/srv/irclogs.linaro.org/2011/10/10/#linaro-kernel.txt

linuswdsaxena, npitre: a few comments on MFD devices...09:37
linuswdsaxena: strange drivers in mfd/ as well as misc/ often mask the underlying problem of creating a new subsystem09:38
linuswdsaxena: as does strange drivers under arch/arm/*09:38
linuswdsaxena: for example drivers/adc drivers/dac and drivers/pwm are badly needed subsystems09:39
linuswdsaxena: so some of the cruft building up in these actually need to be migrated to places that do not exist09:39
linuswdsaxena: I think we need a run-down of "subsystems needing to be implemented" and maybe just go ahead and do them09:40
linuswdsaxena: each one a non-trivial task on its own, but fun!09:40
linuswdsaxena: we also really need to know if drivers/staging/iio* is ever going mainline because right now that thing contains an array of needed subsystems09:41
dmarthi all15:01
dsaxenagood morning/afternoon/evening15:01
mounirhi15:01
shawnguohello everyone15:02
jstultz_vmdsaxena:  hey15:03
dsaxenagreetings shawnguo , jstultz_vm , dmart , mounir15:04
dsaxenaanyone else around?15:04
nheHi...15:04
dsaxenahi nhe15:04
linusw_I'm here15:05
dsaxenagreetings linusw_15:05
dsaxenaok, this might be a short meeting. let's start with a couple of quick announcements15:06
dsaxena1. I've started creating session blueprints for Linaro connect. We have a total of about 15 slots, I've created 8 BPs so far15:06
dsaxenai'll make a few more today and send email to session owners to work on creating an outline15:07
dsaxenawe also have about 15 hours of hacking time available and we should start thinking about how we want to split that.15:09
dsaxenawhich leads to the second announcement.15:09
dsaxenawe're going to be changing how we work in the KWG a little bit starting at Linaro connect15:10
dsaxenacurrently we have everyone working on their own, targeting somewhat completely different deliverables15:10
dsaxenathis is making it hard for Linaro as an organization to say "We will have feature  X up and running on all SOCs by this timeframe"15:11
dsaxenawhat we'll be doing in the future, is setting up small mini-teams to go tackle a specififc problem across all SOCs at the same time15:12
dsaxenafor example, we'll work on getting pinmux supported on all member SOCs15:12
linusw_cool15:12
dsaxenawith linusw_ as the lead for the mini-team.15:12
dsaxenathis wouldn't be everyone's 100% effort. we'll figure out the details at LC15:13
dmartthat generally sounds like a good idea -- this is sort of what happens anyway, but without the team structure15:13
shawnguoyeah, just like the mini-team for device tree led by Grant15:14
dsaxenayep15:14
dsaxenathe hope is also that by having us work in parallel, we increase chances of new infrastructure we work on to get pushed up stream as there is more than one user for it15:14
linusw_so like with DeviceTree I get sleepless nights if Torvalds doesn't pull in my code ;-)15:14
dsaxenaand also to build more of a sense of a team working in unison15:15
dsaxenalinusw_, :)15:16
dsaxenaother items will include transitioning regulator drivers to DT, inter-CPU IPC drivers (once we have a generic solution), and whatever other cleanup/consolidation efforts we dig into15:16
arndwhat about DT support for clocks? is that something we expect to have a patch for before LC, or should we work on that too?15:18
dsaxenaarnd, we'll have to work on that too. gcl mentioned some ideas around clock support @ LPC, we need someone to take lead.15:19
linusw_arnd: do we have a way forward on generic clk device/subsystem now then?15:20
dsaxenalinusw_, mike t. and amit in the pmwg are taking lead of that . we're sharing rooms with them @ connect, so there will likely be a joint hacking session on that15:21
linusw_OK sounds like a plan, is tglx coming?15:22
dsaxenahmm, unsure, I will ping him15:23
dsaxenaok, one last thing on my end is that Mounir, Michael H., myself, and some other tech leads have put together a little document talking about the upstream kernel process15:25
dsaxenaand different approaches for it to map to Linaro's process15:26
dsaxenawe'll send out a link and would really like input15:26
dsaxenawe have a session at LC to talk about this, the hope is to have some agreements across the organization before we get there so it is not just us talking about how the process doesn't work :)15:29
mounirdsaxena, the hope is to get an agreement at connect how to fit upstreaming with Linaro process as best as we can15:31
dsaxenamounir, ack. i'm looking forward to having that discussion15:33
dsaxenaok, let's do a go around, let us know what you're focusing on and if there's any action you need from anyone else to help you15:34
dsaxenai'm going to be finishing up creating LC sessions and will need help on filling out the outlines15:35
dsaxenaalso working on getting more requirements into our requirements system and working on an overall kernel consolidation roadmap15:35
dsaxenaarnd, want to start?15:36
linusw_dsaxena: up in the log I noted something like more consolidation == more subsystems needing to be created15:36
arndsure15:36
arndI have spent a lot of time merging stuff and verifying the results15:37
arndthe randconfig work is starting to pay off for me, I could take the for-next branch and merge it with the randconfig branch and see everything that broke :)15:37
dsaxenaarnd, have you subscribed to the kernel ci build's RSS feed for arm-soc?15:38
arndI've manage to get a few of the randconfig branches into maintainer trees, but there is much more work, and I found another ~30 problems checking linux-next with it15:38
arndyes15:38
arndbut it seems that always breaks with git errors in the last weeks15:39
arndI've also helped out a few people doing patch reviews last week, but nothing spectacular there15:40
arndhopefully we got the ball rolling on the CMA patches again, a few people sent ACKs back to that, but it might be too late for 3.215:40
arndI think that's all for me this time15:41
dsaxenaarnd, tnx. i'll look into the git issues15:42
dsaxenawith the builds15:42
dsaxenadmart, you're it15:42
dsaxenadmart, still around?15:44
dsaxenashawnguo, want to give a quick update on what you're working on?15:44
shawnguodsaxena: sure15:44
shawnguoback from holiday, I'm preparing the pull-request for imx5-dt and imx6q series to Arnd.  And I also start looking into pinctrl subsystem for migrating imx iomuxc to it.15:45
linusw_shawnguo: much appreciated!15:46
dmartdsaxena: yep, sorry -- I'll wait for shawnguo15:46
shawnguolinusw_: I hope Linus T will pull your bits for v3.2 ;)15:47
shawnguodmart: that's it for me15:47
dsaxenashawnguo, tnx!15:48
dsaxenadmart, ping15:49
dmartdsaxena: mostly last week, I was working on the amba driver module autoloading issue15:51
dmartI have acks an about half of the patches15:51
dmartThis would have solved the aaci brokenness in the 11.09 image for vexpress (though really it was a config regression, it exposed the problem)15:52
dmartI also completed finished the Thumb-2 kernel porting advice.  What's the best way to advertise this to the landing teams?15:52
linusw_dmart: what about a Googledoc and share it "view" with everyone?15:53
shawnguodmart: great document, thanks a lot15:53
dmartWe've been building Thumb-2 for all the non-LT kernels for a while now, but except for omap our tree lacks the scariest code15:53
dsaxenadmart, put it on the linaro wiki and post to linaro-dev and cc: techleads@linaro.org15:53
dmartlinusw_: Currently it's on the wiki -- I posted to linaro-dev, but that may tend to be buried in the noise15:53
dmartI'll try the techleads alias15:53
linusw_dmart: maybe post it to linux-arm-kernel@infradead.org? :-)15:54
dsaxenamounir, we should make sure to include that in this week's kernel summary if it didn't make it on last weeks15:54
dsaxenalinusw_, dmart +1 on the l-a-k posting15:54
dmartlinusw_: Hmmm, that could be a good idea15:54
dmartIt's only draft quality -- if anyone has any comments, suggestions or fixes15:54
dmartthen please let me know15:54
mounirdsaxena, ok15:54
dmartThere will also be mistakes, though hopefully not too many :P15:55
dsaxenadmart, will do!15:55
linusw_dmart: don't worry Russell will let you know all mistakes :-)15:55
dmartLast time the topic came up, Russell decided that Thumb-2 was "too weird" for him15:56
linusw_dmart: I can forward a copy to Rabin Vincent @STE he sort of like Thumb2 stuff I think15:56
dmartThere's nothing secret, so feel free to forward links15:56
dsaxenajstultz_vm, how goes android hacking?15:57
dsaxenadmart, tnx for the work on the doc, i think it would also make a good presentation for conferences...15:57
jstultz_vmdsaxena:  ok.. focusing on the scheduler freezer idea right now to try to promote the wakelock alternative idea i had..15:58
jstultz_vmalso having some discussions on the pm list about what userland approaches we could also try.15:58
jstultz_vm(although i'm skeptical)15:59
jstultz_vmdsaxena:  need to re-ping folks on the config merge script.. last submission didn't get any feedback..15:59
jstultz_vmdsaxena:  basically that's all right now..  need to figure out what the story will be for 11.10 android kernel, since nico is thinking 3.1 for the linaro tree.. so i might have to pull the unofficial android fwd port that andy green did16:01
jstultz_vm(or i guess has in his tree, was done internally by TI)16:01
linusw_jstultz_vm: moving ahead of Android official baseline will anyway be necessary for things like Cortex A15 support...16:02
dsaxenais it just an android fwd port or does it have other bits in it too? I always worry a bit about vendor-internal Android trees16:02
jstultz_vmlinusw:  yea.. with android.git.kernel.org still down, its a bad spot.16:03
jstultz_vmdsaxena:  i think its just the forward port..16:03
dsaxenaok, good to hear.16:03
jstultz_vmdsaxena:  anyway, i'm a little worried about stability and bugs, so 11.10 might be a rougher release.16:04
jstultz_vmhopefully 3.1 is solid enough already..16:04
dsaxenanpitre, jstultz_vm can we do an -rc pre-release to play with this week?16:05
dsaxenadmart, sounds like we should pull in your amba patches into 11.10?16:05
jstultz_vmdsaxena:  yea, i was planning on having something ready this week for testing..16:06
dsaxenajstultz_vm, ok, great16:07
dsaxenatnx!16:07
dsaxenalinusw_, what's on your plate this week?16:07
linusw_dsaxena: pin control core + pinmux stabilization for the merge window16:08
linusw_dsaxena: then I'm researching pin control generics for biasing, driving etc.16:08
linusw_dsaxena: going to look up dept. of applied electronics @ Lund University and ASIC dept. @ STE for that stuff16:08
linusw_dsaxena: found out what was blocking NHE:s work on DT16:09
linusw_dsaxena: random reviews of random stuff and thinking out loud as usual16:09
dsaxenalinusw_, ping shawnguo, rajendra, thomas abraham, to get input on other SOCs16:09
dsaxenalinusw_, what was the DT issue?16:09
linusw_dsaxena: compat node didn't match compiled-in value, then the error message was on earlyprint and earlyprint didn't work16:10
linusw_dsaxena: I'm not very good at coordination, how do I know the key person for each member platform?16:11
dsaxenaso past it and we can boot with DT now?16:11
linusw_dsaxena: I sorta need a list to go over or so16:11
dmartdsaxena, npitre: I have no change requests on that series from anyone, and the changes are pretty straightforward -- so I think these could go into 11.1016:11
dsaxenalinusw_, ask me for now. :) we'll sort it out @ connect16:11
linusw_dsaxena: the HREF500 boots with DT I think nhe still has some problem with the Snowball so I guess I fix that next...16:11
dsaxenalinusw_, and the go-to person might change per project too as people rotate into differnt things16:11
dsaxenalinusw_, ok16:12
linusw_dsaxena: OK I guess they will come to me too...16:12
linusw_dsaxena: and if nothing happens I'll force-convert their platforms myself haha16:12
dmartdsaxena, npitre: For now though, it's not very interesting -- we do (or should) build most amba drivers into the kernel anyway.  And once the relevant drivers are fully DT-ised, I believe autoloading should be taken care of the OF way anyhow, at least for fdt-based kernels16:13
linusw_dsaxena: that's all for me.16:13
dsaxenalinusw_, tnx16:13
dsaxenai think that was everyone who's here today?16:15
nheMe also...16:15
dsaxenanhe, hi16:16
dsaxenanhe, sounds like some good progress on DT!16:16
nheShort then, I am on dt for snowball.  Used some three days for some adm that have been stacked up. Unless a Linaro <-> STE presentation goes as Linaro work, then 1.5 days16:16
nhemounir: The bp is updated and the items should be trackable and nice.16:17
mounirnhe: thank you for handling16:17
nheok, next, if there is someone left...16:17
dsaxenanhe, i think that presentation counts, it's important to get intenral teams up and running on DT16:18
dsaxenai think that's it?16:18
nhedsaxena: not dt, Linaro, slightly different...16:18
dsaxenanhe, ah ok16:19
dsaxenaok, we're a bit over 1 hr, so thank you all for your time16:19
linusw_thx16:20
dsaxenagcl, ping21:00

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!