From: Philipp K. K. <pk...@sp...> - 2012-04-24 09:04:02
|
Am 23.04.2012 23:01, schrieb Maarten Brock: > Hi all, > > When I run the regression tests I have a few hanging tests: > > gcc-torture-execute-961017-2.c hangs on ucgbz80 and ucr2k and thus > generates a timeout. It passes for ucz80 and ucz180. This fails with a > proper timeout. > > But gcc-torture-execute-930529-1.c hangs when running test-host. And it > does not timeout! > It fails after printing __prints("Running testTortureExecute\n"); and I > see nothing wrong in the generated asm. > > (gdb) disassemble > Dump of assembler code for function __prints: > 0x080484d0 <+0>: push %ebp > 0x080484d1 <+1>: mov %esp,%ebp > 0x080484d3 <+3>: push %ebx > 0x080484d4 <+4>: sub $0x14,%esp > => 0x080484d7 <+7>: mov 0x8(%ebp),%ebx > 0x080484da <+10>: movzbl (%ebx),%eax > 0x080484dd <+13>: test %al,%al > 0x080484df <+15>: je 0x80484fd <__prints+45> > 0x080484e1 <+17>: lea 0x0(%esi,%eiz,1),%esi > 0x080484e8 <+24>: movsbl %al,%eax > 0x080484eb <+27>: add $0x1,%ebx > 0x080484ee <+30>: mov %eax,(%esp) > 0x080484f1 <+33>: call 0x8048760 <_putchar> > 0x080484f6 <+38>: movzbl (%ebx),%eax > 0x080484f9 <+41>: test %al,%al > 0x080484fb <+43>: jne 0x80484e8 <__prints+24> > 0x080484fd <+45>: add $0x14,%esp > 0x08048500 <+48>: pop %ebx > 0x08048501 <+49>: pop %ebp > 0x08048502 <+50>: ret > End of assembler dump. > > I have gcc version 4.4.3 (Ubuntu 4.4.3-4ubuntu5.1) installed and run > linux inside VirtualBox. Could this be the reason? Is anyone else > experiencing this too? It all passes for me. gcc 4.4.3 was released in 2010. gcc-torture-execute-930529-1.c is a gcc bug from 1993. Could you post the asm output you get for gcc-torture-execute-961017-2.c on ucgbz80 and ucr2k, so I can compare to what I get? Philipp |