Re: usbus428 0.0.7 patch against release alsa-driver 0.9.6 (was Re: tascam gives a go for us428 @ alsa ...)

[Takashi Iwai <tiwai@suse.de>]

At Mon, 11 Aug 2003 01:26:09 +0200,
Martin Langer wrote:
> 
> On Thu, Aug 07, 2003 at 04:44:08PM +0200, Takashi Iwai wrote:
> > At Wed, 30 Jul 2003 21:33:37 +0200,
> > Karsten Wiese wrote:
> > > 
> > > new here is:
> > > - both firmware-downloads are handled now by the snd-usb-us428-dl module (this 
> > > functionality should move further on into userspace via libusb ? volunteers?)
> > 
> > yes, definitely it's the right way to go.
> > 
> 
> I've looked into my US122 and there I've found an EZUSB chip. The US428
> might be the same.
> 
> I think the first download step which also changes the ID has to be the ezusb
> firmware. 
> 
> After comparing that sniffed log with some available windows binaries from
> Tascam, I finally found the data in US122DL.SYS/US428DL.SYS/US224DL.SYS.
> Instead of loading the sniffed file it could also be an extract from 
> USx2xDL.SYS (it should be similar to the MidiSport firmware extract and
> loading procedure).

it's nice.  could you compare the stuff with the one in Karsten's
patch?

> That could be an easy way to move the first firmware part away from the driver
> directory. And we won't need another ok from Tascam for including
> sniffed us224/us122 firmware...

yep.  that would be far much better.


while we're in here: i'm looking at the code now and have some
questions / suggestions to Karsten:

* the comments in usbus428.c looks like the GPL/BSD dual license but
  MODULE_LICENSE() is defined as GPL.  is it correct?

* what's the reason of the independent snd-usb-us428-dl module?
  this module seems doing only feeding the firmware to the interface,
  so why not in snd-usb-us428 itself?  this will reduce the use of
  inter_module_* stuffs which are obsolete in 2.6 kernels.

* usbus428ctls.c defines mmap & co.  they still be needed?
  could they be replaced with the standard ALSA control API?
  (btw, P_us428ctls_sharedmem_u may be unalighed to the page
   when allocated via kmalloc().  so the mmap code looks buggy for
   me anyway.  also, please avoid virt_to_phys() as much as possible.
   this is architecture dependent, and won't run on non-x86 well.)

* 64bit division macros can be found in pcm.h.
  the asm codes in usbus428audio.c can be removed.

* there are still GCC-style of structs initialization in some parts :)

* please follow the coding style described in linux/Documentation/
  CodingStyle.  spaces after a (open) parenthesis are not common...
  (i know it's a matter of tastes, but we should keep the consistent
   style if possible.)
  also, debug codes which are obviously unnecessary any more should be
  removed in the release source codes.

* you don't need to modify configure stuff but just add the following
  to alsa-driver/utils/Modules.dep and run cvscompile again:

	%dir linux/sound/usb
	snd-usb-us428-dl 
	|snd-usb-us428 snd-usb-audio snd-usb-us428-dl


thanks!

Takashi


-------------------------------------------------------
This SF.Net email sponsored by: Free pre-built ASP.NET sites including
Data Reports, E-commerce, Portals, and Forums are available now.
Download today and enter to win an XBOX or Visual Studio .NET.
http://aspnet.click-url.com/go/psa00100003ave/direct;at.aspnet_072303_01/01