On Thu, May 09, 2002, David Brownell <[EMAIL PROTECTED]> wrote:
>     - compiler warnings ... doesn't use set_bit/clear_bit, those
>       were changed (2.5.10?) to use "unsigned long"

While I understand that the API for set_bit/clear_bit has changed, this
patch removes some warnings and replaces them with a bunch of races.

We use clear_bit and set_bit because it's atomic and makes sure we don't
race with the HC updating the same data.

>     - BUG() from the "modprobe uhci; rmmod uhci" sequence.
>       the root hub was getting billed for refcounts it didn't know
>       about, so it couldn't clean up right.

*sigh* This is not a problem with a sane reference counting system.

> I suspect this should apply to the 2.4 version too.  Please
> merge.

Absolutely not. clear_bit/set_bit are just fine in 2.4 and the reference
counting works absolutely fine in 2.4 too.

> p.s. There's another refcount BUG() lurking in uhci still.

I feel like I'm talking to a broken record recently. Ugh.

Greg, please DON'T apply this patch to 2.4 or 2.5. It is not appropriate
for either. I'll work on a patch for 2.5 that fixes the warnings for
set_bit and clear_bit.

JE


_______________________________________________________________

Have big pipes? SourceForge.net is looking for download mirrors. We supply
the hardware. You get the recognition. Email Us: [EMAIL PROTECTED]
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

Reply via email to