I've just been testing GCC4 with the SDL library
downloaded from the autobuilder packaging site.
It appears that the use of threading is causing
the test programs to crash. The test programs
all worked with GCC3.
The SDL threading with RISC OS is just a
wrapper around pthreads.
The crash I got from the simple testhread.c example
provided with the SDL source is:
*testhread
Started thread #1: My thread id is 654408
Thread '#1' is alive!
Thread '#1' is alive!
Thread '#1' is alive!
Fatal signal received: Segmentation fault
Stack backtrace:
Running thread 0x79d18
( a87ef8) pc: 560e0074 lr: 560e063c sp: a87efc __write_backtrace()
( a87fa0) pc: 560e0194 lr: 560e1bf8 sp: a87fa4 __unixlib_raise_signal()
( a87fb0) pc: 560e1aa4 lr: 1fb04 sp: a86f80 __h_cback()
Register dump at 00a87fb4:
a1: 1 a2: 8000 a3: e3a03000 a4: 1
v1: 6df53 v2: 6a134 v3: 68 v4: 3a7d8
v5: 5f708 v6: a86f94 sl: a86210 fp: a86f90
ip: 6aeb8 sp: a86f80 lr: 1fb04 pc: 56149504
cpsr: 20000093
561494f0 : llba : 61626c6c : Undefined instruction
561494f4 : ck.. : 00006b63 : ANDEQ R6,R0,R3,ROR #22
561494f8 : ...ÿ : ff000014 : Undefined instruction
561494fc : .!Ÿå : e59f211c : LDR R2,&56149620
56149500 : . —ç : e7972002 : LDR R2,[R7,R2]
56149504 :
_________________________________________________________________
http://clk.atdmt.com/UKM/go/msnnkmgl0010000007ukm/direct/01/
_______________________________________________
GCCSDK mailing list [email protected]
Bugzilla: http://www.riscos.info/bugzilla/index.cgi
List Info: http://www.riscos.info/mailman/listinfo/gcc
Main Page: http://www.riscos.info/index.php/GCCSDK
