On 04/20/2015 12:38 AM, Alexandre Courbot wrote:
On Sat, Apr 18, 2015 at 12:14 AM, Terje Bergstrom <[email protected]> wrote:
Wouldn't that be just a more complicated way of expressing the same
thing?
Right now we have r->offset |= BIT(priv->iommu_phys_addr_bit -
priv->iommu_pgshift), which doesn't look much simpler. :)
One reason is that you're doing two things in the same bit operation.
The original address has already been shifted, so you're compensating
the shift, and you're setting bit 34.
_______________________________________________
Nouveau mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/nouveau
