Hello Mark
Sorry, somehow I missed your following comment
On 02/14/2017 10:20 AM, Mark Brown wrote:
On Wed, Feb 08, 2017 at 03:20:27PM +0900, Jiada Wang wrote:
This looks basically fine, a couple of fairly minor things here:
+ for (i = 0; i< transfer->len / 4; i++) {
+
Hello Mark
Sorry, somehow I missed your following comment
On 02/14/2017 10:20 AM, Mark Brown wrote:
On Wed, Feb 08, 2017 at 03:20:27PM +0900, Jiada Wang wrote:
This looks basically fine, a couple of fairly minor things here:
+ for (i = 0; i< transfer->len / 4; i++) {
+
On Wed, Feb 08, 2017 at 03:20:27PM +0900, Jiada Wang wrote:
This looks basically fine, a couple of fairly minor things here:
> + for (i = 0; i < transfer->len / 4; i++) {
> + u8 temp;
> +
> + temp = *(buf + i * 4);
> + *(buf + i * 4) = *(buf + i * 4 + 3);
On Wed, Feb 08, 2017 at 03:20:27PM +0900, Jiada Wang wrote:
This looks basically fine, a couple of fairly minor things here:
> + for (i = 0; i < transfer->len / 4; i++) {
> + u8 temp;
> +
> + temp = *(buf + i * 4);
> + *(buf + i * 4) = *(buf + i * 4 + 3);
previously burst length (BURST_LENGTH) is always set to equal
to bits_per_word, causes a 10us gap between each word in
transfer, which significantly affects performance.
This patch uses 32 bits transfer to simulate lower bits transfer,
and adjusts burst length runtimely to use biggeest burst
previously burst length (BURST_LENGTH) is always set to equal
to bits_per_word, causes a 10us gap between each word in
transfer, which significantly affects performance.
This patch uses 32 bits transfer to simulate lower bits transfer,
and adjusts burst length runtimely to use biggeest burst
6 matches
Mail list logo