Bug 21765 - Memory corruption
Summary: Memory corruption
Status: RESOLVED INVALID
Alias: None
Product: xorg
Classification: Unclassified
Component: Server/General (show other bugs)
Version: git
Hardware: Other All
: medium normal
Assignee: Xorg Project Team
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2009-05-16 05:03 UTC by Søren Sandmann Pedersen
Modified: 2009-06-01 22:53 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments
A log file (34.82 KB, text/plain)
2009-05-22 03:37 UTC, Søren Sandmann Pedersen
no flags Details

Description Søren Sandmann Pedersen 2009-05-16 05:03:59 UTC
I happened to have gdb attached (for other reasons) when the server crashed. Unfortunately, the backtrace is not particularly enlightening:

#0  0x00bf8422 in __kernel_vsyscall ()
#1  0x007439d1 in *__GI_raise (sig=6)
    at ../nptl/sysdeps/unix/sysv/linux/raise.c:64
#2  0x007452d4 in *__GI_abort () at abort.c:88
#3  0x00782395 in __libc_message (do_abort=2, 
    fmt=0x857150 "*** glibc detected *** %s: %s: 0x%s ***\n")
    at ../sysdeps/unix/sysv/linux/libc_fatal.c:170
#4  0x007888d1 in malloc_printerr (action=<value optimized out>, 
    str=0x6 <Address 0x6 out of bounds>, ptr=0x8e34460) at malloc.c:6179
#5  0x080b99ae in Xfree (ptr=0x8e34460) at utils.c:1159
#6  0x0806e4e0 in dixFreePrivates (privates=0x8cf1528) at privates.c:217
#7  0x004d0705 in fbDestroyPixmap (pPixmap=0x88d9938) at fbpixmap.c:103
#8  0x00890587 in XAADestroyPixmap (pPix=0x88d9938) at xaaInit.c:472
#9  0x0814522c in damageDestroyPixmap (pPixmap=0x88d9938) at damage.c:1747
#10 0x0034a241 in XvDestroyPixmap (pPix=0x88d9938) at xvmain.c:381
#11 0x08109ad1 in ShmDestroyPixmap (pPixmap=0x88d9938) at shm.c:292
#12 0x0813037f in FreePicture (value=0x8b42f40, pid=0) at picture.c:1530
#13 0x08131d15 in FreeGlyph (glyph=0x93b2788, format=2) at glyph.c:312
#14 0x0813252f in FreeGlyphSet (value=0x8770428, gid=96487318) at glyph.c:546
#15 0x0809db50 in FreeResource (id=96487318, skipDeleteFuncType=0)
    at resource.c:562
#16 0x081365a8 in ProcRenderFreeGlyphSet (client=0x872ba68) at render.c:1009
#17 0x08138ff3 in ProcRenderDispatch (client=0x872ba68) at render.c:2039
---Type <return> to continue, or q <return> to quit---
#18 0x08084e77 in Dispatch () at dispatch.c:432
#19 0x08064bce in main (argc=8, argv=0xbfd27924, envp=0xbfd27948) at main.c:283
(gdb) quit
The program is running.  Quit anyway (and detach it)? (y or n) y
Comment 1 Michel Dänzer 2009-05-18 04:01:23 UTC
Please attach the corresponding log file.
Comment 2 Søren Sandmann Pedersen 2009-05-20 05:39:40 UTC
Sorry, I forgot to save the log file. 
Comment 3 Michel Dänzer 2009-05-20 07:14:30 UTC
Then please attach the current log file, or manually provide the relevant information (driver, versions, ...).
Comment 4 Søren Sandmann Pedersen 2009-05-22 03:37:17 UTC
Created attachment 26097 [details]
A log file

Something similar happened again. Here is a log file that judging from the time stamp seems to have been the active one when the crash happened. I'm generally running git versions of pixman, so it's certainly conceivable that the memory corruption comes from that.

Program received signal SIGABRT, Aborted.
0x00961422 in __kernel_vsyscall ()
(gdb) bt
#0  0x00961422 in __kernel_vsyscall ()
#1  0x006019d1 in *__GI_raise (sig=6) at ../nptl/sysdeps/unix/sysv/linux/raise.c:64
#2  0x006032d4 in *__GI_abort () at abort.c:88
#3  0x00640395 in __libc_message (do_abort=2, 
    fmt=0x715150 "*** glibc detected *** %s: %s: 0x%s ***\n")
    at ../sysdeps/unix/sysv/linux/libc_fatal.c:170
#4  0x006468d1 in malloc_printerr (action=<value optimized out>, 
    str=0x6 <Address 0x6 out of bounds>, ptr=0x96b7b10) at malloc.c:6179
#5  0x0064bf28 in *__GI___libc_realloc (oldmem=0x0, bytes=128) at malloc.c:3739
#6  0x080b9924 in Xrealloc (ptr=0x96b7b10, amount=128) at utils.c:1127
#7  0x0806e247 in dixAllocatePrivate (privates=0x96b7e60, key=0x518704) at privates.c:129
#8  0x0806e368 in dixLookupPrivate (privates=0x96b7e60, key=0x518704) at privates.c:165
#9  0x004ed313 in XAAComposite (op=12 '\f', pSrc=0x96b7ab8, pMask=0x0, pDst=0x97a5f10, xSrc=0, 
    ySrc=0, xMask=0, yMask=0, xDst=155, yDst=0, width=8, height=13) at xaaPict.c:518
#10 0x0050f653 in cwComposite (op=12 '\f', pSrcPicture=0x96b7ab8, pMskPicture=0x0, 
    pDstPicture=0x97a5f10, xSrc=0, ySrc=0, xMsk=0, yMsk=0, xDst=155, yDst=0, width=8, height=13)
    at cw_render.c:274
#11 0x08141146 in damageComposite (op=12 '\f', pSrc=0x96b7ab8, pMask=0x0, pDst=0x97a5f10, 
    xSrc=0, ySrc=0, xMask=0, yMask=0, xDst=155, yDst=0, width=8, height=13) at damage.c:643
#12 0x08130773 in CompositePicture (op=12 '\f', pSrc=0x96b7ab8, pMask=0x0, pDst=0x97a5f10, 
    xSrc=0, ySrc=0, xMask=0, yMask=0, xDst=155, yDst=0, width=8, height=13) at picture.c:1676
#13 0x08132bc6 in miGlyphs (op=3 '\3', pSrc=0x96c5ef8, pDst=0x96f2620, maskFormat=0x8f1d580, 
    xSrc=14, ySrc=154, nlist=0, list=0xbf952898, glyphs=0xbf9524ec) at glyph.c:726
#14 0x004edad1 in XAAGlyphs (op=3 '\3', pSrc=0x96c5ef8, pDst=0x96f2620, maskFormat=0x8f1d580, 
    xSrc=14, ySrc=154, nlist=1, list=0xbf952898, glyphs=0xbf952498) at xaaPict.c:635
#15 0x08141486 in damageGlyphs (op=3 '\3', pSrc=0x96c5ef8, pDst=0x96f2620, 
    maskFormat=0x8f1d580, xSrc=14, ySrc=154, nlist=1, list=0xbf952898, glyphs=0xbf952498)
    at damage.c:721
#16 0x0813285d in CompositeGlyphs (op=3 '\3', pSrc=0x96c5ef8, pDst=0x96f2620, 
    maskFormat=0x8f1d580, xSrc=14, ySrc=154, nlist=1, lists=0xbf952898, glyphs=0xbf952498)
    at glyph.c:632
#17 0x08137465 in ProcRenderCompositeGlyphs (client=0x934e0b8) at render.c:1413
#18 0x08138ff3 in ProcRenderDispatch (client=0x934e0b8) at render.c:2039
#19 0x08084e77 in Dispatch () at dispatch.c:432
#20 0x08064bce in main (argc=8, argv=0xbf952d84, envp=0xbf952da8) at main.c:283
(gdb) quit
Comment 5 Michel Dänzer 2009-05-22 05:31:52 UTC
Running the X server in valgrind might be interesting.
Comment 6 Søren Sandmann Pedersen 2009-06-01 22:53:14 UTC
This hasn't happened in a while, so maybe it was just some flaky intermediate version of pixman. Probably not worth keeping this bug open. If it's real, there is likely no one but me who can track it down anyway.


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.