Processor::init() segfaults (linux)

#1 So you don't think it's a problem with the OO script resource then? I'll see what develops with this thread before I look into it much further.

#2 Sorry for the confusion: until I saw where it was segfaulting it didn't occur to me that I hadn't compiled the release version from the vanilla source, just the debug. Through the process of elimination it looks to be the -O2 optimization, or the optimization in combination with one or both of the other flags (inline and omit frame pointer).

#3 This is the thread you need to fix this issue.

#4 Then its a shame I don't have access to it.

#5 Reposted from that thread...
This has also been fixed in Torque 1.4, so it will be updated into T2D when 1.4 is released.

--- x86UNIXCPUInfo.old.cc   2004-11-15 22:11:32.000000000 -0500
+++ x86UNIXCPUInfo.cc   2004-11-15 22:13:07.000000000 -0500
@@ -29,6 +29,8 @@
 static char vendor[13] = {0,};
 static U32 properties = 0;
 static U32 processor  = 0;
+static U32 timeLo;
+static U32 timeHi;
 
 void Processor::init()
 {
@@ -55,6 +57,9 @@
        Platform::SystemInfo.processor.properties & CPU_PROP_FPU)
    {
       const U32 MS_INTERVAL = 750;
+#if defined(TORQUE_COMPILER_GCC) && ((__GNUC__ >= 3) && (__GNUC_MINOR__ >= 4))
+       asm ("rdtsc" : "=a" (timeLo), "=d" (timeHi));
+#else
       __asm__(
          "pushl  %eax\n"
          "pushl  %edx\n"
@@ -64,10 +69,23 @@
          "popl   %edx\n"
          "popl   %eax\n"
          );
+#endif
       U32 ms = Platform::getRealMilliseconds();
       while ( Platform::getRealMilliseconds() < ms+MS_INTERVAL )
       { /* empty */ }
       ms = Platform::getRealMilliseconds()-ms;
+#if defined(TORQUE_COMPILER_GCC) && ((__GNUC__ >= 3) && (__GNUC_MINOR__ >= 4))
+      asm(
+         "pushl  %eax\n"
+         "pushl  %edx\n"
+         "rdtsc\n"
+         "sub    (timeHi), %edx\n"
+         "sbb    (timeLo), %eax\n"
+         "mov    %eax, (clockticks)\n"
+         "popl   %edx\n"
+         "popl   %eax\n"
+         );
+#else
       asm(
          "pushl  %eax\n"
          "pushl  %edx\n"
@@ -78,6 +96,7 @@
          "popl   %edx\n"
          "popl   %eax\n"
          );
+#endif
       U32 mhz = static_cast<U32>(F32(clockticks) / F32(ms) / 1000.0f);
       
       // catch-22 the timing method used above to calc Mhz is generally

#6 Thanks!

#7 This is just FYI:

I never got around to patching because I was using the debug as modified by the OO resource. I noticed that release 1.0.2 doesn't incorporate the fix (presumably Torque 1.4 hasn't been released yet) so I patched it myself and observe the same behavior as previously. So checking version numbers I've got (GCC) 3.3.5 20050117 (prerelease) (SUSE Linux) -- so the patch is expecting gcc 3.4+ ? Incidentally, the altered code works with the version of gcc shipped with SuSE.

I don't pretend to be familiar with what SuSE has done with gcc, why they aren't using 3.4, nor do I know why the patch thinks its needs 3.4+ -- but regardless for anyone using SuSE (including a completely patched 9.3) the patch as shown won't help. I may be the only linux user of T2D using SuSE, but just in case someone else is...