summaryrefslogtreecommitdiff
path: root/doc
diff options
context:
space:
mode:
authorozaki-r <ozaki-r@NetBSD.org>2017-12-05 03:23:29 +0000
committerozaki-r <ozaki-r@NetBSD.org>2017-12-05 03:23:29 +0000
commitbbed17c2cb3cb1e54c8b7d3a4b21ef9dbae9b992 (patch)
treec3f108e496a37df73e44a1b2ce86f1deec7e21ee /doc
parentf51e41694468f6031b6c00131f138bb9f0185ccb (diff)
Mention the limitation of enabling both NET_MPSAFE and ALTQ
Diffstat (limited to 'doc')
-rw-r--r--doc/TODO.smpnet11
1 files changed, 10 insertions, 1 deletions
diff --git a/doc/TODO.smpnet b/doc/TODO.smpnet
index e0a36417c42..c772a1c9d78 100644
--- a/doc/TODO.smpnet
+++ b/doc/TODO.smpnet
@@ -1,4 +1,4 @@
-$NetBSD: TODO.smpnet,v 1.17 2017/11/20 09:01:20 ozaki-r Exp $
+$NetBSD: TODO.smpnet,v 1.18 2017/12/05 03:23:29 ozaki-r Exp $
MP-safe components
==================
@@ -173,3 +173,12 @@ ec_multiaddrs must be protected by ec_lock. The core of ethernet subsystem is
already MP-safe, however, device drivers that use the data should also be fixed.
A typical change should be to protect manipulations of the data via ETHER_*
macros such as ETHER_FIRST_MULTI by ETHER_LOCK and ETHER_UNLOCK.
+
+ALTQ
+----
+
+If ALTQ is enabled in the kernel, it enforces to use just one Tx queue (if_snd)
+for packet transmissions, resulting in serializing all Tx packet processing on
+the queue. We should probably design and implement an alternative queuing
+mechanism that deals with multi-core systems at the first place, not making the
+existing ALTQ MP-safe because it's just annoying.