[kde-solaris] Solaris 9 64 bit startup problem

[Lars Tunkrans]

robert_kislin at bankone.com wrote:
> 
> Here is the output of my truss when running startkde from within CDE:
> 
> 13065:  door_call(3, 0xFFFFFFFF7FFF55E8)                = 0
> 13054:  poll(0xFFFFFFFF7FFF8E60, 3, 10)                 = 0
> 13054:  ioctl(9, AUDIO_GETINFO, 0x1001712D4)            = 0
> 13054:  lseek(11, 158748, SEEK_SET)                     = 158748
> 13054:  read(11, "E3F4 dF7D4FC1B02A3\t o\0".., 288)     = 288
> 13054:  lseek(11, 159004, SEEK_SET)                     = 159004
> 13054:  read(11, " NFC\fFA gFE >05 [04 :F6".., 288)     = 288
> 13054:  lseek(11, 159260, SEEK_SET)                     = 159260
> 13054:  read(11, " 8F0ACF6 4FB N021904AE05".., 288)     = 288
> 13054:  lseek(11, 159516, SEEK_SET)                     = 159516
> 13054:  read(11, "D201 g\t g07 |040703 QFE".., 288)     = 288
> 13054:  ioctl(9, AUDIO_GETINFO, 0x1001712D4)            = 0
> 13054:  write(9, "FFCBFFCBF88AF88AF4CFF4CF".., 4096)    = 4096
> 13054:  ioctl(9, AUDIO_GETINFO, 0x1001712D4)            = 0
> 13054:  ioctl(9, AUDIO_GETINFO, 0x1001712D4)            = 0
> 13065:      Incurred fault #6, FLTBOUNDS  %pc = 0xFFFFFFFF7A6563E8
> 13065:        siginfo: SIGSEGV SEGV_MAPERR addr=0x7BBB6180
> 13065:      Received signal #11, SIGSEGV [caught]
> 13065:        siginfo: SIGSEGV SEGV_MAPERR addr=0x7BBB6180
> 13065:  sigprocmask(SIG_SETMASK, 0xFFFFFFFF7FFF7490, 0x00000000) = 0
> 13065:  sigaction(SIGALRM, 0xFFFFFFFF7FFF6F40, 0xFFFFFFFF7FFF7100) = 0
> 13065:  alarm(3)                                        = 0
> 13065:  close(10)                                       = 0
> 13065:  getrlimit(RLIMIT_NOFILE, 0xFFFFFFFF7FFF72D0)    = 0
> 13065:  close(0)                                        = 0
> 13065:  close(1)                                        = 0
> 13065:  close(2)
> 

Right,  A segmentation error, the process tries to access memory
outside its allocated memoryspace.

   The crash happends in process 13065 ,  However in the bit
you posted lines from process 13054 is intertwined here so we dont
see the stuff leading up to the crash.

Can you do

   $  grep "13065:" tracefile  >  trace.13065

  so that we can see more of the trace from this particular process
  also tell us if it is kwin.  Check the lines at the beginning of
  this process . the systemcall  "exec"  is this the kwin ?

    Please post a hundred lines or so around the crash.


-- 
========================================================
Lars Tunkrans
--------------------------------------------------------