bnx2x: avoid two atomic ops per page on x86
authorEric Dumazet <edumazet@google.com>
Fri, 20 Jan 2017 16:25:34 +0000 (08:25 -0800)
committerDavid S. Miller <davem@davemloft.net>
Mon, 23 Jan 2017 16:16:27 +0000 (11:16 -0500)
commitb9032741e4f86844d8c4a7c18001ee328dae2f7a
tree8f4858e51e80ccc51e5f0f586f620b9af344c08a
parent41e8c70ee1628a32606bfafd9cfce4f5ae3b5205
bnx2x: avoid two atomic ops per page on x86

Commit 4cace675d687 ("bnx2x: Alloc 4k fragment for each rx ring buffer
element") added extra put_page() and get_page() calls on arches where
PAGE_SIZE=4K like x86

Reorder things to avoid this overhead.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Cc: Yuval Mintz <Yuval.Mintz@cavium.com>
Cc: Ariel Elior <ariel.elior@cavium.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
drivers/net/ethernet/broadcom/bnx2x/bnx2x_cmn.c