On Mon, Jan 09, 2017 at 12:40:07PM +, Tvrtko Ursulin wrote:
>
> On 09/01/2017 12:28, Chris Wilson wrote:
> >On Mon, Jan 09, 2017 at 12:23:37PM +, Tvrtko Ursulin wrote:
> >>
> >>Hi,
> >>
> >>On 20/12/2016 13:42, Tvrtko Ursulin wrote:
> >>>From: Tvrtko Ursulin
> >>>
> >>>Rather than freeing
On 09/01/2017 12:28, Chris Wilson wrote:
On Mon, Jan 09, 2017 at 12:23:37PM +, Tvrtko Ursulin wrote:
Hi,
On 20/12/2016 13:42, Tvrtko Ursulin wrote:
From: Tvrtko Ursulin
Rather than freeing and re-allocating the pages when DMA mapping
in large chunks fails, we can just rebuild the sg ta
On Mon, Jan 09, 2017 at 12:23:37PM +, Tvrtko Ursulin wrote:
>
> Hi,
>
> On 20/12/2016 13:42, Tvrtko Ursulin wrote:
> >From: Tvrtko Ursulin
> >
> >Rather than freeing and re-allocating the pages when DMA mapping
> >in large chunks fails, we can just rebuild the sg table with no
> >coalescing.
Hi,
On 20/12/2016 13:42, Tvrtko Ursulin wrote:
From: Tvrtko Ursulin
Rather than freeing and re-allocating the pages when DMA mapping
in large chunks fails, we can just rebuild the sg table with no
coalescing.
Also change back the page counter to unsigned int because that
is what the sg API s
On 20/12/2016 14:14, Chris Wilson wrote:
On Tue, Dec 20, 2016 at 01:42:47PM +, Tvrtko Ursulin wrote:
From: Tvrtko Ursulin
Rather than freeing and re-allocating the pages when DMA mapping
in large chunks fails, we can just rebuild the sg table with no
coalescing.
You are freeing and real
On Tue, Dec 20, 2016 at 01:42:47PM +, Tvrtko Ursulin wrote:
> From: Tvrtko Ursulin
>
> Rather than freeing and re-allocating the pages when DMA mapping
> in large chunks fails, we can just rebuild the sg table with no
> coalescing.
You are freeing and reallocating the pages - I thought you m
From: Tvrtko Ursulin
Rather than freeing and re-allocating the pages when DMA mapping
in large chunks fails, we can just rebuild the sg table with no
coalescing.
Also change back the page counter to unsigned int because that
is what the sg API supports.
Signed-off-by: Tvrtko Ursulin
Cc: Chris
