Bug#366245: X crashes with Sig 11 when switching from console to+GUI using Alt + Fn keys
On Sat, Dec 23, 2006 at 02:31:56PM -0500, David Nusinow wrote:
That said, let's try to actually fix the problem. The best thing to do
would be to get a full backtrace. You'll want to build the server and
driver with debugging symbols enabled. These are stock debhelper packages,
so building with DEB_BUILD_OPTIONS set to nostrip (as per dh_strip(1)) will
work. Then if you follow the basic instructions at
http://wiki.x.org/wiki/DebuggingTheXserver you can get the backtraces. We
should be able to go from there. Sorry we don't have -dbg packages, but
several people vetoed them as archive bloat, although I'll probably be
making a server one for etch.
It seems be much harder to reproduce when gdb is attached.
All I get from this is, that it seems to have had a 2 second timeout.
I have no idea from which point it starts to count those 2 seconds.
What I notice when switching is that it does take some time before it
actually redraws the screen, and that might be in the order of 2 seconds.
What I always notice when it's crashing is that it only draws the bottom
half of the screen, and the upper half is black.
I'm going to try with a higher timeout, and see what results this gives.
This seems like a good candidate, since it's timing related, not everybody
has the problem, and I might have a slower machine.
Anyway, I finally got a stack trace:
#0 0xb7f3a410 in ?? ()
#1 0xbfc2589c in ?? ()
#2 0x0006 in ?? ()
#3 0x0bc1 in ?? ()
#4 0xb7d51811 in raise () from /lib/tls/i686/cmov/libc.so.6
#5 0xb7d52fb9 in abort () from /lib/tls/i686/cmov/libc.so.6
#6 0x081a299c in FatalError (f=0xb7bbc6fc lockup\n) at ../../os/log.c:556
#7 0xb7b8f83c in I830WaitLpRing (pScrn=0x81fd0d8, n=131064,
timeout_millis=2000) at ../../src/i830_accel.c:115
#8 0xb7b90092 in I830Sync (pScrn=0x81fd0d8) at ../../src/i830_accel.c:166
#9 0xb79bfe2d in XAALeaveVT () from /usr/lib/xorg/modules/libxaa.so
#10 0x080d7e79 in xf86XVLeaveVT (index=0, flags=0)
at ../../../../hw/xfree86/common/xf86xv.c:1278
#11 0xb7c247cf in glxDRILeaveVT ()
from /usr/lib/xorg/modules/extensions/libglx.so
#12 0x0809f548 in AbortDDX () at ../../../../hw/xfree86/common/xf86Init.c:1308
#13 0x081a2473 in AbortServer () at ../../os/log.c:408
#14 0x081a2987 in FatalError (f=0xb7bbc6fc lockup\n) at ../../os/log.c:554
#15 0xb7b8f83c in I830WaitLpRing (pScrn=0x81fd0d8, n=131064,
timeout_millis=2000) at ../../src/i830_accel.c:115
#16 0xb7b90092 in I830Sync (pScrn=0x81fd0d8) at ../../src/i830_accel.c:166
#17 0xb79c6502 in XAACopyAreaFallback () from /usr/lib/xorg/modules/libxaa.so
#18 0xb79c77f7 in XAACopyArea () from /usr/lib/xorg/modules/libxaa.so
#19 0x081566cd in damageCopyArea (pSrc=0x8397af8, pDst=0x8393b60,
pGC=0x8391e28, srcx=0, srcy=0, width=280, height=20, dstx=0, dsty=0)
at ../../../miext/damage/damage.c:790
#20 0x08084da9 in ProcCopyArea (client=0x835c360) at ../../dix/dispatch.c:1743
#21 0x08086cab in Dispatch () at ../../dix/dispatch.c:459
#22 0x0806e699 in main (argc=11, argv=0xbfc261f4, envp=0x0)
at ../../dix/main.c:479
(gdb) frame 6
#6 0x081a299c in FatalError (f=0xb7bbc6fc lockup\n) at ../../os/log.c:556
556 abort();
(gdb) up
#7 0xb7b8f83c in I830WaitLpRing (pScrn=0x81fd0d8, n=131064,
timeout_millis=2000) at ../../src/i830_accel.c:115
115 FatalError(lockup\n);
(gdb) p last_head
$5 = 12
(gdb) p *ring
$7 = {tail_mask = 1, mem = {Start = -268374161, End = -268377405,
Size = -268374161, Physical = 4026593135, Offset = 4026597204,
Alignment = 4026564616, Key = -268374161, Pool = 0xf000fea5},
virtual_start = 0xf000e987 Address 0xf000e987 out of bounds,
head = -268374161, tail = -268374161, space = -268374161}
(gdb) p n
$8 = 131064
(gdb) up
#8 0xb7b90092 in I830Sync (pScrn=0x81fd0d8) at ../../src/i830_accel.c:166
166I830WaitLpRing(pScrn, pI830-LpRing-mem.Size - 8, 0);
(gdb) p pScrn
$17 = (ScrnInfoPtr) 0x81fd0d8
(gdb) p *pScrn
$19 = {driverVersion = 4000, driverName = 0xb7bbc87e i810,
pScreen = 0x8201350, scrnIndex = 0, configured = 1, origIndex = 0,
imageByteOrder = 0, bitmapScanlineUnit = 32, bitmapScanlinePad = 32,
bitmapBitOrder = 0, numFormats = 0, formats = {{depth = 0 '\0',
bitsPerPixel = 0 '\0', scanlinePad = 0 '\0'}, {depth = 0 '\0',
bitsPerPixel = 0 '\0', scanlinePad = 0 '\0'}, {depth = 0 '\0',
bitsPerPixel = 0 '\0', scanlinePad = 0 '\0'}, {depth = 0 '\0',
bitsPerPixel = 0 '\0', scanlinePad = 0 '\0'}, {depth = 0 '\0',
bitsPerPixel = 0 '\0', scanlinePad = 0 '\0'}, {depth = 0 '\0',
bitsPerPixel = 0 '\0', scanlinePad = 0 '\0'}, {depth = 0 '\0',
bitsPerPixel = 0 '\0', scanlinePad = 0 '\0'}, {depth = 0 '\0',
bitsPerPixel = 0 '\0', scanlinePad = 0 '\0'}}, fbFormat = {
depth = 16 '\020', bitsPerPixel = 16 '\020', scanlinePad = 32 ' '},
bitsPerPixel = 16, pixmap24 = Pix24DontCare, depth = 16,
depthFrom = X_CONFIG, bitsPerPixelFrom = X_PROBED, weight = {red = 5,
green = 6, blue = 5}, mask = {red = 63488,
Bug#366245: X crashes with Sig 11 when switching from console to+GUI using Alt + Fn keys
On Mon, Dec 25, 2006 at 04:31:46PM +0100, Kurt Roeckx wrote: It seems be much harder to reproduce when gdb is attached. All I get from this is, that it seems to have had a 2 second timeout. I have no idea from which point it starts to count those 2 seconds. What I notice when switching is that it does take some time before it actually redraws the screen, and that might be in the order of 2 seconds. What I always notice when it's crashing is that it only draws the bottom half of the screen, and the upper half is black. I'm going to try with a higher timeout, and see what results this gives. This seems like a good candidate, since it's timing related, not everybody has the problem, and I might have a slower machine. Anyway, I finally got a stack trace: I've forwarded this along to upstream at https://bugs.freedesktop.org/show_bug.cgi?id=9444 since I'm not sure what the real problem is. Hopefully Alan can help you out. - David Nusinow -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#366245: X crashes with Sig 11 when switching from console to+GUI using Alt + Fn keys
On Mon, Dec 25, 2006 at 04:31:46PM +0100, Kurt Roeckx wrote: All I get from this is, that it seems to have had a 2 second timeout. I have no idea from which point it starts to count those 2 seconds. What I notice when switching is that it does take some time before it actually redraws the screen, and that might be in the order of 2 seconds. What I always notice when it's crashing is that it only draws the bottom half of the screen, and the upper half is black. I'm going to try with a higher timeout, and see what results this gives. This seems like a good candidate, since it's timing related, not everybody has the problem, and I might have a slower machine. So, this didn't help at all, and just seemt have as effect that when it goes wrong, it now still seems be responsive for a while, and then give the error. What i notice when it works is that the screen is mostly something close to white, except the top 80 lines. There is also a white square also about 80 lines or something lower then where it ends up after it has properly redrawn it. When it fails, it already already has the bottom half of the screen, but about 80 lines lower than it should be, top half is black, except the top 80 lines or something, which look about the same as when it works. It then just seems to move everything up 80 lines, and then it hangs. Just before it hangs, and the top half is black, with the delay that I changed, I can perfectly move the mouse into that top half and it properly shows it, but that's probably normal. Kurt -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#366245: X crashes with Sig 11 when switching from console to+GUI using Alt + Fn keys
Hi, I've seen that you've set the severity to important again. Can you please explain why? Why i think this should have a higher severity is: - X crashed at something that I consider to be a normal use case. - When it crashes, you basicly can't use the computer anymore and it needs to be restarted the hard way. - I'm not the only having the problem. Kurt -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#366245: X crashes with Sig 11 when switching from console to+GUI using Alt + Fn keys
Hi Kurt, On Sat, Dec 23, 2006 at 08:20:32PM +0100, Kurt Roeckx wrote: Hi, I've seen that you've set the severity to important again. Can you please explain why? Why i think this should have a higher severity is: - X crashed at something that I consider to be a normal use case. - When it crashes, you basicly can't use the computer anymore and it needs to be restarted the hard way. - I'm not the only having the problem. Sorry, I've been waiting for your mail to hit my mailbox so I could reply to it properly. Stupid slow bts... The standard for grave X bugs is very high. Your bug only affects certain hardware (we can't fully support all hardware) and under certain conditions (they might be normal use cases for you, but not in general). For example, many XSF members have intel chips and none of us is having this issue. Yes, this is a very important bug to fix, but it doesn't meet the standards the team has been using since before I started working on X. If you look at all the important severity bugs, they should be very similar to this one. That said, let's try to actually fix the problem. The best thing to do would be to get a full backtrace. You'll want to build the server and driver with debugging symbols enabled. These are stock debhelper packages, so building with DEB_BUILD_OPTIONS set to nostrip (as per dh_strip(1)) will work. Then if you follow the basic instructions at http://wiki.x.org/wiki/DebuggingTheXserver you can get the backtraces. We should be able to go from there. Sorry we don't have -dbg packages, but several people vetoed them as archive bloat, although I'll probably be making a server one for etch. - David Nusinow -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#366245: X crashes with Sig 11 when switching from console to GUI using Alt + Fn keys
This seems to be fixed in the latest version of xserver-xorg. Please close this bug. Xorg details are as follows: [EMAIL PROTECTED]:~$ dpkg-query -s xserver-xorg Package: xserver-xorg Status: install ok installed Priority: optional Section: x11 Installed-Size: 576 Maintainer: Debian X Strike Force debian-x@lists.debian.org Architecture: all Source: xorg Version: 1:7.0.22 Replaces: xserver-common Depends: xserver-xorg-core, xserver-xorg-video-all | xserver-xorg-video, xserver-xorg-input-all | xserver-xorg-input, debconf, xkb-data | xkb-data-legacy, xbase-clients Pre-Depends: x11-common (= 7.0.0-0ubuntu3) Recommends: laptop-detect, xresprobe, mdetect, discover1 | discover Conflicts: xserver-xfree86 ( 6.8.2.dfsg.1-1), xserver-common Conffiles: /etc/init.d/xserver-xorg 3a12829045ded9a8427be206f7861f5e obsolete Description: the X.Org X server This package depends on the full suite of the server and drivers for the X.Org X server, as well as providing a configuration infrastructure to manage xorg.conf. It does not provide the actual server itself, but removing it is strongly discouraged. = Thanks, Toufeeq -- blog @ http://toufeeq.blogspot.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
