> Joe,
>
> As you have probably read in the bcm43xx list, I cannot test the
> suspend/resume code as my graphics
> card basically locks up when I try to suspend, but others have reported an
> oops on resume. I have
> added the mailing list and one of the people with the problem to the cc's.
>
> Perhaps, the bcm43xx needs a bit more settling time before we try to resume.
> Could you please
> change IRQWAIT_MAX_RETRIES back to the original value and try the following
> patch? If it works,
> perhaps you could try reducing the msleep time. An extra 50 times through the
> loop in chip_init is
> only 500 usec, and it is likely that msleep(1) will be sufficient. If it
> works with the msleep, then
> we won't spin the system the way a udelay does.
>
> Good luck with the 4318 problem.
>
> Thanks,
>
> Larry
Sorry, the msleep, even at 1000, doesn't make a difference. It
seems to be taking about 65 cycles through the loop for the
IRQ_READY flag to show up, with or without the msleep after suspend.
If I unload and reload the module during normal usage, it's only
about 15 cycles. Maybe Johannes has a better idea as to why, he's
much more familiar with the firmware than I am.
-Joe
_______________________________________________
Bcm43xx-dev mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
