lkml.org 
[lkml]   [2013]   [Jun]   [13]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
Subject[PATCH v4 net-next 0/4] net: low latency sockets follow ups
Date
David,

Here are three clean up patches, plus one that adds a socket option for LLS.

Patch 1 converts sysctl_net_ll_poll into an unsigned int
Patch 2 converts low latency sockets to sched_clock.
Patch 3 removes the config menu for NET_LL_RX_POLL and defaults to y.
Patch 4 adds an SO_LL socket option to allow per-socket control of busy polling.

I will reply to this thread with a patch to sockperf that enables
the socket option. (unchanged from v2)

Again, thanks to everyone for their comments.

-Eliezer


Change log:
v4
- Fixed possible overflow, pointed out by Eric Dumazet.
(reordered the patches)

v3
- Suggested by David Miller and Stephen Hemminger:
added a patch to covert from get_cycles() to sched_clock(),
remove dependency on X86_TSC.
- Suggested by Eric Dumazet:
un-exported sysctl_net_ll_poll since for now it's not needed in modules.

v2
- added a patch to convert sysctl_net_ll_poll into an unsigned int.
- suggested by Eric Dimazet: treat sk_ll_usec as a u_int rather than a u_long.
- suggested by Ben Huchings: allow unprivileged user to decrease SO_LL value,
correct return value when operation is not allowed.



\
 
 \ /
  Last update: 2013-06-14 04:21    [W:0.063 / U:2.216 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site