Re: OOPS (kswapd) in 2.4.5 and 2.4.6

[Andrew Morton]

Henry wrote:
> 
> ...
> So far, so good.  There has not been a single oops on the two principle
> servers I patched.
> 
> uptime1:                8:04am  up 18:22,  1 user,  load average: 0.09, 0.15, 0.11
> uptime2:                8:04am  up 18:25,  1 user,  load average: 0.15, 0.20, 0.15

OK, that looks good.

> Andrew my china, you are the _MAN_!

Not only that - I have great legs!

>  We should know by monday afternoon
> (the monday morning/midday crunch should provide some valuable
> feedback).

I wonder why it only affects you.  Is the drive which holds
your swap partition running in PIO mode?  `hdparm' will tell
you.  If it is, then that could easily cause the page to come
unlocked before brw_page() has finished touching the buffer
ring.  Then all it takes is a parallel try_to_free_buffers
on the other CPU.

There's a similar bug in __block_write_full_page().  I'll
send a patch...

-