Thanks for the tips again guys - I've checked out andy-tracking already so it's just a matter of syncing.
Also I had been using OE to build all of my binaries and filesystems using the org.openmoko.stable branch. Stable was giving me (ironically) several build errors, which i had submitted trac tickets and patches for. It's cumbersome to say the least. Just to ensure that I am on the right track, where is most of the current work happening? org.openmoko.dev ? FSO ? A different branch ? @ Lars - I'll check out your code as well - if I manage to make an improvement then I'lll request write access to your repo. Many thanks, Chris On Sat, Jan 31, 2009 at 8:18 AM, Lars-Peter Clausen <l...@metafoo.de> wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Hi > > Graeme Gregory wrote: >> Lars-Peter Clausen wrote: >>> I have a git repository with a cleaned up version of graemes xorg >>> driver located at [1]. Its quite stable(survied a 1 hour >>> `x11perf -all`) and exa acceleration is working, but you'll have >>> to apply [2] to your xserver to get it working. What still has to >>> be done to have a good and simple 2d accelerated driver is proper >>> randr support and some cleanups to the driver setup code. >>> >>> If you want to i can give you write permission to that >>> repository. >>> >>> Lars >>> >>> [1] >>> http://github.com/larsclausen/xf86-video-glamo/tree/xf86-video-glamo >>> [2] >>> > https://dev.openwrt.org/browser/packages/Xorg/xorg/xserver/xorg-server/patches/010-sync-dev-mem.patch >>> >> Hi, I haven't run it yet but a quick read of your code looks >> amazing. >> >> Can you explain the patch in [2] and I can get it into OE. >> Basically I need to know the "why?" and "will it break any >> platforms?". I can then get a recipe in OE to build your code. > It opens /dev/mem synchronously. Thus every write will block until the > data has been written to the actual hardware. This is necessary for > mmio, because reads/writes depend on that the hardware has received > previous commands. It wont break other plattforms, but may slow them > down. The patch is neither complete nor am I completely sure if its > right to make it depend on no mtrr available. The easiest solution > properly is to leave xorg untouched and add some code to the glamo > driver which mmap the mmio. > >> >> I think we should see about getting you write access on the OM git. >> Are you one of the guys who have an NDA? >> > yes, i recently signed the nda. > > Lars > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v1.4.9 (GNU/Linux) > Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org > > iEYEARECAAYFAkmET7MACgkQBX4mSR26RiNZrgCeNy5bnpPnLg/XizoDJbkaHIoS > YQ4AnAlTui5ub0v12j5aH/iPQ7Y5UnAR > =c87a > -----END PGP SIGNATURE----- > > > _______________________________________________ > devel mailing list > devel@lists.openmoko.org > https://lists.openmoko.org/mailman/listinfo/devel > _______________________________________________ devel mailing list devel@lists.openmoko.org https://lists.openmoko.org/mailman/listinfo/devel
