tcp: fix/cleanup inet_ehash_locks_alloc()
authorEric Dumazet <edumazet@google.com>
Tue, 26 May 2015 14:55:34 +0000 (07:55 -0700)
committerDavid S. Miller <davem@davemloft.net>
Tue, 26 May 2015 23:48:46 +0000 (19:48 -0400)
commit095dc8e0c3686d586a01a50abc3e1bb9ac633054
tree5b6b15dcfcc64e553af3587a83d70b5f859f8a45
parentf3903bcc0091df871ac64261f65ed2e4c3519d39
tcp: fix/cleanup inet_ehash_locks_alloc()

If tcp ehash table is constrained to a very small number of buckets
(eg boot parameter thash_entries=128), then we can crash if spinlock
array has more entries.

While we are at it, un-inline inet_ehash_locks_alloc() and make
following changes :

- Budget 2 cache lines per cpu worth of 'spinlocks'
- Try to kmalloc() the array to avoid extra TLB pressure.
  (Most servers at Google allocate 8192 bytes for this hash table)
- Get rid of various #ifdef

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
include/net/inet_hashtables.h
net/ipv4/inet_hashtables.c