[pca] Patch 144500-19 and matlab on SPARC

[Willi Burmeister]

Hi all,

this is not an pca issue but hopefully someone has any idea what's
going on.

We are using matlab R2009b (the last version for Solaris SPARC) on
several SunRayServer. After applying patch 144500-19 matlab is simpy
dumping core with a SEGV:

----------------------------------------------------------------------
2962:   sigaction(SIGCANCEL, 0xFFFFFFFF7FFFECC0, 0x00000000) = 0
2962:   sysconfig(_CONFIG_STACK_PROT)                   = 3
2962:   mmap(0x00000000, 8388608, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_NORESERVE|MAP_ANON, -1, 0) = 0xFFFFFFFF6C800000
2962:   lwp_create(0xFFFFFFFF7FFFEE50, LWP_SUSPENDED, 0xFFFFFFFF7FFFEE4C) = 2
2962/2:         lwp_create()    (returning as new lwp ...)      = 0
2962/1:         lwp_continue(2)                                 = 0
2962/2:         setustack(0xFFFFFFFF76900AC0)
2962/2:         schedctl()                                      = 0xFFFFFFFF7F724010
2962/2:         mmap(0x00000000, 2080768, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_NORESERVE|MAP_ANON, -1, 0) = 0xFFFFFFFF6C600000
2962/2:         lwp_create(0xFFFFFFFF6CFFF9C0, LWP_SUSPENDED, 0xFFFFFFFF6CFFF9BC) = 3
2962/2:         lwp_continue(3)                                 = 0
2962/3:         lwp_create()    (returning as new lwp ...)      = 0
2962/3:         setustack(0xFFFFFFFF769012C0)
2962/3:         schedctl()                                      = 0xFFFFFFFF7F724020
2962/3:         resolvepath("/home/matlab/matlab-R2009b/bin/sol64", "/home/matlab/matlab-R2009b/bin/sol64", 1024) = 36
2962/3:             Incurred fault #6, FLTBOUNDS  %pc = 0xFFFFFFFF76F80398
2962/3:               siginfo: SIGSEGV SEGV_MAPERR addr=0xFFFFFFFF6C5FE000
2962/3:             Received signal #11, SIGSEGV [default]
2962/3:               siginfo: SIGSEGV SEGV_MAPERR addr=0xFFFFFFFF6C5FE000
----------------------------------------------------------------------

Does anyone else see the same behaviour?

I know matlab on SPARC is now longer supported but we still need these 
SPARC systems for a while and not patching at all is not really an 
option.

Any hints are welcome.

Thanks

Willi