound framework called status SUSPEND when
suspend, but called START when resume. My sound driver has
START/STOP status support, but doesn't have SUSPEND/RESUME. It
solved if I added SUSPEND/RESUME. I would like to post this patch
too, but..
<morimoto> --end-- [16:03]
<morimoto> I'm looking forward to meet you in Berlin
<pinchartl> so am I :-) [16:04]
<pinchartl> any particular plan for the next two weeks, or just continuing the
work in progress ? [16:05]
<morimoto> I will meet ALSA SoC people on ELCE timing, I have no other plan
for next 2 weeks
<morimoto> And pinchartl can join to it ? [16:06]
*** horms (~horms@217.111.208.18) has joined channel #periperi
<pinchartl> yes I'll join, it's in my calendar
<pinchartl> note that ELCE is in more than two weeks ;-)
<morimoto> Nice to know
<morimoto> I will goto LinuxCon Euro too
<morimoto> I'm in Berlin from 3rd to 14th Oct [16:07]
<pinchartl> ah right
<pinchartl> have fun there :-)
<pinchartl> any blocker or issue ?
<morimoto> I will :)
<geertu> pinchartl: Exactly two weeks
<morimoto> I don't know how to ride plane
<morimoto> (joke :) [16:08]
<kbingham-> morimoto: Just make sure you get business class and you'll be fine
:D
<morimoto> Hehe :)
<pinchartl> :-)
<pinchartl> next, Niklas
<neg> a) Update HGT testcase to work with all hue area configurations [16:09]
<neg> b) Refresh the current VIN Gen3 work ontop of media_tree and start M3-W
work
<neg> c) Still no review on VIN Gen3 patches
<neg> thats about it [16:10]
<pinchartl> by M3-W work, do you mean VIN support ?
<neg> ohh yes, sorry was a bit unclear [16:11]
*** horms (~horms@217.111.208.18) has quit: Ping timeout: 255 seconds
<pinchartl> for the Gen3 VIN patches I'm very sorry for the delay :-/
<pinchartl> I'll try to address it this week but it's a bit unlikely as I'm
attending kernel recipes [16:12]
<pinchartl> but I will review them before ELCE in any case
<kbingham-> I have a lot of 'train time' today - so I can cast my eye through
the VIN patches if it helps. Perhaps won't be as good an eye as
pinchartl - but it might help.
<pinchartl> kbingham-: that would certainly be appreciated
<neg> np, at this time maybe it's best if you hold off a few days so I can
resend the refreshed patches on top of media_tree, we got a lot of good
stuff for the VIN driver merged \o/ [16:13]
<neg> kbingham-: thanks I will CC you when I resend the H3 version of the VIN
Gen3 patches
<kbingham-> neg :D [16:14]
<pinchartl> next, uli___
<uli___> 1) finished up gen2 vin integration [16:15]
<uli___> 2) try and fit soc_device_match() into gen3 hdmi out
<uli___> 3) nothing
<uli___> (i just noticed that 1) needs a trivial fix, gonna resend that rsn)
[16:16]
<pinchartl> thanks
<pinchartl> on a general note
<pinchartl> we've received testing reports from Jinzai for the 9/M tasks
[16:17]
<pinchartl> they noticed quite a few failures, due to two issues
<pinchartl> the first one was missing patches
<pinchartl> Jinzai tested the latest renesas-drivers release
<pinchartl> so please make sure you get your patches merged in renesas-drivers
when you send the task report
<pinchartl> the second issue was due to missing config options in the kernel
[16:18]
<pinchartl> they used the ARM64 defconfig
<pinchartl> to avoid test failures, I propose including in future test reports
the kernel config options required for testing of each task
[16:19]
<pinchartl> does anyone have any objection there ?
<pinchartl> or any concern ?
<kbingham-> sounds reasonable to specify required kernel config
parameters. Just have to remember them when we add them - or diff
our local configs against arm64defconfig. [16:20]
<pinchartl> you can
<pinchartl> make ARCH=arm64 defconfig [16:21]
<pinchartl> sorry
<pinchartl> make ARCH=arm64 savedefconfig
<pinchartl> that will generate a minimal diff
<kbingham-> Aha - that's new to me! [16:22]
<pinchartl> and you can manually filter out anything related to your setup
* kbingham- tests locally.
<pinchartl> next topic, schedule for the multimedia meeting in Berlin [16:25]
<pinchartl> the meeting is scheduled on Sunday morning
<pinchartl> I want to send a draft of the agenda next weekend [16:26]
<pinchartl> so please make sure to e-mail proposals for topics you want to
discuss by then
<pinchartl> or add them directly to the wiki page [16:27]
<pinchartl>
https://osdr.renesas.com/projects/linux-kernel-development/wiki/Miniperi-2016-10
<pinchartl> so far, we have four topics
<pinchartl> three of them are status reports for tasks
<pinchartl> I'll group them as a single status report topic [16:28]
<pinchartl> the last one is "DU/VIN DT style difference ES1.x/ES2.0"
*** horms (~horms@217.111.208.18) has joined channel #periperi
* kbingham- gets a 403 -ENOAUTH on the wiki :(
<morimoto> kbingham-: oops, please wait
<pinchartl> kbingham-: I assume you now have an account for osdr.renesas.com ?
[16:29]
<kbingham-> pinchartl: I have an account- but no permissions :D
<pinchartl> ah, and there's a fifth topic proposed by Niklas, I'll add it to
the wiki
<pinchartl> regarding V4L2 pad-aware subdev operations
<morimoto> kbingham-: you are "Kieran Bingham" correct ?
<kbingham-> morimoto: Yes :)
<pinchartl> scratch that, it's for the V4L2 meeting on Monday
<morimoto> kbingham-: can you now ?
<kbingham-> morimoto: Ack :) Yes thanks [16:30]
<morimoto> nice :)
<pinchartl> we will also discuss additional 12/M tasks in Berlin, as well as
plans for the next quarter
<pinchartl> again, if you have identified areas that you think need some love,
please speak out now or by e-mail [16:31]
<pinchartl> next topic, 11/M tasks [16:32]
<pinchartl> I've submitted the detailed descriptions of all tasks to Magnus
<pinchartl> they should follow the usual process [16:33]
<pinchartl> and I hope to see SoWs coming soon
<pinchartl> I believe that it's safe to start working on them at this point,
but you're free to wait for the SoW
<pinchartl> that's all for me. does someone have additional topic proposals ?
[16:35]
<morimoto> Do you know where we can meet for MiniPeriCon ?
<morimoto> Wolfram control ? [16:36]
<pinchartl> that's a good question, I don't know
<morimoto> OK, no one knows :)
<pinchartl> that seems to be a central problem, not specific to multimedia
[16:37]
<pinchartl> isn't it the job of the core group to handle that kind of issues ?
;-)
<morimoto> Hehe :)
<neg> also not related to periperi multimedia, but do you know any time or
location for the V4L2 meeting on Monday?
<pinchartl> I assume it will start at 9:00 [16:38]
<pinchartl> as for the location, Hans said he got a room from the Linux
Foundation
<pinchartl> I suppose he will announce it
<kbingham-> pinchartl: He has announced it on linux-media [16:39]
<kbingham-> I presume it is : '[ANN] Codec & Request API Brainstorm meeting
Oct 10 & 11'
<kbingham-> in room 'Salon 13 Paris' on level 1 of the Maritim Berlin
<neg> nice, thanks (I need to read linux-media more carefull) [16:41]
<pinchartl> the next meeting will be held in Berlin on Sunday the 9th [16:42]
<pinchartl> with this said, I propose adjourning this meeting. does anyone
second ?
<neg> second
<morimoto> second [16:43]
<neg> (so I have time to brew coffy before core meeting :-)
<kbingham-> no time for more tea for me. I have to pack and head to a train
station :)