xf86-video-sis issueshttps://gitlab.freedesktop.org/xorg/driver/xf86-video-sis/-/issues2018-08-10T20:47:08Zhttps://gitlab.freedesktop.org/xorg/driver/xf86-video-sis/-/issues/1Switching to text console and back hangs X server2018-08-10T20:47:08ZBugzilla Migration UserSwitching to text console and back hangs X server## Submitted by Pawel Salek
Assigned to **Xorg Project Team**
**[Link to original bug (#5842)](https://bugs.freedesktop.org/show_bug.cgi?id=5842)**
## Description
It appears that this is probably a bug in the SiS driver not
saving...## Submitted by Pawel Salek
Assigned to **Xorg Project Team**
**[Link to original bug (#5842)](https://bugs.freedesktop.org/show_bug.cgi?id=5842)**
## Description
It appears that this is probably a bug in the SiS driver not
saving/restoring properly. Switching the virtual console to text mode and back
leaves the Xorg process unresponsive and the screen white. This problem has
appearead in Fedora rawhide in the last month, I believe. The box as such is
alive and one can connect via ssh to it. See also
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=180465 for the original report.
After the hang, the Xorg process gets stuck. It utilizes 100% CPU. When I
connect gdb to Xorg, I get following stack trace:
```
(gdb) where
#0 0x00000036286bd900 in __write_nocancel () from /lib64/libc.so.6
#1 0x0000003628667793 in _IO_new_file_write () from /lib64/libc.so.6
#2 0x0000003628668bef in _IO_new_file_xsputn () from /lib64/libc.so.6
#3 0x000000362865e89b in fwrite () from /lib64/libc.so.6
#4 0x00000000005666a2 in LogVWrite ()
#5 0x000000000056717d in LogVMessageVerb ()
#6 0x0000000000481f2c in xf86VDrvMsgVerb ()
#7 0x00002b59ab2893c2 in sisRestoreExtRegisterLock ()
from /usr/lib64/xorg/modules/drivers/sis_drv.so
#8 0x00002b59ab28b2f5 in sisSaveUnlockExtRegisterLock ()
from /usr/lib64/xorg/modules/drivers/sis_drv.so
#9 0x00002b59ab28f4d9 in SiS_SetSISTVedgeenhance ()
from /usr/lib64/xorg/modules/drivers/sis_drv.so
#10 0x00002b59ab29b3cd in SISAdjustFrame ()
from /usr/lib64/xorg/modules/drivers/sis_drv.so
#11 0x00002b59ab29b850 in SISAdjustFrame ()
from /usr/lib64/xorg/modules/drivers/sis_drv.so
#12 0x00002b59ab5fa69b in xf86ForceHWCursor ()
from /usr/lib64/xorg/modules/libramdac.so
#13 0x000000000047f7d1 in xf86InitFBManagerArea ()
#14 0x000000000048c152 in xf86XVScreenInit ()
#15 0x0000000000479714 in xf86Wakeup ()
#16 0x000000000044e2a5 in WakeupHandler ()
#17 0x0000000000559cb4 in WaitForSomething ()
#18 0x000000000044a33a in Dispatch ()
#19 0x0000000000432c95 in main ()
```
Version: 7.0.0https://gitlab.freedesktop.org/xorg/driver/xf86-video-sis/-/issues/3(sis) xterm standout mode not drawn properly when another window overlaps xterm2018-08-10T20:47:15ZBugzilla Migration User(sis) xterm standout mode not drawn properly when another window overlaps xterm## Submitted by raf
Assigned to **Xorg Project Team**
**[Link to original bug (#12003)](https://bugs.freedesktop.org/show_bug.cgi?id=12003)**
## Description
software versions:
ubuntu-7.0-4
xorg 7.2-0ubuntu11
xserver-xorg-core 1....## Submitted by raf
Assigned to **Xorg Project Team**
**[Link to original bug (#12003)](https://bugs.freedesktop.org/show_bug.cgi?id=12003)**
## Description
software versions:
ubuntu-7.0-4
xorg 7.2-0ubuntu11
xserver-xorg-core 1.2.0-3ubuntu8
xserver-xorg-video-sis 0.9.1-4ubuntu1
xterm 223-1 (and xterm-229 compiled from source)
problem description:
when a program that is running in xterm uses standout/reverse
mode and there is another window overlapping the xterm window
such that the top or bottom edge of the overlapping window
intersects with the region of standout mode in the xterm window,
then xterm incorrectly draws that part of the standout region
that is below the intersecting horizontal edge of the
overlapping window.
this only seems to happen on ubuntu hosts with sis video adapters
and it happens with ubuntu's packaged xterm (223) as well as the
latest version of xterm (229) running from both ubuntu and from solaris
as long as the window is displayed on the ubuntu/xorg host with the
sis driver.
in order to help you understand the problemn description above, i have
put some screenshots on a website:
http://raf.org/dump1.png - shows mutt in a little xterm on the left
with everything looking correct.
http://raf.org/dump2.png - shows what happens after the little xterm
on the right is moved to overlap with the little xterm on the left.
then, in the little xterm on the left, i pressed Control-R to redraw
it. notice that the the first row (in standout mode) in that xterm
is not redrawn properly. the part of that line where the y-coord is
below the y-coord of the top of the overlapping window is drawn in black.
http://raf.org/dump3.png - shows what happens after resizing the overlapping
xterm on the right so that it's lower edge also intersects with a line in
standout mode in the little xterm window on the left. after pressing
Control-R in the xterm on the left, both standout lines that intersect
with the horizontal edges of the overlapping window are drawn incorrectly.
http://raf.orgdump4.png - shows what happens while/after dragging the
overlapping window to the right and down a bit. the standout mode is
redrawn correctly where the overlapping window was, but the original,
icorrectly drawn parts aren't redrawn (until Control-R is pressed in
the xterm on the left)
http://raf.org/.xresources - my xterm-related x resources
notes:
it doesn't matter what window is overlapping the xterm window
(i.e. it doesn't also have to be xterm) but this only happens
when the overlapped window is xterm.
it doesn't matter what window manager is being used. this bug
happens both with twm and with whatever gnome's default
window manager is.
it didn't happen with the latest xterm (229) compiled on solaris
and displayed on a sunray terminal so i doubt that it is an
xterm bug.
it happened on two ubuntu hosts with sis video adapters but not
on other ubuntu hosts with different video adapters.
so i suspect that it is either a bug in xorg's sis driver or
a bug that exists between xterm and the sis driver.
workarounds:
never let any window overlap an xterm window if you use
the xorg's sis driver.
get a different video adapter :)https://gitlab.freedesktop.org/xorg/driver/xf86-video-sis/-/issues/4Support for SiS 662/6712018-08-10T20:47:18ZBugzilla Migration UserSupport for SiS 662/671## Submitted by Calorì Alessandro
Assigned to **Xorg Project Team**
**[Link to original bug (#14848)](https://bugs.freedesktop.org/show_bug.cgi?id=14848)**
## Description
Created attachment 14873
Intel patch against X.org SiS driv...## Submitted by Calorì Alessandro
Assigned to **Xorg Project Team**
**[Link to original bug (#14848)](https://bugs.freedesktop.org/show_bug.cgi?id=14848)**
## Description
Created attachment 14873
Intel patch against X.org SiS driver version 0.9.1
The current SiS driver, as well as older versions, doesn't provide good support for SiS 662/671-based video adapters. I own an Intel D201GLY2 with SiS 662FX video adapter and, when starting X.org with this driver, multiple flickering vertical lines are shown on the screen. The only way to use X.org without them is either to reduce the resolution to 800x600 or the color depth to 8 bit.
Intel has released a modified driver that works good even at high resolutions (1680x1050) and with high color depths (24 bit).
The problem occurs on different platforms (I've tried x86 and x86_64), on different OS (I've tried various Linux distro and FreeBSD) and different driver versions (I've tried 0.9.1, 0.9.3 and 0.9.4).
I attach a patch for X.org driver version 0.9.1 made against the Intel driver, based on that exact version (so that only Intel changes are shown).
It is also being reported that the problem occurs on Intel D201GLY.
**Patch 14873**, "Intel patch against X.org SiS driver version 0.9.1":
[intel.patch.bz2](/uploads/413abb6e2b3887cb61148b72647f4e03/intel.patch.bz2)
### See also
* http://bugs.freedesktop.org/show_bug.cgi?id=9409https://gitlab.freedesktop.org/xorg/driver/xf86-video-sis/-/issues/5sis: hardware cursor broken on some part of the screen2018-08-10T20:47:21ZBugzilla Migration Usersis: hardware cursor broken on some part of the screen## Submitted by Brice Goglin `@bgoglin`
Assigned to **Xorg Project Team**
**[Link to original bug (#15024)](https://bugs.freedesktop.org/show_bug.cgi?id=15024)**
## Description
Bug reported by "Daniel" a while ago on the Debian BT...## Submitted by Brice Goglin `@bgoglin`
Assigned to **Xorg Project Team**
**[Link to original bug (#15024)](https://bugs.freedesktop.org/show_bug.cgi?id=15024)**
## Description
Bug reported by "Daniel" a while ago on the Debian BTS.
"with a screen samsung 1680x1050 when I move the mouse on the 10cm left vertical edge of the screen, it disappear. and when the mouse arrives has 1cm higher edge (over all its length), it changes to become a small line vertical and remains has 1cm of this edge."
Config and log available at the URL above.
SWcursor fixes the problem.
He was using the sis driver 0.9.3 on
XGI - Xabre Graphics Inc Volari V3XT/V5/V8 (rev 02)
By the way, he also tried the xgi driver from git:
"this driver can not use 1680x1050 60Hz or 1280x1024 75Hz just 1280x1024
60Hz but the graphical and colors is better. i remain with the "sis" driver for the moment"
Brice
Version: 7.2 (2007.02)https://gitlab.freedesktop.org/xorg/driver/xf86-video-sis/-/issues/6[SiS] 771/671 PCIE VGA Display Adapter [1039:6351]2018-08-10T20:47:24ZBugzilla Migration User[SiS] 771/671 PCIE VGA Display Adapter [1039:6351]## Submitted by Mirko
Assigned to **Xorg Project Team**
**[Link to original bug (#15190)](https://bugs.freedesktop.org/show_bug.cgi?id=15190)**
## Description
I have a [SiS] 771/671 PCIE VGA Display Adapter [1039:6351] (rev 10) dr...## Submitted by Mirko
Assigned to **Xorg Project Team**
**[Link to original bug (#15190)](https://bugs.freedesktop.org/show_bug.cgi?id=15190)**
## Description
I have a [SiS] 771/671 PCIE VGA Display Adapter [1039:6351] (rev 10) driver from xorg doesn't have support for my card but I've found http://downloadcenter.intel.com/filter_results.aspx?strTypes=all&ProductID=2773&OSFullName=Linux*&lang=eng&strOSs=39&submit=Go%21 that supports my card but driver I have to force it using ignoreABI and is not stable.https://gitlab.freedesktop.org/xorg/driver/xf86-video-sis/-/issues/7Xorg is not accepting a good display mode.2018-08-10T20:47:28ZBugzilla Migration UserXorg is not accepting a good display mode.## Submitted by Bradley Pursley
Assigned to **Xorg Project Team**
**[Link to original bug (#22499)](https://bugs.freedesktop.org/show_bug.cgi?id=22499)**
## Description
Created attachment 27180
xorg.conf used by my system
I am us...## Submitted by Bradley Pursley
Assigned to **Xorg Project Team**
**[Link to original bug (#22499)](https://bugs.freedesktop.org/show_bug.cgi?id=22499)**
## Description
Created attachment 27180
xorg.conf used by my system
I am using a wide-screen monitor but Xorg is claiming that the 1366x768 display mode is invalid but when I use the xvidtune program to set the display to the same settings I have in the xorg.conf file, it works just fine.
Here is the relevant excerpt from the xorg.0.log:
(**) SIS(0): Option "NoDDC"
(==) SIS(0): Min pixel clock is 5 MHz
(--) SIS(0): Max pixel clock is 168 MHz
(II) SIS(0): Monitor0: Using hsync range of 31.50-50.00 kHz
(II) SIS(0): Monitor0: Using vrefresh range of 25.00-65.00 Hz
(II) SIS(0): Clock range: 5.00 to 168.00 MHz
(II) SIS(0): Not using mode "1366x768" (illegal horizontal timings)
I am using FC11 but when I was using FC8 (I just upgraded) it accepted the wider video mode just fine. Since the hardware configuration did not change between the two versions, there must be a bug with the newer software.
Bradley
**Attachment 27180**, "xorg.conf used by my system":
[xorg.conf](/uploads/5a59f85701cbf722099da187481ae4f9/xorg.conf)https://gitlab.freedesktop.org/xorg/driver/xf86-video-sis/-/issues/8sis 661fx x crash when using color picker2018-08-10T20:47:30ZBugzilla Migration Usersis 661fx x crash when using color picker## Submitted by ash
Assigned to **Xorg Project Team**
**[Link to original bug (#23885)](https://bugs.freedesktop.org/show_bug.cgi?id=23885)**
## Description
use color picker in any kde application (custom colors window or color pi...## Submitted by ash
Assigned to **Xorg Project Team**
**[Link to original bug (#23885)](https://bugs.freedesktop.org/show_bug.cgi?id=23885)**
## Description
use color picker in any kde application (custom colors window or color picker desktop widget). immediate x crash follows
may be related to the same bug : on the same box (and only on it) when logging out and choosing shut down it just logs out to kdm. then to shut down you use the menu in kdm and it works. maybe its actually an x crash which happens when closing kde and then x restarts and goes to kdm
affected system
sis 661fx onboard graphics (epox mainboard)
1680x1050 monitor (i think 24 bit color) on vga port
xorg-server 1.6
xorg-server-utils 7.4
xf86-video-sis 0.10.2
kde 4.3.1
arch linux last updated in 09.2009
xorg.conf made with X -configure (contains modules dri dri2 dbe extmod glx record) and then added
Option "AccelMethod" "EXA"
Option "MigrationHeuristic" "greedy"
Version: 7.4 (2008.09)https://gitlab.freedesktop.org/xorg/driver/xf86-video-sis/-/issues/9corrupted graphics screen 1024x768 SiS driver with xorg on EPSON PT-2100 unde...2018-08-10T20:47:32ZBugzilla Migration Usercorrupted graphics screen 1024x768 SiS driver with xorg on EPSON PT-2100 under gentoo linux## Submitted by miy
Assigned to **Xorg Project Team**
**[Link to original bug (#28863)](https://bugs.freedesktop.org/show_bug.cgi?id=28863)**
## Description
Dear xorg community,
Trying to set up Gentoo on an Epson Endeavor PT-210...## Submitted by miy
Assigned to **Xorg Project Team**
**[Link to original bug (#28863)](https://bugs.freedesktop.org/show_bug.cgi?id=28863)**
## Description
Dear xorg community,
Trying to set up Gentoo on an Epson Endeavor PT-2100
http://www2.ttcn.ne.jp/~TATSUYUKI-SATO/rpc-pt2100.htm
and I've been stuck with setting up the X server for weeks now.
Card: SiS 630/730
Monitor: 15" TFT XGA 1024x768 True Color
startx enters graphics mode without much complaint, but the graphics
screen is condensed horizontally to about one third and moved to the
right so that I see about half of the screen on the very left edge of my
monitor, flickering, unreadable. The rest, about five sixth of my
monitor stays black. Here is a photo (2MB):
http://math.usask.ca/~auer/gentoo/screen1024x768c.jpg
My xorg.conf is this:
http://math.usask.ca/~auer/gentoo/xorg.conf
X's answer in Xorg.0.log reads
http://math.usask.ca/~auer/gentoo/Xorg.0.log
Tried 800x600 and 640x480, too, mostly with DefaultDepth 24 and with all
kinds of ranges for HorizSync and VertRefresh. Never obtain anything
like a recognisable graphics screen. After so many attempts I am about
to give up. Please help.
As you can see, I have also tried to tamper with the modes.
What is the Option "InternalModes" for? Do I have to switch it off, in
order to push my own mode.
sisfb and fb0 are modules rather than built into the linux kernel.
Does the line
(WW) Open ACPI failed (/var/run/acpid.socket) (No such file or directory)
indicate anything critical?
What does the warning
(WW) Falling back to old probe method for sis
mean?
It seems, that SiS recognizes my TFT monitor
(--) SIS(0): Detected LCD/plasma panel (1024x768, 12, non-exp., RGB18
[c2e7ff])
But what does "12, non-exp., RGB18 [c2e7ff]" mean? Is the color depth 12
or 18 rather than 24?
And why does SiS spell out the modes
(**) SIS(0): *Default mode "1024x768" (1024x768) (For CRT device:
65.1 MHz, 48.5 kHz, 60.1 Hz)
as if my monitor were a CRT?
So many questions. Any clue?
Thanks in advance, Rolandhttps://gitlab.freedesktop.org/xorg/driver/xf86-video-sis/-/issues/10[Ubuntu 10.04.4 LTS 32-bit] SiS 6326 AGP graphics on a mainboard with Intel 8...2018-08-10T20:47:34ZBugzilla Migration User[Ubuntu 10.04.4 LTS 32-bit] SiS 6326 AGP graphics on a mainboard with Intel 815E chipset will not boot from a hard drive, but is okay with DVD boot## Submitted by myp..@..il.com
Assigned to **Xorg Project Team**
**[Link to original bug (#56341)](https://bugs.freedesktop.org/show_bug.cgi?id=56341)**
## Description
Hi,
I will like to report that when I try to boot Ubuntu 10.0...## Submitted by myp..@..il.com
Assigned to **Xorg Project Team**
**[Link to original bug (#56341)](https://bugs.freedesktop.org/show_bug.cgi?id=56341)**
## Description
Hi,
I will like to report that when I try to boot Ubuntu 10.04.4 LTS 32-bit from a hard drive on a system with Diamond Multimedia Speedstar A50 (SiS 6326 AGP 8MB SDRAM, AGP 3.3V) graphics card and Intel D815EAA2 mainboard (Intel 815E chipset), the system will crash during boot in a few seconds.
Floppy disk drive indicator light goes on and stays on.
I won't even see the Ubuntu graphical logo on the screen.
However, if I boot Ubuntu 10.04.4 LTS 32-bit from a DVD, then it will boot fine.
I also happened to own a PCI version of SiS 6326 AGP graphics card called Palit AGP6326 graphics card (SiS 6326 AGP, PCI card, 4 MB FPM or EDO DRAM).
This card will boot fine with either DVD boot or hard drive boot.
Based on these results, I suspect there is a bug in either Intel AGP GART device driver or Intel PATA device driver for ICH2 southbridge.
System Configuration 1:
- Ubuntu 10.04.4 LTS 32-bit
* All updates issued as of 10/15/2012 have been applied.
- Intel Pentium III 1 GHz
- Intel D815EEA2 mainboard
* BIOS Version P21-0039 (last BIOS release)
* Power -> ACPI -> ACPI: S3 State
* Power -> ACPI -> Video RePOST: Off
- 512 MB PC133 SDRAM
* 256 MB module
* 128 MB module
* 128 MB module
- Diamond Multimedia Speedstar A50 graphics card
* SiS 6236 AGP
* Chip Revision H0
* 3.3V AGP card
* 8 MB SDRAM
* VGA BIOS Version 1.22a
- Gateway EV700 17 inch VGA monitor
- Seagate ST320414A 20 GB PATA hard drive
- Pioneer DVR-105 PATA DVD-RW drive
- Toshiba XM-6402B PATA CD-ROM drive
- PS/2 keyboard
- PS/2 mouse
System Configuration 2:
- Palit 6326AGP graphics card
* SiS 6236 AGP
* Chip Revision H0
* 5V PCI card
* 4 MB FPM or EDO DRAM
* VGA BIOS Version 1.28q
- Other than the graphics card, the system configuration is identical to System Configuration 1.
Note that if I am correct, there are SiS 6326/6326 AGP/6326 DVD.
Both of the graphics card had a SiS 6326 AGP chip even though one of them was a PCI card.
Regards,
fpgahardwareengineerhttps://gitlab.freedesktop.org/xorg/driver/xf86-video-sis/-/issues/11[Ubuntu 12.04.1 LTS 32-bit] Several SiS 661GX / 761GX (Real256E graphics) issues2018-08-10T20:47:37ZBugzilla Migration User[Ubuntu 12.04.1 LTS 32-bit] Several SiS 661GX / 761GX (Real256E graphics) issues## Submitted by myp..@..il.com
Assigned to **Xorg Project Team**
**[Link to original bug (#56803)](https://bugs.freedesktop.org/show_bug.cgi?id=56803)**
## Description
Hi,
I will like to report that SiS 661GX's integrated graphic...## Submitted by myp..@..il.com
Assigned to **Xorg Project Team**
**[Link to original bug (#56803)](https://bugs.freedesktop.org/show_bug.cgi?id=56803)**
## Description
Hi,
I will like to report that SiS 661GX's integrated graphics (Real256E graphics) has number of issues (bugs) under Ubuntu 12.04 LTS 32-bit.
1) Ubuntu logo doesn't get displayed until the last moment, and it is at a low screen resolution
In a typical system, during boot, Ubuntu displays its graphical logo with red dots flashing on the screen.
In a SiS 661GX-based system, the screen will have a dark purple background pretty much throughout the boot up, and Ubuntu graphical logo will not get displayed until just right before the login screen gets displayed.
2) The graphics starts out in interlaced mode if I use Gateway EV700 CRT monitor with it
I often use this computer (see System Configuration 1) with Gateway EV700 CRT monitor.
Unfortunately, Ubuntu will choose interlaced mode for the screen (87 Hz Interlaced).
Since Ubuntu 12.04 doesn't have a direct way to alter the screen refresh rate from its control panel like menu, the screen quality will be quite poor.
Since the monitor itself it capable of 1024 X 768 @ 75 Hz Non-Interlaced, shouldn't this behavior (starting out in interlaced mode) be considered a bug?
3) Mouse cursor disappears if there is rendering happening underneath it
I don't know if the SiS x.org display device driver uses hardware cursor feature of SiS graphics (I am assuming that SiS graphics supports hardware cursor.), but I notice that it disappears when there is some rendering happening underneath it.
Particularly noticable in Firefox.
I didn't observe this with NVIDIA or ATI Technologies graphics so I assume this is a bug.
4) Screen horizontal and/or vertical frequency goes nuts during restart or shut down
When I try to restart or shut down this computer, the monitor's horizontal and/or vertical frequency goes out of range.
I don't know which one goes out of range, and I have no way of knowing this.
I will say that this is potentially dangerous since it can damage a monitor, although if I am correct, most monitors have ways to detect this kind of condition, and will display a warning message to prevent damage to the monitor.
Regardless, it is not a desirable condition, and I believe this should be considered a bug.
Overall, these issues mentioned here didn't exist in Ubuntu 10.04 LTS 32-bit.
All of them appeared in Ubuntu 12.04 LTS 32-bit.
Regarding this SiS Real256E integrated graphics, I have another mainboard called ECS 761GX-M754 (SiS 761GX/964 chipset).
The integrated graphics inside SiS 761GX is identical to SiS 661GX.
SiS 761GX also exhibits issues similar to what was described here.
System Configuration 1:
- Intel Pentium 4 2.8 GHz
* 800 MHz FSB
* 130 nm device (Northwood)
* Hyperthreading enabled
- ASUS P4S8X-MX mainboard
* SiS 661GX northbridge
* SiS 964 southbridge
* BIOS Revision 0808 (last release)
- 1 GB DDR SDRAM
- SiS Real256E integrated graphics
* Using 16MB for graphics
- Hitachi IC35L060AVV207-0 60GB PATA hard drive
- Hitachi-LG Data Storage GCC-4481B PATA CD-RW/DVD-ROM drive
- USB multi-card reader
- USB keyboard
- USB mouse
- Gateway EV700 CRT monitor
System Configuration 2:
- AMD Sempron 2800+
- ECS 761GX-M754 mainboard
* SiS 761GX Northbridge
* SiS 964 Southbridge
* BIOS Version 070227
- 1 GB DDR SDRAM
- SiS Real256E integrated graphics
* Using 32 MB for graphics
- Emprex 1008IM PATA DVD-RW drive
- USB multi-card reader
- PS/2 keyboard
- PS/2 mouse
- Gateway EV700 CRT monitor
Regards,
fpgahardwareengineerhttps://gitlab.freedesktop.org/xorg/driver/xf86-video-sis/-/issues/12No X with X.Org X Server 1.13.* and SiS video driver2018-08-10T20:47:41ZBugzilla Migration UserNo X with X.Org X Server 1.13.* and SiS video driver## Submitted by Marja van Waes
Assigned to **Xorg Project Team**
**[Link to original bug (#60891)](https://bugs.freedesktop.org/show_bug.cgi?id=60891)**
## Description
Since X.Org X Server 1.13.0 (and now using 1.13.2), X segfault...## Submitted by Marja van Waes
Assigned to **Xorg Project Team**
**[Link to original bug (#60891)](https://bugs.freedesktop.org/show_bug.cgi?id=60891)**
## Description
Since X.Org X Server 1.13.0 (and now using 1.13.2), X segfaults when trying to start it while using the SiS video driver. Vesa driver works fine.
SiS video driver is version 0.10.7
Finding [bug 35763](https://bugs.freedesktop.org/show_bug.cgi?id=35763) helped me to find a better workaround than using the Vesa driver:
adding
Option "NoAccel" "true"
to the Section "Device" about SiS in /etc/X11/xorg.conf
made the SiS driver usable again.
Of course, the best solution would be if someone would fix [bug 35763](https://bugs.freedesktop.org/show_bug.cgi?id=35763) :)
### Depends on
* [Bug 35763](https://bugs.freedesktop.org/show_bug.cgi?id=35763)https://gitlab.freedesktop.org/xorg/driver/xf86-video-sis/-/issues/13sis_dac.c:1528: bad switch statement ?2018-08-10T20:47:46ZBugzilla Migration Usersis_dac.c:1528: bad switch statement ?## Submitted by dcb..@..il.com
Assigned to **Xorg Project Team**
**[Link to original bug (#67095)](https://bugs.freedesktop.org/show_bug.cgi?id=67095)**
## Description
I just ran the static analysis checker "cppcheck" over
the sou...## Submitted by dcb..@..il.com
Assigned to **Xorg Project Team**
**[Link to original bug (#67095)](https://bugs.freedesktop.org/show_bug.cgi?id=67095)**
## Description
I just ran the static analysis checker "cppcheck" over
the source code of the Redhat Fedora package xorg-x11-drv-sis
It said
xorg-x11-drv-sis.spec.out:[sis_dac.c:1528] -> [sis_dac.c:
1532]: (warning) Variable 'max' is reassigned a value before the old one has bee
n used. 'break;' missing?
Source code is
case PCI_CHIP_SIS660:
if((pSiS->ChipType >= SIS_660) &&
(pSiS->ChipFlags & SiSCF_760LFB)) {
magic = magicDED[bus/64];
} else {
magic = magicINT[bus/64];
}
max = 680000;
case PCI_CHIP_SIS340:
case PCI_CHIP_XGIXG40:
magic = magicDED[bus/64];
max = 800000;
break;
Suggest add case statement to first case.https://gitlab.freedesktop.org/xorg/driver/xf86-video-sis/-/issues/16Xorg error - SiS Driver works with Arch linux-lts 4.4.23-1 but fails with Arc...2023-02-23T17:43:31ZBugzilla Migration UserXorg error - SiS Driver works with Arch linux-lts 4.4.23-1 but fails with Arch linux 4.6.3-1 and later. libpciaccess?## Submitted by James
Assigned to **Xorg Project Team**
**[Link to original bug (#98158)](https://bugs.freedesktop.org/show_bug.cgi?id=98158)**
## Description
Arch Linux:
xf86-video-sis 0.10.8-2
xorg-server 1.18.4-1
/usr/lib/libpc...## Submitted by James
Assigned to **Xorg Project Team**
**[Link to original bug (#98158)](https://bugs.freedesktop.org/show_bug.cgi?id=98158)**
## Description
Arch Linux:
xf86-video-sis 0.10.8-2
xorg-server 1.18.4-1
/usr/lib/libpciaccess.so.0.11.1
Build Source:
http://xorg.freedesktop.org/releases/individual/driver/xf86-video-sis-0.10.8.tar.bz2
Hardware:
01:00.0 VGA compatible controller [0300]: Silicon Integrated Systems [SiS] 661/741/760 PCI/AGP or 662/761Gx PCIE VGA Display Adapter [1039:6330]
00:00.0 Host bridge [0600]: Silicon Integrated Systems [SiS] 661FX/M661FX/M661MX Host [1039:0661] (rev 11)
This is on an older generic Winbook laptop, circa 2005, with an Intel Celeron M and all SiS chipset.
The driver works fine with the linux-lts 4.4.23-1 kernel, but fails with, at least, the linux-4.6.3-1 kernel, and later kernels. I've also tested with linux 4.7.6-1 and linux 4.8.0-1.
The only thing changing here is the kernel version. The Xorg server and the SiS driver are exactly the same in each case.
The error occurs right after some "VESA BIOS detected" reporting, and is always:
```
(EE) SIS(0): Unable to map FB aperture. Invalid argument (22)
(EE) SIS(0): **************************************************
(EE) SIS(0): ERROR:
(EE) SIS(0): SiSMapMem() failed
(EE) SIS(0): END OF MESSAGE
(EE) SIS(0): **************************************************
```
which is from
http://code.metager.de/source/xref/xorg/driver/xf86-video-sis/src/sis_driver.c
/usr/lib/xorg/modules/drivers/sis_drv.so
```
7278 int err = pci_device_map_range(pSiS->PciInfo,
7279 (ULong)pSiS->realFbAddress,
7280 pSiS->FbMapSize,
7281 PCI_DEV_MAP_FLAG_WRITABLE |
7282 PCI_DEV_MAP_FLAG_WRITE_COMBINE,
7283 (void *)&pSiSEnt->FbBase);
```
or - cannot tell which code path, since the error message is the same -
```
7305 int err = pci_device_map_range(pSiS->PciInfo,
7306 (ULong)pSiS->realFbAddress,
7307 pSiS->FbMapSize,
7308 PCI_DEV_MAP_FLAG_WRITABLE |
7309 PCI_DEV_MAP_FLAG_WRITE_COMBINE,
7310 (void *)&pSiS->FbBase);
```
which is in /xorg/lib/libpciaccess/src/common_interface.c,
http://code.metager.de/source/xref/xorg/lib/libpciaccess/src/common_interface.c
```
235/**
236 * Map the specified memory range so that it can be accessed by the CPU.
237 *
238 * Maps the specified memory range for access by the processor. The pointer
239 * to the mapped region is stored in \c addr. In addition, the
240 * \c pci_mem_region::memory pointer for the BAR will be updated.
241 *
242 * \param dev Device whose memory region is to be mapped.
243 * \param base Base address of the range to be mapped.
244 * \param size Size of the range to be mapped.
245 * \param map_flags Flag bits controlling how the mapping is accessed.
246 * \param addr Location to store the mapped address.
247 *
248 * \return
249 * Zero on success or an \c errno value on failure.
250 *
251 * \sa pci_device_unmap_range
252 */
253 int
254 pci_device_map_range(struct pci_device *dev, pciaddr_t base,
255 pciaddr_t size, unsigned map_flags,
256 void **addr)
257 {
...
314 if (dev->regions[region].memory == NULL) {
315 err = (*pci_sys->methods->map_range)(dev,
316 &mappings[devp->num_mappings]);
317 }
...
```
http://code.metager.de/source/xref/xorg/lib/libpciaccess/src/pciaccess_private.h
```
184 extern struct pci_system * pci_sys;
...
159struct pci_system {
160 /**
161 * Platform dependent implementations of specific API routines.
162 */
163 const struct pci_system_methods * methods;
...
59 struct pci_system_methods {
...
64 int (*map_range)(struct pci_device *dev, struct pci_device_mapping *map);
```
which, presumably, is either from
http://code.metager.de/source/xref/xorg/lib/libpciaccess/src/x86_pci.c
```
815 static const struct pci_system_methods x86_pci_methods = {
...
819 .map_range = pci_device_x86_map_range,
...
639 pci_device_x86_map_range(struct pci_device *dev,
640 struct pci_device_mapping *map)
641 {
...
```
or from
http://code.metager.de/source/xref/xorg/lib/libpciaccess/src/linux_sysfs.c
```
969 static const struct pci_system_methods linux_sysfs_methods = {
...
974 .map_range = pci_device_linux_sysfs_map_range,
...
539 pci_device_linux_sysfs_map_range(struct pci_device *dev,
540 struct pci_device_mapping *map)
541 {
...
```
Somewhere, the pci_device_map_range{} function is not compatible with the newer kernel versions.
Is this really a SiS driver issue, or is this ultimately a libpciaccess issue?
Version: 7.7 (2012.06)