Menu

Main Menu
Talk Get Daily Search

Member's Online

    User Name
    Password

    Arch Linux ARM on N900

    Reply
    Page 32 of 64 | Prev | 22   30     31   32   33     34   42 | Next | Last
    Skry | # 311 | 2013-04-09, 19:24 | Report

    re Xorg issues..

    It seems the same issue is present for raspi folk, and even someone using chromebook, there's an issue opened at archlinuxarm issue tracker for it. So, nothing to do but wait for update that fixes it or revert to 1.13.

    EDIT:
    https://github.com/archlinuxarm/PKGBUILDs/issues/446
    http://www.raspberrypi.org/phpBB3/vi...38568&p=320464

    Edit | Forward | Quote | Quick Reply | Thanks

    Last edited by Skry; 2013-04-09 at 21:56.
    The Following 5 Users Say Thank You to Skry For This Useful Post:
    AhmadMhd, caveman, handaxe, int_ua, PanzerSajt

     
    int_ua | # 312 | 2013-04-09, 21:50 | Report

    Originally Posted by Skry View Post
    there's an issue opened at archlinuxarm issue tracker for it.
    Can you please add the link?

    Edit | Forward | Quote | Quick Reply | Thanks
    The Following User Says Thank You to int_ua For This Useful Post:
    misiak

     
    syrjala | # 313 | 2013-04-10, 22:00 | Report

    Originally Posted by Skry View Post
    Also, I noticed that X crashes appear even with the regular fbdev driver. Not sure what to do about this, might need to revert to 1.13.
    I'm surprised that you got the fbdev-sgx driver to work w/ more recent X servers at all. Mainly I would expect that you'd need to forward port all the dri2 changes from the harmattan xserver that never made it upstream.

    Edit | Forward | Quote | Quick Reply | Thanks

     
    Skry | # 314 | 2013-04-10, 22:18 | Report

    Forward port for 1.13 was done by Marko Saukko, all I did was a tweak for 1.14 and corrected some paths here and there to match dss changes in more recent kernels, packaging etc. For kernel driver, I bluntly removed dependency to dss bits that appeared to be not upstreamed. I'm not sure at all it's done right, or if it's enough but at least it's been usable for 2D, and the few GLES tests I've done, did work.
    So, we have usable driver, which is somewhat faster and more featurefull than the plain fbdev to use until someone comes up with a better one. Unfortunately I don't possess the skills to do that.

    Edit | Forward | Quote | Quick Reply | Thanks
    The Following User Says Thank You to Skry For This Useful Post:
    caveman

     
    syrjala | # 315 | 2013-04-11, 00:09 | Report

    Originally Posted by Skry View Post
    Forward port for 1.13 was done by Marko Saukko, all I did was a tweak for 1.14 and corrected some paths here and there to match dss changes in more recent kernels, packaging etc. For kernel driver, I bluntly removed dependency to dss bits that appeared to be not upstreamed. I'm not sure at all it's done right, or if it's enough but at least it's been usable for 2D, and the few GLES tests I've done, did work.
    So, we have usable driver, which is somewhat faster and more featurefull than the plain fbdev to use until someone comes up with a better one. Unfortunately I don't possess the skills to do that.
    I was refering to dri2 patches in the xserver itself, not in the driver. I suspect the result is less than good if the xserver's dri2 code is taken from upstream. But it's possible that the problems won't become visible unless you have a GLES based compositor.

    As far as the kernel goes, IIRC the whole vsync infrastructure never got fully upstreamed. That may mean that page flipping either won't work, or blows up as well

    Edit | Forward | Quote | Quick Reply | Thanks
    The Following 2 Users Say Thank You to syrjala For This Useful Post:
    int_ua, Wikiwide

     
    Skry | # 316 | 2013-04-11, 05:20 | Report

    Originally Posted by syrjala View Post
    I was refering to dri2 patches in the xserver itself, not in the driver. I suspect the result is less than good if the xserver's dri2 code is taken from upstream. But it's possible that the problems won't become visible unless you have a GLES based compositor.
    Yes, less than good seems to be the case here. Actually, compiling with -DCOMPOSITE etc does not even work currently, so the support is not there.

    Originally Posted by syrjala View Post
    As far as the kernel goes, IIRC the whole vsync infrastructure never got fully upstreamed. That may mean that page flipping either won't work, or blows up as well
    That is exactly what the state currently is However, I've seen it cause (visual) trouble only with hildon-desktop for far.

    Edit | Forward | Quote | Quick Reply | Thanks
    The Following 2 Users Say Thank You to Skry For This Useful Post:
    int_ua, Wikiwide

     
    nokiabot | # 317 | 2013-04-14, 03:43 | Report

    whats going up?? update a bit abt achivments and still to be achived

    Edit | Forward | Quote | Quick Reply | Thanks
    The Following User Says Thank You to nokiabot For This Useful Post:
    Wikiwide

     
    Skry | # 318 | 2013-04-14, 09:46 | Report

    Originally Posted by nokiabot View Post
    whats going up?? update a bit abt achivments and still to be achived
    I'm busy and currently have very little time to put into this, if at all.

    Some goals to achieve / ideas / todo:

    Cleanup all PKGBUILDS etc and put them to github so my absence can be compensated if necessary.

    Package Xorg 1.13. Fbdev is broken apparently upstream, so this needs to be done to restore functionality. I will try to do that at some point if no one else is going to.

    Update all neon builds.

    In addition to neon builds of packages, provide also egl/gles2 builds (like Cairo, Enlightenment, etc)

    Provide Weston with fbdev backend, and maybe something else too supporting framebuffer at compile time, like netsurf.

    Sort Pulseaudio mess. Currently call audio works but anything else makes PA go 80% CPU time. Needs packaging.

    SGX binaries need repackaging. This will probably end in me compiling mesa too.

    New rootfs, especially since netctl is out.

    Get people involved, I can't handle this currently by myself.

    So, what I'm doing? I'm working on Awesome configuration for touchscreen devices, partially inspired by Hildon desktop. I've also been doing some theme work in that department, and otherwise.

    Kernel side is completely in other hands, so we're going with the current 3.5 until upstream kernel is usable.

    I guess that's about it. Comments, questions and suggestions are very welcome.

    Edit | Forward | Quote | Quick Reply | Thanks
    The Following 15 Users Say Thank You to Skry For This Useful Post:
    AhmadMhd, ayazpak, caveman, Estel, fw190, int_ua, madry72, MetalGearSolid, mr_pingu, nokiabot, PanzerSajt, preflex, Sohil876, Wikiwide, Älä hakkaa

     
    Android_808 | # 319 | 2013-04-14, 12:04 | Report

    skry, are the android drivers for sgx in better shape? just wondering if maybe using them with libhybris would sort out some of your issues.

    Edit | Forward | Quote | Quick Reply | Thanks
    The Following User Says Thank You to Android_808 For This Useful Post:
    Wikiwide

     
    Skry | # 320 | 2013-04-14, 19:07 | Report

    Originally Posted by Android_808 View Post
    skry, are the android drivers for sgx in better shape? just wondering if maybe using them with libhybris would sort out some of your issues.
    No idea, I can take a look.

    Edit:
    If I understood correctly, this would require us to use Android kernel.

    Edit | Forward | Quote | Quick Reply | Thanks

    Last edited by Skry; 2013-04-14 at 20:27.

     
    Page 32 of 64 | Prev | 22   30     31   32   33     34   42 | Next | Last
vBulletin® Version 3.8.8
Normal Logout