Re: [patch 0/6] lightweight robust futexes: -V3

[Ingo Molnar <mingo@elte.hu>]

* Paul Jackson <pj@sgi.com> wrote:

> [ ... nice writeup of the robust-futex ABI ... ]

can i put this into Documentation/robust-futex-ABI.txt?

> Other details ...
> 
> Nit ...
> 
>     +If a futex is found to be held at exit time, the kernel sets the highest
>     +bit of the futex word:
>     +
>     +	#define FUTEX_OWNER_DIED        0x40000000
> 
>     Contrary to the comment, that doesn't look like the "highest bit."

ok, i fixed this in the text.

> Confusion ...
> 
>     +The list is guaranteed to be private and per-thread, so it's lockless.
> 
>     This statement seems like it is stretching the truth a bit.
>     As best as I can tell, the 'head' is private per-thread, but the
>     elements on the list are shared by all contending threads, and so
>     adding and removing these elements from a given threads list requires
>     some sort of contention handling mechanism, which the code provides.

well, from the kernel's perspective, the list _as it exists_ is private 
and per-thread, so it can be accessed in a lockless way.

from the userspace perspective you are right, it's only private if the 
list entry is manipulated after acquiring the lock.

i fixed up the text to reflect this.

	Ingo