summaryrefslogtreecommitdiff
path: root/.gitignore
diff options
context:
space:
mode:
authorEric Dumazet <dada1@cosmosbay.com>2009-04-28 04:43:42 -0700
committerDavid S. Miller <davem@davemloft.net>2009-04-28 04:43:42 -0700
commit6a321cb370ad3db4ba6e405e638b3a42c41089b0 (patch)
tree2e7c25cff76fc6b439afb04e24ab1133e42667b8 /.gitignore
parent8555a5948467e941acc34e8b1c30df4c1f2f5862 (diff)
downloadlinux-3.10-6a321cb370ad3db4ba6e405e638b3a42c41089b0.tar.gz
linux-3.10-6a321cb370ad3db4ba6e405e638b3a42c41089b0.tar.bz2
linux-3.10-6a321cb370ad3db4ba6e405e638b3a42c41089b0.zip
net: netif_tx_queue_stopped too expensive
netif_tx_queue_stopped(txq) is most of the time false. Yet its cost is very expensive on SMP. static inline int netif_tx_queue_stopped(const struct netdev_queue *dev_queue) { return test_bit(__QUEUE_STATE_XOFF, &dev_queue->state); } I saw this on oprofile hunting and bnx2 driver bnx2_tx_int(). We probably should split "struct netdev_queue" in two parts, one being read mostly. __netif_tx_lock() touches _xmit_lock & xmit_lock_owner, these deserve a separate cache line. Signed-off-by: Eric Dumazet <dada1@cosmosbay.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to '.gitignore')
0 files changed, 0 insertions, 0 deletions