> I also worked on that part, although I have to admitt that I did not
> got that part of the manpage. Instead I invented the require_context
> attribute.
Heh.
> Here is the second patch. It applies on top of the first one...
I don't really have time to look through it right now, sorry.
> I hope that we get the good ideas of our two works combined and
> accepted into sparse...
That would be good :)
> +# define __cond_lock(x,c) ((c) ? ({ __acquire(x); 1; }) : 0)
> +# define __protected_by(x) __attribute__((require_context(x,1,1,"rdwr")))
> +# define __protected_read_by(x) __attribute__((require_context(x,1,1,"read")))
> +# define __protected_write_by(x) __attribute__((require_context(x,1,1,"write")))
> +# define __must_hold(x) __attribute__((context(x,1,1), require_context(x,1,1,"call")))
What's this "rdwr" etc. for? And "call"? Also, how are you planning to
handle nested contexts, where 1,1 doesn't cut it any >0,>0 is really
more like what we need?
johannes
Attachment:
signature.asc
Description: This is a digitally signed message part
