On Fri, 25 Mar 2005, Pavel Machek wrote:
> > +unsigned int sysctl_scrub_start = 100; /* Min percentage of zeroed
> > free pages per zone (~10% default) */
> > +unsigned int sysctl_scrub_stop = 300; /* Max percentage of zeroed
> > free pages per zone (~30% default) */
> > +unsigned int s
Hi!
> Changelog:
> - Drop clear_pages and the approach to zero pages of higher order
> first
> - Zero a percentage of pages from all orders to avoid fragmentation
>
> Adds management of ZEROED and NOT_ZEROED pages and a background daemon
> called scrubd. /proc/sys/vm/scrubd_load, /proc/sys/vm_s
On Thu, 17 Mar 2005 18:09:11 -0800 (PST), Christoph Lameter
<[EMAIL PROTECTED]> wrote:
> On Thu, 17 Mar 2005, Jason Uhlenkott wrote:
>
> > On Thu, Mar 17, 2005 at 05:36:50PM -0800, Christoph Lameter wrote:
> > > +while (avenrun[0] >= ((unsigned long)sysctl_scrub_load <<
> > > FSHIFT)) {
>
On Thu, 2005-03-17 at 15:06 -0800, Christoph Lameter wrote:
> I want to sleep 30 seconds because the system load is unlikely to change
> frequently.
Ugh ? That sounds like a magic number coming right from your hat or from
your test scenario ...
Ben.
-
To unsubscribe from this list: send the lin
On Thu, 17 Mar 2005, Jason Uhlenkott wrote:
> On Thu, Mar 17, 2005 at 05:36:50PM -0800, Christoph Lameter wrote:
> > +while (avenrun[0] >= ((unsigned long)sysctl_scrub_load << FSHIFT))
> > {
> > + set_current_state(TASK_UNINTERRUPTIBLE);
> > + schedule_timeout(30*HZ);
On Thu, Mar 17, 2005 at 05:36:50PM -0800, Christoph Lameter wrote:
> +while (avenrun[0] >= ((unsigned long)sysctl_scrub_load << FSHIFT)) {
> + set_current_state(TASK_UNINTERRUPTIBLE);
> + schedule_timeout(30*HZ);
> + }
This should probably be TASK_INTERRUPTIBLE.
Here is the fixed up zeroing patch with management of hot/cold zeroed
pages.
If quicklists would like the use this then they need to use
free_hot_zeroed_page(page)
and
get_zeroed_page(GFP)
for their management of hot zeroed pages. If the pool is empty then it
will be replenished either from th
On Thu, 17 Mar 2005, Andrew Morton wrote:
> > I switched off the page-zeroing hardware for the tests.
>
> What tests?
For the results on the darn URL.
> See, a speedup in a simple malloc+memset could be due to either a simple
> transfer of load from user to kscrubd, or it could be due to leverag
Christoph Lameter <[EMAIL PROTECTED]> wrote:
>
> On Thu, 17 Mar 2005, Andrew Morton wrote:
>
> > > http://oss.sgi.com/projects/page_fault_performance/
> >
> > Oh no, not that page again ;)
>
> Yes indeed!
>
> > Seems to say that prezeroing makes negligible difference to kernel builds,
> > but s
On Thu, 17 Mar 2005, Andrew Morton wrote:
> > http://oss.sgi.com/projects/page_fault_performance/
>
> Oh no, not that page again ;)
Yes indeed!
> Seems to say that prezeroing makes negligible difference to kernel builds,
> but speeds up a big malloc+memset by 3x to 4x, yes?
Correct.
> Are the
On Thu, 17 Mar 2005, Andrew Morton wrote:
> Christoph Lameter <[EMAIL PROTECTED]> wrote:
> >
> > > And given that we have separate buddy structures for zeroed and not-zeroed
> > > pages, why is this tagging needed at all?
> >
> > Because the buddy pointers may point to a page of the different ki
Christoph Lameter <[EMAIL PROTECTED]> wrote:
>
> On Thu, 17 Mar 2005, Andrew Morton wrote:
>
> > > > It's hard to know what to think about this without benchmarking
> numbers.
>
> http://oss.sgi.com/projects/page_fault_performance/
Oh no, not that page again ;)
Seems to say that prezeroing m
Christoph Lameter <[EMAIL PROTECTED]> wrote:
>
> > And given that we have separate buddy structures for zeroed and not-zeroed
> > pages, why is this tagging needed at all?
>
> Because the buddy pointers may point to a page of the different kind. Then
> a merge is not possible.
In that case I s
On Thu, 17 Mar 2005, Andrew Morton wrote:
> > > It's hard to know what to think about this without benchmarking numbers.
http://oss.sgi.com/projects/page_fault_performance/
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
Mor
On Thu, 17 Mar 2005, Andrew Morton wrote:
> OK, so we're splitting each zone's buddy structure into two: one for zeroed
> pages and one for not-zeroed pages, yes?
Right.
> It's not obvious what the page->private of freed pages are being used for.
> Please comment that.
Ok.
> What's all this (z
Christoph Lameter <[EMAIL PROTECTED]> wrote:
>
> On Thu, 17 Mar 2005, Andrew Morton wrote:
>
> > Christoph Lameter <[EMAIL PROTECTED]> wrote:
> > >
> > > Adds management of ZEROED and NOT_ZEROED pages and a background daemon
> > > called scrubd. /proc/sys/vm/scrubd_load, /proc/sys/vm_scrubd_start
On Thu, 17 Mar 2005, Nish Aravamudan wrote:
> > + if (system_state != SYSTEM_RUNNING)
> > + return;
> > +
> > +while (avenrun[0] >= ((unsigned long)sysctl_scrub_load << FSHIFT))
> > + schedule_timeout(30*HZ);
>
> This is a busy-loop, unless you set the sta
On Thu, 17 Mar 2005 13:43:47 -0800 (PST), Christoph Lameter
<[EMAIL PROTECTED]> wrote:
> Changelog:
> - Drop clear_pages and the approach to zero pages of higher order
> first
> - Zero a percentage of pages from all orders to avoid fragmentation
>
> Adds management of ZEROED and NOT_ZEROED pages
On Thu, 17 Mar 2005, Andrew Morton wrote:
> Christoph Lameter <[EMAIL PROTECTED]> wrote:
> >
> > Adds management of ZEROED and NOT_ZEROED pages and a background daemon
> > called scrubd. /proc/sys/vm/scrubd_load, /proc/sys/vm_scrubd_start and
> > /proc/sys/vm_scrubd_stop control the scrub daemon.
Christoph Lameter <[EMAIL PROTECTED]> wrote:
>
> Adds management of ZEROED and NOT_ZEROED pages and a background daemon
> called scrubd. /proc/sys/vm/scrubd_load, /proc/sys/vm_scrubd_start and
> /proc/sys/vm_scrubd_stop control the scrub daemon. See Documentation/vm/
> scrubd.txt
It's hard to know
20 matches
Mail list logo
