Poulsbo update

Just a quick update on the Poulsbo stuff.

It's in review for RPM Fusion, now - see here. The current packages in the repo should be pretty damn bulletproof, up to kernel 2.6.30. I'm running them on 2.6.30 on my Vaio P now and they're solid, it stays up forever and everything works.

I starting fooling around trying to get something going on Fedora 12 (Rawhide) this morning, but no joy. I've patched the kernel module and X driver to build on 2.6.31 and X server 1.7 respectively (with a little help from Adam Jackson - thanks), but they don't actually work; hangs on X startup with a blank screen, at least for me. I've been testing on a live USB stick so no way to grab the X log and see what's going wrong, for now. I tried a different approach for a bit - building X server 1.6 for Rawhide - but it's too icky, there's too many changes to underlying libraries and I'm not sure I want to rebuild the entire X stack. I may wind up having to do that in the end, but I'm not going to spend the time yet.

If anyone wants to take a look at the Rawhide builds, they're available: binaries here, sources here.


lucazade wrote on 2009-08-29 08:07:
Yes you can! :) The black screen could be solved, at least here, with IgnoreAcpi in xorg.conf.
adamw wrote on 2009-08-29 19:14:
Different problem. I've already patched the driver package to enable that option by default.
toalex77 wrote on 2009-09-03 20:00:
I've downloaded the src files, and I've compiled and installed libdrm-poulsbo, libva, psb-firmware, psb-kmod, xorg-x11-drv-psb and xpsb-glx packages on my openSuSE. Until 3 days ago, everything works properly (with 3D acceleration). In the last 3 days, instead, every time X server, starts, the server crash with this message: Backtrace: 0: X(xf86SigHandler+0x79) [0x80cbc79] 1: [0xffffe400] Fatal server error: Caught signal 11. Server aborting If I start Xorg manually, all works fine, and I can start xterm, glxgears, plasma, dolphin, kdesktop, .... The problem seems related to KDM. OpenSuSE has KDE4. Ah, if I add Option "ShadowFB" "true", the X server starts, but every time I launch a 3D application I obtain this error [drm:psb_xhw_add] *ERROR* No Xpsb 3D extension available. with no 3D at all.
adamw wrote on 2009-09-03 20:40:
I'm not sure I can help with that, not much info there. Perhaps you could switch to gdm as a workaround?
toalex77 wrote on 2009-09-04 21:56:
Today, I've tried with xdm, with no success. Then I've tried to start manually a complete session of KDE, and I've discovered that, if I launch /usr/bin/X -nolisten tcp and then /etc/X11/xinit/xinitrc KDE4 starts normally (with some problems derived from the absence of a X Authorization file). But, if I launch /usr/bin/X -nolisten tcp -auth xauthfile where xauthfile is generated with echo add :0 . $(mcookie) | xauth -f xauthfile and then /etc/X11/xinit/xinitrc the X server crash with the message that I've reported in previous comment. I've tried also with a basic version of xorg,conf with only Device section with no options at all. If you need more info, please, tell me what you need, and I'll try to provides it. Thanks.
pabr wrote on 2009-09-04 23:01:
toalex77, I have the same problem starting Xorg on Fedora. I thought it had something to do with whether the first X client is a windows manager or a user session; I hadn't noticed the link with xauth. Does your Xorg also die when waking up from standby ?
toalex77 wrote on 2009-09-07 12:38:
Ok, I've "solved". I've updated Xorg server from an additional repository with latest patches for 7.4 version (openSuSE 11.1 has Xorg 7.4-17.6.1 by default, but exist a repository with Xorg 7.4-70.2 updated at september 4, 2009), and then I've recompiled packages for FC12 (except the kmod from FC11). Now all works... ehm... "fine" with 3D enabled. I have to disable the IgnoreACPI and put it to false, otherwise, when the display goes sleep, it doesn't resume anyway. Beyond this, the only broken things, are the no-KDE4 icons (KDE3/GTK) in KDE4 system tray. They are all blacks.
teetee1 wrote on 2009-09-10 03:13:
I have this little box called fit-pc2 http://fit-pc2.com/wiki/index.php?title=Fit-PC2_Specifications and I installed Fedora11 and updated all the packages up to 9/7. I was hoping to use this computer for doing some sequence alignment tasks and thought that the xorg/GUI is not important but later realize that lots of reports requires GUI interface to visualize the results. After I installed the driver according to http://www.happyassassin.net/2009/08/10/intel-gma500-poulsbo-on-fedora-11-repository-with-working-3d-compiz-support/ and followed the comments to modify the xorg.conf, and then startx, the X could not start and left me these error messages below: (EE) PSB(0): the stolenBase is:0x3f800000 (EE) PSB(0): screnIndex is:0;fbPhys is:0x3f800000; fbsize is:0x007bf000 (EE) PSB(0): Probe Address LVDSBLC_B Slave 0x58 failed. (EE) PSB(0): has_fbdev is true (EE) PSB(0): Probe Address LVDSBLC_B Slave 0x58 failed. (EE) PSB(0): Probe Address LVDSBLC_B Slave 0x58 failed. (EE) PSB(0): Probe Address LVDSBLC_B Slave 0x58 failed. (EE) PSB(0): Probe Address LVDSBLC_B Slave 0x58 failed. (EE) PSB(0): Probe Address LVDSBLC_B Slave 0x58 failed. Here is the system info: Loaded modules: Module Size Used by ipv6 235656 10 vfat 8600 1 fat 40628 1 vfat dm_multipath 14036 0 uinput 6676 0 snd_hda_codec_realtek 193372 1 mmc_block 9420 0 snd_hda_intel 24272 0 snd_hda_codec 59428 2 snd_hda_codec_realtek,snd_hda_intel snd_hwdep 6724 1 snd_hda_codec snd_pcm 62796 2 snd_hda_intel,snd_hda_codec psb 133464 1 snd_timer 17864 1 snd_pcm video 18744 0 snd 50336 6 snd_hda_codec_realtek,snd_hda_intel,snd_hda_codec,snd_hwdep,snd_pcm,snd_timer drm 152600 1 psb output 2476 1 video soundcore 5476 1 snd sdhci_pci 6972 0 sdhci 17496 1 sdhci_pci i2c_isch 6984 0 snd_page_alloc 7776 2 snd_hda_intel,snd_pcm r8169 28412 0 mmc_core 48464 2 mmc_block,sdhci i2c_algo_bit 4852 1 psb mii 4044 1 r8169 pcspkr 2168 0 serio_raw 5088 0 i2c_core 25024 4 psb,drm,i2c_isch,i2c_algo_bit ata_generic 4288 0 pata_acpi 3644 0 pata_sch 3156 4 lspci: 00:00.0 Host bridge: Intel Corporation System Controller Hub (SCH Poulsbo) (rev 07) 00:02.0 VGA compatible controller: Intel Corporation System Controller Hub (SCH Poulsbo) Graphics Controller (rev 07) 00:1b.0 Audio device: Intel Corporation System Controller Hub (SCH Poulsbo) HD Audio Controller (rev 07) 00:1c.0 PCI bridge: Intel Corporation System Controller Hub (SCH Poulsbo) PCI Express Port 1 (rev 07) 00:1c.1 PCI bridge: Intel Corporation System Controller Hub (SCH Poulsbo) PCI Express Port 2 (rev 07) 00:1d.0 USB Controller: Intel Corporation System Controller Hub (SCH Poulsbo) USB UHCI #1 (rev 07) 00:1d.1 USB Controller: Intel Corporation System Controller Hub (SCH Poulsbo) USB UHCI #2 (rev 07) 00:1d.2 USB Controller: Intel Corporation System Controller Hub (SCH Poulsbo) USB UHCI #3 (rev 07) 00:1d.7 USB Controller: Intel Corporation System Controller Hub (SCH Poulsbo) USB EHCI #1 (rev 07) 00:1e.0 SD Host controller: Intel Corporation System Controller Hub (SCH Poulsbo) SDIO Controller #1 (rev 07) 00:1e.1 SD Host controller: Intel Corporation System Controller Hub (SCH Poulsbo) SDIO Controller #2 (rev 07) 00:1f.0 ISA bridge: Intel Corporation System Controller Hub (SCH Poulsbo) LPC Bridge (rev 07) 00:1f.1 IDE interface: Intel Corporation System Controller Hub (SCH Poulsbo) IDE Controller (rev 07) 02:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8111/8168B PCI Express Gigabit Ethernet controller (rev 02) 03:00.0 Network controller: RaLink RT3090 Wireless 802.11n 1T/1R PCIe uname Linux #1 SMP Thu Aug 27 21:34:36 EDT 2009 i686 i686 i386 GNU/Linux
teetee1 wrote on 2009-09-10 03:27:
In my Xorg errors above, there was also this message: -- Fatal server error: Couldn't find PLL settings for mode! -- I tried to use keyword to search online but still could not figure out where the real cause is. I only have a feeling that xorg was trying to detect the 2nd monitor (hence the word "slave" in the error msg) but failed. I just wish I know more about coding to be able to troubleshoot this problem better. By the way, I just want to say thank you for spending the time to port the driver into Fedora repository.
adamw wrote on 2009-09-14 22:20:
That does look like a problem with mode detection, yeah. However, the set of messages relating to 'LVDS' is unlikely to be to do with a second monitor, because LVDS always refers to the internal panel of a laptop. I wonder if psb is confused by the fact this _isn't_ a laptop and doesn't have an LVDS... you might want to look at the various parameters available for the psb kernel module, driver options available for the psb X driver, and xorg.conf options which can be used to disable certain outputs (see http://wiki.debian.org/XStrikeForce/HowToRandR12 ). Some combination of those may do the job of letting you circumvent the problem.
Sparky wrote on 2009-11-02 18:25:
Kudos Adam. Thanks to you I now have a stable Xserver on my brand new AO75h. I am not sure it is stable when plugging/unplugging USB devices. It was an easy YUM update. My xorg.conf file was updates, looks good. I have been testing it and I have a list of concerns I would like to ask about. 1. Scrolling in Firefox is sluggish. 2. Dragging windows by their decoration is sluggish. 3. I am not sure how fast glxgears should run (~60fps at 1366 x 768) 4. Video is sluggish (~1-2fps with hulu) 5. I did not notice any compile time needed by akmod. 6. I don't bother with compiz not sure what to test for. 7. Not sure if I need some sort of mplayer compile. 8. I get some errors and warnings in Xorg.log.0 This makes me think that my xserver is psb: $ lsmod|grep psb psb 133304 3 drm_psb 152224 3 psb i2c_algo_bit 4844 1 psb i2c_core 25032 4 psb,drm_psb,i2c_algo_bit,i2c_isch I am not sure what to make of these Xorg.log.0 Messages: $ grep '^([WE][WE])' /var/log/Xorg.0.log (WW) Falling back to old probe method for psb (WW) PSB: No matching Device section for instance (BusID PCI:0@0:0:0) found (WW) PSB: No matching Device section for instance (BusID PCI:0@0:28:0) found (WW) PSB: No matching Device section for instance (BusID PCI:0@0:28:1) found (WW) PSB: No matching Device section for instance (BusID PCI:0@0:31:0) found (EE) PSB(0): the stolenBase is:0x7f800000 (EE) PSB(0): screnIndex is:0;fbPhys is:0x7f800000; fbsize is:0x007bf000 (WW) PSB(0): BIOS panel mode data doesn't match probed data, continuing with probed. (WW) PSB(0): Unknown vendor-specific block f (EE) PSB(0): First SDVO output reported failure to sync or input is not trainded!!! (WW) PSB(0): Option "DRI" is not used (EE) PSB(0): has_fbdev is true (EE) PSB(0): First SDVO output reported failure to sync or input is not trainded!!! (WW) PSB(0): Unknown vendor-specific block f ... I get these errors but the xerver works. Some people with Xserver failures may have seen them too and may have been mislead. Is this as good as it gets or did I miss something?
adamw wrote on 2009-11-02 18:33:
Sparky: can you just check that Option 'MigrationHeuristic' 'greedy' got written into your /etc/X11/xorg.conf , as it should've done? If so, then performance ought to be as good as it gets, yeah. Your X log indicates it's definitely the psb driver running. I find desktop performance is rather better if you use Compiz rather than a regular window manager. Flash-based video playback will always be slow with Poulsbo in Linux at present because the CPU just isn't powerful enough to handle it, and the Adobe Flash plugin for Linux has no video playback acceleration support (so it can't use the Poulsbo's video playback acceleration capabilities). So that's normal. The WW / EE lines in your X log are fairly standard and don't look like anything to worry about. akmod won't show any compile time if there was a pre-built kmod package available for the kernel you're running. akmod only kicks in to do the compile if a pre-built kmod package isn't already installed.
Sparky wrote on 2009-11-02 20:24:
Wow Adam, that was quick, thanks. I got greedy. Compiz sounds like it might be worth checking out. I'll be watching to see how F12 goes. Would it help if I wrote Intel a letter?
Sparky wrote on 2009-11-18 02:23:
Hi Adam, Woot F12 is out! I'm not clear on Poulsbo. I saw that Moblin was added. I saw that Moblin might provided added support for Poulsbo. I saw some possible support for Flash acceleration with Poulsbo. Maybe when you get back from Seattle you could update us on the status of Poulsbo and Fedora 12.
sreejiraj wrote on 2009-11-21 05:26:
I'm installing the f12 psb driver from your f12 repo on a fresh install.. let's see.