On 11/27/21 5:05 PM, Jonathan Cameron wrote:
Non-coherent mapping with no cache sync:
- fileio:
read: 156 MiB/s
write: 123 MiB/s
- dmabuf:
read: 234 MiB/s (capped by sample rate)
write: 182 MiB/s
Non-coherent reads with no cache sync + write-combine writes:
On Thu, 25 Nov 2021 17:29:58 +
Paul Cercueil wrote:
> Hi Jonathan,
>
> Le dim., nov. 21 2021 at 17:43:20 +, Paul Cercueil
> a écrit :
> > Hi Jonathan,
> >
> > Le dim., nov. 21 2021 at 15:00:37 +, Jonathan Cameron
> > a écrit :
> >> On Mon, 15 Nov 2021 14:19:21 +
> >> Paul
On Sun, 21 Nov 2021 17:43:20 +
Paul Cercueil wrote:
> Hi Jonathan,
>
> Le dim., nov. 21 2021 at 15:00:37 +, Jonathan Cameron
> a écrit :
> > On Mon, 15 Nov 2021 14:19:21 +
> > Paul Cercueil wrote:
> >
> >> We can be certain that the input buffers will only be accessed by
> >>
Hi Jonathan,
Le dim., nov. 21 2021 at 17:43:20 +, Paul Cercueil
a écrit :
Hi Jonathan,
Le dim., nov. 21 2021 at 15:00:37 +, Jonathan Cameron
a écrit :
On Mon, 15 Nov 2021 14:19:21 +
Paul Cercueil wrote:
We can be certain that the input buffers will only be accessed by
user
Hi Jonathan,
Le dim., nov. 21 2021 at 15:00:37 +, Jonathan Cameron
a écrit :
On Mon, 15 Nov 2021 14:19:21 +
Paul Cercueil wrote:
We can be certain that the input buffers will only be accessed by
userspace for reading, and output buffers will mostly be accessed by
userspace for wr
On Mon, 15 Nov 2021 14:19:21 +
Paul Cercueil wrote:
> We can be certain that the input buffers will only be accessed by
> userspace for reading, and output buffers will mostly be accessed by
> userspace for writing.
Mostly? Perhaps a little more info on why that's not 'only'.
>
> Therefor
Hi,
Le lun., nov. 15 2021 at 14:19:21 +, Paul Cercueil
a écrit :
We can be certain that the input buffers will only be accessed by
userspace for reading, and output buffers will mostly be accessed by
userspace for writing.
Therefore, it makes more sense to use only fully cached input buff
We can be certain that the input buffers will only be accessed by
userspace for reading, and output buffers will mostly be accessed by
userspace for writing.
Therefore, it makes more sense to use only fully cached input buffers,
and to use the write-combine cache coherency setting for output buffe
