Re: [PATCH] Re: today's futex changes

[Rusty Russell]

In message <Pine.LNX.4.44.0309081746180.7008-100000@xxxxxxxxxxxxxxxxxxxxx> you write:
> Most of it (the futex_wait tweaks) looked fine to me -
> though I look forward to the first report of that BUG().

Me too.  But at least we'll *get* a report if someone does spurious
wakeups.

> Part 2, requiring VM_WRITE and removing the comment on VM_MAYSHARE,
> seems a regression to me.  Perhaps I misinterpreted Linus' action in
> taking Jamie's patch: I took that to mean he relented a little on his
> hardline position about VM_SHARED, and now accepts that in this context
> VM_MAYSHARE is more appropriate (easier to document).  I know I argued
> that readonly futices are pointless, but I thought Jamie gave a good
> picture of how a readonly view could still be used.  I'd rather that
> part were a separate patch, so Linus can merge or not as he wishes.

Sure, I jumboed them together for feedback from you guys.

All users I am currently aware of won't care either way.  Current
test-5 is:

				Sees Changes		Sees FUTEX_WAKE
			   from another MAP_SHARED   from another MAP_SHARED

MAP_PRIVATE read-only:		Y			N
MAP_PRIVATE read-write:		Y*			N
MAP_SHARED read-only:		Y			Y
MAP_SHARED read-write:		Y			Y

[* Only until page is written to, after which COW splits them ]

Previously, the FUTEX_WAKE column was identical to the first column.
Now, IMHO, this new semantic is more sensible, but I don't really
mind.

But I don't recall anything about believable use of RO futexes: Jamie?

Cheers,
Rusty.
--
  Anyone who quotes me in their sig is an idiot. -- Rusty Russell.
-
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/