Jens Axboe wrote:
On Fri, Jul 08 2005, Andrew Morton wrote:
Jens Axboe <[EMAIL PROTECTED]> wrote:
Some more investigation - it appears to be broken read-ahead, actually.
hdparm does repeated read(), lseek() loops which causes the read-ahead
logic to mark the file as being in cache (sin
Andrew Morton wrote:
Nick Piggin <[EMAIL PROTECTED]> wrote:
AS basically does its own TCQ strangulation, which IIRC involves things
> > like completing all reads before issuing new writes, and completing all
> > reads from one process before reads from another. As well as the
> > fun
Andrew Morton wrote:
Nick Piggin <[EMAIL PROTECTED]> wrote:
- The effects of tcq on AS are much less disastrous than I thought they
> were. Do I have the wrong workload? Memory fails me. Or did we fix the
> anticipatory scheduler?
>
>
Yes, we did fix it ;)
Quite a long time ago, so m
Ram wrote:
On Wed, 2005-03-02 at 11:08, Oleg Nesterov wrote:
..snip...
@@ -527,7 +527,7 @@ page_cache_readahead(struct address_spac
}
out:
- return newsize;
+ return ra->prev_page + 1;
This change introduces one key behavioural change in
page_cache_readahead(). Instead of returning
Ram wrote:
Andrew,
I have verified the patches against my standard benchmarks
and did not see any bad effects.
Also I have reviewd the patch and it looked clean and correct.
RP
I have not had a chance to benchmark, but visual inspection looks good.
Steve
On Thu, 2005-02-24 at 11:37,
Oleg Nesterov wrote:
This patch introduces make_ahead_window() function for
simplification of page_cache_readahead.
Signed-off-by: Oleg Nesterov <[EMAIL PROTECTED]>
--- 2.6.11-rc2/mm/readahead.c~ 2005-01-27 22:14:39.0 +0300
+++ 2.6.11-rc2/mm/readahead.c 2005-01-29 15:51:04.0 +030
Ram wrote:
On Tue, 2005-01-25 at 03:59, Oleg Nesterov wrote:
Cases "no ahead window" and "crossed into ahead window"
can be unified.
No. There is a reason why we had some duplication. With your patch,
we will end up reading-on-demand instead of reading ahead.
When we notice a sequential re
No problem with this patch either. Again, not sure it buys much but it
should work fine.
Steve
Oleg Nesterov wrote:
I think that do_page_cache_readahead() can be inlined
in blockable_page_cache_readahead(), this makes the
code a bit more readable in my opinion.
Also makes check_ra_success() stat
Not sure how much better this is, but it doesn't hurt anything.
Steve
Oleg Nesterov wrote:
This patch moves some code into the get_next_ra_size()
and renames it into 'set_next_ahead_window'.
Signed-off-by: Oleg Nesterov <[EMAIL PROTECTED]>
--- 2.6.11-rc2/mm/readahead.c~ 2005-01-25 15:17:13.000
I like this one, especially getting rid of the large duplicate comment.
No functional difference.
Signed-off-by: Steven Pratt <[EMAIL PROTECTED]>
Oleg Nesterov wrote:
Cases "no ahead window" and "crossed into ahead window"
can be unified.
Signed-off-by: Oleg
Looks fine, thought we had some reason for it in the past, but it will
definitly be overwritten.
Signed-off-by: Steven Pratt <[EMAIL PROTECTED]>
Oleg Nesterov wrote:
There is no point in setting ra->prev_page before 'goto out',
it will be overwritten anyway.
Signed-off-by: Ol
Back in April there was some discussion about the race condition where a
call to zap_page_range followed by a call to flush_tlb_range allows for
a page which has been freed to be re-allocated on a different cpu and
referenced via a tlb on a third cpu before the tlb is actually flushed.
Below is a
Andrea Arcangeli wrote:
>
> On Mon, Oct 30, 2000 at 03:31:22PM -0600, Steve Pratt/Austin/IBM wrote:
> > [..] no patch ever
> > appeared. [..]
>
> You didn't followed l-k closely enough as the strict fix was submitted two
> times but it got not merged. (maybe because it had an #ifdef __s390__ tha
13 matches
Mail list logo
