[22:24:35]  <mjg59> Twenty more lines of code should get me to the point
   where it either works or crashes immediately...
   [22:25:09]  <halfline> faster! faster!
   [22:27:25]  <dcbw> anyone have any idea why the following code wouldn't
   actually trap SIGILL?
   [22:28:10]  <dcbw>         sa.sa_flags = SA_SIGINFO;
   [22:28:11]  <dcbw>         sigemptyset(&sa.sa_mask);
   [22:28:11]  <dcbw>         sa.sa_sigaction = vmx_test;
   [22:28:11]  <dcbw>         sigaction(SIGILL, &sa, &osa);
   [22:28:11]  <dcbw>         jmp_result = setjmp (jump_env);
   [22:28:12]  <dcbw>         if (jmp_result == 0) {
   [22:28:13]  <dcbw>             asm volatile ( "vor 0, 0, 0" );
   [22:28:15]  <dcbw>         }
   [22:28:17]  <dcbw>         sigaction(SIGILL, &osa, NULL);
   [22:28:19]  <dcbw>         have_vmx = (jmp_result == 0);
   [22:28:45]  <dcbw> static void vmx_test(int sig, siginfo_t *si, void
   *unused) {
   [22:28:45]  <dcbw>     longjmp (jump_env, 1);
   [22:28:45]  <dcbw> }
   [22:28:47]  <airlied> I'm going to go with because someone broke signals
   on ppc :)
   [22:28:58]  <dcbw^ that could well be :)
   [22:30:12]  <halfline> well
   [22:30:12]  <dcbw> airlied: if I #if 0 that block out and set have_vmx to
   0, stuff works just fine
   [22:30:22]  <halfline> longjmp from a signal handler is bad news anyway
   [22:30:31]  <halfline> that's what siglongjmp is for i think
   [22:30:34]  <mjg59> Oh, hey, at least the drm comes up without failing
   [22:31:35]  <halfline> dcbw: i'd try switching it to use the right calls
   and see if that fixes thing
   [22:31:36]  <halfline> s
   [22:41:08]  ***  caillon has left chat #fedora-desktop (Ping timeout: 600
   seconds).
   [22:44:28]  ***  sankarshan (~sankarsha@202.80.58.210) has joined chat
   #fedora-desktop.
   [22:52:04]  <dcbw> well here goes with Apple's suggested method
   [22:52:04]  <dcbw>
   http://developer.apple.com/hardwaredrivers/ve/g3_compatibility.html
   [22:52:27]  <dcbw> nope
   [22:54:52]  <halfline> okay if sigsetjmp/siglongjmp didn't help things
   then i'm with airlied
   has been screwed up
   [23:02:59]  <dcbw> halfline: I'll see if dwmw2 will take the bait :)
   [23:03:23]  <dcbw> or ajax; one of them is will probably be a tool and get
   suckered into fixing my bug
   ...
   [16:07:20]  <dcbw> ssp: ssp_: still around anywhere?
   [16:07:34]  <ssp_^ Yeah
   [16:07:37]  <dcbw^ git clone git://fedorapeople.org/~dcbw/pixman.git
   [16:07:54]  <dcbw> that'll have dwmw2's patch
   [16:09:01]  <dcbw> mind if I build an F-10 RPM?
   [16:09:15]  <ssp_^ No, go ahead
   [16:09:31]  <ssp_> Did you find a ppc to test on?
   [16:09:39]  <dcbw^ yup, got a G5
   [16:10:01]  <dcbw> tested on the G3 (doesn't crash anymore) and on the G5
   (altivec reports as enabled)
   [16:10:43]  ***  nirik has left chat #fedora-desktop (Ping timeout: 600
   seconds).
   [16:11:22]  ***  kenvandine has left chat #fedora-desktop (Ex-Chat).
   [16:14:31]  <ssp_> dcbw: It's in master now
   [16:14:56]  <dcbw^ great, thanks!
   [16:35:58]  <dcbw> ssp_: all built and F-10 tag request filed
   [16:36:01]  ***  bpepple|lt