[lkml]   [2000]   [Nov]   [10]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: Oddness in i_shared_lock and page_table_lock nesting hierarchies ?

>Let me save you some time, below is the fix I sent to
>Linus this evening:

Yes it does :-) Thanks.

I had started making changes along similar lines, and then as
I realized all the places it'd affect, I got a little diverted trying to
see if there was a way to do it without having to bring the locks
out of the common insert routines; whether it was possible to
avoid having to hold i_shared_lock and page_table_lock
simultaneously (other than in vmtruncate of course), through a
careful sequencing of steps, since we really need to serialize
only with truncate and the page stealer (as the mm sem takes
care of most other cases). Unmap does that now as it crosses
multiple objects (unlike the other cases where luckily we just
have one object at a time, which makes it possible to bring
the i_shared_lock to the top).

But it gets kind of complicated and harder to verify correctness
or generalize such an approach.
So your fix looks like a natural and consistent way to do this.

Now I can get back to what I was originally working on when I
hit this :-)

- Suparna

Suparna Bhattacharya
Systems Software Group, IBM Global Services, India
E-mail :
Phone : 91-80-5267117, Extn : 2525

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
Please read the FAQ at

 \ /
  Last update: 2005-03-22 12:45    [W:0.023 / U:14.088 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site