Re: [patch 2.6.13-rc4] fix get_user_pages bug

[Nick Piggin <nickpiggin@yahoo.com.au>]

Hugh Dickins wrote:

> Stupidity was the reason I thought handle_mm_fault couldn't be inline:
> I was picturing it static inline within mm/memory.c, failed to make the
> great intellectual leap you've achieved by moving it to include/linux/mm.h.
> 

Well it was one of my finer moments, so don't be too hard on
yourself.

> 
> No, I don't think it would break anything: it's just an historic oddity,
> used to be -1 for failure, and only got given a name recently, I think
> when wli added the proper major/minor counting.
> 
> Your version of the patch looks less hacky to me (not requiring
> VM_FAULT_WRITE_EXPECTED arg), though we could perfectly well remove
> that at leisure by adding VM_FAULT_WRITE case into all the arches in
> 2.6.14 (which might be preferable to leaving the __inline obscurity?).
> 

Well depends on what they want I suppose. Does it even make sense
to expose VM_FAULT_WRITE to arch code if it will just fall through
to VM_FAULT_MINOR?

With my earlier VM_FAULT_RACE thing, you can squint and say that's
a different case to VM_FAULT_MINOR - and accordingly not increment
the minor fault count. Afterall, minor faults *seem* to track count
of modifications made to the pte entry minus major faults, rather
than the number of times the hardware traps. I say this because we
increment the number in get_user_pages too.

But...

> I don't mind either way, but since you've not yet found an actual
> error in mine, I'd prefer you to make yours a tidyup patch on top,
> Signed-off-by your own good self, and let Linus decide whether he
> wants to apply yours on top or not.  Or perhaps the decision rests
> for the moment with Robin, whether he gets his customer to test
> yours or mine - whichever is tested is the one which should go in.
> 

I agree that for the moment we probably just want something that
works. I'd be just as happy to go with your patch as is for now.

Nick

-- 
SUSE Labs, Novell Inc.

Send instant messages to your online friends http://au.messenger.yahoo.com 
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>