linux/net/sched
Herbert Xu 932ff279a4 [NET]: Add netif_tx_lock
Various drivers use xmit_lock internally to synchronise with their
transmission routines.  They do so without setting xmit_lock_owner.
This is fine as long as netpoll is not in use.

With netpoll it is possible for deadlocks to occur if xmit_lock_owner
isn't set.  This is because if a printk occurs while xmit_lock is held
and xmit_lock_owner is not set can cause netpoll to attempt to take
xmit_lock recursively.

While it is possible to resolve this by getting netpoll to use
trylock, it is suboptimal because netpoll's sole objective is to
maximise the chance of getting the printk out on the wire.  So
delaying or dropping the message is to be avoided as much as possible.

So the only alternative is to always set xmit_lock_owner.  The
following patch does this by introducing the netif_tx_lock family of
functions that take care of setting/unsetting xmit_lock_owner.

I renamed xmit_lock to _xmit_lock to indicate that it should not be
used directly.  I didn't provide irq versions of the netif_tx_lock
functions since xmit_lock is meant to be a BH-disabling lock.

This is pretty much a straight text substitution except for a small
bug fix in winbond.  It currently uses
netif_stop_queue/spin_unlock_wait to stop transmission.  This is
unsafe as an IRQ can potentially wake up the queue.  So it is safer to
use netif_tx_disable.

The hamradio bits used spin_lock_irq but it is unnecessary as
xmit_lock must never be taken in an IRQ handler.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
2006-06-17 21:30:14 -07:00
..
act_api.c [NET_SCHED]: act_api: fix skb leak in error path 2006-03-12 20:39:36 -08:00
act_gact.c
act_ipt.c [NETFILTER]: ipt action: use xt_check_target for basic verification 2006-04-24 17:27:34 -07:00
act_mirred.c
act_pedit.c
act_police.c [PKT_SCHED] act_police: Rename methods. 2006-04-09 22:25:46 -07:00
act_simple.c
cls_api.c
cls_basic.c
cls_fw.c
cls_route.c
cls_rsvp6.c
cls_rsvp.c
cls_rsvp.h
cls_tcindex.c
cls_u32.c [NET_SCHED]: cls_u32: remove unnecessary NULL-ptr check 2006-03-23 01:16:48 -08:00
em_cmp.c
em_meta.c
em_nbyte.c
em_text.c
em_u32.c
ematch.c
estimator.c
Kconfig [PKT_SCHED]: Let NET_CLS_ACT no longer depend on EXPERIMENTAL 2006-03-20 22:44:24 -08:00
Makefile
sch_api.c
sch_atm.c [PKT_SCHED]: Dump child qdisc handle in sch_{atm,dsmark} 2006-03-20 19:01:06 -08:00
sch_blackhole.c
sch_cbq.c
sch_dsmark.c [PKT_SCHED]: Dump child qdisc handle in sch_{atm,dsmark} 2006-03-20 19:01:06 -08:00
sch_fifo.c
sch_generic.c [NET]: Add netif_tx_lock 2006-06-17 21:30:14 -07:00
sch_gred.c
sch_hfsc.c [NET_SCHED]: HFSC: fix thinko in hfsc_adjust_levels() 2006-05-11 12:22:03 -07:00
sch_htb.c
sch_ingress.c
sch_netem.c [PKT_SCHED] netem: fix loss 2006-04-29 18:33:12 -07:00
sch_prio.c [PKT_SCHED]: Qdisc drop operation is optional 2006-03-20 19:00:49 -08:00
sch_red.c [PKT_SCHED]: Convert sch_red to a classful qdisc 2006-03-20 19:20:44 -08:00
sch_sfq.c [PKT_SCHED]: Keep backlog counter in sch_sfq 2006-03-20 19:01:38 -08:00
sch_tbf.c [PKT_SCHED]: Restore TBF change semantic 2006-03-20 19:01:21 -08:00
sch_teql.c [NET]: Add netif_tx_lock 2006-06-17 21:30:14 -07:00