> On Fri, Aug 30, 2013 at 12:24:17PM +0900, YAMAMOTO Takashi wrote:
>> > Batching reduces overheads and enables upto 4 times the upcall processing
>> > performance in a specialized test case.
>> >
>> > Signed-off-by: Jarno Rajahalme
>> > ---
>> > v2: Do not wake consumer threads if they are alrea
On Fri, Aug 30, 2013 at 12:24:17PM +0900, YAMAMOTO Takashi wrote:
> > Batching reduces overheads and enables upto 4 times the upcall processing
> > performance in a specialized test case.
> >
> > Signed-off-by: Jarno Rajahalme
> > ---
> > v2: Do not wake consumer threads if they are already keepi
> Batching reduces overheads and enables upto 4 times the upcall processing
> performance in a specialized test case.
>
> Signed-off-by: Jarno Rajahalme
> ---
> v2: Do not wake consumer threads if they are already keeping the queue
> level low enough.
if the queue level is high, the consumer
FWIW I'm happy with the final version as well. More obviously correct.
Ethanl
On Wed, Aug 28, 2013 at 3:00 PM, Ben Pfaff wrote:
> On Wed, Aug 28, 2013 at 12:06:07PM -0700, Jarno Rajahalme wrote:
>> Batching reduces overheads and enables upto 4 times the upcall processing
>> performance in a spe
On Wed, Aug 28, 2013 at 12:06:07PM -0700, Jarno Rajahalme wrote:
> Batching reduces overheads and enables upto 4 times the upcall processing
> performance in a specialized test case.
>
> Signed-off-by: Jarno Rajahalme
> ---
> v2: Do not wake consumer threads if they are already keeping the queue
Batching reduces overheads and enables upto 4 times the upcall processing
performance in a specialized test case.
Signed-off-by: Jarno Rajahalme
---
v2: Do not wake consumer threads if they are already keeping the queue
level low enough.
Be more conservative about locking while cond_signa
