* Sam Fourman Jr. [2007-09-07 04:44]:
traditionally, and interrupt cards intterupts once per packet it has
received (or successfully sent, but let us ignore the tx path for now).
interrupt mitigation, coalescing, *insert name-of-the-day for int
mitgation here* etc all describe teh same thing, a certain heuristic to
not interrupt once per packet, but once for a bunch. this is usually
achieved by a combination of amount of packets (i. e. "interrupt every
25 packets") or bytes and a timer (to make sure there is no packet
sitting in the rx queue for too long).
--
Henning Brauer, hb@bsws.de, henning@openbsd.org
BS Web Services, http://bsws.de
Full-Service ISP - Secure Hosting, Mail and DNS Services
Dedicated Servers, Rootservers, Application Hosting - Hamburg & Amsterdam
| Heiko Carstens | Re: -mm merge plans for 2.6.23 -- sys_fallocate |
| Linus Torvalds | Linux 2.6.21-rc4 |
| Michael Kerrisk | nanosleep() uses CLOCK_MONOTONIC, should be CLOCK_REALTIME? |
| Linus Torvalds | Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3 |
git: | |
| Gary Thomas | Marvell 88E609x switch? |
| Gerrit Renker | [PATCH 15/37] dccp: Set per-connection CCIDs via socket options |
| David Miller | [GIT]: Networking |
| David Miller | Re: [PATCH] pkt_sched: Destroy gen estimators under rtnl_lock(). |
