I should also add that there's a ton of work that needs to be done
in the kernel for BGL removal in general, particularly the I/O
paths (the network path being only part of the larger picture).
There's plenty of thesis material there, particularly because our
cpu-localization approach is very different from the tact that
other OS's have taken.
-Matt
Matthew Dillon
<dillon@backplane.com>
| Pavel Machek | Re: [AppArmor 39/45] AppArmor: Profile loading and manipulation, pathname matching |
| David Newall | Re: Slow DOWN, please!!! |
| Mark Weber | hdparm standby timeout not working for WD raptors? |
| Andrea Arcangeli | [PATCH 01 of 11] mmu-notifier-core |
git: | |
| David Kastrup | Empty directories... |
| linux | [DRAFT] Branching and merging with git |
| Peter Stahlir | Git as a filesystem |
| Junio C Hamano | Re: irc usage.. |
| Darrin Chandler | Re: bcw(4) is gone |
| Jacob Yocom-Piatt | Re: Real men don't attack straw men |
| Siju George | Re: Real men don't attack straw men |
| Ihar Hrachyshka | Re: That whole "Linux stealing our code" thing |
| YAMAMOTO Takashi | yamt-km branch |
| Martin Husemann | Re: iic(4) device discovery |
| Andrew Doran | Thread benchmarks, round 2 |
| Jonathan Stone | Re: fixing send(2) semantics (kern/29750) |
