Subject: Re: [linux-audio-dev] Fragment size question.
From: Benno Senoner (sbenno_AT_gardena.net)
Date: Sat Sep 16 2000 - 00:51:02 EEST
Ah, so you are in fact emulating AFMT_U8, right ?
In that case I would report back 8bits and a fragsize of 256
and the number of frags (-1 to comply with the OSS/Free bug),
the user requested.
But I don't think AFMT_U8 is THAT important to get emulated,
since most apps (which one does not ?) support the 16bit mode.
(the OSS specs does not seem to require the support of certain
audio bit sizes)
Is that still the famous Mac HW audio driver ?
(just for curiousity: is it usable for decent audio (at least comparable to
a good 16bit PC audio card ?)
PS: I'm just printing out intel's FPU manuals ... guess why ? :-))
cheers,
Benno.
On Fri, 15 Sep 2000, Iain Sandoe wrote:
> (OSS kernel driver Q)
>
> I'm working on a driver which, internally, only deals with 16bit stereo
> data.
>
> So when you set AFMT_U8 (for example) and request a frag size of 256 the
> driver figures that (internally) it needs 256 * 2 * 2 bytes to work in - and
> reports this as the frag size for GETCAPS.
>
> We are not dealing with mmap IO, the buffers are all OK (pre-allocated at
> init) and there is no actual crash involved. The question is:
>
> Should I fix it to report 256 back - or is it the expected behaviour to
> report back the frag size used internally?
>
> TIA,
> Iain.
This archive was generated by hypermail 2b28 : Sat Sep 16 2000 - 00:18:46 EEST