On Sun, 21 Mar 2004 15:55:05 -0600 (CST) "Daniel Hansen"
<[EMAIL PROTECTED]> babbled:

> For a while I have been having problems when compiling mmx support into
> evas.  If I compile without any cpu optimizations, I have no problems, but
> once I compile in mmx, (since I have a p2) any efl app run at the same
> time as entrance crashes with an illegal instruction.  If I login in with
> xdm, these problems go away.  So it would appear the issue is between evas
> and entrance.  Here is my output from gdb of evas_software_x11_test while
> entrance is running:
> 
> [Switching to Thread 16384 (LWP 12599)]
> evas_common_cpu_mmx2_test () at evas_cpu.c:39
> 39         mmx_r2m(movntq, mm0, data);
> 
> and
> 
> (gdb) backtrace
> #0  evas_common_cpu_mmx2_test () at evas_cpu.c:39
> #1  0x10000000 in ?? ()
> #2  0x404e8c1c in __libc_start_main () from /lib/libc.so.6

ok... first.. this sint a crash - evas checks cpu features by trying to execute
an instruction and "catching" the SIGILL (illegal instruction) signal - if it
gets a SIGILL it knows that instruction is not supported. all that's happening
is gdb is getting the SIGILL instead... just type:

cont

and let it continue. it may do other SIGILL's depending on your cpu, but this is
not a bug - it is, in-fact a feature and is not the problem (if any). just cont
until your bt isn't inside the evas_common_cpu_*_test() routines.

also i fail to see how entrance can cause another application to segv. if this
is happening i definitely see a system issue (kernel?) on the horizon - linxu is
a memory protected system - one app cannto cause another to segv (directly)
unless theres an implicit protocol/resource shared that the app is handling
badly in certain cases - which is not the case with evas (unless its x that is
screwing up?)... anyway - try it again but "cont" out of the cpu checks.

> Also, my p2 doesnt support mmx2, only mmx, and so I think the problem
> could be related to that.  Then again, I'm not a programmer, so I really
> don't know.  Anyway, I hope that helps and I hope the bug is fixed :)
> 
> Daniel Hansen
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> enlightenment-devel mailing list
> [EMAIL PROTECTED]
> https://lists.sourceforge.net/lists/listinfo/enlightenment-devel


-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
The Rasterman (Carsten Haitzler)    [EMAIL PROTECTED]
熊耳 - 車君 (数田)                  [EMAIL PROTECTED]
Tokyo, Japan (東京 日本)


-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
_______________________________________________
enlightenment-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/enlightenment-devel

Reply via email to