On Tue, Jul 22, 2003 at 05:38:15PM +0200, Marcel Holtmann wrote:
> Hi Manuel,
> 
> > > I installed linux-2.6.0-test1-ac2 and tried to port my driver for the
> > > BlueFRITZ! USB Bluetooth dongle to 2.6. This device needs a firmware
> > > download and I want to use the new firmware class for getting the
> > > firmware file from userspace. After reading the documentation and
> > > testing the driver samples I got the results that I expected.
> > > 
> > > My problem is now that the firmware loader is not working with my
> > > firmware file and it seems that this is a problem of the file size,
> > > because copying small files through the same interface is working fine.
> > > This is the file I want to load:
> > > 
> > > -rw-r--r--  1 holtmann staff  418352 Jul 11 12:38 bfubase.frm
> > > 
> > > I have written my own firmware.agent hotplug script, which looks in
> > > general something like this:
> > > 
> > > 	echo 1 > $LOADING
> > > 	cp bfubase.frm $DATA
> > > 	echo 0 > $LOADING
> > > 
> > > Loading the above firmware file through this interface results in
> > > different behaviours. The results are complete freezes, instant reboots,
> > > X server crashes with black screens and sometimes I see an oops about
> > > virtual memory, but it goes bye bye too fast to let me do anything
> > > useful with it.
> > 
> >  Could you send me a tarball with a sample showing the problem. If
> >  possible I would like to do "make test" and have it compile and crash
> >  the system appropriately :)
> 
> I tracked down the problem to request_firmware() or a sysfs problem.
> With the firmware included in a header file the driver itself works
> perfect.

 You are right, the problem was in sysfs, attached goes a patch that
 WorksForMe (tm), please test and report.

> Attached is a sample of how I use the request_firmware() and from the
> documentation it seems correct to me.

 Not what I was asking for, but it seams OK.

 About the patch:
	- undo recent change, made in the believe that "buffer" was the
	  size of the whole file, it is just PAGE_SIZE in size.
		- Since files are allowed to have unknown sizes, by
		  setting their size to 0, we can't preallocate a buffer
		  of their size on open.

	- don't use any offset when handling buffer
		- simplifies code
		- since it is temporary storage it doesn't really matter

	- undo relevant changes to request_firmware() code.

	- hopefully adapt drivers/pci/pci-sysfs.c to this changes
		- Please double check, I didn't look very carefully on
		  this.

 Have a nice day

 	Manuel
-- 
--- Manuel Estrada Sainz <ranty@debian.org>
                         <ranty@bigfoot.com>
			 <ranty@users.sourceforge.net>
------------------------ <manuel.estrada@hispalinux.es> -------------------
Let us have the serenity to accept the things we cannot change, courage to
change the things we can, and wisdom to know the difference.