Skip to contentSkip to conversation

+ net-uninline-xfrm_selector_match.patch added to -mm tree

DISCUSSION David Miller 2 participants 3 messages Nov 6, 2006 1 message not in the archive

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

#1 lore raw

From: [email protected]
Date: Fri, 03 Nov 2006 11:34:12 -0800

Quoted 7 lines

Subject: net: uninline xfrm_selector_match()
From: Andrew Morton <[email protected]>

Six callsites, huge.

Cc Arnaldo Carvalho de Melo <[email protected]>

You can't implement it like this :-)

xfrm_user.c is a bad place to put the uninlined version because this can be built modular, whereas the callsites in places such as xfrm_policy.c will be built statically into the kernel.

#2 lore raw
Quoted 8 lines · On Mon, 06 Nov 2006 14:38:31 -0800 (PST) David Miller <[email protected]>…

From: [email protected]
Date: Fri, 03 Nov 2006 11:34:12 -0800

Quoted 5 lines

Subject: net: uninline xfrm_selector_match()
From: Andrew Morton <[email protected]>

Six callsites, huge.

Quoted 8 lines
Quoted 2 lines
Cc Arnaldo Carvalho de Melo <[email protected]>

You can't implement it like this :-)

xfrm_user.c is a bad place to put the uninlined version because this can be built modular, whereas the callsites in places such as xfrm_policy.c will be built statically into the kernel.

I would have found that out when I got around to compiling it ;)

Where should the out-of-line function be placed, or should I just drop it?

#3 lore raw

From: Andrew Morton <[email protected]>
Date: Mon, 6 Nov 2006 14:49:32 -0800

Quoted 3 lines

I would have found that out when I got around to compiling it ;)

Where should the out-of-line function be placed, or should I just drop it?

Either xfrm_policy.c or xfrm_state.c would work fine.