[Openmcl-devel] Clozure CL 1.3-RC1 available

[Gary Byers]

On Mon, 16 Feb 2009, John Stoneham wrote:

>
>> Great, thanks.  There are a couple of things that I'm curious about.
>>
>> 1) In the function 'illegal() in the mach32.c function that I sent out
>> the other day, there's a commented-out
>>
>> /*  __asm__ volatile("int $0xd5");  */
>>
>> If you remove the /* */ comments and recompile, do you get:
>>
>> Mach exception: 1 (EXC_BAD_ACCESS) with 2 codes: 0xd 0x6aa
>>
>> (I'd think so.)
>>
>
> Yes, that's exactly what I get.

Just out of curiousity, if you change that "int $0xd5" to "int $0xc5",
do you still get  EXC_BAD_ACCESS with the first code 0xd and the
second (I think) 0x62a ?

The point where you're now dying on Tiger

> Unhandled exception 4 at 0x88adc77, context->regs at #xbffff5cc

- where the "exception 4" is SIGILL - is

0x88adc77:	int    $0xc5

If all unimplemented 'int $n" instructions raised EXC_BAD_ACCESS,
then that wouldn't have been affected by a change to the EXC_BAD_INSTRUCTION
clause.  (So either the assumption that all such 'int $n' instructions
are created equal is invalid, some other clause was changed by mistake,
or there's another, even weirder explanation ...

The last thing that I can think of to ask you to try involves changing
both the EXC_BAD_ACCESS and EXC_BAD_INSTRUCTION clauses, to try to
generate SIGSEGV if the exception is either EXC_BAD_ACCESS or
EXC_BAD_INSTRUCTION, the first code is EXC_I386_GPFLT (= 0xd), and
the program counter is pointing at an 'int $n' instruction.

       switch (exception) {
       case EXC_BAD_ACCESS:
         if (code == EXC_I386_GPFLT) {
           signum = SIGILL;
           if (*((pc)ts_pc(ts)) == INTN_OPCODE) {
             signum = SIGSEGV;
           }
         } else {
           signum = SIGBUS;
         }
         break;

       case EXC_BAD_INSTRUCTION:
         signum = SIGILL;
         if ((code == EXC_I386_GPFLT) &&
             *((pc)ts_pc(ts)) == INTN_OPCODE) {
           signum = SIGSEGV;
         }
         break;

If that works, it was a lucky guess.  If not, I probably won't be able
to come up with a better guess until I get that Core Duo iMac back (a
week or two ?) and can see what's going on here.


>
>> 2) Does changing the following case clause in
>> lisp-kernel/x86-exceptions.c (and rebuilding the kernel) fix things
>> for you, without obviously breaking anything else ?  The clause in
>> quesion is around line 3190, and the proposed change is to
>> unconditionally set 'signum' to SIGILL, e.g., to change:
>>
>>        case EXC_BAD_INSTRUCTION:
>>          if (code == EXC_I386_GPFLT) {
>>            signum = SIGSEGV;
>>          } else {
>>            signum = SIGILL;
>>          }
>>          break;
>>
>>
>> to:
>>
>>        case EXC_BAD_INSTRUCTION:
>>          signum = SIGILL;
>>          break;
>
> Nope, didn't work, but Tiger and Leopard were different this time. I changed the code above and ran "make" in the lisp-kernel/darwinx8632 directory. It compiled fine without errors on both. On Leopard, when I then ran ccl with the new kernel it behaved the same as it did before: QUIT still threw an error, and rebuilding ccl borked at loading the newly compiled image. The backtraces were the same as they were before, as near as I can tell. The only difference was that ^D didn't cause it to freeze, it just went to a new user prompt. But here's the thing. When I tried it on Tiger, ccl wouldn't even get to the repl. This is what I get (with B, T, and L info):
>
> $ ccl
> Unhandled exception 4 at 0x88adc77, context->regs at #xbffff5cc
> ? for help
> [528] Clozure CL kernel debugger: B
> current thread: tcr = 0x300480, native thread ID = 0x113, interrupts disabled
> [528] Clozure CL kernel debugger: T
> Current Thread Context Record (tcr) = 0x300480
> Control (C) stack area:  low = 0xbfe9c000, high = 0xbffff930
> Value (lisp) stack area: low = 0x405000, high = 0x0
> Exception stack pointer = 0x547fec
> [528] Clozure CL kernel debugger: L
> %ebx (arg_z) = #<Anonymous Function #x089B1C3E>
> %esi (arg_y) = #<Anonymous Function #x089B0EBE>
> ------
> %edi (fn) = #<Function (:INTERNAL SAVE-IMAGE) #x088ADC3E>
> ------
> Unhandled exception 10 at 0x2cdc7, context->regs at #xbfffeebc
> Exception occurred while executing foreign code
> at sprint_vector + 12
> ? for help
> [528] Clozure CL kernel debugger:
>
>
>
>
>
>