Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
X crashes on startup with new kernel
View unanswered posts
View posts from last 24 hours

 
Reply to topic    Gentoo Forums Forum Index Desktop Environments
View previous topic :: View next topic  
Author Message
multix
Apprentice
Apprentice


Joined: 28 Nov 2004
Posts: 181
Location: Italy

PostPosted: Tue Apr 15, 2014 7:58 am    Post subject: X crashes on startup with new kernel Reply with quote

Hi,

I did a deep emerge update (all packages, revdep-rebuild, new kernel, etc etc). Now X doesn't startup anymore. It worked before...

Code:

Initializing built-in extension DRI2
Loading extension GLX
    [10f] 320 x 200, 70Hz
    [112] 640 x 480, 60Hz, 72Hz, 75Hz, 85Hz, 100Hz
    [115] 800 x 600, 60Hz, 72Hz, 75Hz, 85Hz, 100Hz
    [118] 1024 x 768, 60Hz, 70Hz, 75Hz, 85Hz, 100Hz
    [11b] 1280 x 1024, 60Hz, 75Hz, 85Hz
    [11e] 640 x 400, 70Hz
    [124] 1600 x 1200, 60Hz, 75Hz, 85Hz
    [134] 320 x 240, 72Hz
    [13e] 1400 x 1050, 60Hz, 75Hz
    [144] 400 x 300, 72Hz
    [154] 512 x 384, 70Hz
    [175] 720 x 480, 75Hz
    [17f] 720 x 576, 75Hz
(EE)
(EE) Backtrace:
(EE) 0: /usr/bin/X (xorg_backtrace+0x50) [0x8242ad0]
(EE) 1: /usr/bin/X (0x8048000+0x1ff2d6) [0x82472d6]
(EE) 2: /usr/bin/X (0x8048000+0x20fc3e) [0x8257c3e]
(EE) 3: linux-gate.so.1 (__kernel_rt_sigreturn+0x0) [0xb77a340c]
(EE) 4: /usr/lib/libpixman-1.so.0 (0xb74d9000+0x525e8) [0xb752b5e8]
(EE) 5: /usr/lib/libpixman-1.so.0 (0xb74d9000+0x6e434) [0xb7547434]
(EE) 6: /usr/lib/libpixman-1.so.0 (pixman_fill+0x5d) [0xb74e03bd]
(EE) 7: /usr/lib/xorg/modules/libfb.so (fbFill+0x301) [0xb2d0ac31]
(EE) 8: /usr/lib/xorg/modules/libfb.so (fbPolyFillRect+0x1c8) [0xb2d0b488]
(EE) 9: /usr/lib/xorg/modules/libexa.so (0xb2cd7000+0x17100) [0xb2cee100]
(EE) 10: /usr/lib/xorg/modules/libexa.so (0xb2cd7000+0xee29) [0xb2ce5e29]
(EE) 11: /usr/bin/X (0x8048000+0x1597ab) [0x81a17ab]
(EE) 12: /usr/bin/X (miPaintWindow+0x21c) [0x82204cc]
(EE) 13: /usr/bin/X (miWindowExposures+0x22e) [0x8220eee]
(EE) 14: /usr/bin/X (DRIWindowExposures+0x9a) [0x820121a]
(EE) 15: /usr/bin/X (0x8048000+0xa8ec4) [0x80f0ec4]
(EE) 16: /usr/bin/X (MapWindow+0x393) [0x80ba113]
(EE) 17: /usr/bin/X (InitRootWindow+0x13a) [0x80ba27a]
(EE) 18: /usr/bin/X (0x8048000+0x39950) [0x8081950]
(EE) 19: /usr/bin/X (0x8048000+0x1efee) [0x8066fee]
(EE) 20: /lib/libc.so.6 (__libc_start_main+0xef) [0xb72b4a3f]
(EE) 21: /usr/bin/X (0x8048000+0x1f014) [0x8067014]
(EE)
(EE) Bus error at address 0xb2d9d000
(EE)
Fatal server error:
(EE) Caught signal 7 (Bus error). Server aborting
(EE)
(EE)


I don't know if I should find a core file?

The setup should start a plain and simple windowmaker. When exporting display to another X server, windowmaker works, this would point to a local Xorg problem.

The laptop, an older Thinkpad with a S3 Savage.

Checking the X log I see this:
Code:

[   866.011] (II) SAVAGE(0): Reserved 8528 kb for textures at offset 0x78b000
[   866.011] (II) EXA(0): Offscreen pixmap area of 2072576 bytes
[   866.011] (II) EXA(0): Driver registered support for the following operations:
[   866.011] (II)         Solid
[   866.011] (II)         Copy
[   866.011] (II)         UploadToScreen
[   866.011] (II) SAVAGE(0): Savage EXA Acceleration enabled.
[   866.011] (==) SAVAGE(0): Backing store enabled
[   866.024] (==) SAVAGE(0): DPMS enabled
[   866.024] (II) SAVAGE(0): [DRI] installation complete
[   866.024] (II) SAVAGE(0): [junkers]pSAVAGEDRIServer:
[   866.024] (II) SAVAGE(0): [junkers]  reserved_map_agpstart:0x00000000
[   866.024] (II) SAVAGE(0): [junkers]  reserved_map_idx:0x00000000
[   866.024] (II) SAVAGE(0): [junkers]  sarea_priv_offset:0x00000000
[   866.024] (II) SAVAGE(0): [junkers]  chipset:0x00000000
[   866.024] (II) SAVAGE(0): [junkers]  sgram:0x00000000
[   866.024] (II) SAVAGE(0): [junkers]  frontbufferSize:0x001d5000
[   866.025] (II) SAVAGE(0): [junkers]  frontOffset:0x00000000
[   866.025] (II) SAVAGE(0): [junkers]  frontPitch:0x00000c80
[   866.025] (II) SAVAGE(0): [junkers]  backbufferSize:0x001db000
[   866.025] (II) SAVAGE(0): [junkers]  backOffset:0x003d5000
[   866.025] (II) SAVAGE(0): [junkers]  backPitch:0x00000c80
[   866.025] (II) SAVAGE(0): [junkers]  depthbufferSize:0x001db000
[   866.025] (II) SAVAGE(0): [junkers]  depthOffset:0x005b0000
[   866.025] (II) SAVAGE(0): [junkers]  depthPitch:0x00000c80
[   866.025] (II) SAVAGE(0): [junkers]  textureOffset:0x0078b000
[   866.025] (II) SAVAGE(0): [junkers]  textureSize:0x00854000
[   866.025] (II) SAVAGE(0): [junkers]  textureSize:0x00854000
[   866.025] (II) SAVAGE(0): [junkers]  logTextureGranularity:0x00000000
[   866.025] (II) SAVAGE(0): [junkers]  agp:handle:0x00000001
[   866.025] (II) SAVAGE(0): [junkers]  agp:offset:0x01000000
[   866.025] (II) SAVAGE(0): [junkers]  agp:size:0x01000000
[   866.025] (II) SAVAGE(0): [junkers]  agp:map:0x00000000
[   866.025] (II) SAVAGE(0): [junkers]  registers:handle:0xc0100000
[   866.025] (II) SAVAGE(0): [junkers]  registers:offset:0x00000000
[   866.025] (II) SAVAGE(0): [junkers]  registers:size:0x00080000
[   866.025] (II) SAVAGE(0): [junkers]  registers:map:0x00000000
[   866.025] (II) SAVAGE(0): [junkers]  status:handle:0x00000000
[   866.025] (II) SAVAGE(0): [junkers]  status:offset:0x00000000
[   866.025] (II) SAVAGE(0): [junkers]  status:size:0x00000000
[   866.025] (II) SAVAGE(0): [junkers]  status:map:0x00000000
[   866.025] (II) SAVAGE(0): [junkers]  agpTextures:handle:0xd0000000
[   866.025] (II) SAVAGE(0): [junkers]  agpTextures:offset:0x00000000
[   866.025] (II) SAVAGE(0): [junkers]  agpTextures:size:0x01000000
[   866.025] (II) SAVAGE(0): [junkers]  apgTextures:map:0x00000000
[   866.025] (II) SAVAGE(0): [junkers]  logAgpTextureGranularity:0x00000000
[   866.025] (II) SAVAGE(0): [junkers]  cmdDma:handle:0x00000000
[   866.025] (II) SAVAGE(0): [junkers]  cmdDma:offset:0x00000000
[   866.025] (II) SAVAGE(0): [junkers]  cmdDma:size:0x00000000
[   866.025] (II) SAVAGE(0): [junkers]  cmdDma:map:0x00000000
[   866.025] (II) SAVAGE(0): [junkers]pSAVAGEDRI:
[   866.025] (II) SAVAGE(0): [junkers]  chipset:0x00000007
[   866.026] (II) SAVAGE(0): [junkers]  width:0x00000320
[   866.026] (II) SAVAGE(0): [junkers]  height:0x00000258
[   866.026] (II) SAVAGE(0): [junkers]  mem:0x01000000
[   866.026] (II) SAVAGE(0): [junkers]  cpp:4
[   866.026] (II) SAVAGE(0): [junkers]  zpp:4
[   866.026] (II) SAVAGE(0): [junkers]  agpMode:1
[   866.026] (II) SAVAGE(0): [junkers]  bufferSize:65536
[   866.026] (II) SAVAGE(0): [junkers]  frontbufferSize:0x001d5000
[   866.026] (II) SAVAGE(0): [junkers]  frontOffset:0x00000000
[   866.026] (II) SAVAGE(0): [junkers]  backbufferSize:0x001db000
[   866.026] (II) SAVAGE(0): [junkers]  backOffset:0x003d5000
[   866.026] (II) SAVAGE(0): [junkers]  depthbufferSize:0x001db000
[   866.026] (II) SAVAGE(0): [junkers]  depthOffset:0x005b0000
[   866.026] (II) SAVAGE(0): [junkers]  textureOffset:0x0078b000
[   866.026] (II) SAVAGE(0): [junkers]  textureSize:0x00800000
[   866.026] (II) SAVAGE(0): [junkers]  logTextureGranularity:0x00000014
[   866.026] (II) SAVAGE(0): [junkers]  agpTextureHandle:0xd0000000
[   866.026] (II) SAVAGE(0): [junkers]  agpTextureSize:0x01000000
[   866.026] (II) SAVAGE(0): [junkers]  logAgpTextureGranularity:0x00000014
[   866.026] (II) SAVAGE(0): [junkers]  apertureHandle:0xe4000000
[   866.026] (II) SAVAGE(0): [junkers]  apertureSize:0x05000000
[   866.026] (II) SAVAGE(0): [junkers]  aperturePitch:0x00002000
[   866.026] (II) SAVAGE(0): [junkers]  statusHandle:0x00000000
[   866.026] (II) SAVAGE(0): [junkers]  statusSize:0x00000000
[   866.026] (II) SAVAGE(0): [junkers]  sarea_priv_offset:0x00000898
[   866.026] (II) SAVAGE(0): Direct rendering enabled
[   866.026] (WW) SAVAGE(0): Option "PreferredMode" is not used
[   866.026] (==) RandR enabled
[   866.057] (II) AIGLX: Screen 0 is not DRI2 capable
[   866.057] (EE) AIGLX: reverting to software rendering
[   866.740] (II) AIGLX: Loaded and initialized swrast
[   866.740] (II) GLX: Initialized DRISWRAST GL provider for screen 0
[   866.829] (EE)
[   866.829] (EE) Backtrace:
[   866.853] (EE) 0: /usr/bin/X (xorg_backtrace+0x50) [0x8242ad0]


It looks as if first EXA gets enabled, DRI gets disabled (I suppose that's normal? I htink it happened before..) and then crashes.

I tried disabling the loading of glx, but it did not help, it looks it still gets loaded!

Thanks - R
_________________
Unix is user friendly. It is only careful selecting its friends.


Last edited by multix on Fri May 16, 2014 8:40 am; edited 1 time in total
Back to top
View user's profile Send private message
chithanh
Developer
Developer


Joined: 05 Aug 2006
Posts: 1745
Location: Berlin, Germany

PostPosted: Tue Apr 15, 2014 10:34 am    Post subject: Reply with quote

Code:
(EE) Backtrace:
(EE) 0: /usr/bin/X (xorg_backtrace+0x50) [0x8242ad0]
...
(EE) Bus error at address 0xb2d9d000
This means that your X server crashed, probably a software bug. See if upgrading to the latest unstable or going back to the previous version helps (you'll need to rebuild all xf86-video-* and xf86-input-* packages too). To see which was the previous version, use "qlop -lu xorg-server" from portage-utils package.

If it still happens, such issues can sometimes be triggered by the kernel, like console framebuffer. Disable them all (savagefb, vesafb, uvesafb etc.).

If the latest xorg-server version is still affected, you can report a bug with a full stack trace. See http://www.x.org/wiki/Development/Documentation/ServerDebugging/ for instructions. You will need to build xorg-server by adding FEATURES="splitdebug" and CFLAGS="-g".
Back to top
View user's profile Send private message
multix
Apprentice
Apprentice


Joined: 28 Nov 2004
Posts: 181
Location: Italy

PostPosted: Wed Apr 16, 2014 7:30 am    Post subject: Reply with quote

Hi,

yes, i got that it crashed :) I was wondering if it was a bug, an option or something.

I don't have any framebuffer enabled in the kernel:

Quote:

kernel /boot/vmlinuz-3.14.0-gentoo root=/dev/sda5
#kernel /boot/vmlinuz-3.12.0-gentoo root=/dev/sda5 video=savagefb vga=791
#kernel /boot/vmlinuz-3.12.0-gentoo root=/dev/sda5 video=vesa vga=791


Is this enough? I had to do many experiments, when switching from XAA to EXA essentially the Xorg with savage developed many problems, so I had to disable anything or I would get either garbage in X or get a broken console when exiting X, etc etc.

the last
Code:
qlop -lu xorg-server
show:
Code:

<..>
Thu Dec  5 03:58:38 2013 >>> x11-base/xorg-server-1.14.99.903-r1
Fri Dec 20 17:24:20 2013 <<< x11-base/xorg-server-1.14.99.903-r1
Fri Dec 20 17:24:32 2013 >>> x11-base/xorg-server-1.14.99.904
Thu Jan  9 10:00:41 2014 <<< x11-base/xorg-server-1.14.99.904
Thu Jan  9 10:00:51 2014 >>> x11-base/xorg-server-1.15.0
Sun Apr 13 23:39:10 2014 <<< x11-base/xorg-server-1.15.0
Sun Apr 13 23:39:16 2014 >>> x11-base/xorg-server-1.15.0


I suppose thus that the last working versions were the xorg-server-1.14.99.* series

If I run
Code:
emerge -Dua world
I see:
Code:

<..>
[ebuild  NS    ] sys-kernel/gentoo-sources-3.14.1 [3.9.3, 3.10.0, 3.10.4, 3.10.5, 3.10.5-r1, 3.11.1, 3.12.6, 3.14.0] USE="-build -deblob -experimental -symlink"
[ebuild  r  U  ] x11-base/xorg-server-1.15.1 [1.15.0]
[ebuild  rR    ] x11-drivers/xf86-input-keyboard-1.7.0
[ebuild  rR    ] x11-drivers/xf86-input-evdev-2.8.2
[ebuild  rR    ] x11-drivers/xf86-input-mouse-1.9.0
<..>



I will first update then and see if the problems solve by updating, else I'll ask for more detailed help in how to revert all the X stuff! thanks.
[/quote]
_________________
Unix is user friendly. It is only careful selecting its friends.
Back to top
View user's profile Send private message
chithanh
Developer
Developer


Joined: 05 Aug 2006
Posts: 1745
Location: Berlin, Germany

PostPosted: Wed Apr 16, 2014 8:12 am    Post subject: Reply with quote

Quote:
I don't have any framebuffer enabled in the kernel:
To be sure, look at /proc/fb as some framebuffer drivers activate automatically when built into the kernel.

In order to disable one specific driver, the kernel parameter "video=savagefb:off" or so is needed.
Back to top
View user's profile Send private message
multix
Apprentice
Apprentice


Joined: 28 Nov 2004
Posts: 181
Location: Italy

PostPosted: Thu Apr 17, 2014 3:01 pm    Post subject: Reply with quote

/proc/fb is empty, thus I am sure not to run any framebuffer, right?

I upated kernel, xorg, anything... but I still get the crash.

I think that it is correct first to report the stacktrace so that it can be solved and then to revert my packages so that I "unbrick" my laptop from being a console-only (without even framebuffer).

1) could you give me a bit more detailed on how to build the debug version of xorg-server in gentoo?
2) what would be the best way to revert xorg-server and all the related xf86-stuff? The first think that comes to my mind is masking the xorg-server and re-emerge it. Perhaps the related xf86 packages jsut need to be re-built? or do I need to mask them all manually?

Thanks.
_________________
Unix is user friendly. It is only careful selecting its friends.
Back to top
View user's profile Send private message
chithanh
Developer
Developer


Joined: 05 Aug 2006
Posts: 1745
Location: Berlin, Germany

PostPosted: Thu Apr 17, 2014 3:23 pm    Post subject: Reply with quote

"qlop -lu" will tell you which packages were installed or upgraded recently.

For getting X to run again, it should be sufficient to add
Code:
Option "NoAccel" "true"
to the device section in xorg.conf (if you don't have xorg.conf then just create one with only a device section). If that fails, use xf86-video-vesa instead of savage.
Back to top
View user's profile Send private message
multix
Apprentice
Apprentice


Joined: 28 Nov 2004
Posts: 181
Location: Italy

PostPosted: Thu Apr 17, 2014 4:14 pm    Post subject: Reply with quote

I disabled accel as you mention and xommented the AccelMethod EXA line too, it still crashes.
I disabled DRI, it crashes.
I totally removed my xorg.conf, it crashes.

How do I build the server to generate a useful stacktrace?

You mention I need to add FEATURES="splitdebug" and CFLAGS="-g" , wher edo I do that? I suppose CFLAGS also need to override any optimization options in make.conf. Can I do that per-package?

Riccardo
_________________
Unix is user friendly. It is only careful selecting its friends.
Back to top
View user's profile Send private message
chithanh
Developer
Developer


Joined: 05 Aug 2006
Posts: 1745
Location: Berlin, Germany

PostPosted: Thu Apr 17, 2014 4:43 pm    Post subject: Reply with quote

xorg will default to savage driver. You need a device section in xorg.conf if you want to use a different driver.

You can either set FEATURES and CFLAGS in make.conf, but then it will apply to all packages (which you probably don't want). Another option is via package.env:
/etc/portage/env/splitdebug:
FEATURES="splitdebug"
CFLAGS="${CFLAGS} -g"
CXXFLAGS="${CXXFLAGS} -g"

/etc/portage/package.env:
x11-base/xorg-server splitdebug
Back to top
View user's profile Send private message
multix
Apprentice
Apprentice


Joined: 28 Nov 2004
Posts: 181
Location: Italy

PostPosted: Fri Apr 25, 2014 7:36 pm    Post subject: Reply with quote

Debugging is proving very difficult, compared to the howto you mentioned. Since the server crashes by itself and ot when executing a certain task/program, I cannot just "attach" to it.

If I directly try to gdb, like this:

I get errors:
Quote:
Dwarf Error: wrong version in compilation unit header (is 4, should be 2) [in module /usr/lib/debug/usr/bin/Xorg.debug]


if I do run, I only get this stacktrace, which is quite ugly:
Code:
(gdb) bt
#0  0xb74b35e8 in ?? () from /usr/lib/libpixman-1.so.0
#1  0xb74cf434 in ?? () from /usr/lib/libpixman-1.so.0
#2  0xb74683bd in pixman_fill () from /usr/lib/libpixman-1.so.0
#3  0xb2ca7355 in fbFill () from /usr/lib/xorg/modules/libfb.so
#4  0xb2d38000 in ?? ()
#5  0x00000800 in ?? ()
#6  0x00000020 in ?? ()
#7  0x00000000 in ?? ()


I suppose not very useful to report! also it looks incoherent with the trace that the Xorg.log originally reported. I don't think it is good for a bug report at all.
_________________
Unix is user friendly. It is only careful selecting its friends.
Back to top
View user's profile Send private message
multix
Apprentice
Apprentice


Joined: 28 Nov 2004
Posts: 181
Location: Italy

PostPosted: Mon Apr 28, 2014 10:22 am    Post subject: Reply with quote

I reverted to a previous version of xorg-server (1.4.6, even if I was using 1.4.99...) I did that by

adding to
Quote:
/etc/portage/package.mask


Code:

>x11-base/xorg-server-1.14.99.904


I then rebuilt mesa, input drivers and also the savage driver, even if the version did not change (so I did not block it to revert to a previous one).

I still get a crash, this is then mysterious, right? If the problem is xorg-server, this is unexpected.

Code:

(EE) 0: /usr/bin/X (xorg_backtrace+0x49) [0x81d5739]
(EE) 1: /usr/bin/X (0x8048000+0x191394) [0x81d9394]
(EE) 2: linux-gate.so.1 (__kernel_rt_sigreturn+0x0) [0xb779040c]
(EE) 3: /usr/lib/libpixman-1.so.0 (0xb74c7000+0x525e8) [0xb75195e8]
(EE) 4: /usr/lib/libpixman-1.so.0 (0xb74c7000+0x6e434) [0xb7535434]
(EE) 5: /usr/lib/libpixman-1.so.0 (pixman_fill+0x5d) [0xb74ce3bd]
(EE) 6: /usr/lib/xorg/modules/libfb.so (fbFill+0x2f5) [0xb2f43475]
(EE) 7: /usr/lib/xorg/modules/libfb.so (fbPolyFillRect+0x1a0) [0xb2f43c60]
(EE) 8: /usr/lib/xorg/modules/libexa.so (0xb2f17000+0x11066) [0xb2f28066]
(EE) 9: /usr/lib/xorg/modules/libexa.so (0xb2f17000+0xa551) [0xb2f21551]
(EE) 10: /usr/bin/X (0x8048000+0x112b0b) [0x815ab0b]
(EE) 11: /usr/bin/X (miPaintWindow+0x206) [0x81b67d6]
(EE) 12: /usr/bin/X (miWindowExposures+0x20c) [0x81b714c]
(EE) 13: /usr/bin/X (DRIWindowExposures+0x8c) [0x819f6fc]
(EE) 14: /usr/bin/X (0x8048000+0x8c034) [0x80d4034]
(EE) 15: /usr/bin/X (MapWindow+0x263) [0x80a6ed3]
(EE) 16: /usr/bin/X (InitRootWindow+0x124) [0x80a7024]
(EE) 17: /usr/bin/X (0x8048000+0x1ea90) [0x8066a90]
(EE) 18: /lib/libc.so.6 (__libc_start_main+0xef) [0xb72a5a3f]
(EE) 19: /usr/bin/X (0x8048000+0x1eed8) [0x8066ed8]

_________________
Unix is user friendly. It is only careful selecting its friends.
Back to top
View user's profile Send private message
Navar
Guru
Guru


Joined: 20 Aug 2012
Posts: 333

PostPosted: Mon Apr 28, 2014 10:39 pm    Post subject: Reply with quote

chithanh wrote:
you'll need to rebuild all xf86-video-* and xf86-input-* packages too

chithanh wrote:
(add) Option "NoAccel" "true" to the device section in xorg.conf (if you don't have xorg.conf then just create one with only a device section). If that fails, use xf86-video-vesa instead of savage.


^--- What they said. I don't see where you rebuilt via:

emerge -1a @x11-module-rebuild or
emerge -1a $(qlist -IC x11-drivers) or
x11-drivers/xf86-video-savage directly.

Also try your prior functioning kernel to rule that out. Reverting your xorg server didn't eliminate the regression, so it's most likely either in x11-libs (in your case I'd start with x11-libs/pixman and rebuilds of its dependants), x11-drivers, or kernel.

What was the output of cat /proc/fb?
Back to top
View user's profile Send private message
multix
Apprentice
Apprentice


Joined: 28 Nov 2004
Posts: 181
Location: Italy

PostPosted: Fri May 16, 2014 7:19 am    Post subject: Reply with quote

Navar wrote:

emerge -1a @x11-module-rebuild or
emerge -1a $(qlist -IC x11-drivers) or
x11-drivers/xf86-video-savage directly.

Also try your prior functioning kernel to rule that out. Reverting your xorg server didn't eliminate the regression, so it's most likely either in x11-libs (in your case I'd start with x11-libs/pixman and rebuilds of its dependants), x11-drivers, or kernel.

What was the output of cat /proc/fb?


Sorry, I wrote that I "rebuilt mesa, input drivers and also the savage driver". In any case, I executed:

Code:
emerge -1a @x11-module-rebuild

and set:

Code:
        Option     "NoAccel"           "true"


I still get the crash on startup.

As I wrote, my /proc/fb is empty, here the proof:
Code:

multix@think ~ $ cat /proc/fb
multix@think ~ $


I'm going to try a previous version of pximan too.

(Sorry for the delay, i was AFK and used my other NetBSD laptop in the meanwhile we get this one back working)
_________________
Unix is user friendly. It is only careful selecting its friends.
Back to top
View user's profile Send private message
multix
Apprentice
Apprentice


Joined: 28 Nov 2004
Posts: 181
Location: Italy

PostPosted: Fri May 16, 2014 7:35 am    Post subject: Reply with quote

I am checking the last updates of pixman and savage:

Quote:
Sat Nov 9 01:51:47 2013 <<< x11-libs/pixman-0.30.2
Sat Nov 9 01:51:50 2013 >>> x11-libs/pixman-0.30.2-r1
Thu Dec 5 00:50:57 2013 <<< x11-libs/pixman-0.30.2-r1
Thu Dec 5 00:51:01 2013 >>> x11-libs/pixman-0.32.4


Quote:

Sun Nov 4 10:25:31 2012 >>> x11-drivers/xf86-video-savage-2.3.6
Sun Oct 6 17:36:41 2013 <<< x11-drivers/xf86-video-savage-2.3.6
Sun Oct 6 17:36:44 2013 >>> x11-drivers/xf86-video-savage-2.3.7
Tue Oct 8 20:13:25 2013 <<< x11-drivers/xf86-video-savage-2.3.7
Tue Oct 8 20:13:28 2013 >>> x11-drivers/xf86-video-savage-2.3.7
Wed Dec 11 19:56:39 2013 <<< x11-drivers/xf86-video-savage-2.3.7
Wed Dec 11 19:56:43 2013 >>> x11-drivers/xf86-video-savage-2.3.7
Thu Jan 9 11:58:52 2014 <<< x11-drivers/xf86-video-savage-2.3.7
Thu Jan 9 11:58:56 2014 >>> x11-drivers/xf86-video-savage-2.3.7
Sat Apr 26 09:51:15 2014 <<< x11-drivers/xf86-video-savage-2.3.7
Sat Apr 26 09:51:18 2014 >>> x11-drivers/xf86-video-savage-2.3.7
Fri May 16 09:01:48 2014 <<< x11-drivers/xf86-video-savage-2.3.7
Fri May 16 09:02:24 2014 >>> x11-drivers/xf86-video-savage-2.3.7


You can see, that except the rebuilds, there were no versions of both since 2013, while in April 2014 I had a working X setup at 1024x768 using EXA acceleration. Sadly not always as good as XAA, but for certain tasks better.
_________________
Unix is user friendly. It is only careful selecting its friends.
Back to top
View user's profile Send private message
multix
Apprentice
Apprentice


Joined: 28 Nov 2004
Posts: 181
Location: Italy

PostPosted: Fri May 16, 2014 7:48 am    Post subject: Reply with quote

if I mask pixman with:

Code:
>x11-libs/pixman-0.30.2-r1


when I emerge, I get:
Quote:

Calculating dependencies... done!
[ebuild R #] x11-libs/pixman-0.32.4

The following mask changes are necessary to proceed:
(see "package.unmask" in the portage(5) man page for more details)
# required by pixman (argument)
# /etc/portage/package.mask:
=x11-libs/pixman-0.32.4


I suppose I cannot install an older version in this case?
_________________
Unix is user friendly. It is only careful selecting its friends.
Back to top
View user's profile Send private message
multix
Apprentice
Apprentice


Joined: 28 Nov 2004
Posts: 181
Location: Italy

PostPosted: Fri May 16, 2014 9:15 am    Post subject: Reply with quote

Some important update regarding the kernel.

while running the older xorg-server version, 1.4.6, I reverted to the 3.13.6-gentoo I still had on disk. Voil# xstarts again, in full 1024x768 splendour.


I then update xorg-server again to current, and X still starts, albeit in 800x600, something I hope that can be fixed, but no crash.

To sum up:
3.13.6-gentoo -> works more or or less depending on Xorg version
3.14.0-gentoo -> crash
3.14.1-gentoo -> crash

It could be the kernel or perhaps some slight copnfig adjustments I had to make while updaging is configuration? how can I know?
_________________
Unix is user friendly. It is only careful selecting its friends.
Back to top
View user's profile Send private message
chithanh
Developer
Developer


Joined: 05 Aug 2006
Posts: 1745
Location: Berlin, Germany

PostPosted: Fri May 16, 2014 12:58 pm    Post subject: Reply with quote

As this is almost certainly a bug in the savage driver or X server, it is probably a good idea to report this at https://bugs.freedesktop.org/ including a proper X server backtrace.
Back to top
View user's profile Send private message
multix
Apprentice
Apprentice


Joined: 28 Nov 2004
Posts: 181
Location: Italy

PostPosted: Fri May 16, 2014 6:00 pm    Post subject: Reply with quote

chithanh wrote:
As this is almost certainly a bug in the savage driver or X server, it is probably a good idea to report this at https://bugs.freedesktop.org/ including a proper X server backtrace.


I did some further tests, reverting to the older xorg-server.

It works with an older kernel, but fails with the newer one. I also update to latest 3.14.4 kernel and it fails.

However, my fb is disabled, as /proc/fb says

I thus fear some kind of kernel compatibilit more than server? it looks mesys.

As I wrote before, I failed to generate a proper stacktrace, I posted what I got, it seems quite useless? My gdb doesn't like what I run...
_________________
Unix is user friendly. It is only careful selecting its friends.
Back to top
View user's profile Send private message
multix
Apprentice
Apprentice


Joined: 28 Nov 2004
Posts: 181
Location: Italy

PostPosted: Wed Jun 11, 2014 6:33 am    Post subject: Reply with quote

Isn't this a kernel problem then? if it works with an older kernel 3.13.x? I rebuild my new 3.14 kernel again by updating my config file from the working 3.13 series, but no luck.
_________________
Unix is user friendly. It is only careful selecting its friends.
Back to top
View user's profile Send private message
Navar
Guru
Guru


Joined: 20 Aug 2012
Posts: 333

PostPosted: Wed Jun 18, 2014 6:36 am    Post subject: Reply with quote

Welcome to the world of abandoned hardware. I know it well.

https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-savage/+bug/1083032
http://lists.x.org/archives/xorg/2013-March/055511.html
https://forums.gentoo.org/viewtopic-t-952908.html

DRI/mesa/Xorg abandonment issues. I suspect it's not a kernel 'bug' but a direction change away from support.

Try the xorg.conf Section "device" area Option "DisableTile" as mentioned in those links.
Back to top
View user's profile Send private message
chithanh
Developer
Developer


Joined: 05 Aug 2006
Posts: 1745
Location: Berlin, Germany

PostPosted: Wed Jun 18, 2014 11:00 am    Post subject: Reply with quote

I am not sure what exactly broke at which time. You may be able to find out the offending commit with git bisect.

Navar may be right that this driver is mostly abandoned. However, if you do not properly report your issue to the upstream developers, it will most likely never be fixed.
Back to top
View user's profile Send private message
multix
Apprentice
Apprentice


Joined: 28 Nov 2004
Posts: 181
Location: Italy

PostPosted: Sun Jun 29, 2014 8:58 am    Post subject: Reply with quote

Yes, recently, X drivers tend to make manyolder laptops problematic, while the use of Linux & BSD was always extending life in those venerable machines! I got problems on my mythical G3 iBook too..

Here however, I do not know how to report teh problem.. if I do not know how to generate a useful stacktrace!

Anyway... I compiled other kernels. I upgraded the 3.12 series up to the more recent 3.12.22 and it works fine.

All 3.14 series seem to be problematic, but a first attempt with 3.15 seems to have solved the problem!

I need now to check that this is not a false positive and check that everything is up-to-date and configured correctly, but perhaps this time it got fixed.
_________________
Unix is user friendly. It is only careful selecting its friends.
Back to top
View user's profile Send private message
chithanh
Developer
Developer


Joined: 05 Aug 2006
Posts: 1745
Location: Berlin, Germany

PostPosted: Mon Jun 30, 2014 8:12 am    Post subject: Reply with quote

multix wrote:
Here however, I do not know how to report teh problem.. if I do not know how to generate a useful stacktrace!

Well I told you about how to get a stack trace in an earlier post in this thread. You will need a second computer (or mobile device etc.) with ssh client that can connect your computer.
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Desktop Environments All times are GMT
Page 1 of 1

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum