Migration from Solaris 9 to Solaris 10: Memory fault problem

Sorry i should have mentioned it as migration from Solaris 8 to Solaris 10.

Other info: In solaris 8 we compiled using -m32 and on Solaris 10 we are using -m64.

We had never tried on Solaris 9.

i just tried compiling on Solaris 9, i got the same error as in Solaris 10.

I tried to find using dbx but it crashes at different points for different runs.

First Run by pointing connection string - schema 1

-

dbx m2 core

detected a multithreaded program

program terminated by signal BUS (invalid address alignment)

(/Teamware/bin/../WS5.0/bin/sparcv9/dbx) where

current thread: t@1

=>[1] AddClient(0xffffffff7ffff0c0, 0xffffffff7ffff2f0, 0xffffffff7ffff1d4, 0x373300, 0xff0000000000, 0x8080808080808080), at

0x100025ca8

[2] DBSERV_RadReadProxyClients(0x1, 0x10002599c, 0x0, 0x0, 0x0, 0x0), at 0x10006106c

[3] ClientInitialise(0x1, 0x0, 0x0, 0x0, 0x0, 0x0), at 0x1000263e8

[4] InitialiseDetails(0x1, 0x1, 0x0, 0xffffffff7ffff600, 0xffffffff7ffff5f8, 0xffffffff7ffff5f0), at 0x10002bd4c

[5] ReadDatabaseIntoSharedMem(0x1000a9dd8, 0x2aa5, 0x10cf000, 0x20000000, 0x1, 0x8), at 0x10000c5d0

Second run by pointing Connect-string to Schema 2:

-

detected a multithreaded program

program terminated by signal BUS (invalid address alignment)

(/Teamware/bin/../WS5.0/bin/sparcv9/dbx) where

current thread: t@1

=>[1] AddReplyMessage(0xffffffff7ffff1d8, 0x6564006d69732d63, 0xffffffff7ffff200, 0x656400, 0xff0000000000,

0x8080808080808080), at 0x10002e458

[2] DBSERV_RadReadRadiusReplyMessages(0x10002e348, 0x0, 0x0, 0x0, 0x0, 0x0), at 0x1000700a4

[3] RadiusReplyMessageInitialise(0x1000b0538, 0x0, 0x0, 0x0, 0x0, 0x0), at 0x10002e2dc

[4] InitialiseDetails(0x1, 0x1, 0x0, 0xffffffff7ffff5f0, 0xffffffff7ffff5e8, 0xffffffff7ffff5e0), at 0x10002bc3c

[5] ReadDatabaseIntoSharedMem(0x1000a9dd8, 0x2c29, 0x1be400, 0x20000000, 0x1, 0x8), at 0x10000c5d0

steve.mj.barrona at 2007-7-13 0:47:02 >

compilation with -m32 on solaris 10 is throwing SIGSEGV

Incurred fault #6, FLTBOUNDS %pc = 0xFF3E6534

siginfo: SIGSEGV SEGV_MAPERR addr=0x00000000

Received signal #11, SIGSEGV [default]

siginfo: SIGSEGV SEGV_MAPERR addr=0x00000000

dbx m2 core

gives

=>[1] __align_cpy_1(0x1, 0x1084, 0x93394, 0xfe140b00, 0xfe3c2000, 0x1000), at 0xff3e6624

[2] radius_regcmp(0x100bf920, 0x0, 0x420, 0x20, 0x100bf920, 0x100bf920), at 0x3ce6c

[3] DBSERV_ReadCustomMatch(0xffbff1dc, 0xffbff08b, 0xffbfef8d, 0xffeb2d5e, 0x3d, 0xff0000), at 0x6956c

[4] DBSERV_RadReadProxyRules(0x1, 0x32188, 0x242450, 0xfffffffd, 0x0, 0x269605), at 0x66ac8

[5] RuleInitialise(0x1, 0x242450, 0x0, 0x0, 0x0, 0x0), at 0x31394

2nd run pointing to different connect-string

(/Teamware/bin/../WS5.0/bin/sparcv9/dbx) where

current thread: t@1

=>[1] __align_cpy_1(0x101c9424, 0x0, 0x43c, 0x4, 0x0, 0x101c9424), at 0xff3e6534

[2] AddProxyRule(0xffbff0b4, 0x1, 0x3d8410, 0xffd21b46, 0x3d, 0xff0000), at 0x325ac

[3] DBSERV_RadReadProxyRules(0x1, 0x32188, 0x2494d0, 0xfffffffd, 0x0, 0x3fa875), at 0x66bc0

Any catch seeing the trace?

Thanks & Regards

steve.mj.barrona at 2007-7-13 0:47:02 >