2008-08-18 00:28:20

by Dmitry Baryshkov

[permalink] [raw]
Subject: [PATCH] lockdep: fix spurious 'inconsistent lock state' warning

Since f82b217e3513fe3af342c0f3ee1494e86250c21c lockdep can output spurious
warnings related to hwirqs due to hardirq_off shrinkage from int to bit-sized
flag. Guard it with double negation to fix the warning.

Signed-off-by: Dmitry Baryshkov <[email protected]>
---
kernel/lockdep.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/kernel/lockdep.c b/kernel/lockdep.c
index 1aa91fd..b298888 100644
--- a/kernel/lockdep.c
+++ b/kernel/lockdep.c
@@ -2584,7 +2584,7 @@ static int __lock_acquire(struct lockdep_map *lock, unsigned int subclass,
hlock->trylock = trylock;
hlock->read = read;
hlock->check = check;
- hlock->hardirqs_off = hardirqs_off;
+ hlock->hardirqs_off = !!hardirqs_off;
#ifdef CONFIG_LOCK_STAT
hlock->waittime_stamp = 0;
hlock->holdtime_stamp = sched_clock();
--
1.5.6.3


2008-08-18 07:42:46

by Ingo Molnar

[permalink] [raw]
Subject: Re: [PATCH] lockdep: fix spurious 'inconsistent lock state' warning


* Dmitry Baryshkov <[email protected]> wrote:

> Since f82b217e3513fe3af342c0f3ee1494e86250c21c lockdep can output
> spurious warnings related to hwirqs due to hardirq_off shrinkage from
> int to bit-sized flag. Guard it with double negation to fix the
> warning.
>
> Signed-off-by: Dmitry Baryshkov <[email protected]>

good spotting! Applied to tip/core/urgent, thanks Dmitry.

I'm wondering, is there any way to teach gcc some sanity here - a safer
variant of bitfields, that is just not allowed to overflow into or
corrupt nearby fields? The fact that a benign looking hlock->state = 15
can corrupt other fields worries me quite a bit. Valid C semantics or
not, this is a totally dangerous construct. The space savings are very
real though, so it would be nice to get 'safer bitfields', somehow.

Ingo