This is the mail archive of the
ecos-discuss@sources.redhat.com
mailing list for the eCos project.
Re: RE: Interrupt vectors for FIQ
- To: miketc at voomtech dot com
- Subject: Re: [ECOS] RE: Interrupt vectors for FIQ
- From: Gary Thomas <gthomas at redhat dot com>
- Date: 14 Nov 2001 09:06:47 +0900
- Cc: eCos Discussion <ecos-discuss at sources dot redhat dot com>
- References: <000201c16c9e$de566100$6701a8c0@VOOM1>
[Get raw message]
On Wed, 2001-11-14 at 08:56, Mike Checky wrote:
> I believe this is a question for 'Gary Thomas <gthomas at redhat dot
> com>'.
>
> Is there any particular reason the you didn't include FIQ and SWIRQ in
> AT91EB40 interrupt processing?
>
No. Just that it was not going to be tested and in general, we try and
discourage use of FIQ via the standard ISR/DSR mechanisms (since they
tend to obviate the whole reason for using FIQ in the first place).
> I see no reason why it would not work.
>
> In relation to 'wu shaojie <w_shaojie at hotmail dot com>' question
> about adding FIQ processing to the AT91EB40, I think the
> 'hal_platform_ints.h' file should be changed to:
>
> #define CYGNUM_HAL_INTERRUPT_FIQ 0
> #define CYGNUM_HAL_INTERRUPT_SWIRQ 1
>
> Now for a general question: The ARM interrupt processing is rather
> heavy weight, is there any plans to streamline this? Not that I need
> it, I'm just curious.
How would you streamline it? What's in there that could be [safely]
removed? The only thing I can think of is possibly adjusting the
registers that are saved/restored to be a more minimal set.