RE: Floating point corruption

One of the 2 kernels that shows the problem has that call in it...
This may be related to upper registers at least in the f32-f38 range (maybe
more)

Mike

> -----Original Message-----
> From: David Mosberger [mailto:[email protected]]
> Sent: Thursday, May 01, 2003 4:36 PM
> To: [email protected]
> Subject: Re: Floating point corruption
> 
> >>>>> On Thu, 1 May 2003 16:27:35 -0500, [email protected] said:
> 
>   Mike> Before I get bitten by spending too much time on a "known"
>   Mike> issue...  Are there any known problems with 2.4.19 related to
>   Mike> the failure to properly store floating point registers during
>   Mike> a context switch.  (or values that get clobbered) (not related
>   Mike> to the floating point corruption problems associated with
>   Mike> fpswa 1.12)
> 
> There was a problem where firmware calls (PAL, SAL, and EFI calls)
> could potentially corrupt floating-point registers f10-f15.  I don't
> think the problem was ever observed with EFI and even for PAL/SAL it
> was a rare event.
> 
> In any case, check the kernel source file
> linux/include/asm-ia64/sal.h.  If it has calls to
> "ia64_save_scratch_fpregs()", you should be fine.
> 
> 	--david
> 
> 
> 
> _______________________________________________
> ia64-list mailing list
> [email protected]
> https://listman.redhat.com/mailman/listinfo/ia64-list