Re: firmware separation filesystem (fwfs)

[Manuel Estrada Sainz]

On Wed, Apr 16, 2003 at 07:00:00PM -0700, Perez-Gonzalez, Inaky wrote:
> 
> > From: David Gibson [mailto:[e-mail]]
> > 
> > Incidentally another approach that also avoids nasty ioctl()s would be
> > to invoke the userland helper with specially set up FD 1, which lets
> > the kernel capture the program's stdout. 
> 
> I think this makes too many assumptions specially taking into
> account that most hotplug stuff are shell scripts - they are
> probably going to be writing all kinds of stuff to stdout.

 Well, FD 3 could be used instead and "cat firmware_image >&3". shell
 scripts should not be writing to FD 3.
 
> With the risk of repeating myself (again) and being a PITA,
> I really think it'd be easier to copy the firmware file to a 
> /sysfs binary file registered by the device driver during 
> initialization; then the driver can wait for the file to be
> written with a valid firmware before finishing the init
> sequence. The infrastructure is already there (or isn't ... 
> is it?).

 I don't know that much about sysfs, after a little investigation, it
 seams like sysfs entries are restricted in size to PAGE_SIZE, which on
 i386 is 4K, and ezusb firmware is already 6.9K in size.

 I would really appreciate someone more knowledgeable than myself
 commenting on the possibility of extending sysfs to fill this gap.

 Have a nice day

 	ranty

-- 
--- Manuel Estrada Sainz <[e-mail]>
                         <[e-mail]>
			 <[e-mail]>
------------------------ <[e-mail]> -------------------
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.