Re: Floating Point Exception
From: Steve (steve_at_steve.com)
Date: 10/11/04
- Next message: Måns Rullgård: "Re: Floating Point Exception"
- Previous message: Scott J. McCaughrin: "exception support for FPU"
- In reply to: Steve: "Floating Point Exception"
- Next in thread: Måns Rullgård: "Re: Floating Point Exception"
- Reply: Måns Rullgård: "Re: Floating Point Exception"
- Reply: Grant Edwards: "Re: Floating Point Exception"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Mon, 11 Oct 2004 22:47:23 +0100
Thanks for all the replies. I wasn't expecting so many.
I made a typo in my message. The source code has the correct values of
0xAAAAAAAA and 0x55555555. I'm not compiling with any sort of optimization
as this is just a test. None the less, I have now made the pointer
volatile. When I do:
*p = 0x12345678;
printf("*p : %X", *p);
It communicates with the PCI card correctly and the PCI card displays /
returns the correct values.
I was wandering if some sort of buffer exists for the PCI bus. When I write
my values in a loop :
while(1)
{
*p = 0xAAAAAAAA;
*p = 0x55555555;
}
maybe the buffer is receiving data faster than it's sending data.
Therefore, generating some sort of buffer overrun. Is this plausable?
Would it generate a floating point exception? Any ideas on what I can about
it?
Is using a pointer to the memory mapped PCI space the problem? Should I be
using some specific functions?
Thanks again,
- Next message: Måns Rullgård: "Re: Floating Point Exception"
- Previous message: Scott J. McCaughrin: "exception support for FPU"
- In reply to: Steve: "Floating Point Exception"
- Next in thread: Måns Rullgård: "Re: Floating Point Exception"
- Reply: Måns Rullgård: "Re: Floating Point Exception"
- Reply: Grant Edwards: "Re: Floating Point Exception"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Relevant Pages
|