Re: [patch - v3] epoll ready set loops diet ...

2007-03-01 Thread Ingo Molnar
* Davide Libenzi wrote: > > I was wrong about the size of epitem : it is now 68 bytes instead of > > 72. At least we now use/dirty one cache line instead of two per > > epitem. > > > > Do you have another brilliant idea to shrink 4 more bytes ? :) > > I don't think we can cleanly shove more

Re: [patch - v3] epoll ready set loops diet ...

2007-02-28 Thread Davide Libenzi
On Wed, 28 Feb 2007, Eric Dumazet wrote: > On Wednesday 28 February 2007 19:37, Davide Libenzi wrote: > > > + list_del(&epi->rdllink); > > + if (!(epi->event.events & EPOLLET) && (revents & > > epi->event.events)) > > + list_add_tail(&epi->rdllink, &injlist)

Re: [patch - v3] epoll ready set loops diet ...

2007-02-28 Thread Eric Dumazet
On Wednesday 28 February 2007 19:37, Davide Libenzi wrote: > + list_del(&epi->rdllink); > + if (!(epi->event.events & EPOLLET) && (revents & > epi->event.events)) > + list_add_tail(&epi->rdllink, &injlist); > + else { Is the ( ... & epi->ev

[patch - v3] epoll ready set loops diet ...

2007-02-28 Thread Davide Libenzi
ChangeLog: v3) Removed the "revents" field from the epoll item structure, as suggested by Eric Dumazet v2) In v1, I was trying to avoid to get the spinlock twice WRT yesterday patch, but it turns out I can't since the ready list will be travelling through a path w/out the ep->sem