On Thu, 2007-07-05 at 11:10 +0200, Peter Zijlstra wrote: > > +/* both _name and _key must not be NULL */ Heh actually I got confused. The requirement that _key is not NULL comes from the fact that I need to copy it for the workqueue stuff, not from lockdep. Hence, only _name must be non-NULL, I'll change the comment to reflect that. > > +#define STATIC_LOCKDEP_MAP_INIT(_name, _key) \ > > + { .name = (_name), .key = (_key), } > > looks sane enough :-) :) I'll repost with these minor changes directly to akpm tomorrow then; no lockdep support on my powerbook that I'm using right now so I can't even compile-test. johannes
Attachment:
signature.asc
Description: This is a digitally signed message part
- References:
- [PATCH] debug work struct cancel deadlocks with lockdep
- From: Johannes Berg <[email protected]>
- Re: [PATCH] debug work struct cancel deadlocks with lockdep
- From: Ingo Molnar <[email protected]>
- Re: [PATCH] debug work struct cancel deadlocks with lockdep
- From: Johannes Berg <[email protected]>
- Re: [PATCH] debug work struct cancel deadlocks with lockdep
- From: Peter Zijlstra <[email protected]>
- Re: [PATCH] debug work struct cancel deadlocks with lockdep
- From: Johannes Berg <[email protected]>
- Re: [PATCH] debug work struct cancel deadlocks with lockdep
- From: Peter Zijlstra <[email protected]>
- [PATCH] debug work struct cancel deadlocks with lockdep
- Prev by Date: Re: [linux-pm] Re: [PATCH] Remove process freezer from suspend to RAM pathway
- Next by Date: Re: platform_driver_register() ??
- Previous by thread: Re: [PATCH] debug work struct cancel deadlocks with lockdep
- Next by thread: Re: [PATCH] debug work struct cancel deadlocks with lockdep
- Index(es):