On Tue, 2 Sep 2003, David Brownell wrote:

> [EMAIL PROTECTED] wrote:
> > diff -Nru linux-2.5.68/drivers/usb/core/usb.c 
> > linux-2.5.68/drivers/usb_patch/core/usb.c
> 
> Here's an updated version, against 2.6.0-test4.
> It updates the doc and also adds a bit of paranoia
> in case the altsettings are out of order.
> 
> This has gotten some discussion from time to time,
> I saw agreement that inactive altsettings must be
> ignored.  Potentially affects user-mode drivers,
> ones relying on usbfs misbehavior.

I think this patch isn't correct.  It includes code that iterates through 
the list of interfaces, searching for the one whose bAlternateSetting 
value is equal to intf->act_altsetting.  But act_altsetting is documented 
in usb.h as "index of current altsetting" -- _not_ the bAlternateSetting 
value of the current altsetting.

There are several other places in the core where this fact is used.  See 
for example usb_enable_interface() and usb_set_interface() in message.c.

Alan Stern





-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

Reply via email to