Re: [RFT][PATCH] 2.3.10 pre5 SMP/vm fixes

[kanoj@google.engr.sgi.com (Kanoj Sarcar)]

>
>
>
> On Thu, 8 Jul 1999, Kanoj Sarcar wrote:
> >
> > *****************************************************************
> > 3. In mm/memory.c, a new comment claims:
> > "The adding of pages is protected by the MM semaphore"
> > which is not quite correct, since swapoff does not hold this semaphore.
>
> The fix is to fix swapoff, not to change the comment.
> 

Couldn't agree more ... that is why I sent a patch for this (ie to
fix swapoff) on linux-mm, and reiterated the pointer to the patch 
in this thread. Here it is again:

    http://humbolt.nl.linux.org/lists/linux-mm/1999-06/msg00075.html

Let me know if you want me to create it against pre5. Or if you 
see a problem with it ... we talked about it quite a bit on
linux-mm.


> page_table_lock is _not_ going to be added to this path - I refuse to add> locks to common cases just to protect against things that never run in
> practice.
>

I assume you are talking about the fault path here ... none of the 
patches I posted adds page_table_lock to that path, right? Of
course without the lock, there might be other problems, I am 
not sure yet, will let you know if I see something ...

Kanoj

>               Linus
>

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://humbolt.geo.uu.nl/Linux-MM/