mbox series

[v16,0/9] mm / virtio: Provide support for free page reporting

Message ID 20200103210509.29237.18426.stgit@localhost.localdomain (mailing list archive)
Headers show
Series mm / virtio: Provide support for free page reporting | expand

Message

Alexander Duyck Jan. 3, 2020, 9:16 p.m. UTC
This series provides an asynchronous means of reporting free guest pages
to a hypervisor so that the memory associated with those pages can be
dropped and reused by other processes and/or guests on the host. Using
this it is possible to avoid unnecessary I/O to disk and greatly improve
performance in the case of memory overcommit on the host.

When enabled we will be performing a scan of free memory every 2 seconds
while pages of sufficiently high order are being freed. In each pass at
least one sixteenth of each free list will be reported. By doing this we
avoid racing against other threads that may be causing a high amount of
memory churn.

The lowest page order currently scanned when reporting pages is
pageblock_order so that this feature will not interfere with the use of
Transparent Huge Pages in the case of virtualization.

Currently this is only in use by virtio-balloon however there is the hope
that at some point in the future other hypervisors might be able to make
use of it. In the virtio-balloon/QEMU implementation the hypervisor is
currently using MADV_DONTNEED to indicate to the host kernel that the page
is currently free. It will be zeroed and faulted back into the guest the
next time the page is accessed.

To track if a page is reported or not the Uptodate flag was repurposed and
used as a Reported flag for Buddy pages. We walk though the free list
isolating pages and adding them to the scatterlist until we either
encounter the end of the list, processed as many pages as were listed in
nr_free prior to us starting, or have filled the scatterlist with pages to
be reported. If we fill the scatterlist before we reach the end of the
list we rotate the list so that the first unreported page we encounter is
moved to the head of the list as that is where we will resume after we
have freed the reported pages back into the tail of the list.

Below are the results from various benchmarks. I primarily focused on two
tests. The first is the will-it-scale/page_fault2 test, and the other is
a modified version of will-it-scale/page_fault1 that was enabled to use
THP. I did this as it allows for better visibility into different parts
of the memory subsystem. The guest is running with 32G for RAM on one
node of a E5-2630 v3. The host has had some features such as CPU turbo
disabled in the BIOS.

Test                   page_fault1 (THP)    page_fault2
Name            tasks  Process Iter  STDEV  Process Iter  STDEV
Baseline            1    1012402.50  0.14%     361855.25  0.81%
                   16    8827457.25  0.09%    3282347.00  0.34%

Patches Applied     1    1007897.00  0.23%     361887.00  0.26%
                   16    8784741.75  0.39%    3240669.25  0.48%

Patches Enabled     1    1010227.50  0.39%     359749.25  0.56%
                   16    8756219.00  0.24%    3226608.75  0.97%

Patches Enabled     1    1050982.00  4.26%     357966.25  0.14%
 page shuffle      16    8672601.25  0.49%    3223177.75  0.40%

Patches Enabled     1    1003238.00  0.22%     360211.00  0.22%
 shuffle w/ RFC    16    8767010.50  0.32%    3199874.00  0.71%

The results above are for a baseline with a linux-next-20191219 kernel,
that kernel with this patch set applied but page reporting disabled in
virtio-balloon, the patches applied and page reporting fully enabled, the
patches enabled with page shuffling enabled, and the patches applied with
page shuffling enabled and an RFC patch that makes used of MADV_FREE in
QEMU. These results include the deviation seen between the average value
reported here versus the high and/or low value. I observed that during the
test memory usage for the first three tests never dropped whereas with the
patches fully enabled the VM would drop to using only a few GB of the
host's memory when switching from memhog to page fault tests.

Any of the overhead visible with this patch set enabled seems due to page
faults caused by accessing the reported pages and the host zeroing the page
before giving it back to the guest. This overhead is much more visible when
using THP than with standard 4K pages. In addition page shuffling seemed to
increase the amount of faults generated due to an increase in memory churn.
The overhead is reduced when using MADV_FREE as we can avoid the extra
zeroing of the pages when they are reintroduced to the host, as can be seen
when the RFC is applied with shuffling enabled.

The overall guest size is kept fairly small to only a few GB while the test
is running. If the host memory were oversubscribed this patch set should
result in a performance improvement as swapping memory in the host can be
avoided.

A brief history on the background of free page reporting can be found at:
https://lore.kernel.org/lkml/29f43d5796feed0dec8e8bb98b187d9dac03b900.camel@linux.intel.com/

Changes from v14:
https://lore.kernel.org/lkml/20191119214454.24996.66289.stgit@localhost.localdomain/
Renamed "unused page reporting" to "free page reporting"
  Updated code, kconfig, and patch descriptions
Split out patch for __free_isolated_page
  Renamed function to __putback_isolated_page
Rewrote core reporting functionality
  Added logic to reschedule worker in 2 seconds instead of run to completion
  Removed reported_pages statistics
  Removed REPORTING_REQUESTED bit used in zone flags
  Replaced page_reporting_dev_info refcount with state variable
  Removed scatterlist from page_reporting_dev_info
  Removed capacity from page reporting device
  Added dynamic scatterlist allocation/free at start/end of reporting process
  Updated __free_one_page so that reported pages are not always added to tail
  Added logic to handle error from report function
Updated virtio-balloon patch that adds support for page reporting
  Updated patch description to try and highlight differences in approaches
  Updated logic to reflect that we cannot limit the scatterlist from device
  Added logic to return error from report function
Moved documentation patch to end of patch set

Changes from v15:
https://lore.kernel.org/lkml/20191205161928.19548.41654.stgit@localhost.localdomain/
Rebased on linux-next-20191219
Split out patches for budget and moving head to last page processed
  Updated budget code to reduce how much memory is reported per pass
  Added logic to also rotate the list if we exit due a page isolation failure
Added migratetype as argument in __putback_isolated_page

---

Alexander Duyck (9):
      mm: Adjust shuffle code to allow for future coalescing
      mm: Use zone and order instead of free area in free_list manipulators
      mm: Add function __putback_isolated_page
      mm: Introduce Reported pages
      virtio-balloon: Pull page poisoning config out of free page hinting
      virtio-balloon: Add support for providing free page reports to host
      mm: Rotate free list so reported pages are moved to the tail of the list
      mm: Add budget limit to how many pages can be reported per list per pass
      mm: Add free page reporting documentation


 Documentation/vm/free_page_reporting.rst |   41 +++
 drivers/virtio/Kconfig                   |    1 
 drivers/virtio/virtio_balloon.c          |   87 +++++++
 include/linux/mmzone.h                   |   44 ----
 include/linux/page-flags.h               |   11 +
 include/linux/page_reporting.h           |   26 ++
 include/uapi/linux/virtio_balloon.h      |    1 
 mm/Kconfig                               |   11 +
 mm/Makefile                              |    1 
 mm/internal.h                            |    2 
 mm/page_alloc.c                          |  164 ++++++++++----
 mm/page_isolation.c                      |    6 
 mm/page_reporting.c                      |  364 ++++++++++++++++++++++++++++++
 mm/page_reporting.h                      |   54 ++++
 mm/shuffle.c                             |   12 -
 mm/shuffle.h                             |    6 
 16 files changed, 725 insertions(+), 106 deletions(-)
 create mode 100644 Documentation/vm/free_page_reporting.rst
 create mode 100644 include/linux/page_reporting.h
 create mode 100644 mm/page_reporting.c
 create mode 100644 mm/page_reporting.h

--

Comments

Nitesh Narayan Lal Jan. 8, 2020, 7:57 a.m. UTC | #1
On 1/3/20 4:16 PM, Alexander Duyck wrote:
> This series provides an asynchronous means of reporting free guest pages
> to a hypervisor so that the memory associated with those pages can be
> dropped and reused by other processes and/or guests on the host. Using
> this it is possible to avoid unnecessary I/O to disk and greatly improve
> performance in the case of memory overcommit on the host.
>
> When enabled we will be performing a scan of free memory every 2 seconds
> while pages of sufficiently high order are being freed. In each pass at
> least one sixteenth of each free list will be reported. By doing this we
> avoid racing against other threads that may be causing a high amount of
> memory churn.
>
> The lowest page order currently scanned when reporting pages is
> pageblock_order so that this feature will not interfere with the use of
> Transparent Huge Pages in the case of virtualization.
>
> Currently this is only in use by virtio-balloon however there is the hope
> that at some point in the future other hypervisors might be able to make
> use of it. In the virtio-balloon/QEMU implementation the hypervisor is
> currently using MADV_DONTNEED to indicate to the host kernel that the page
> is currently free. It will be zeroed and faulted back into the guest the
> next time the page is accessed.
>
> To track if a page is reported or not the Uptodate flag was repurposed and
> used as a Reported flag for Buddy pages. We walk though the free list
> isolating pages and adding them to the scatterlist until we either
> encounter the end of the list, processed as many pages as were listed in
> nr_free prior to us starting, or have filled the scatterlist with pages to
> be reported. If we fill the scatterlist before we reach the end of the
> list we rotate the list so that the first unreported page we encounter is
> moved to the head of the list as that is where we will resume after we
> have freed the reported pages back into the tail of the list.
>
> Below are the results from various benchmarks. I primarily focused on two
> tests. The first is the will-it-scale/page_fault2 test, and the other is
> a modified version of will-it-scale/page_fault1 that was enabled to use
> THP. I did this as it allows for better visibility into different parts
> of the memory subsystem. The guest is running with 32G for RAM on one
> node of a E5-2630 v3. The host has had some features such as CPU turbo
> disabled in the BIOS.
>
> Test                   page_fault1 (THP)    page_fault2
> Name            tasks  Process Iter  STDEV  Process Iter  STDEV
> Baseline            1    1012402.50  0.14%     361855.25  0.81%
>                    16    8827457.25  0.09%    3282347.00  0.34%
>
> Patches Applied     1    1007897.00  0.23%     361887.00  0.26%
>                    16    8784741.75  0.39%    3240669.25  0.48%
>
> Patches Enabled     1    1010227.50  0.39%     359749.25  0.56%
>                    16    8756219.00  0.24%    3226608.75  0.97%
>
> Patches Enabled     1    1050982.00  4.26%     357966.25  0.14%
>  page shuffle      16    8672601.25  0.49%    3223177.75  0.40%
>
> Patches Enabled     1    1003238.00  0.22%     360211.00  0.22%
>  shuffle w/ RFC    16    8767010.50  0.32%    3199874.00  0.71%

Just to be sure that I understand your test setup correctly:
- You have a 32GB guest with a single node affined to a single node of your host
(E5-2630).
- You have THP in both host and the guest enabled and set to 'madvise'.
- On top of the default x86_64 config and other virtio config options you have
CONFIG_SLAB_FREELIST_RANDOM and CONFIG_SHUFFLE_PAGE_ALLOCATOR enabled for the
third observation (Patches Enabled page shuffle).
did I miss anything?

Can you also remind me of the reason you have skipped recording the number of
threads count reported as part of page_fault tests? Was it because you were
observing different values with every fresh boot?


> The results above are for a baseline with a linux-next-20191219 kernel,
> that kernel with this patch set applied but page reporting disabled in
> virtio-balloon, the patches applied and page reporting fully enabled, the
> patches enabled with page shuffling enabled, and the patches applied with
> page shuffling enabled and an RFC patch that makes used of MADV_FREE in
> QEMU. These results include the deviation seen between the average value
> reported here versus the high and/or low value. I observed that during the
> test memory usage for the first three tests never dropped whereas with the
> patches fully enabled the VM would drop to using only a few GB of the
> host's memory when switching from memhog to page fault tests.

Do you mean that in the later case you run the page fault tests after memhog?
If so how much memory do you pass to memhog?
Alexander Duyck Jan. 8, 2020, 4:29 p.m. UTC | #2
On Wed, 2020-01-08 at 02:57 -0500, Nitesh Narayan Lal wrote:
> On 1/3/20 4:16 PM, Alexander Duyck wrote:

<snip>

> > Below are the results from various benchmarks. I primarily focused on two
> > tests. The first is the will-it-scale/page_fault2 test, and the other is
> > a modified version of will-it-scale/page_fault1 that was enabled to use
> > THP. I did this as it allows for better visibility into different parts
> > of the memory subsystem. The guest is running with 32G for RAM on one
> > node of a E5-2630 v3. The host has had some features such as CPU turbo
> > disabled in the BIOS.
> > 
> > Test                   page_fault1 (THP)    page_fault2
> > Name            tasks  Process Iter  STDEV  Process Iter  STDEV
> > Baseline            1    1012402.50  0.14%     361855.25  0.81%
> >                    16    8827457.25  0.09%    3282347.00  0.34%
> > 
> > Patches Applied     1    1007897.00  0.23%     361887.00  0.26%
> >                    16    8784741.75  0.39%    3240669.25  0.48%
> > 
> > Patches Enabled     1    1010227.50  0.39%     359749.25  0.56%
> >                    16    8756219.00  0.24%    3226608.75  0.97%
> > 
> > Patches Enabled     1    1050982.00  4.26%     357966.25  0.14%
> >  page shuffle      16    8672601.25  0.49%    3223177.75  0.40%
> > 
> > Patches Enabled     1    1003238.00  0.22%     360211.00  0.22%
> >  shuffle w/ RFC    16    8767010.50  0.32%    3199874.00  0.71%
> 
> Just to be sure that I understand your test setup correctly:
> - You have a 32GB guest with a single node affined to a single node of your host
> (E5-2630).
> - You have THP in both host and the guest enabled and set to 'madvise'.
> - On top of the default x86_64 config and other virtio config options you have
> CONFIG_SLAB_FREELIST_RANDOM and CONFIG_SHUFFLE_PAGE_ALLOCATOR enabled for the
> third observation (Patches Enabled page shuffle).
> did I miss anything?

So the only things I think you overlooked was that CPU turbo was disbled
int eh BIOS. Without that my numbers were much more unpredictable as the
CPUs were turboing up and down and me and giving me inconsistent results.

Also one thing I forgot to mention is that I had to modify the grub kernel
command line to include page_alloc.shuffle=Y so that the page shuffling
was actually active.

> Can you also remind me of the reason you have skipped recording the number of
> threads count reported as part of page_fault tests? Was it because you were
> observing different values with every fresh boot?

Mainly because the threads test gave me data that was all over the place
at higher task counts and because it doesn't scale as well as the
processes test case. The averages between the two worked out to be about
the same, but the standard deviation was maxing out at 7% for the baseline
and 8% for the patches enabled case. However the differences in the
averages is still less than 1%.

So for example the same data using the threads values for Baseline vs
Patches enabled comes out as follows:
Baseline                1    1133900.25  0.24%     358395.25  0.30%
                       16    5848684.75  6.96%    2181989.00  1.69%

Patches Enabled         1    1132748.50  0.20%     356615.00  0.11%
                       16    5796647.00  8.38%    2160475.50  1.84%

> > The results above are for a baseline with a linux-next-20191219 kernel,
> > that kernel with this patch set applied but page reporting disabled in
> > virtio-balloon, the patches applied and page reporting fully enabled, the
> > patches enabled with page shuffling enabled, and the patches applied with
> > page shuffling enabled and an RFC patch that makes used of MADV_FREE in
> > QEMU. These results include the deviation seen between the average value
> > reported here versus the high and/or low value. I observed that during the
> > test memory usage for the first three tests never dropped whereas with the
> > patches fully enabled the VM would drop to using only a few GB of the
> > host's memory when switching from memhog to page fault tests.
> 
> Do you mean that in the later case you run the page fault tests after memhog?
> If so how much memory do you pass to memhog?

For every test I would run memhog 32g in the guest to make sure all memory
was allocated at least once before running the page fault tests. I was
using that to make certain that the page reporting was working before
running the test.

That way the baseline gives more consistent results as we don't have to
worry about there being any memory the guest has yet to fault in.
Alexander Duyck Jan. 15, 2020, 11:08 p.m. UTC | #3
On Fri, 2020-01-03 at 13:16 -0800, Alexander Duyck wrote:
> This series provides an asynchronous means of reporting free guest pages
> to a hypervisor so that the memory associated with those pages can be
> dropped and reused by other processes and/or guests on the host. Using
> this it is possible to avoid unnecessary I/O to disk and greatly improve
> performance in the case of memory overcommit on the host.

<snip>

> 
> Changes from v15:
> https://lore.kernel.org/lkml/20191205161928.19548.41654.stgit@localhost.localdomain/
> Rebased on linux-next-20191219
> Split out patches for budget and moving head to last page processed
>   Updated budget code to reduce how much memory is reported per pass
>   Added logic to also rotate the list if we exit due a page isolation failure
> Added migratetype as argument in __putback_isolated_page

It's been about a week and a half since I posted the set and haven't
really gotten much feedback other than a suggestion of a slight tweak to
the titles for patches 7 & 8 to mention page_reporting. I'm mainly looking
for input on patches 3, 4, 7 and 8 since those are the ones that contain
most of the changes based on recent feedback.

I'm wondering if there is any remaining concerns or if these patches are
in a state where they are ready to be pulled into the MM tree?

Thanks.

- Alex