Re: Regression with wait_event_timeout in next-20140226

From: Peter Zijlstra
Date: Wed Apr 09 2014 - 07:17:02 EST


On Wed, Feb 26, 2014 at 02:25:34PM -0800, Andrew Morton wrote:
> Is there anything we can do to make all this clearer? Simply using a
> distinctive variable name ("__wait_var__"?) in place of __ret (and
> documenting it) would help a lot.
>
> Some __ret's are long and some are int. Maybe that's a glitch, maybe
> it's because some __ret's are used for inter-macro communications and
> some are not, which just makes things worse.
>
> I started to do a patch, got all confused and gave up. We've made
> quite a tangly mess in there, alas.

Something like so?

---
Subject: wait: Explain the shadowing and type inconsistencies
From: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Date: Wed Apr 9 12:50:34 CEST 2014

Stick in a comment before someone else tries to fix the sparse warning
this generates.

Requested-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
Signed-off-by: Peter Zijlstra <peterz@xxxxxxxxxxxxx>
Link: http://lkml.kernel.org/n/tip-o2ro6f3vkxklni0bc8f7m68s@xxxxxxxxxxxxxx
---
include/linux/wait.h | 14 +++++++++++++-
1 file changed, 13 insertions(+), 1 deletion(-)

--- a/include/linux/wait.h
+++ b/include/linux/wait.h
@@ -191,11 +191,23 @@ wait_queue_head_t *bit_waitqueue(void *,
(!__builtin_constant_p(state) || \
state == TASK_INTERRUPTIBLE || state == TASK_KILLABLE) \

+/*
+ * The below macro ___wait_event() has an explicit shadow of the __ret
+ * variable when used from the wait_event_*() macros.
+ *
+ * This is so that both can use the ___wait_cond_timeout() construct
+ * to wrap the condition.
+ *
+ * The type inconsistency of the wait_event_*() __ret variable is also
+ * on purpose; we use long where we can return timeout values and int
+ * otherwise.
+ */
+
#define ___wait_event(wq, condition, state, exclusive, ret, cmd) \
({ \
__label__ __out; \
wait_queue_t __wait; \
- long __ret = ret; \
+ long __ret = ret; /* explicit shadow */ \
\
INIT_LIST_HEAD(&__wait.task_list); \
if (exclusive) \
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/