SGI: Development

errors after patches

Geoman
Addict
Who joined May 26, 2005, 3:37 a.m.
and authored 367 notes

Wrote the following at March 24, 2010, 10:53 a.m...

I guess it has something to do with the new kernel, that has been compiled after installing the newest IRIX patches. Starting Quake3, I get the following error:

Code:

Octane2 2% ioquake3

ioq3 1.35_SVN1323M irix-mips Apr 19 2008

Zone data failed to allocate 24 megs

recursive error after: Zone data failed to allocate 24 megs

recursive error after: Zone data failed to allocate 24 megs

It's Quake3 Version: 1.35 SVN 1323 from Nekoware.

Strange.. Perhaps one of you know what's this about....

update:
seems like IRIX does not like nekoware any longer (all the commercial stuff runs..)

Code:

Octane2 1% abiword

Segmentation fault (core dumped)

Octane2 2% audacity

Bus error

Octane2 5% fgfs

Abort (core dumped)

_________________
:Octane2:

2xR12000 400MHz, 4GB RAM, V12 GFX
SGI - the legend will never die!!

PymbleSoftware
Addict
Who joined July 15, 2008, 5:48 p.m.
and authored 664 notes

Wrote the following at March 24, 2010, 1:12 p.m...

Geoman wrote:

I guess it has something to do with the new kernel, that has been compiled after installing the newest IRIX patches. Starting Quake3, I get the following error:

Code:

Octane2 2% ioquake3

ioq3 1.35_SVN1323M irix-mips Apr 19 2008

Zone data failed to allocate 24 megs

recursive error after: Zone data failed to allocate 24 megs

recursive error after: Zone data failed to allocate 24 megs

It's Quake3 Version: 1.35 SVN 1323 from Nekoware.

Strange.. Perhaps one of you know what's this about....

update:
seems like IRIX does not like nekoware any longer (all the commercial stuff runs..)

Code:

Octane2 1% abiword

Segmentation fault (core dumped)

Octane2 2% audacity

Bus error

Octane2 5% fgfs

Abort (core dumped)

In a winterm session (UNIX Shell from toolchest) try this:

Code:

echo $LD_LIBRARY_PATH

Code:

dbx /usr/nekoware/bin/abiword

run

(wait for crash)

Code:

where

quit

Post the results back here.

_________________
死の神はりんごだけ食べる

アレゲはアレゲ以上のなにものでもなさげ -- アレゲ研究家

:Onyx2R:

<->

J5600,
3 x SUN, 2 x Mac, Alpha DS20E, Alpha 800 5/550, 2 x RS/6000, Amiga 4000 VideoToaster, Amiga4000 -030, 733MHz Sam440 AmigaOS 4.1 update 1. Tandem Himalaya S-Series Nonstop S72000 ServerNet.

Geoman
Addict
Who joined May 26, 2005, 3:37 a.m.
and authored 367 notes

Wrote the following at March 25, 2010, 8:23 a.m...

Cool! Now I know how to debug those!

Here comes the output:

Code:

Octane2 1% echo $LD_LIBRARY_PATH

/usr/aw/alias/lib32:/usr/aw/alias/lib32

Octane2 2% dbx /usr/nekoware/bin/abiword

dbx version 7.3.7 (96015_Nov16 MR) Nov 16 2004 07:34:16

*STRIPPED* Executable /usr/nekoware/bin/abiword

(dbx) run

Process  1344 (AbiWord-2.2) started

Process  1344 Thread 0x10000 (AbiWord-2.2) stopped on signal SIGSEGV: Segmentation violation (default) at [FcObjectFindByName:162 ,0xed914f0]

Source (of /usr/people/giulio/Sandbox/fontconfig-2.8.0/src/fcname.c) not available for Process  1344

(dbx) where

Thread 0x10000

>  0 FcObjectFindByName(0x7ffefac1, 0x1, 0xedac77c, 0x0, 0x68, 0x66, 0x69, 0x0) ["/usr/people/giulio/Sandbox/fontconfig-2.8.0/src/fcname.c":162, 0xed914f0]

1 FcObjectFromName(0x7ffefac3, 0xedab230, 0xedac77c, 0x0, 0x68, 0x66, 0x69, 0x0) ["/usr/people/giulio/Sandbox/fontconfig-2.8.0/src/fcname.c":322, 0xed91c58]

2 FcParseEdit(0x7ffefac3, 0xedab230, 0xedac77c, 0x0, 0x68, 0x66, 0x69, 0x0) ["/usr/people/giulio/Sandbox/fontconfig-2.8.0/src/fcxml.c":1797, 0xed9c4d8]

3 FcEndElement(0x7ffef8e0, 0xedab230, 0xedac77c, 0x0, 0x68, 0x66, 0x69, 0x0) ["/usr/people/giulio/Sandbox/fontconfig-2.8.0/src/fcxml.c":2187, 0xed9d1c8]

4 <Unknown>() [< unknown >, 0x3c1ae74]

(dbx) quit

_________________
:Octane2:

2xR12000 400MHz, 4GB RAM, V12 GFX
SGI - the legend will never die!!

Geoman
Addict
Who joined May 26, 2005, 3:37 a.m.
and authored 367 notes

Wrote the following at March 25, 2010, 9:36 a.m...

looks like fontconfig 2.8.0 is somehow corrupted:

Code:

Octane2 10# ./fc-cat

Bus error (core dumped)

Code:

Octane2 11# rm core

Octane2 12# dbx fc-cat

dbx version 7.3.7 (96015_Nov16 MR) Nov 16 2004 07:34:16

*STRIPPED* Executable /usr/nekoware/bin/fc-cat

(dbx) run

Process  4795 (fc-cat) started

Process  4795 (fc-cat) stopped on signal SIGBUS: Bus error (default) at [FcObjectFindByName:162 ,0xed914f0]

Source (of /usr/people/giulio/Sandbox/fontconfig-2.8.0/src/fcname.c) not available for Process  4795

(dbx) where

>  0 FcObjectFindByName(0x7ffefaf1, 0x1, 0xfb4f990, 0x0, 0x72, 0x67, 0x62, 0x0) ["/usr/people/giulio/Sandbox/fontconfig-2.8.0/src/fcname.c":162, 0xed914f0]

1 FcObjectFromName(0x7ffefaf5, 0x1, 0xfb4f990, 0x0, 0x72, 0x67, 0x62, 0x0) ["/usr/people/giulio/Sandbox/fontconfig-2.8.0/src/fcname.c":322, 0xed91c58]

2 FcParseEdit(0x7ffefaf5, 0x1, 0xfb4f990, 0x0, 0x72, 0x67, 0x62, 0x0) ["/usr/people/giulio/Sandbox/fontconfig-2.8.0/src/fcxml.c":1797, 0xed9c4d8]

3 FcEndElement(0x7ffef910, 0x1, 0xfb4f990, 0x0, 0x72, 0x67, 0x62, 0x0) ["/usr/people/giulio/Sandbox/fontconfig-2.8.0/src/fcxml.c":2187, 0xed9d1c8]

4 <Unknown>() [< unknown >, 0x3c1ae74]

(dbx)

_________________
:Octane2:

2xR12000 400MHz, 4GB RAM, V12 GFX
SGI - the legend will never die!!

Geoman
Addict
Who joined May 26, 2005, 3:37 a.m.
and authored 367 notes

Wrote the following at March 25, 2010, 9:54 a.m...

I downgraded to fontconfig 2.3.2 and the apps start again w/o core dumping.
But now I've got such squares instead of characters. fontconfig is still messed up - got to repair that.

So Fontconfig2.8 of nekoware_beta is broken and should be removed from distribution.

_________________
:Octane2:

2xR12000 400MHz, 4GB RAM, V12 GFX
SGI - the legend will never die!!

foetz
Old Salt
Who joined April 14, 2003, 4:34 a.m.
and authored 4411 notes

Wrote the following at March 25, 2010, 10:57 a.m...

tried to rebuild the fc cache?

_________________
r-a-c.de

Geoman
Addict
Who joined May 26, 2005, 3:37 a.m.
and authored 367 notes

Wrote the following at March 25, 2010, 12:03 p.m...

not yet - I simply deinstalled all beta nekoware and the problems vanished - the characters are now displayed correctly.

The only (less important) problem that still remains is:

Code:

Octane2 1% dbx /usr/nekoware/bin/ioquake3

dbx version 7.3.7 (96015_Nov16 MR) Nov 16 2004 07:34:16

*STRIPPED* Executable /usr/nekoware/bin/ioquake3

(dbx) run

Process  8354 (ioquake3) started

Interrupt

Process  8354 (ioquake3) stopped on signal SIGINT: Interrupt (default) at [_syssgi:8 +0x8,0xfb7d4e4]

Source (of /xlv22/ficus-jan23/work/irix/lib/libc/libc_n32_M3_ns/sys/syssgi.s) not available for Process  8354

(dbx) where

>  0 _syssgi(0x44, 0x6, 0x7fff0db0, 0x0, 0x73, 0x1, 0x0, 0x2e) ["/xlv22/ficus-jan23/work/irix/lib/libc/libc_n32_M3_ns/sys/syssgi.s":8, 0xfb7d4e4]

1 elfmap(0x7fff1440, 0x6, 0x7fff0db0, 0x0, 0x73, 0x1, 0x0, 0x2e) ["/j7/mtibuild/v744/workarea/v7.4.4m/rld/elfmap.c":365, 0xfb60c54]

2 found_dso_using_path(0x11ea342d, 0xfbd50e0, 0x7fff0db0, 0xfbdacb8, 0x73, 0x1, 0x0, 0x2e) ["/j7/mtibuild/v744/workarea/v7.4.4m/rld/rld_load_object.c":458, 0xfb696f8]

3 attempt_map_on_paths(0x11ea342d, 0x7fff1440, 0x1000, 0xfbdacb8, 0x73, 0x1, 0x0, 0x2e) ["/j7/mtibuild/v744/workarea/v7.4.4m/rld/rld_load_object.c":723, 0xfb69aec]

4 map_object_into_mem_and_init_object_info(0xfbdacb8, 0xfbd8468, 0x7fff0db0, 0xfbd7c58, 0x73, 0x1, 0x0, 0x2e) ["/j7/mtibuild/v744/workarea/v7.4.4m/rld/rld_load_object.c":1132, 0xfb6a39c]

5 build_obj_list_by_liblist_index(0xfbd8468, 0x3, 0x7fff2d60, 0x0, 0x73, 0x1, 0x0, 0x2e) ["/j7/mtibuild/v744/workarea/v7.4.4m/rld/rld_startup.c":2226, 0xfb71c2c]

6 build_obj_list(0xfbd8468, 0x6, 0x7fff0db0, 0x0, 0x73, 0x1, 0x0, 0x2e) ["/j7/mtibuild/v744/workarea/v7.4.4m/rld/rld_startup.c":2074, 0xfb71a34]

7 build_entire_object_list(0xfbd7c58, 0x2, 0x7fff2d60, 0x0, 0x73, 0x1, 0x0, 0x2e) ["/j7/mtibuild/v744/workarea/v7.4.4m/rld/rld_startup.c":521, 0xfb70040]

8 read_aux_vec_and_build_objlist(0x2, 0x10000000, 0x7fff0db0, 0x0, 0x73, 0x1, 0x0, 0x2e) ["/j7/mtibuild/v744/workarea/v7.4.4m/rld/rld_startup.c":1508, 0xfb71104]

9 sgi_main(0x0, 0x7fff2f14, 0x7fff2f14, 0x0, 0x73, 0x1, 0x0, 0x2e) ["/j7/mtibuild/v744/workarea/v7.4.4m/rld/rld_startup.c":1648, 0xfb71344]

10 newmain() ["/j7/mtibuild/v744/workarea/v7.4.4m/rld/rld_newmain.s":68, 0xfb60184]

(dbx) quit

Let's see what's up with libc now

_________________
:Octane2:

2xR12000 400MHz, 4GB RAM, V12 GFX
SGI - the legend will never die!!

foetz
Old Salt
Who joined April 14, 2003, 4:34 a.m.
and authored 4411 notes

Wrote the following at March 25, 2010, 12:04 p.m...

i guess that's why that nekoware is 'beta'

_________________
r-a-c.de

Geoman
Addict
Who joined May 26, 2005, 3:37 a.m.
and authored 367 notes

Wrote the following at March 25, 2010, 2:08 p.m...

That's true-and thanks to PymbleSoftware I learned a lot today ;-)

_________________
:Octane2:

2xR12000 400MHz, 4GB RAM, V12 GFX
SGI - the legend will never die!!

hamei
Old Salt
Who joined Feb. 24, 2004, 5:10 p.m.
and authored 6608 notes

Wrote the following at March 26, 2010, 2:01 a.m...

Hmmm. I've had the beta fontconfig 2.8.0 running here for a while on 6.5.30 with no difficulties. You say you applied some Irix patches ?

PymbleSoftware
Addict
Who joined July 15, 2008, 5:48 p.m.
and authored 664 notes

Wrote the following at March 26, 2010, 3:26 a.m...

Geoman wrote:

That's true-and thanks to PymbleSoftware I learned a lot today ;-)

Good. Now become a God...
http://www.amazon.com/PANIC-System-Anal ... 948&sr=8-1

I spent weeks causing panics in AIX and moving my panic() in the driver code, rebuilding and rebooting and analysing what the driver data structures contained at the time of the panic... I eventually moved some driver code from the lower driver to the upper driver but I can't recall why.

... oh and maybe read the man page on dbx if you have it.

_________________
死の神はりんごだけ食べる

アレゲはアレゲ以上のなにものでもなさげ -- アレゲ研究家

:Onyx2R:

<->

J5600,
3 x SUN, 2 x Mac, Alpha DS20E, Alpha 800 5/550, 2 x RS/6000, Amiga 4000 VideoToaster, Amiga4000 -030, 733MHz Sam440 AmigaOS 4.1 update 1. Tandem Himalaya S-Series Nonstop S72000 ServerNet.

Geoman
Addict
Who joined May 26, 2005, 3:37 a.m.
and authored 367 notes

Wrote the following at March 26, 2010, 5:40 a.m...

@PymbleSoftware: I will spend some time on this topic - it's very interesting, and there are still some unsolved problems.

@hamei: The problem is: what caused the problems first? Unfortunately I made an "unhappy patchday": updating Nekoware and installing SGI patches all at once together. rather doing it one after one and checking for errors.

The downgrade to fontconfig 2.3.2 resolves the problems with nekoware that depends on it, but ioquake3 and fgfs(flightgear) still panic.
dbx says: libc from sgi0007202 causes it. After removal of this patch, dbx says "syssgi.s" is the cause. but such a file does not exists. Unfortunately, even after uninstalling all patches, this error persists.
Perhaps I have to invoke a recompiling of the kernel, since removing the patches in swmgr did not revert the kernel rebuild that took place after installing them.

IRIX Patch 7235: IRIX ntp 4.1.1 patch 08-Mar-2010
IRIX Patch 7234: named fix for IRIX 6.5.[28..30] 06-Nov-2009
IRIX Patch 7233: networking updates for 6.5.30 15-Oct-2009
IRIX Patch 7232: OpenSSH security patch 03-Aug-2009
IRIX Patch 7217: OpenSSL security patch 09-Sep-2008
IRIX Patch 7207: TimeZone processing change and 2007 zonefiles update 16-Jan-2008

were the installed ones.

_________________
:Octane2:

2xR12000 400MHz, 4GB RAM, V12 GFX
SGI - the legend will never die!!

hamei
Old Salt
Who joined Feb. 24, 2004, 5:10 p.m.
and authored 6608 notes

Wrote the following at March 26, 2010, 7:46 a.m...

^^ You've scared me off. I was only interested in the networking updates but maybe someone else will be the guinea pig this time.

I kind of wonder why SGI does not release patches here first. Several private OS/2 users used to get patches from an IBM test group because if they blew up on us, at least we weren't paying customers

Geoman
Addict
Who joined May 26, 2005, 3:37 a.m.
and authored 367 notes

Wrote the following at March 26, 2010, 9:48 a.m...

Hehe

Well, they somehow made my system behave strange - when logging in for example, it takes a minute for the desktop-icons to appear... and flightgear does not want any longer...

_________________
:Octane2:

2xR12000 400MHz, 4GB RAM, V12 GFX
SGI - the legend will never die!!

foetz
Old Salt
Who joined April 14, 2003, 4:34 a.m.
and authored 4411 notes

Wrote the following at March 26, 2010, 12:23 p.m...

i applied that patches on 6.5.30 also and i'm fine so far

_________________
r-a-c.de

Zerolapse
Enthusiast
Who joined Sept. 12, 2005, 8:32 p.m.
and authored 275 notes

Wrote the following at March 31, 2010, 11:36 a.m...

I've been running patched for quite some time on all of my systems and have not had any issues.

Regan: Awesome book suggestion, just bought a copy, they're only: $0.01 on amazon :-)

-Jesse

_________________
:ChallengeL:

(x6 Linked) :Octane2:

(x2)

(4x Challenge S) :PI: