The dirty_ratio was silently limited to >= 5%. This is not a user
expected behavior. Let's rip it.
It's not likely the user space will depend on the old behavior.
So the risk of breaking user space is very low.
CC: Jan Kara <[email protected]>
CC: Neil Brown <[email protected]>
Signed-off-by: Wu Fengguang <[email protected]>
---
mm/page-writeback.c | 10 ++--------
1 file changed, 2 insertions(+), 8 deletions(-)
--- linux-next.orig/mm/page-writeback.c 2010-08-20 10:55:17.000000000 +0800
+++ linux-next/mm/page-writeback.c 2010-08-20 10:56:36.000000000 +0800
@@ -415,14 +415,8 @@ void global_dirty_limits(unsigned long *
if (vm_dirty_bytes)
dirty = DIV_ROUND_UP(vm_dirty_bytes, PAGE_SIZE);
- else {
- int dirty_ratio;
-
- dirty_ratio = vm_dirty_ratio;
- if (dirty_ratio < 5)
- dirty_ratio = 5;
- dirty = (dirty_ratio * available_memory) / 100;
- }
+ else
+ dirty = (vm_dirty_ratio * available_memory) / 100;
if (dirty_background_bytes)
background = DIV_ROUND_UP(dirty_background_bytes, PAGE_SIZE);
On 08/19/2010 11:25 PM, Wu Fengguang wrote:
> The dirty_ratio was silently limited to>= 5%. This is not a user
> expected behavior. Let's rip it.
>
> It's not likely the user space will depend on the old behavior.
> So the risk of breaking user space is very low.
>
> CC: Jan Kara<[email protected]>
> CC: Neil Brown<[email protected]>
> Signed-off-by: Wu Fengguang<[email protected]>
Acked-by: Rik van Riel <[email protected]>
--
All rights reversed
> The dirty_ratio was silently limited to >= 5%. This is not a user
> expected behavior. Let's rip it.
>
> It's not likely the user space will depend on the old behavior.
> So the risk of breaking user space is very low.
>
> CC: Jan Kara <[email protected]>
> CC: Neil Brown <[email protected]>
> Signed-off-by: Wu Fengguang <[email protected]>
Thank you.
Reviewed-by: KOSAKI Motohiro <[email protected]>
On Fri, 20 Aug 2010 02:13:25 pm KOSAKI Motohiro wrote:
> > The dirty_ratio was silently limited to >= 5%. This is not a user
> > expected behavior. Let's rip it.
> >
> > It's not likely the user space will depend on the old behavior.
> > So the risk of breaking user space is very low.
> >
> > CC: Jan Kara <[email protected]>
> > CC: Neil Brown <[email protected]>
> > Signed-off-by: Wu Fengguang <[email protected]>
>
> Thank you.
> Reviewed-by: KOSAKI Motohiro <[email protected]>
I have tried to do this in the past, and setting this value to 0 on some
machines caused the machine to come to a complete standstill with small
writes to disk. It seemed there was some kind of "minimum" amount of data
required by the VM before anything would make it to the disk and I never
quite found out where that blockade occurred. This was some time ago (3 years
ago) so I'm not sure if the problem has since been fixed in the VM since
then. I suggest you do some testing with this value set to zero before
approving this change.
Regards,
--
-ck
On Fri, Aug 20, 2010 at 01:50:54PM +0800, Con Kolivas wrote:
> On Fri, 20 Aug 2010 02:13:25 pm KOSAKI Motohiro wrote:
> > > The dirty_ratio was silently limited to >= 5%. This is not a user
> > > expected behavior. Let's rip it.
> > >
> > > It's not likely the user space will depend on the old behavior.
> > > So the risk of breaking user space is very low.
> > >
> > > CC: Jan Kara <[email protected]>
> > > CC: Neil Brown <[email protected]>
> > > Signed-off-by: Wu Fengguang <[email protected]>
> >
> > Thank you.
> > Reviewed-by: KOSAKI Motohiro <[email protected]>
>
> I have tried to do this in the past, and setting this value to 0 on some
> machines caused the machine to come to a complete standstill with small
> writes to disk. It seemed there was some kind of "minimum" amount of data
> required by the VM before anything would make it to the disk and I never
> quite found out where that blockade occurred. This was some time ago (3 years
> ago) so I'm not sure if the problem has since been fixed in the VM since
> then. I suggest you do some testing with this value set to zero before
> approving this change.
Good point. I'll do more homework. Thanks for the reminding!
Thanks,
Fengguang
On Fri, 20 Aug 2010 15:50:54 +1000
Con Kolivas <[email protected]> wrote:
> On Fri, 20 Aug 2010 02:13:25 pm KOSAKI Motohiro wrote:
> > > The dirty_ratio was silently limited to >= 5%. This is not a user
> > > expected behavior. Let's rip it.
> > >
> > > It's not likely the user space will depend on the old behavior.
> > > So the risk of breaking user space is very low.
> > >
> > > CC: Jan Kara <[email protected]>
> > > CC: Neil Brown <[email protected]>
> > > Signed-off-by: Wu Fengguang <[email protected]>
> >
> > Thank you.
> > Reviewed-by: KOSAKI Motohiro <[email protected]>
>
> I have tried to do this in the past, and setting this value to 0 on some
> machines caused the machine to come to a complete standstill with small
> writes to disk. It seemed there was some kind of "minimum" amount of data
> required by the VM before anything would make it to the disk and I never
> quite found out where that blockade occurred. This was some time ago (3 years
> ago) so I'm not sure if the problem has since been fixed in the VM since
> then. I suggest you do some testing with this value set to zero before
> approving this change.
>
If it is appropriate to have a lower limit, that should be imposed where
the sysctl is defined in kernel/sysctl.c, not imposed after the fact where
the value is used.
As we now have dirty_bytes which over-rides dirty_ratio, there is little
cost in having a lower_limit for dirty_ratio - it could even stay at 5% -
but it really shouldn't be silent. Writing a number below the limit to the
sysctl file should fail.
NeilBrown
On Mon, Aug 23, 2010 at 12:42:48PM +0800, Neil Brown wrote:
> On Fri, 20 Aug 2010 15:50:54 +1000
> Con Kolivas <[email protected]> wrote:
>
> > On Fri, 20 Aug 2010 02:13:25 pm KOSAKI Motohiro wrote:
> > > > The dirty_ratio was silently limited to >= 5%. This is not a user
> > > > expected behavior. Let's rip it.
> > > >
> > > > It's not likely the user space will depend on the old behavior.
> > > > So the risk of breaking user space is very low.
> > > >
> > > > CC: Jan Kara <[email protected]>
> > > > CC: Neil Brown <[email protected]>
> > > > Signed-off-by: Wu Fengguang <[email protected]>
> > >
> > > Thank you.
> > > Reviewed-by: KOSAKI Motohiro <[email protected]>
> >
> > I have tried to do this in the past, and setting this value to 0 on some
> > machines caused the machine to come to a complete standstill with small
> > writes to disk. It seemed there was some kind of "minimum" amount of data
> > required by the VM before anything would make it to the disk and I never
> > quite found out where that blockade occurred. This was some time ago (3 years
> > ago) so I'm not sure if the problem has since been fixed in the VM since
> > then. I suggest you do some testing with this value set to zero before
> > approving this change.
You are right, vm.dirty_ratio=0 will block applications for ever..
>
> If it is appropriate to have a lower limit, that should be imposed where
> the sysctl is defined in kernel/sysctl.c, not imposed after the fact where
> the value is used.
>
> As we now have dirty_bytes which over-rides dirty_ratio, there is little
> cost in having a lower_limit for dirty_ratio - it could even stay at 5% -
> but it really shouldn't be silent. Writing a number below the limit to the
> sysctl file should fail.
How about imposing an explicit bound of 1%? That's more natural and
its risk of breaking user space should be lower than 5%.
Thanks,
Fengguang
---
writeback: remove the internal 5% low bound on dirty_ratio
The dirty_ratio was silently limited in global_dirty_limits() to >= 5%. This
is not a user expected behavior. And it's inconsistent with calc_period_shift(),
which uses the plain vm_dirty_ratio value. So let's rip the internal bound.
At the same time, force a user visible low bound of 1% for the vm.dirty_ratio
interface. Applications trying to write 0 will be rejected with -EINVAL. This
will break user space applications if they
1) try to write 0 to vm.dirty_ratio
2) and check the return value
That is very weird combination, so the risk of breaking user space is low.
CC: Jan Kara <[email protected]>
CC: Neil Brown <[email protected]>
CC: Rik van Riel <[email protected]>
CC: Con Kolivas <[email protected]>
CC: Peter Zijlstra <[email protected]>
CC: KOSAKI Motohiro <[email protected]>
Signed-off-by: Wu Fengguang <[email protected]>
---
kernel/sysctl.c | 2 +-
mm/page-writeback.c | 10 ++--------
2 files changed, 3 insertions(+), 9 deletions(-)
--- linux-next.orig/mm/page-writeback.c 2010-08-20 20:14:11.000000000 +0800
+++ linux-next/mm/page-writeback.c 2010-08-23 10:31:01.000000000 +0800
@@ -415,14 +415,8 @@ void global_dirty_limits(unsigned long *
if (vm_dirty_bytes)
dirty = DIV_ROUND_UP(vm_dirty_bytes, PAGE_SIZE);
- else {
- int dirty_ratio;
-
- dirty_ratio = vm_dirty_ratio;
- if (dirty_ratio < 5)
- dirty_ratio = 5;
- dirty = (dirty_ratio * available_memory) / 100;
- }
+ else
+ dirty = (vm_dirty_ratio * available_memory) / 100;
if (dirty_background_bytes)
background = DIV_ROUND_UP(dirty_background_bytes, PAGE_SIZE);
--- linux-next.orig/kernel/sysctl.c 2010-08-23 14:06:11.000000000 +0800
+++ linux-next/kernel/sysctl.c 2010-08-23 14:07:30.000000000 +0800
@@ -1029,7 +1029,7 @@ static struct ctl_table vm_table[] = {
.maxlen = sizeof(vm_dirty_ratio),
.mode = 0644,
.proc_handler = dirty_ratio_handler,
- .extra1 = &zero,
+ .extra1 = &one,
.extra2 = &one_hundred,
},
{
On Mon, 23 Aug 2010 04:23:59 pm Wu Fengguang wrote:
> On Mon, Aug 23, 2010 at 12:42:48PM +0800, Neil Brown wrote:
> > On Fri, 20 Aug 2010 15:50:54 +1000
> >
> > Con Kolivas <[email protected]> wrote:
> > > On Fri, 20 Aug 2010 02:13:25 pm KOSAKI Motohiro wrote:
> > > > > The dirty_ratio was silently limited to >= 5%. This is not a user
> > > > > expected behavior. Let's rip it.
> > > > >
> > > > > It's not likely the user space will depend on the old behavior.
> > > > > So the risk of breaking user space is very low.
> > > > >
> > > > > CC: Jan Kara <[email protected]>
> > > > > CC: Neil Brown <[email protected]>
> > > > > Signed-off-by: Wu Fengguang <[email protected]>
> > > >
> > > > Thank you.
> > > > Reviewed-by: KOSAKI Motohiro <[email protected]>
> > >
> > > I have tried to do this in the past, and setting this value to 0 on
> > > some machines caused the machine to come to a complete standstill with
> > > small writes to disk. It seemed there was some kind of "minimum" amount
> > > of data required by the VM before anything would make it to the disk
> > > and I never quite found out where that blockade occurred. This was some
> > > time ago (3 years ago) so I'm not sure if the problem has since been
> > > fixed in the VM since then. I suggest you do some testing with this
> > > value set to zero before approving this change.
>
> You are right, vm.dirty_ratio=0 will block applications for ever..
Indeed. And while you shouldn't set the lower limit to zero to avoid this
problem, it doesn't answer _why_ this happens. What is this "minimum write"
that blocks everything, will 1% be enough, and is it hiding another real bug
somewhere in the VM?
Regards,
Con
--
-ck
On Mon, Aug 23, 2010 at 02:30:40PM +0800, Con Kolivas wrote:
> On Mon, 23 Aug 2010 04:23:59 pm Wu Fengguang wrote:
> > On Mon, Aug 23, 2010 at 12:42:48PM +0800, Neil Brown wrote:
> > > On Fri, 20 Aug 2010 15:50:54 +1000
> > >
> > > Con Kolivas <[email protected]> wrote:
> > > > On Fri, 20 Aug 2010 02:13:25 pm KOSAKI Motohiro wrote:
> > > > > > The dirty_ratio was silently limited to >= 5%. This is not a user
> > > > > > expected behavior. Let's rip it.
> > > > > >
> > > > > > It's not likely the user space will depend on the old behavior.
> > > > > > So the risk of breaking user space is very low.
> > > > > >
> > > > > > CC: Jan Kara <[email protected]>
> > > > > > CC: Neil Brown <[email protected]>
> > > > > > Signed-off-by: Wu Fengguang <[email protected]>
> > > > >
> > > > > Thank you.
> > > > > Reviewed-by: KOSAKI Motohiro <[email protected]>
> > > >
> > > > I have tried to do this in the past, and setting this value to 0 on
> > > > some machines caused the machine to come to a complete standstill with
> > > > small writes to disk. It seemed there was some kind of "minimum" amount
> > > > of data required by the VM before anything would make it to the disk
> > > > and I never quite found out where that blockade occurred. This was some
> > > > time ago (3 years ago) so I'm not sure if the problem has since been
> > > > fixed in the VM since then. I suggest you do some testing with this
> > > > value set to zero before approving this change.
> >
> > You are right, vm.dirty_ratio=0 will block applications for ever..
>
> Indeed. And while you shouldn't set the lower limit to zero to avoid this
> problem, it doesn't answer _why_ this happens. What is this "minimum write"
> that blocks everything, will 1% be enough, and is it hiding another real bug
> somewhere in the VM?
Good question.
This simple change will unblock the application even with vm_dirty_ratio=0.
# echo 0 > /proc/sys/vm/dirty_ratio
# echo 0 > /proc/sys/vm/dirty_background_ratio
# vmmon nr_dirty nr_writeback nr_unstable
nr_dirty nr_writeback nr_unstable
0 444 1369
37 37 326
0 0 37
74 772 694
0 0 19
0 0 1406
0 0 23
0 0 0
0 370 186
74 1073 1221
0 12 26
0 703 1147
37 0 999
37 37 1517
0 888 63
0 0 0
0 0 20
37 0 0
37 74 1776
0 0 8
37 629 333
0 12 19
Even with it, the 1% explicit bound still looks reasonable for me.
Who will want to set it to 0%? That would destroy IO inefficient.
Thanks,
Fengguang
---
--- a/mm/page-writeback.c
+++ b/mm/page-writeback.c
@@ -542,8 +536,8 @@ static void balance_dirty_pages(struct address_space *mapping,
* the last resort safeguard.
*/
dirty_exceeded =
- (bdi_nr_reclaimable + bdi_nr_writeback >= bdi_thresh)
- || (nr_reclaimable + nr_writeback >= dirty_thresh);
+ (bdi_nr_reclaimable + bdi_nr_writeback > bdi_thresh)
+ || (nr_reclaimable + nr_writeback > dirty_thresh);
if (!dirty_exceeded)
break;
> writeback: remove the internal 5% low bound on dirty_ratio
>
> The dirty_ratio was silently limited in global_dirty_limits() to >= 5%. This
> is not a user expected behavior. And it's inconsistent with calc_period_shift(),
> which uses the plain vm_dirty_ratio value. So let's rip the internal bound.
>
> At the same time, force a user visible low bound of 1% for the vm.dirty_ratio
> interface. Applications trying to write 0 will be rejected with -EINVAL. This
> will break user space applications if they
> 1) try to write 0 to vm.dirty_ratio
> 2) and check the return value
> That is very weird combination, so the risk of breaking user space is low.
I'm ok this one too. because I bet nobody use 0% dirty ratio on their production
server and/or their own desktop. (i.e. I don't mind lab machine crash)
Reviewed-by: KOSAKI Motohiro <[email protected]>
>
> CC: Jan Kara <[email protected]>
> CC: Neil Brown <[email protected]>
> CC: Rik van Riel <[email protected]>
> CC: Con Kolivas <[email protected]>
> CC: Peter Zijlstra <[email protected]>
> CC: KOSAKI Motohiro <[email protected]>
> Signed-off-by: Wu Fengguang <[email protected]>
> ---
> kernel/sysctl.c | 2 +-
> mm/page-writeback.c | 10 ++--------
> 2 files changed, 3 insertions(+), 9 deletions(-)
>
> --- linux-next.orig/mm/page-writeback.c 2010-08-20 20:14:11.000000000 +0800
> +++ linux-next/mm/page-writeback.c 2010-08-23 10:31:01.000000000 +0800
> @@ -415,14 +415,8 @@ void global_dirty_limits(unsigned long *
>
> if (vm_dirty_bytes)
> dirty = DIV_ROUND_UP(vm_dirty_bytes, PAGE_SIZE);
> - else {
> - int dirty_ratio;
> -
> - dirty_ratio = vm_dirty_ratio;
> - if (dirty_ratio < 5)
> - dirty_ratio = 5;
> - dirty = (dirty_ratio * available_memory) / 100;
> - }
> + else
> + dirty = (vm_dirty_ratio * available_memory) / 100;
>
> if (dirty_background_bytes)
> background = DIV_ROUND_UP(dirty_background_bytes, PAGE_SIZE);
> --- linux-next.orig/kernel/sysctl.c 2010-08-23 14:06:11.000000000 +0800
> +++ linux-next/kernel/sysctl.c 2010-08-23 14:07:30.000000000 +0800
> @@ -1029,7 +1029,7 @@ static struct ctl_table vm_table[] = {
> .maxlen = sizeof(vm_dirty_ratio),
> .mode = 0644,
> .proc_handler = dirty_ratio_handler,
> - .extra1 = &zero,
> + .extra1 = &one,
> .extra2 = &one_hundred,
> },
> {
On Wed, Aug 25, 2010 at 04:40:00PM +0800, Con Kolivas wrote:
> On Tue, 24 Aug 2010 05:14:40 pm you wrote:
> > On Tue, Aug 24, 2010 at 02:20:54PM +0800, Con Kolivas wrote:
> > > On Mon, 23 Aug 2010 05:15:35 pm you wrote:
> > > > On Mon, Aug 23, 2010 at 02:30:40PM +0800, Con Kolivas wrote:
> > > > > On Mon, 23 Aug 2010 04:23:59 pm Wu Fengguang wrote:
> > > > > > On Mon, Aug 23, 2010 at 12:42:48PM +0800, Neil Brown wrote:
> > > > > > > On Fri, 20 Aug 2010 15:50:54 +1000
> > > > > > >
> > > > > > > Con Kolivas <[email protected]> wrote:
> > > > > > > > On Fri, 20 Aug 2010 02:13:25 pm KOSAKI Motohiro wrote:
> > > > > > > > > > The dirty_ratio was silently limited to >= 5%. This is not
> > > > > > > > > > a user expected behavior. Let's rip it.
> > > > > > > > > >
> > > > > > > > > > It's not likely the user space will depend on the old
> > > > > > > > > > behavior. So the risk of breaking user space is very low.
> > > > > > > > > >
> > > > > > > > > > CC: Jan Kara <[email protected]>
> > > > > > > > > > CC: Neil Brown <[email protected]>
> > > > > > > > > > Signed-off-by: Wu Fengguang <[email protected]>
> > > > > > > > >
> > > > > > > > > Thank you.
> > > > > > > > > Reviewed-by: KOSAKI Motohiro
> > > > > > > > > <[email protected]>
> > > > > > > >
> > > > > > > > I have tried to do this in the past, and setting this value to
> > > > > > > > 0 on some machines caused the machine to come to a complete
> > > > > > > > standstill with small writes to disk. It seemed there was some
> > > > > > > > kind of "minimum" amount of data required by the VM before
> > > > > > > > anything would make it to the disk and I never quite found out
> > > > > > > > where that blockade occurred. This was some time ago (3 years
> > > > > > > > ago) so I'm not sure if the problem has since been fixed in the
> > > > > > > > VM since then. I suggest you do some testing with this value
> > > > > > > > set to zero before approving this change.
> > > > > >
> > > > > > You are right, vm.dirty_ratio=0 will block applications for ever..
> > > > >
> > > > > Indeed. And while you shouldn't set the lower limit to zero to avoid
> > > > > this problem, it doesn't answer _why_ this happens. What is this
> > > > > "minimum write" that blocks everything, will 1% be enough, and is it
> > > > > hiding another real bug somewhere in the VM?
> > > >
> > > > Good question.
> > > > This simple change will unblock the application even with
> > > > vm_dirty_ratio=0.
> > > >
> > > > # echo 0 > /proc/sys/vm/dirty_ratio
> > > > # echo 0 > /proc/sys/vm/dirty_background_ratio
> > > > # vmmon nr_dirty nr_writeback nr_unstable
> > > >
> > > > nr_dirty nr_writeback nr_unstable
> > > > 0 444 1369
> > > > 37 37 326
> > > > 0 0 37
> > > > 74 772 694
> > > > 0 0 19
> > > > 0 0 1406
> > > > 0 0 23
> > > > 0 0 0
> > > > 0 370 186
> > > > 74 1073 1221
> > > > 0 12 26
> > > > 0 703 1147
> > > > 37 0 999
> > > > 37 37 1517
> > > > 0 888 63
> > > > 0 0 0
> > > > 0 0 20
> > > > 37 0 0
> > > > 37 74 1776
> > > > 0 0 8
> > > > 37 629 333
> > > > 0 12 19
> > > >
> > > > Even with it, the 1% explicit bound still looks reasonable for me.
> > > > Who will want to set it to 0%? That would destroy IO inefficient.
> > >
> > > Thanks for your work in this area. I'll experiment with these later.
> > > There are low latency applications that would benefit with it set to
> > > zero.
> >
> > It might be useful to some users. Shall we give the rope to users, heh?
> >
> > Note that for these applications, they may well use
> > /proc/sys/vm/dirty_bytes for more fine grained control. That interface only
> > imposes a low limit of 2 pages.
>
> I don't see why there needs to be a limit. Users fiddling with sysctls should
> know what they're messing with, and there may well be a valid use out there
> somewhere for it.
OK, the following patch gives users the full freedom. I tested 1
single dirtier and 9 parallel dirtiers, the system remains alive, but
with much slower IO throughput. Maybe not all users care IO performance
in all situations?
Thanks,
Fengguang
---
writeback: remove the internal 5% low bound on dirty_ratio
The dirty_ratio was silently limited in global_dirty_limits() to >= 5%.
This is not a user expected behavior. And it's inconsistent with
calc_period_shift(), which uses the plain vm_dirty_ratio value.
Let's rip the internal bound.
At the same time, fix balance_dirty_pages() to work with the
dirty_thresh=0 case. This allows applications to proceed when
dirty+writeback pages are all cleaned.
CC: Jan Kara <[email protected]>
CC: Neil Brown <[email protected]>
CC: Con Kolivas <[email protected]>
CC: Rik van Riel <[email protected]>
CC: Peter Zijlstra <[email protected]>
CC: KOSAKI Motohiro <[email protected]>
Signed-off-by: Wu Fengguang <[email protected]>
---
mm/page-writeback.c | 14 ++++----------
1 file changed, 4 insertions(+), 10 deletions(-)
--- linux-next.orig/mm/page-writeback.c 2010-08-26 08:37:31.000000000 +0800
+++ linux-next/mm/page-writeback.c 2010-08-26 08:37:55.000000000 +0800
@@ -415,14 +415,8 @@ void global_dirty_limits(unsigned long *
if (vm_dirty_bytes)
dirty = DIV_ROUND_UP(vm_dirty_bytes, PAGE_SIZE);
- else {
- int dirty_ratio;
-
- dirty_ratio = vm_dirty_ratio;
- if (dirty_ratio < 5)
- dirty_ratio = 5;
- dirty = (dirty_ratio * available_memory) / 100;
- }
+ else
+ dirty = (vm_dirty_ratio * available_memory) / 100;
if (dirty_background_bytes)
background = DIV_ROUND_UP(dirty_background_bytes, PAGE_SIZE);
@@ -542,8 +536,8 @@ static void balance_dirty_pages(struct a
* the last resort safeguard.
*/
dirty_exceeded =
- (bdi_nr_reclaimable + bdi_nr_writeback >= bdi_thresh)
- || (nr_reclaimable + nr_writeback >= dirty_thresh);
+ (bdi_nr_reclaimable + bdi_nr_writeback > bdi_thresh)
+ || (nr_reclaimable + nr_writeback > dirty_thresh);
if (!dirty_exceeded)
break;
On Thu, 26 Aug 2010 09:29:45 +0800
Wu Fengguang <[email protected]> wrote:
> On Wed, Aug 25, 2010 at 04:40:00PM +0800, Con Kolivas wrote:
> > On Tue, 24 Aug 2010 05:14:40 pm you wrote:
> > > On Tue, Aug 24, 2010 at 02:20:54PM +0800, Con Kolivas wrote:
> > > > On Mon, 23 Aug 2010 05:15:35 pm you wrote:
> > > > > On Mon, Aug 23, 2010 at 02:30:40PM +0800, Con Kolivas wrote:
> > > > > > On Mon, 23 Aug 2010 04:23:59 pm Wu Fengguang wrote:
> > > > > > > On Mon, Aug 23, 2010 at 12:42:48PM +0800, Neil Brown wrote:
> > > > > > > > On Fri, 20 Aug 2010 15:50:54 +1000
> > > > > > > >
> > > > > > > > Con Kolivas <[email protected]> wrote:
> > > > > > > > > On Fri, 20 Aug 2010 02:13:25 pm KOSAKI Motohiro wrote:
> > > > > > > > > > > The dirty_ratio was silently limited to >= 5%. This is not
> > > > > > > > > > > a user expected behavior. Let's rip it.
> > > > > > > > > > >
> > > > > > > > > > > It's not likely the user space will depend on the old
> > > > > > > > > > > behavior. So the risk of breaking user space is very low.
> > > > > > > > > > >
> > > > > > > > > > > CC: Jan Kara <[email protected]>
> > > > > > > > > > > CC: Neil Brown <[email protected]>
> > > > > > > > > > > Signed-off-by: Wu Fengguang <[email protected]>
> > > > > > > > > >
> > > > > > > > > > Thank you.
> > > > > > > > > > Reviewed-by: KOSAKI Motohiro
> > > > > > > > > > <[email protected]>
> > > > > > > > >
> > > > > > > > > I have tried to do this in the past, and setting this value to
> > > > > > > > > 0 on some machines caused the machine to come to a complete
> > > > > > > > > standstill with small writes to disk. It seemed there was some
> > > > > > > > > kind of "minimum" amount of data required by the VM before
> > > > > > > > > anything would make it to the disk and I never quite found out
> > > > > > > > > where that blockade occurred. This was some time ago (3 years
> > > > > > > > > ago) so I'm not sure if the problem has since been fixed in the
> > > > > > > > > VM since then. I suggest you do some testing with this value
> > > > > > > > > set to zero before approving this change.
> > > > > > >
> > > > > > > You are right, vm.dirty_ratio=0 will block applications for ever..
> > > > > >
> > > > > > Indeed. And while you shouldn't set the lower limit to zero to avoid
> > > > > > this problem, it doesn't answer _why_ this happens. What is this
> > > > > > "minimum write" that blocks everything, will 1% be enough, and is it
> > > > > > hiding another real bug somewhere in the VM?
> > > > >
> > > > > Good question.
> > > > > This simple change will unblock the application even with
> > > > > vm_dirty_ratio=0.
> > > > >
> > > > > # echo 0 > /proc/sys/vm/dirty_ratio
> > > > > # echo 0 > /proc/sys/vm/dirty_background_ratio
> > > > > # vmmon nr_dirty nr_writeback nr_unstable
> > > > >
> > > > > nr_dirty nr_writeback nr_unstable
> > > > > 0 444 1369
> > > > > 37 37 326
> > > > > 0 0 37
> > > > > 74 772 694
> > > > > 0 0 19
> > > > > 0 0 1406
> > > > > 0 0 23
> > > > > 0 0 0
> > > > > 0 370 186
> > > > > 74 1073 1221
> > > > > 0 12 26
> > > > > 0 703 1147
> > > > > 37 0 999
> > > > > 37 37 1517
> > > > > 0 888 63
> > > > > 0 0 0
> > > > > 0 0 20
> > > > > 37 0 0
> > > > > 37 74 1776
> > > > > 0 0 8
> > > > > 37 629 333
> > > > > 0 12 19
> > > > >
> > > > > Even with it, the 1% explicit bound still looks reasonable for me.
> > > > > Who will want to set it to 0%? That would destroy IO inefficient.
> > > >
> > > > Thanks for your work in this area. I'll experiment with these later.
> > > > There are low latency applications that would benefit with it set to
> > > > zero.
> > >
> > > It might be useful to some users. Shall we give the rope to users, heh?
> > >
> > > Note that for these applications, they may well use
> > > /proc/sys/vm/dirty_bytes for more fine grained control. That interface only
> > > imposes a low limit of 2 pages.
> >
> > I don't see why there needs to be a limit. Users fiddling with sysctls should
> > know what they're messing with, and there may well be a valid use out there
> > somewhere for it.
>
> OK, the following patch gives users the full freedom. I tested 1
> single dirtier and 9 parallel dirtiers, the system remains alive, but
> with much slower IO throughput. Maybe not all users care IO performance
> in all situations?
>
> Thanks,
> Fengguang
> ---
> writeback: remove the internal 5% low bound on dirty_ratio
>
> The dirty_ratio was silently limited in global_dirty_limits() to >= 5%.
> This is not a user expected behavior. And it's inconsistent with
> calc_period_shift(), which uses the plain vm_dirty_ratio value.
>
> Let's rip the internal bound.
>
> At the same time, fix balance_dirty_pages() to work with the
> dirty_thresh=0 case. This allows applications to proceed when
> dirty+writeback pages are all cleaned.
And ">" fits with the name "exceeded" better than ">=" does. I think it is
an aesthetic improvement as well as a functional one.
Reviewed-by: NeilBrown <[email protected]>
Thanks,
NeilBrown
>
> CC: Jan Kara <[email protected]>
> CC: Neil Brown <[email protected]>
> CC: Con Kolivas <[email protected]>
> CC: Rik van Riel <[email protected]>
> CC: Peter Zijlstra <[email protected]>
> CC: KOSAKI Motohiro <[email protected]>
> Signed-off-by: Wu Fengguang <[email protected]>
> ---
> mm/page-writeback.c | 14 ++++----------
> 1 file changed, 4 insertions(+), 10 deletions(-)
>
> --- linux-next.orig/mm/page-writeback.c 2010-08-26 08:37:31.000000000 +0800
> +++ linux-next/mm/page-writeback.c 2010-08-26 08:37:55.000000000 +0800
> @@ -415,14 +415,8 @@ void global_dirty_limits(unsigned long *
>
> if (vm_dirty_bytes)
> dirty = DIV_ROUND_UP(vm_dirty_bytes, PAGE_SIZE);
> - else {
> - int dirty_ratio;
> -
> - dirty_ratio = vm_dirty_ratio;
> - if (dirty_ratio < 5)
> - dirty_ratio = 5;
> - dirty = (dirty_ratio * available_memory) / 100;
> - }
> + else
> + dirty = (vm_dirty_ratio * available_memory) / 100;
>
> if (dirty_background_bytes)
> background = DIV_ROUND_UP(dirty_background_bytes, PAGE_SIZE);
> @@ -542,8 +536,8 @@ static void balance_dirty_pages(struct a
> * the last resort safeguard.
> */
> dirty_exceeded =
> - (bdi_nr_reclaimable + bdi_nr_writeback >= bdi_thresh)
> - || (nr_reclaimable + nr_writeback >= dirty_thresh);
> + (bdi_nr_reclaimable + bdi_nr_writeback > bdi_thresh)
> + || (nr_reclaimable + nr_writeback > dirty_thresh);
>
> if (!dirty_exceeded)
> break;
> > writeback: remove the internal 5% low bound on dirty_ratio
> >
> > The dirty_ratio was silently limited in global_dirty_limits() to >= 5%.
> > This is not a user expected behavior. And it's inconsistent with
> > calc_period_shift(), which uses the plain vm_dirty_ratio value.
> >
> > Let's rip the internal bound.
> >
> > At the same time, fix balance_dirty_pages() to work with the
> > dirty_thresh=0 case. This allows applications to proceed when
> > dirty+writeback pages are all cleaned.
>
> And ">" fits with the name "exceeded" better than ">=" does. I think it is
> an aesthetic improvement as well as a functional one.
>
> Reviewed-by: NeilBrown <[email protected]>
I agree :)
Reviewed-by: KOSAKI Motohiro <[email protected]>