Summary: | siliconmotion driver locks-up notebook computer | ||
---|---|---|---|
Product: | xorg | Reporter: | James <james> |
Component: | Driver/siliconmotion | Assignee: | Xorg Project Team <xorg-team> |
Status: | RESOLVED INVALID | QA Contact: | Xorg Project Team <xorg-team> |
Severity: | critical | ||
Priority: | highest | CC: | billr, currojerez, domenico, erik.andren, ickers |
Version: | git | ||
Hardware: | x86 (IA32) | ||
OS: | Linux (All) | ||
Whiteboard: | |||
i915 platform: | i915 features: | ||
Attachments: |
Description
James
2004-11-30 15:38:27 UTC
Try using NoAccel *and* NoMTRR options. I have to do that to get my old laptop with a siliconmotion chipset to function properly. NoMTRR is critical, NoAccel will make things display properly. Thanks. Tried it - no joy. NoMTRR is not actually a valid option for the Silicon Motion driver, version 1.3.1. The option seems to be ignored by the driver. I did notice one oddity in the log file. Where the driver reports the probed size of "videoram" and doesn't accept a "videoram" option, the size reported is 4096kB. (--) Silicon Motionvideoram: 4096kB The "fbdev" driver, on the other hand, reports only 3072kB for this hardware, which is exactly 4 bytes per pixel for the 1024x768 LCD display. (II) FBDEV(0): Hardware: VESA VGA (vidmem: 3072k) How much memory is there, really? There's a problem if 3072k is all. Are you still experiencing this problem with a current release of xorg? Well, I haven't tried this again with a current release of Xorg, but, unless you believe in fairies, nothing will have changed. Don't get me wrong - fairies would be nice, but I think I would have heard. What I know is that the old XFree86 version 3 driver works, and that the version 4 driver is broken. The version 4 driver became the basis for the Xorg driver, and so is also broken. I had also noted a similar difference with the Trident driver, which I was trying to use on an old IBM Thinkpad 365XD. The driver would not start properly, though it did not crash the system. I spent some time emailing back and forth with Egbert Eich <eich@freedesktop.org>, trying to track down the problem, running the old version 3 trident driver in the newer version 4 XFree86 build. What I discovered is simply that whatever was changed, the problem was _not_ in the trident driver itself, but somewhere in the rest of XFree86 version 4. There seem to be four other version 4 modules that could be causing the problem, but I got discouraged, and haven't followed up. The non-driver modules that I might suspect - simply because they seem to run on startup and I don't know any better - are (from the log): (II) Loading /usr/X11R6/lib/modules/libvgahw.so (II) Loading /usr/X11R6/lib/modules/libramdac.so (II) Loading /usr/X11R6/lib/modules/libvbe.so (II) Loading /usr/X11R6/lib/modules/linux/libint10.so Now, I mention all of that because I suspect that whatever broke between XFree86 version 3 and version 4 is _also_ causing the problem with the Silicon Motion driver. The fact that the problem is not in the trident driver itself also supports this idea. Most likely, this same problem has "broken" some of the other older chipset drivers, but I haven't checked them. If you have some _factual_ basis to believe that someone has addressed this initialization problem in a specific module, I will be glad to test out the new software. No, I don't have any facts or proof. I'm just doing a routine checkup on old bugs to bring them up to date and see if the submitter is still alive. Marking broken (status null/blank) bugs in xorg with no activity in a long time as fixed. Please reopen if you think it's necessary, but first do a search if a similar bug report is already filed and in a NEW/ASSIGNED state. These bugs do not currently show in most search results as they do not have any status. Sorry for this janitorial spam, you know where to send hate mails to when your inbox gets full of bugs you're subscribed to. Just tried latest version of driver from git (v1.5.1), and am seeing behaviour as per initial bug reporter, so this one has yet to be squashed. :-( I am seeing a similar problem on a Cougar3DR based board. The system locks up solid after starting X. Console mode is fine. This same hardware worked with Fedora Core 4 and X11R6.8.2, but locks up solid on Fedora 6, Fedora 7, and Mandriva 2008, all of which are running X11R7.1 or later. I have tried the current git and still no joy. Created attachment 15717 [details]
X log file from misbehaving siliconmotion driver v1.6.0 & X server v1.4.0.90
Have just tried the Silicon Motion driver v1.6.0 with X.Org X Server v1.4.0.90 (this is what Slax 6.04 is using) and am still having the same issue - a complete lockup of my laptop.
In the attached server log file there is a new error appearing which I've not observed in logs from previous driver and/or server versions:
"Silicon MotionV_BIOS address 0xb1e70 out of range"
No idea if this is significant or even relevant to the issue though.
this may be a result of the libpciaccess port. Created attachment 15776 [details]
X log file from working(!) siliconmotion driver v1.6.0 & X server v1.4.0.90
Successful start of the X server v1.4.0.90 with siliconmotion driver v1.6.0 compiled with "SMI_DEBUG" defined.
Log created by increasing X server verbosity and redirecting stdout to file on synchronous mount. It seems the slowing effect of the synchronous writing of stdout during X server startup somehow makes the server behave differently(?!).
Created attachment 15778 [details]
full gdb backtrace of misbehaving siliconmotion driver just prior to point of system lockup/freeze
Lockup occurs at line 1666 of smi_driver.c in function SMI_WriteMode() calling macro WRITE_CPR, itself defined as a wrapper around macro MMIO_OUT32.
WRITE_CPR(pSmi, 0x00, restore->CPR00);
Prodding around with GDB just before this point reveals (see attached backtrace too):
(gdb) print restore->CPR00
$1 = 0
(gdb) print pSmi->CPRBase
$2 = (CARD8 *) 0xb7a76000 <Address 0xb7a76000 out of bounds>
(gdb) print pSmi->VPRBase
$4 = (CARD8 *) 0xb7a74000 <Address 0xb7a74000 out of bounds>
(gdb) print pSmi->DPRBase
$5 = (CARD8 *) 0xb7a70000 <Address 0xb7a70000 out of bounds>
(gdb) print pSmi->FBBase
$7 = (unsigned char *) 0xb75af000 <Address 0xb75af000 out of bounds>
Assuming it is likely that "Address 0x???? out of bounds" is significant to the problem at hand...
hello we have the same issue on our PCI LVDS card for the embedded mediastation keyboards. we use the silicon motion SM722 8Mb version, the LYNX3DM8+ This frooze issue happen only with the new Mainboards where is available the PCIexpress slot and with the SATA2 HD connection. With the old mainboard with the AGP VGA connector and SATA1, the silicon motion work really nice ) kernel 2.6.17-RT) When the silicon motion PCI card is installed on new generation mainboard, the standard booting is full working but then Xorg seem not reconize or inizialize well, we get some display linees and then sometime some is working. Same issue happen with the all new linux distribution, Ubuntu, kubuntu ..8.0.4.. Someone there can help us to fix this bug under payment too? we can support with the all hardware ready to debug and also with ssh login in remote. waiting asap one reply or contact me direct by email. regards Domenico anyone there to help us? (In reply to comment #14) > anyone there to help us? > hello we have now a new silicon mtion driver and now is working under ubuntu 8.04 Still the problem for recall multiple display, SM722 + Nvidia, Xorg is able only to inizialize one display at time but not in combination. with a payment, can someone there help us to fix this Xorg issue? (In reply to comment #15) > (In reply to comment #14) > > anyone there to help us? > > > > hello > we have now a new silicon mtion driver and now is working under ubuntu 8.04 > Still the problem for recall multiple display, SM722 + Nvidia, Xorg is able > only to inizialize one display at time but not in combination. > > with a payment, can someone there help us to fix this Xorg issue? > This sounds like an unrelated problem and you should better file another bug report. BTW, does the original problem still apply to the current version? There have been quite a bit of driver changes since then. (In reply to comment #16) > (In reply to comment #15) > > (In reply to comment #14) > > > anyone there to help us? > > > > > > > hello > > we have now a new silicon mtion driver and now is working under ubuntu 8.04 > > Still the problem for recall multiple display, SM722 + Nvidia, Xorg is able > > only to inizialize one display at time but not in combination. > > > > with a payment, can someone there help us to fix this Xorg issue? > > > > This sounds like an unrelated problem and you should better file another bug > report. > > BTW, does the original problem still apply to the current version? There have > been quite a bit of driver changes since then. > Yes, this problem is still apply on the current 8.04 version and still is not fixed. Silicon Motion released us a new drive and the source code for the SM722 and in stand alone is working much better, But still Xorg is not able to running the multiple monitor, SM722+ native binary Nvidia driver. Are you interested for fix us this issue? We have also ready one PC desktop hardware, for developement, complete with the SM722 display, Mainboard AM2 with the double VGA Nvidia and already installed the official Ubuntu 8.04. Let me know if someone there can help us. (In reply to comment #17) > > Yes, this problem is still apply on the current 8.04 version and still is not > fixed. > Silicon Motion released us a new drive and the source code for the SM722 and in > stand alone is working much better, But still Xorg is not able to running the > multiple monitor, SM722+ native binary Nvidia driver. > > Are you interested for fix us this issue? > We have also ready one PC desktop hardware, for developement, complete with the > SM722 display, Mainboard AM2 with the double VGA Nvidia and already installed > the official Ubuntu 8.04. > Let me know if someone there can help us. > I think the driver version packaged in Ubuntu 8.04 is actually quite old, you should better build the development version from sources. The GIT repository is located at: > git://anongit.freedesktop.org/xorg/driver/xf86-video-siliconmotion Also, I think it is unlikely that you will get any support here for the privative nvidia driver. Did you have any luck with the "nv" driver? (In reply to comment #18) > (In reply to comment #17) > > > > Yes, this problem is still apply on the current 8.04 version and still is not > > fixed. > > Silicon Motion released us a new drive and the source code for the SM722 and in > > stand alone is working much better, But still Xorg is not able to running the > > multiple monitor, SM722+ native binary Nvidia driver. > > > > Are you interested for fix us this issue? > > We have also ready one PC desktop hardware, for developement, complete with the > > SM722 display, Mainboard AM2 with the double VGA Nvidia and already installed > > the official Ubuntu 8.04. > > Let me know if someone there can help us. > > > > I think the driver version packaged in Ubuntu 8.04 is actually quite old, you > should better build the development version from sources. The GIT repository is > located at: > > git://anongit.freedesktop.org/xorg/driver/xf86-video-siliconmotion > > Also, I think it is unlikely that you will get any support here for the > privative nvidia driver. Did you have any luck with the "nv" driver? > Also, we made today again the new test, compiled the new Xorg 7.4 and the silicon motion git version that you seggest us. With this version the silicon motion do not start at all, no way to try to run it. Also made some more different test. the Xorg is also NOT able to running multiple VGA with the ATI PCIe card. Example: booting the ubuntu 8.04 with the PCIe ATI card, will work nice, BUT then can NOT inizialize the Nvidia ( or the NV) driver too. The open source NV drive is able to inizialize ONLY one Nvidia VGA port BUT not both: VGA + DVI Original binary Nvidia drive can inizialize the both VGA outputs, VGA+DVI. Booting from the Nvidia VGA, al is working fine , but then Xorg is NOT able to inizialize the ATI VGA. All same Xorg bugs IF are used two different brand VGA. the Xorg seem able to inizialize only 1 card VGA brand at time. suggestion? (In reply to comment #16) > (In reply to comment #15) > > (In reply to comment #14) > > > anyone there to help us? > > > > > > > hello > > we have now a new silicon mtion driver and now is working under ubuntu 8.04 > > Still the problem for recall multiple display, SM722 + Nvidia, Xorg is able > > only to inizialize one display at time but not in combination. > > > > with a payment, can someone there help us to fix this Xorg issue? > > > > This sounds like an unrelated problem and you should better file another bug > report. > > BTW, does the original problem still apply to the current version? There have > been quite a bit of driver changes since then. > Folks, I have not tried an updated driver since I posted the last logs and gdb backtraces. Unfortunately the hardware bricked itself not long after this, and with no sign of any activity on this bug I've had no reason to see if it is repairable. Would it be worth my while to try get the hardware up and running again? Or is this particular hardware so old and obscure that no one really cares... (In reply to comment #20) > (In reply to comment #16) > > (In reply to comment #15) > > > (In reply to comment #14) > > > > anyone there to help us? > > > > > > > > > > hello > > > we have now a new silicon mtion driver and now is working under ubuntu 8.04 > > > Still the problem for recall multiple display, SM722 + Nvidia, Xorg is able > > > only to inizialize one display at time but not in combination. > > > > > > with a payment, can someone there help us to fix this Xorg issue? > > > > > > > This sounds like an unrelated problem and you should better file another bug > > report. > > > > BTW, does the original problem still apply to the current version? There have > > been quite a bit of driver changes since then. > > > > Folks, > > I have not tried an updated driver since I posted the last logs and gdb > backtraces. Unfortunately the hardware bricked itself not long after this, and > with no sign of any activity on this bug I've had no reason to see if it is > repairable. Would it be worth my while to try get the hardware up and running > again? Or is this particular hardware so old and obscure that no one really > cares... > hi there After we get the last SM722 silicon motion driver working, i really dont think is this hardware the problem. we have tested many different AM2 and Intel mainboards and get the same issue. Just try to install the Ubuntu 8.04 ( 8.10 more issue..) with TWO different brand VGA installed ( you can use the VGA integrated on mainboard OR to use two PCI/PCIe/AGP VGA) but with different brand chipset. Like one VGA ATI and one VGA Nvidia. You can then booting from each VGA brand and will work fine. try to boot from one VGA brand and leave X configure the second VGA brand, you get the same bugs. With the OLD linux Debiab Etch we was able to running the 3 display really good, never get one problems. I think is only the new Xorg on Ubuntu that is not able on this features. How I told before, we have one PC destopk ready with Ubuntu, Nvidia VGA and silicon motion SM722 with LCD: http://www.lionstracs.com/demo/PCsm722.JPG if you need the original hardware configuration, we can shippig it. (In reply to comment #20) > (In reply to comment #16) > > (In reply to comment #15) > > > (In reply to comment #14) > > > > anyone there to help us? > > > > > > > > > > hello > > > we have now a new silicon mtion driver and now is working under ubuntu 8.04 > > > Still the problem for recall multiple display, SM722 + Nvidia, Xorg is able > > > only to inizialize one display at time but not in combination. > > > > > > with a payment, can someone there help us to fix this Xorg issue? > > > > > > > This sounds like an unrelated problem and you should better file another bug > > report. > > > > BTW, does the original problem still apply to the current version? There have > > been quite a bit of driver changes since then. > > > > Folks, > > I have not tried an updated driver since I posted the last logs and gdb > backtraces. Unfortunately the hardware bricked itself not long after this, and > with no sign of any activity on this bug I've had no reason to see if it is > repairable. Would it be worth my while to try get the hardware up and running > again? Or is this particular hardware so old and obscure that no one really > cares... > Well, it's true that your hardware is somewhat obscure, don't try very hard on getting it working again. :-) If you happen to do it, I'm willing to help a bit on sorting this out. (In reply to comment #21) > hi there > After we get the last SM722 silicon motion driver working, i really dont think > is this hardware the problem. > > we have tested many different AM2 and Intel mainboards and get the same issue. > Just try to install the Ubuntu 8.04 ( 8.10 more issue..) with TWO different > brand VGA installed ( you can use the VGA integrated on mainboard OR to use two > PCI/PCIe/AGP VGA) but with different brand chipset. Like one VGA ATI and one > VGA Nvidia. > You can then booting from each VGA brand and will work fine. > try to boot from one VGA brand and leave X configure the second VGA brand, you > get the same bugs. > > With the OLD linux Debiab Etch we was able to running the 3 display really > good, never get one problems. > I think is only the new Xorg on Ubuntu that is not able on this features. > > How I told before, we have one PC destopk ready with Ubuntu, Nvidia VGA and > silicon motion SM722 with LCD: > http://www.lionstracs.com/demo/PCsm722.JPG > if you need the original hardware configuration, we can shippig it. > Support for multiple screens has been slightly broken on the core X server for a while, and it might work or not depending which card combination you are using. Does the GIT siliconmotion driver fail with the other cards unplugged? It might be useful if you filed another bug report attaching your config and log files. (In reply to comment #23) > (In reply to comment #21) > > hi there > > After we get the last SM722 silicon motion driver working, i really dont think > > is this hardware the problem. > > > > we have tested many different AM2 and Intel mainboards and get the same issue. > > Just try to install the Ubuntu 8.04 ( 8.10 more issue..) with TWO different > > brand VGA installed ( you can use the VGA integrated on mainboard OR to use two > > PCI/PCIe/AGP VGA) but with different brand chipset. Like one VGA ATI and one > > VGA Nvidia. > > You can then booting from each VGA brand and will work fine. > > try to boot from one VGA brand and leave X configure the second VGA brand, you > > get the same bugs. > > > > With the OLD linux Debiab Etch we was able to running the 3 display really > > good, never get one problems. > > I think is only the new Xorg on Ubuntu that is not able on this features. > > > > How I told before, we have one PC destopk ready with Ubuntu, Nvidia VGA and > > silicon motion SM722 with LCD: > > http://www.lionstracs.com/demo/PCsm722.JPG > > if you need the original hardware configuration, we can shippig it. > > > > Support for multiple screens has been slightly broken on the core X server for > a while, and it might work or not depending which card combination you are > using. > > Does the GIT siliconmotion driver fail with the other cards unplugged? It might > be useful if you filed another bug report attaching your config and log files. > We try also the GIT siliconmotion driver with the all other cards unplugged,but no luck. Here are the new official silicon motion drivers that current work on ubuntu 8.04: http://apt.64studio.com/backports/pool/main/x/xserver-xorg-video-siliconmotion/ http://apt.64studio.com/backports/pool/main/x/xserver-xorg-video-siliconmotion/xserver-xorg-video-siliconmotion_2.2.11.orig.tar.gz and here is our current xorg configuration: http://people.64studio.com/~free/src/64studio/mediastation/trunk/themes/xorg/xorg.conf.mediastation Let me know if you need some else. (In reply to comment #24) > We try also the GIT siliconmotion driver with the all other cards > unplugged,but no luck. Which are the precise symptoms? Could you also attach the generated log files? They are probably written to /var/log/Xorg.0.log. I think the most useful would be to run the X server specifying " -logverbose 8" on the command line. > > Here are the new official silicon motion drivers that current work on ubuntu > 8.04: > http://apt.64studio.com/backports/pool/main/x/xserver-xorg-video-siliconmotion/ > http://apt.64studio.com/backports/pool/main/x/xserver-xorg-video-siliconmotion/xserver-xorg-video-siliconmotion_2.2.11.orig.tar.gz > > and here is our current xorg configuration: > http://people.64studio.com/~free/src/64studio/mediastation/trunk/themes/xorg/xorg.conf.mediastation > > Let me know if you need some else. > I sent you a private email on your account, do you get it? Mass closure: This bug has been untouched for more than six years, and is not obviously still valid. Please reopen this bug or file a new report if you continue to experience issues with current releases. |
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.