Yes, I have taken fiq.c as an example to write DAI driver. However, when I
enable the device, unmask the interrupt. The device fiq routine is called
'VERY' frequently and the machine halts nearly. Of course, I have tried fill
in the FIFO data queue in the fiq routine to prevent it to generate anymore
'empty' or 'half-empty' interrupt. But the data queue consumes too
quickly.....
Do you have such experience?
Anyway, thx for your reply. The DAI driver source will be opened if I can
finished it. :)
Jacky Lam
-----Original Message-----
From: Hugo Fiennes <[EMAIL PROTECTED]>
To: Jacky Lam <[EMAIL PROTECTED]>; Kong Gang Feng <[EMAIL PROTECTED]>;
umesh <[EMAIL PROTECTED]>; [EMAIL PROTECTED]
<[EMAIL PROTECTED]>
Cc: [EMAIL PROTECTED] <[EMAIL PROTECTED]>
Date: Wednesday, June 28, 2000 11:16 PM
Subject: Re: FIQ support on ARM-Linux
>> I also have this trouble before while I am trying to implement some
>drivers
>> using FIQ. It seems that the overhead of FIQ is too large and called so
>> freqently that the kenrel cannot handle it.
>
>I assume you mean the IRQ overhead is too large?
>
>We use FIQs in various empeg units, both for sound buffer filling (on the
>Cirrus 7212) and things like USB/consumer IR. It all works fine, you just
>have to make sure the source is enabled & unmasked & the registers are
setup
>correctly.
>
>The acorn floppy driver is a good example of fiq usage - if you've not done
>fiq code before it's worth reading the ARM architecture ref manual about
fiq
>code so you're aware of the restrictions/things to look out for.
>
>Hugo
>empeg
>
>
>
>unsubscribe: body of `unsubscribe linux-arm' to [EMAIL PROTECTED]
>++ Please use [EMAIL PROTECTED] for
++
>++ kernel-related discussions.
++
>
unsubscribe: body of `unsubscribe linux-arm' to [EMAIL PROTECTED]
++ Please use [EMAIL PROTECTED] for ++
++ kernel-related discussions. ++
