Basically that's a layer translating 32-bit userspace calls intoI _really_ don't want to add new ioctls to usbfs. And as an added penalty, if anyone can convince me to add one, they must write the 64 bit arch specific thunking layer code too (see sparc64 for an example of the pain usbfs's ioctls cause that platform.)
64-bit kernel calls. (Or maybe the other way around, either works.
I seem to recall Solaris started out with the 32-bit kernel to keep
kernel data structures smaller.) It's an ABI compatibility layer,
and I'd suspect x86-64 would want one of those too ... yep, there it
is, linux-2.5.44/arch/x86_64/ia32/ia32_ioctl.c is very similar.
But at least it's straightforward, except for that async call model.
For async, usermode is basically trying to manage urbs, and I think
the #if 0 and comments suggest that those calls won't work in that
kind of environment today.
That's a place where the "usbfs" model is just wrong. Someday
it needs replacing ... like with a model where each non-control
endpoint shows up as a single read-only or write-only fd, and if
folk want async I/O they use standard async infrastructure.
- Dave
-------------------------------------------------------
This sf.net email is sponsored by: Influence the future of Java(TM) technology. Join the Java Community Process(SM) (JCP(SM)) program now. http://ads.sourceforge.net/cgi-bin/redirect.pl?sunm0004en
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
