> Here's one minor fix for your last patch, and documentation for the
> FILL_*_URB macros. Comments?
Looks good to me, but I have yet to run it against kerneldoc.
I'd probably make that minor fix read "USB_DEVICE_*()" to
make it clear(er) that there's a family of USB_DEVICE() macros.
Umm ... I hesitate to suggest kerneldoc for those, but ... ;-)
> No one is using the FILL_BULK_URB_TO and FILL_INT_URB_TO
> macros, should they just be deleted?
I'd say yes, at least for 2.5. And not just because I dislike those
MACROS SHOUTING AT ME ... mostly because they're
unused and not particularly useful. :) Inline functions for the
rest would make me happier, FWIW.
Similarly -- at least in the 2.5 tree -- for the HID-specific calls
that are in "usb.h" and "usb.c". If any are used inside the HID
code, that's where the code should live. Several weren't used
even there.
For the calls that are only used by the hub driver, does anyone
know the details of those exporting changes planned for 2.5?
Normally I'd want to see those declarations removed from "usb.h"
as well as the EXPORT_* section in "usb.c", and just rely on
standard C linking semantics to resolve them in "usbcore.o".
(Probably with an 'extern' decl in hub.c or hub.h, depending
on where they're used.)
Was that something that was going to change? Or is it that only
symbols explicitly listed in an EXPORT_* declaration will
be exported from a module?
- Dave
p.s. Code cleanup and audit -- seems like a good set of things
to start the 2.5 work! Quickly backportable, too, assuming
nothing gets destabilized.
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
