Linux MIPS Architecture development
 help / color / mirror / Atom feed
* Float crash. Fix in exit_thread()
@ 2002-06-06 15:06 Kjeld Borch Egevang
  2002-06-06 18:21 ` Kevin D. Kissell
  0 siblings, 1 reply; 3+ messages in thread
From: Kjeld Borch Egevang @ 2002-06-06 15:06 UTC (permalink / raw)
  To: linux-mips mailing list

I'm running on a system with no FPU. Now, I've got a program do_float 
which seems to work fine. But if I enter:

/bin/echo hello; ./do_float

it terminates with a bus error or FP error. The bug is in exit_thread(). 
Can anyone explain to me, what good the "cfc1 $0,$31" does?


Here is the patch:

RCS file: /cvs/linux/arch/mips/kernel/process.c,v
retrieving revision 1.36
diff -u -r1.36 process.c
--- process.c   2002/05/29 18:36:28     1.36
+++ process.c   2002/06/06 14:51:32
@@ -54,9 +54,11 @@
 void exit_thread(void)
 {
        /* Forget lazy fpu state */
-       if (last_task_used_math == current && mips_cpu.options & MIPS_CPU_FPU) {
-               __enable_fpu();
-               __asm__ __volatile__("cfc1\t$0,$31");
+       if (last_task_used_math == current) {
+               if (mips_cpu.options & MIPS_CPU_FPU) {
+                       __enable_fpu();
+                       __asm__ __volatile__("cfc1\t$0,$31");
+               }
                last_task_used_math = NULL;
        }
 }
@@ -64,9 +66,11 @@
 void flush_thread(void)
 {
        /* Forget lazy fpu state */
-       if (last_task_used_math == current && mips_cpu.options & MIPS_CPU_FPU) {
-               __enable_fpu();
-               __asm__ __volatile__("cfc1\t$0,$31");
+       if (last_task_used_math == current) {
+               if (mips_cpu.options & MIPS_CPU_FPU) {
+                       __enable_fpu();
+                       __asm__ __volatile__("cfc1\t$0,$31");
+               }
                last_task_used_math = NULL;
        }
 }



/Kjeld


-- 
_    _ ____  ___                       Mailto:kjelde@mips.com
|\  /|||___)(___    MIPS Denmark       Direct: +45 44 86 55 85
| \/ |||    ____)   Lautrupvang 4 B    Switch: +45 44 86 55 55
  TECHNOLOGIES      DK-2750 Ballerup   Fax...: +45 44 86 55 56
                    Denmark            http://www.mips.com/

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: Float crash. Fix in exit_thread()
  2002-06-06 15:06 Float crash. Fix in exit_thread() Kjeld Borch Egevang
@ 2002-06-06 18:21 ` Kevin D. Kissell
  2002-06-06 21:00   ` Kjeld Borch Egevang
  0 siblings, 1 reply; 3+ messages in thread
From: Kevin D. Kissell @ 2002-06-06 18:21 UTC (permalink / raw)
  To: Kjeld Borch Egevang, linux-mips mailing list

It's been a while since I worked on the code,
but I'm not sure why last_task_used_math
needs to be cleared if there is no FPU.
The way the FPU emulator was integrated,
the FPU register storage *is* the thread
context, so if there is no FPU, there was no 
FPU context switching, lazy or otherwise.  
Sounds like someone broke this.

Beyond that, the CFC1 instruction is presumably
there because in the R4000 (at least) it is specified
as the means to force the FP pipeline to drain
before the context switch.  I would suppose this
would be done in Linux to avoid mis-attribution
of FP exceptions. (See chapter 6 of the R4000 
User's manual, page 160 in the Second Edition).

            Kevin K.

----- Original Message ----- 
From: "Kjeld Borch Egevang" <kjelde@mips.com>
To: "linux-mips mailing list" <linux-mips@oss.sgi.com>
Sent: Thursday, June 06, 2002 5:06 PM
Subject: Float crash. Fix in exit_thread()


> I'm running on a system with no FPU. Now, I've got a program do_float 
> which seems to work fine. But if I enter:
> 
> /bin/echo hello; ./do_float
> 
> it terminates with a bus error or FP error. The bug is in exit_thread(). 
> Can anyone explain to me, what good the "cfc1 $0,$31" does?
> 
> 
> Here is the patch:
> 
> RCS file: /cvs/linux/arch/mips/kernel/process.c,v
> retrieving revision 1.36
> diff -u -r1.36 process.c
> --- process.c   2002/05/29 18:36:28     1.36
> +++ process.c   2002/06/06 14:51:32
> @@ -54,9 +54,11 @@
>  void exit_thread(void)
>  {
>         /* Forget lazy fpu state */
> -       if (last_task_used_math == current && mips_cpu.options & MIPS_CPU_FPU) {
> -               __enable_fpu();
> -               __asm__ __volatile__("cfc1\t$0,$31");
> +       if (last_task_used_math == current) {
> +               if (mips_cpu.options & MIPS_CPU_FPU) {
> +                       __enable_fpu();
> +                       __asm__ __volatile__("cfc1\t$0,$31");
> +               }
>                 last_task_used_math = NULL;
>         }
>  }
> @@ -64,9 +66,11 @@
>  void flush_thread(void)
>  {
>         /* Forget lazy fpu state */
> -       if (last_task_used_math == current && mips_cpu.options & MIPS_CPU_FPU) {
> -               __enable_fpu();
> -               __asm__ __volatile__("cfc1\t$0,$31");
> +       if (last_task_used_math == current) {
> +               if (mips_cpu.options & MIPS_CPU_FPU) {
> +                       __enable_fpu();
> +                       __asm__ __volatile__("cfc1\t$0,$31");
> +               }
>                 last_task_used_math = NULL;
>         }
>  }
> 
> 
> 
> /Kjeld
> 
> 
> -- 
> _    _ ____  ___                       Mailto:kjelde@mips.com
> |\  /|||___)(___    MIPS Denmark       Direct: +45 44 86 55 85
> | \/ |||    ____)   Lautrupvang 4 B    Switch: +45 44 86 55 55
>   TECHNOLOGIES      DK-2750 Ballerup   Fax...: +45 44 86 55 56
>                     Denmark            http://www.mips.com/
> 
> 

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: Float crash. Fix in exit_thread()
  2002-06-06 18:21 ` Kevin D. Kissell
@ 2002-06-06 21:00   ` Kjeld Borch Egevang
  0 siblings, 0 replies; 3+ messages in thread
From: Kjeld Borch Egevang @ 2002-06-06 21:00 UTC (permalink / raw)
  To: Kevin D. Kissell; +Cc: linux-mips mailing list

On Thu, 6 Jun 2002, Kevin D. Kissell wrote:

> It's been a while since I worked on the code,
> but I'm not sure why last_task_used_math
> needs to be cleared if there is no FPU.
> The way the FPU emulator was integrated,
> the FPU register storage *is* the thread
> context, so if there is no FPU, there was no 
> FPU context switching, lazy or otherwise.  
> Sounds like someone broke this.

If you check do_cpu in traps.c you'll find:

fp_emul:
        if (last_task_used_math != current) {
                if (!current->used_math) {
                        fpu_emulator_init_fpu();
                        current->used_math = 1;
                }
        }
        sig = fpu_emulator_cop1Handler(0, regs, &current->thread.fpu.soft);
        last_task_used_math = current;
        if (sig)


fpu_emulator_init_fpu() is not called, when two processes are created like 
I described. Alternatively the test "if (last_task_used_math != current)" 
could be removed.

> Beyond that, the CFC1 instruction is presumably
> there because in the R4000 (at least) it is specified
> as the means to force the FP pipeline to drain
> before the context switch.  I would suppose this
> would be done in Linux to avoid mis-attribution
> of FP exceptions. (See chapter 6 of the R4000 
> User's manual, page 160 in the Second Edition).

Great. It was exactly something like that I was looking for. Perhaps a 
comment would be nice in the code here ;-)

/Kjeld

-- 
_    _ ____  ___                       Mailto:kjelde@mips.com
|\  /|||___)(___    MIPS Denmark       Direct: +45 44 86 55 85
| \/ |||    ____)   Lautrupvang 4 B    Switch: +45 44 86 55 55
  TECHNOLOGIES      DK-2750 Ballerup   Fax...: +45 44 86 55 56
                    Denmark            http://www.mips.com/

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2002-06-06 20:58 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2002-06-06 15:06 Float crash. Fix in exit_thread() Kjeld Borch Egevang
2002-06-06 18:21 ` Kevin D. Kissell
2002-06-06 21:00   ` Kjeld Borch Egevang

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox