Testdriver from http://www.fairlite.demon.co.uk/intel.html gives screen corruption on my Motion Computing M1200 Tablet PC with 830MG graphics chipset. * The driver starts at 1024x768 which according to Alan it shouldn't do, due to the Video BIOS settings. This is probably because the built-in LCD panel has a resolution of 1024x768. * The lower half of the screen is black. Though the mouse cursor remains visible in this area. * The top ~20 pixels of the screen contain a vertical interlacing green-black line pattern with on top of it symbols from the Matrix written in red (well, they look like that.. probably some data tables). * Below that is a dark area * At 1/3 of the screen a part of the in memory remaining screenbuffer is displayed. After a couple of reboots this shows as red noise. At first this looked like KDM with only a couple of red pixels (which confused me into thinking KDM was still running). * The mouse cursor disappears after ~5 sec and doesn't come back. * Clicking seems to crash X. It powers down the screen and comes back up with (the same?) screen corruption. * Pressing Ctrl+Alt+F1 to switch to the console doesn't seem to anything. * Lateron I found out that you could bring a corrupted console on the screen by starting KDM over SSH (the init.d script reported that KDM was not running). You can still login, and see that you are typing, but you can't really see what's written. xorg.conf can be found under Bug 3879 comment 2 Commenting out the ForceBios option also shows screen corruption. Setting the NoAccel option has the same impact. xorg.0.log can be found under Bug 3879 comment 4 This is a log with ForceBios enabled, I will supply one with this option disabled. Bug 387 comment 4 also contains MD5 checksums of Alan's driver. If Alan could supply his checksums we could check if my download wasn't corrupted. 'wget' marked this driver as being from 11 july 2005 00:11 (my timezone Amsterdam/The Netherlands, which is currently UTC+2).
The comment link is incorrect and should point to Bug 3879 comment 4, sorry.
Created attachment 3161 [details] Xorg config file.
O.k. I've spotted something even more screwy with your Video BIOS. It can't do 640x480 either. In fact the only two resolutions it can set is 800x600 or 1024x600 at all bit depths and 640x400 at 8bpp. It's definately a weird BIOS. I've updated my test driver which might help overcome this corruption.
Okay, this fixes the screen corruption when X is running at 800x600. Though it still shows before the checkerboard patter is shown, and after you quit a KDE session and return to KDM. And it breaks when you switch to a console and back to the running X session (aka Ctrl+Alt+F1, type something, Ctrl+Alt+F7)
"Breaks" as in, the screen is frozen..
From a remote SSH session: MotionM1200:/home/henk# /etc/init.d/kdm stop Stopping K Display Manager: kdm not responding to TERM signal (pid 6046).
You might want to try adding Option "VBERestore" to your Device Section.
Created attachment 3163 [details] [review] Make application/mathematica a subclass of text/plain It decides all by itself that it should be disabled (doesn't matter if I add "true" or not): (**) I810(0): Option "VBERestore" "true" .. (**) I810(0): VBE Restore workaround: disabled. Maybe this is as expected? Anyways it still hangs on switch back to X.
You should have specified "false" not "true"
Can you post a log with and without VBErestore specifed including your VT switch. So you might have to ssh in and get the logs after the VT switch.
Created attachment 3164 [details] [review] mesa_radeon_cubemap_3tmu_20050903.diff.txt btw, I added the synaptics driver to my config, sorry if that's confusing. ..and SSH over Wifi is not fun :-/ (laaatency)
Found the problem in my test driver, will need to work out a fix for this bizarre BIOS.
You might want to upgrade to xorg 7.0 and try the latest test driver.
(In reply to comment #13) > You might want to upgrade to xorg 7.0 and try the latest test driver. I haven't been able to get it to work at all with the i810 driver. I get artifacts as described in 800x600 and 1024x768 at any bit depth. I can get into I suppose 640x400 using vga driver. vesa driver fails. I've done this on both X.org 6.8 and 7.0. Tried Option "VBERestore" "False" in the device section under X.org 7.0, same result.
Can you upload a log with 7.0 ?
Hey Nathan Curry, do you also have a Motion Computing M1200? Or is this another screen corruption bug? Anyways, thanks Alan for checking back on this bug. I've tried running the driver on the X.org 6.9 as supplied by Debian/unstable. It just crashes. Maybe I'll find some time to do the test-fetch-logfile-and-reboot cycle in next few days.
Any improvements using the 7.1 release of xorg?
Created attachment 5878 [details] X.org 7.0.0 i810 1.5.1 crash Fatal server error: lockup Error in I830WaitLpRing() [..]
I don't think X.org 7.1 is available for Debian/Unstable at the moment. But I do have a logfile (see previous comment), as since X.org 7.0.0 doesn't crash the entier machine anymore.
http://www.fairlite.demon.co.uk/intel.html doesn't seem to available anymore, where to get a testdriver now? btw, this log was after a fresh boot to kdm. At first I can see noise with a KDE cursor, then X seems to restart and shows the X-cursor, then another black screen and and I only see a black screen with some red noise at the top.
Created attachment 5880 [details] current xorg.conf
I've just reinstated i810_drv.so - try that.
That one doesn't work with my X.org version: [..] (II) Module i810: vendor="X.Org Foundation" compiled for 7.0.0, module version = 1.6.0 Module class: X.Org Video Driver ABI class: X.Org Video Driver, version 1.0 (EE) module ABI major version (1) doesn't match the server's version (0) (II) UnloadModule: "i810" (II) Unloading /usr/lib/xorg/modules/drivers/i810_drv.so (EE) Failed to load module "i810" (module requirement mismatch, 0) [..]
you could try the -ignoreABI flag to the Xserver.
Created attachment 5881 [details] log with -ignoreABI Xorg -ignoreABI seems to run more or less correct. I can see the X-cursor with garbage in the background. Cursor can move. xorg.0.log looks normal to me. But when [Ctrl]+[C]'ing Xorg, error messages get added to the log: (II) I810(0): [drm] removed 1 reserved context for kernel [..etc..]
btw, I started X.org over ssh, so [Ctrl]+[C] didn't involve VT-switching.
I've uploaded another - try again.
Created attachment 5882 [details] Xorg0.log Just ran XFCE4, long time no see on this Tablet PC. I think this driver works, at least in 800x600. Will test in other resolutions next.
The only modes that work according to your BIOS are 800x600 and 1024x600 and that's it. Pretty borked BIOS. Closing this now.
Indeed, other resoltions don't work. 1024x600 still crashes the Video BIOS.
There is some work going on in the git repository for native modesetting that will remove the reliance on the Video BIOS. That should help you.
For DELL Inspiron BIOS A07, the video is identical to the ones described below, all mode for 800x640 and 1024x800, and only 8bpp for 640x480. Using Xorg 7.0.0, it is choppy and non-viewable. DOesn't crash... Shall I open a new bug? Or there is one related to the i810? # lspci 00:02.0 VGA compatible controller: Intel Corporation 82815 CGC [Chipset Graphics Controller] (rev 11) # lspci -n 00:02.0 0300: 8086:1132 (rev 11) # lshw *-display description: VGA compatible controller product: 82815 CGC [Chipset Graphics Controller] vendor: Intel Corporation physical id: 2 bus info: pci@00:02.0 logical name: /dev/fb0 version: 11 size: 64MB width: 32 bits clock: 66MHz capabilities: vga bus_master cap_list fb accelerated configuration: depth=8 driver=i810fb frequency=60.24Hz mode=640x480 visual=pseudocolor xres=640 yres= 480 resources: iomemory:f8000000-fbffffff iomemory:f4000000-f407ffff irq:10
open a new bug as the i810/i815 don't use the BIOS.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.