Re: [Syllable-developer] Audigy: Problem solved! :)

[Henrik I. <he...@is...>]

Glad it works for someone... :)
Now we only have to figure out why it doesn't work on Vander's machine...
I'll put some printk's in it and send a new version to him.

/Henrik

----- Original Message -----
From: "Rick Caudill" <ca...@cu...>
To: <syl...@li...>
Sent: Sunday, March 30, 2003 7:44 PM
Subject: RE: [Syllable-developer] Audigy: Problem solved! :)


> My SB Live Card works great now!  No more Bad Noise(TM) :)!  The only
> time I have problems is when I move Windows and stuff, the song will
> stop playing for a second or two!  Hey I will take that any day over Bad
> Noise(TM) :)!  Great Job Henrik :)!
>
> Rick Caudill
> Development Site: http://www.syllable-desk.tk/
>
>
> -----Original Message-----
> From: syl...@li...
> [mailto:syl...@li...] On Behalf Of
> Henrik Isaksson
> Sent: Saturday, March 29, 2003 10:01 AM
> To: syl...@li...
> Subject: Re: [Syllable-developer] Audigy: Problem solved! :)
>
> > > The sound card expects to get page aligned addresses (ie. a
> mulptiple of
> > > 4096), but Syllable's kmalloc returns ( page + 24 ) byte aligned
> addresses.
> > > I suppose that the kernel adds some information at the start of the
> page.
> >
> > Nice investigation!  I'm also glad I'm not mental, I never would have
> guessed
> > that was the problem! :D
>
> If I hadn't found it, I would definitely had gone mental, this was the
> weirdest bug I've ever encountered... :o)
>
> > > I have solved it very temporarily by allocating an extra page of
> memory
> and
> > > then move the pointer to where the next page starts. (crystal clear
> sound
> > > with no Bad Noise(tm), but it crashes when the memory is supposed to
> be
> > > freed, at the end of the song...)
> >
> > Hey, better than *pop* *pop* *pop* *pop* *pop* all the way through the
> song ;)
>
> Hmm... some songs sound like that anyway... :)
>
> > > I suppose there are two alternatives for a more permanent solution:
> > >
> > > 1. change kmalloc somehow
> >
> > I'd rather not; there could be other parts of the kernel or drivers
> that
> > either rely on that behavour, or work around it.  Hmm...
> >
> > Actually, we could add a MEMF_ALLIGNED flag and just implement it that
> way,
> > but we'll have to work out what to do with those 24bits the kernel
> seems
> to
> > be using.
>
> It's 24 bytes, but only 8 comes with each block you allocate, and the
> rest
> with the page. (The kmalloced memory isn't aligned at all, it just
> looked
> that way because I always allocated 4096 bytes at a time). I don't think
> there's much to do about this, kfree needs this information to be able
> to
> free the memory later.
>
> But fear not, for I have found another memory allocation function pair
> in
> the kernel:
> get_free_pages() and free_pages(). Despite their confusing names, one of
> them allocates pages and the other one frees them.
>
> These functions aren't declared anywhere in the include files, so I
> don't
> know whether it's legitimate to use them. (but I used them anyway, and
> it
> does seem to work fine)
>
> You'll find the source code attached.
>
> /Henrik
>
> ---
> Incoming mail is certified Virus Free.
> Checked by AVG anti-virus system (http://www.grisoft.com).
> Version: 6.0.463 / Virus Database: 262 - Release Date: 3/17/2003
>
>
>
> ---
> Outgoing mail is certified Virus Free.
> Checked by AVG anti-virus system (http://www.grisoft.com).
> Version: 6.0.463 / Virus Database: 262 - Release Date: 3/17/2003
>
>
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by:
> The Definitive IT and Networking Event. Be There!
> NetWorld+Interop Las Vegas 2003 -- Register today!
> http://ads.sourceforge.net/cgi-bin/redirect.pl?keyn0001en
> _______________________________________________
> Syllable-developer mailing list
> Syl...@li...
> https://lists.sourceforge.net/lists/listinfo/syllable-developer
>
>