Kenji Kaneshige wrote:
Phil Endecott wrote:
Kenji Kaneshige wrote:
This patch adds polling mechanism for Data Link Layer Link Active bit
after turning power on, instead of waiting for 1000 msec. This would
reduce the unnecessary long wait.
Hi Kenji,
Thanks, this does indeed seem to work on my Eee 901 to remove those waits.
As for the other waits when it turns the slot off, would it be possible
to make that delay asynchronous? Presumably the spec only requires that
no other activity is attempted on the same slot for one second. The
kernel could get on with other things.
I don't understand what you mean. The kernel can get on with other things
even with msleep(1000).
Well, not really. If you have pciehp built in, the kernel will call
its init function sequentially with all the other driver init
functions. The next driver will not start its initialisation until
after pciehp initialisation has finised, after the msleep(). The
situation is similar with modules.
By the way, I noticed that msleep(1000) after turning power off is not
needed if a slot doesn't implement power controller (as yours) because
power cannot be controlled by software (power would be automatically
removed when the adapter is removed). I'll make a fix about this.
That sounds like a very good solution. I'll look forward to it.
Thanks again!
Phil.
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
