[lkml]   [2001]   [May]   [7]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: page_launder() bug

On Mon, 7 May 2001, David S. Miller wrote:
> Here, let's talk code a little bit so there are no misunderstandings,
> I really want to put this to rest:
> Calculate dead_swap_page outside of lock.

NO. That's not what you're doing at all. You're calculating something
completely different that "dead swap page". You're calculating "do we have
a swap cache entry that is not mapped into any virtual memory"?

> If dead_swap_page, ignore referenced bit heuristics.

Which is complete crap. Those reference bits are valid and important
data. You have not computed anything that says otherwise. You have
computed a random number that doesn't tell you anything about whether the
page is dead or not.

> Really, what does this have to do with swap counts and page counts?
> It's a heuristic. In fact it even seems stupid to me to recalculate
> dead_swap_page after we get the lock just for the sake of these
> heuristics.


> Maybe I should have diguised this bit as:
> if (dead_swap_page)
> do_writepage_first_pass = 1;

So tell me: what does the above help?

I repeat: your "dead_swap_page" variable is a random number with
absolutely no meaning. ANYTHING that uses it is buggy. It doesn't help in
the least if you use the first random state to set another random
state: the amount of randomness does not increase or decrease.


> To divert people's brains to what the intent was :-)

I can see the intent.

I can also see that the code doesn't match up to the intent.

I call that a bug. You don't. Fine.

But that code isn't coming anywhere _close_ to my tree until the two
match. And I stand by my assertion that it should be reverted from Alans
tree too.


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

 \ /
  Last update: 2005-03-22 12:52    [W:1.993 / U:0.228 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site