On 2019-06-24 22:38, Linus Torvalds wrote:
On Mon, Jun 24, 2019 at 10:42 PM Roman Penyaev
wrote:
So harvesting events from userspace gives 15% gain. Though bench_http
is not ideal benchmark, but at least it is the part of libevent and
was
easy to modify.
Worth to mention that uepoll is ver
On 2019-06-25 02:24, Eric Wong wrote:
Roman Penyaev wrote:
Hi all,
+cc Jason Baron
** Limitations
4. No support for EPOLLEXCLUSIVE
If device does not pass pollflags to wake_up() there is no way to
call poll() from the context under spinlock, thus special work is
schedule
Roman Penyaev wrote:
> Hi all,
+cc Jason Baron
> ** Limitations
> 4. No support for EPOLLEXCLUSIVE
> If device does not pass pollflags to wake_up() there is no way to
> call poll() from the context under spinlock, thus special work is
> scheduled to offload polling. In this s
On Mon, Jun 24, 2019 at 10:42 PM Roman Penyaev wrote:
>
> So harvesting events from userspace gives 15% gain. Though bench_http
> is not ideal benchmark, but at least it is the part of libevent and was
> easy to modify.
>
> Worth to mention that uepoll is very sensible to CPU, e.g. the gain above
Hi all,
This is v5 which introduces pollable epoll from userspace.
v5:
- Do not support architectures with reduced set of atomic ops,
namely arc-plat-eznps, sparc32, parisc. Return -EOPNOTSUP on
attempt to create uepoll on these archs.
- Explicitly hint compiler, that return value o
5 matches
Mail list logo