Wednesday 24 Nov 2010
- slangasek: vacation
- Review action items from last meeting
- Blueprint status
Action Items from this Meeting
- dmart to find out who's in charge of omap kernels now
- Wookey to discuss about gcc-defaults-armel-cross with hrw
- jcrigby to upload/get uploaded omap kernel
Action Items from Previous Meeting
- npitre to write up more formal documentation of what our kernel tree is (and isn't)
- aviksil to talk to Mathieu about creating an lttng "for-linaro" branch that will merge cleanly
- wookey to chair next week's foundations meeting
cross compilers for archive
- discussed with doko about dropping update-alternatives use
- wrote gcc-defaults-armel-cross 1.4 which does proper symlinks for cross compilers
- wrote gcc-4.5-armel-cross 1.41 which removes update-alternatives support
- wrote gcc-4.4-armel-cross 1.37 which removes update-alternatives support
- wrote armel-cross-toolchain-base 1.53 which has all updates which I had
- sent all of them to Steve for review
- Status of changes:
- default version of armel cross compiler will be 4.5 like it is in Natty
- both 4.4 and 4.5 will be provided as it is for native
- any traces of update-alternatives use should be removed
- Needs to be done:
- adding conflicts on older cross compilers to gcc-defaults-armel-cross
- Order of upload to archive:
- Took a look at powerdebug packaging by Amit Arora and sent him few suggestions
- Test builds of qt4-x11 packages
- Checked few old bugs do they still apply:
- Bugs #646729, #637454, #671455 are done with armel-cross-toolchain-base 1.52 (landed in maverick-proposed)
* posted various patches for Thumb-2 kernel enablement * (probably) ported SMP-on-UP to be Thumb-2 compatible -- looking good so far, but needs to be tested on a real uniprocessor platform.
* spec drafting * started looking at Freescale's i.MX BSP code for 2.6.35
* spec drafting
* spec drafting
* brief testing of the 11.05 final release on vexpress
* set up a kernel git tree on git.linaro.org * reposted outstanding perf tools patch to linux kernel mailing list
* Test and post SMP-on-uniprocessor Thumb-2 compatibility patches * Feed back on Freescale BSP
* 2010-12-09 until early January (to be confirmed)
This week I worked on mostly kernel packaging.
Wrote a how to package a kernel wikipage with accompanying template packaging tree: https://wiki.linaro.org/PackageYourOwnKernel
- Produced the first natty linux kernel and linux meta package source
- New approach to applying ubuntu sauce should make it easier to validate what sauce/security fixes we have.
- Includes new autogenerated per-flavour source packages
Next week I plan on
- Doing a first natty u-boot release
- Starting to upstream ST-Ericsson UX500 U-Boot
- Built more packages; 5 packages (and their dependencies) remain to be built
Dependencies include qt & perl
- Two stage build approach showing promise
Raised bugs on https://launchpad.net/cross-patches for the packages cross patched so far
- Built more packages; 5 packages (and their dependencies) remain to be built
- tried installing the yocto plugin - failed on dependencies
- Listened to several planning meetings
armel eclipse does not install - version differs from that in AMD64/i686, won't build on efika SmartBook
- Linaro alip contains an additional 700 packages.
- Continue package building with xdeb
- Add packages in the linaro alip-ael missing from the current list
- Revisit cross patches, when patch acceptable, raise "also affects" against package concerned
- Import existing projects
- Build eclipse with xdeb
- Xmas hols 13th Dec 3rd January inclusive
- Pondering about the actual usefulness of the linaro-next tree. Proposition of a new kernel process where the Linaro stable branch closely follows mainline releases and would also accept stuff not in mainline yet as long as it has a roadmap to mainline.
- More discussion about the Linaro kernel vs Ubuntu kernel, or the Linaro kernel in general. Queries are coming to me from different independent sources about what the Linaro kernel actually is. In connection to the above item this needs to be more formally documented.
- Review/discussion of clock source vs sched_clock() implementations.
- Gave Steve Langasek a few recipes to extract Linaro related Git stats for the kernel.
Meeting opened by wookey at 16:00
<wookey> woo - it works
<wookey> just going down corridor to kick missing types
wookey puts of steve-shaped hat.
<wookey> OK, Review action items from last meeting
<wookey> npitre to write up more formal documentation of what our kernel tree is (and isn't)
<wookey> did that happen at all?
<wookey> ah yes - thank you
<wookey> do feel free to do any missing chairmanning as we go along
<npitre> this didn't happen yet
<wookey> that gets bumped
<wookey> aviksil: is allegeddly not here this week
<wookey> pipe up if you are really
<wookey> and chariing is indeed occuring
<aviksil> i'm here
<wookey> ok did you talk to Mathieu about creating an lttng "for-linaro" branch that will merge cleanly
<aviksil> yes. He agreed to create "lttng-for-linaro" rebase branch that will be based on linux-linaro-2.6.xx
<wookey> OK, so that's sorted then?
<wookey> right. on to the blueprint excitement
<wookey> echo $(shuf -e hrw npitre jcrigby dmart wookey slangasek ppearse JamieBennett aviksil)
<wookey> dmart wookey JamieBennett hrw slangasek aviksil jcrigby npitre ppearse
wookey suddenly realises he doesn't actually know who's doing what (you can say anything you like this week
npitre doesn't always know what he's doing either
<dmart> this week I have been mostly hacking Thumb-2
<wookey> for the kernel?
<dmart> I can now build a single kernel on Thumb-2 which boots SMP on RealView PBX-A9 and uniprocessor on PB-A8
<dmart> ^ yes
<dmart> I don't have a working omap kernel yet ... I'm trying to get to that stage so as to have something linaro folks can actually test
<dmart> At the moment, there seem to be dragons in some low-level power management code
<dmart> Pinged Richard Woodruff for comments, but I haven't heard back yet.
<wookey> and he just left - don;t know if that matters
<dmart> ah ... I wasn't aware of that. That may be why he's quiet.
<wookey> (email only a few mins ago)
<dmart> Anyone know if he will still be a maintainer for omap stuff in the kernel?
<wookey> apparently no-one here
<npitre> no idea -- the announcement is thin on details
<hrw> dmart: was not it Tony?
<dmart> I guess -- I should try him.
<dmart> Richard wrote the actual file I'm having fun with
<npitre> hrw: Tony doesn't work for TI, does he?
<dmart> MAINTAINERS lists him for OMAP
<wookey> [ACTION] dmart to find out who's in charge of omap kernels now
dmart to find out who's in charge of omap kernels now
<hrw> npitre: I track kernel randomly.
<wookey> (perhapsa slightly gratuitous action, but I have a new toy
<npitre> Tony is sorta the TI "community" maintainer
<dmart> Does anyone have a feel for the priority of getting things like ftrace and kprobes working?
<npitre> but he used to work for Nokia
<dmart> Thumb-2 currently breaks these
<wookey> OK, Tony Lindgren?
<dmart> I think he's the next person to ask, so I will
<npitre> ftrace should be fairly simple compared to kprobes
<dmart> npitre: I'd had the same suggestion from others.
<wookey> dmart: you mean priority of getting them working in thumb2 or that work instead of thumb2 work?
<dmart> In addition - thumb2 is the base requirement for the standard architecture spec
<dmart> I was wondering which is considered more important/useful. Any excuse not to loot at kprobes will be very welcome
<wookey> I have no idea. Steve might.
<wookey> ask on the list, I suggest
<dmart> I'll ping a mail to linaro-dev
<npitre> I'd start with ftrace as it should be fixable quickly
<wookey> or do the one you prefer
<wookey> OK, anything else?
<dmart> OK, so it's ftrace or ftrace
<dmart> I have a pending task to review Freescale's 3.6.35 bsp which I still need to feed back on.
<dmart> Otherwise, that's about it.
<dmart> erm, 2.6.35
<dmart> (unless you use my time machine)
<wookey> me next
<wookey> refined xdeb UI spec some more, got a sore head thinking about stuff and decided time to stop thinking and start coding
<wookey> xdeb uploaded to debian - needs some work to remove distro-specific config for deb sources (started but not working right yet)
<wookey> currently in new queue
<wookey> Started adding multiarch support. Currently trying to determine if spec is actually sufficient.
<wookey> lots of distractions:
<wookey> emdebian total server loss (had to sort out a new one)
<wookey> organising arm/emdebian sprint at ARM
<wookey> grokking desktop files/mime types
<wookey> Fixed up ArmHardFloatPort wiki page after meego people use it to choose wrong armhf name
<wookey> beware ancient arguments lying about on wikis
<wookey> ok, that's me - any questions?
<wookey> otherwise: JamieBennett:
<JamieBennett> No blueprint news from me. The past week has been the MeeGo conf, working on private projects, tracking project. The coming week will be more project work, more tracking work, launchpad changes in setup for the natty cycle, enabling natty builds and preparing the alpha-1 release, so not much to do
<wookey> did you see the efficiency there? comapre with that verbose fool wookey
<wookey> ok, hrw:
<hrw> working on updating cross compiler in natty
<hrw> prepared all components, got hit by bzr<>git exchange
<hrw> lool reviewed my code == I need to redone some of changes
<hrw> npitre: bzr is pain
<hrw> working on bug 676454
<ubot2> Launchpad bug 676454 in gcc-defaults-armel-cross (Ubuntu) (and 2 other projects) "switch to gcc-4.5 and symlinks for default versions (affects: 1) (heat: 8)" [Undecided,In progress] https://launchpad.net/bugs/676454
<hrw> few other bugs reported/fixed like byobu or bzr-git
<hrw> thats all
<wookey> OK, thanks for that. I'll talk to you about the defaults setup sometime, but not now
<wookey> slangasek skiving and eating turkey
<aviksil> Nothing much to update, as I'm busy with training and conference. Just executed my action item.
<aviksil> thats all
<wookey> OK, that was easy, we should be able to slack of early for coffee
<hrw> [action] Wookey to discuss about gcc-defaults-armel-cross with hrw
<jcrigby> finished first natty kernel release but no upload yet
<hrw> hmm.. chairmaster needs to do it or it has to be uppercase?
<wookey> heh - only I can do actions
<hrw> [ACTION] Wookey to discuss about gcc-defaults-armel-cross with hrw
<wookey> [action] Wookey to discuss about gcc-defaults-armel-cross with hrw
Wookey to discuss about gcc-defaults-armel-cross with hrw
<jcrigby> omap kernel works on both omap3 and omap4
<jcrigby> got confused about ehci on omap4 but anand set me straight
<jcrigby> thats about it
<npitre> jcrigby: I just merged and pushed out the OMAP4 ehci patches
<jcrigby> npitre, thanks I'll grab your latest
<jcrigby> I guess we will have a second natty release soon, still not sure about upload
<jcrigby> slangasek is out
<jcrigby> maybe I will ask lool
<JamieBennett> jcrigby: Yes, please get it uploaded asap
<jcrigby> JamieBennett, yes sir
<JamieBennett> jcrigby: We have alpha soon
wookey just found https://wiki.linaro.org/Platform/Foundations/Status, but it seems to be maverick vintage?
<wookey> JamieBennett: is alpha the point when we stop syncing from debian?
<hrw> wookey: yes, it is maverick
<lool> jcrigby: hola
<JamieBennett> wookey: No
<JamieBennett> That's 30th
<wookey> lool: jcrigby needs you to be steve for uploads
dmart has to disappear - does anyone need anything else?
<wookey> you can sort that out between you I assume.
<wookey> dmart: no, bugger off
<wookey> [action] jcrigby to upload/get uploaded omap kernel
jcrigby to upload/get uploaded omap kernel
<wookey> (is that right?)
<wookey> npitre, your turn I think
<npitre> been reviewing a bunch of patches, including dmart's thumb2 ones
<npitre> merged 18.104.22.168 into the linaro 2.6.36 kernel, also a few additional bugfixes, and the OMAP4 ehci
<npitre> answered a few other queries for what the Linaro kernel actually is (meaning I should really write that document)
<wookey> keep asking him qeustions people - it's the way to get good docs
<npitre> I did send slangasek a few Git recipes to create Linaro statistics (this was a previous action of mine)
<npitre> yeah... by answering all those people I'll simply have to cut'n paste from my sent email folder
<wookey> ppearse: your up
<ppearse> OK - still x building. Broke my qt/poppler/cups loop. The last 5 packages (of alip-ael) are the hardest
<wookey> you're (shame)
<ppearse> Each seems to have 10's of dependencies, all built with different tools. I have to learn to speak bjam & scons....
<ppearse> Pushed some cross patches up, got push back.
<ppearse> Have to try krw's latest toolchain to see if it solves any library searches.....
<wookey> I'll come over and share your pain a little - I'm sure it's interesting
<hrw> ppearse: newer packages for compiler mostly adapts to natty versions
<ppearse> ? to watch others suffer?
<wookey> pushback saying 'don;t fix it like that'?
<ppearse> hrw: Ah - I forgot. OK those patches can wait till later
<ppearse> wookey: Mainly "Why can't your cross compiler find those?"
<wookey> right? anything else from anyone?
<ppearse> That's it from me
<wookey> well, I'll let you all get back to work. Thanx for being such an amenable audience.
Meeting closed at 16:47
Platform/DevPlatform/Meetings/2010-11-24 (last modified 2011-03-11 11:34:13)