xenbus/backend: Add a memory pressure handler callback
Subsystems (3)
v7 cover xenbus/backend: Add a memory pressure handler callback
Granting pages consumes backend system memory. In systems configured with insufficient spare memory for those pages, it can cause a memory pressure situation. However, finding the optimal amount of the spare memory is challenging for large systems having dynamic resource utilization patterns. Also, such a static configuration might lack flexibility.
To mitigate such problems, this patchset adds a memory reclaim callback to 'xenbus_driver' (patch 1) and use it to mitigate the problem in 'xen-blkback' (patch 2). The third patch is a trivial cleanup of variable names.
Base Version
------------
This patch is based on v5.4. A complete tree is also available at my public git repo:
https://github.com/sjp38/linux/tree/blkback_squeezing_v7
Patch History
-------------
Changes from v6
(https://lore.kernel.org/linux-block/[email protected]/)
- Remove more unnecessary prefixes (suggested by Roger Pau Monné)
- Constify a variable (suggested by Roger Pau Monné)
- Rename 'reclaim' into 'reclaim_memory' (suggested by Roger Pau Monné)
- More wordsmith of the commit message (suggested by Roger Pau Monné)
Changes from v5
(https://lore.kernel.org/linux-block/[email protected]/)
- Wordsmith the commit messages (suggested by Roger Pau Monné)
- Change the reclaim callback return type (suggested by Roger Pau Monné)
- Change the type of the blkback squeeze duration variable (suggested by Roger Pau Monné)
- Add a patch for removal of unnecessary static variable name prefixes (suggested by Roger Pau Monné)
- Fix checkpatch.pl warnings
Changes from v4
(https://lore.kernel.org/xen-devel/[email protected]/)
- Remove domain id parameter from the callback (suggested by Juergen Gross)
- Rename xen-blkback module parameter (suggested by Stefan Nuernburger)
Changes from v3
(https://lore.kernel.org/xen-devel/[email protected]/)
- Add general callback in xen_driver and use it (suggested by Juergen Gross)
Changes from v2
(https://lore.kernel.org/linux-block/[email protected])
- Rename the module parameter and variables for brevity (aggressive shrinking -> squeezing)
Changes from v1
(https://lore.kernel.org/xen-devel/[email protected]/)
- Adjust the description to not use the term, arbitrarily (suggested by Paul Durrant)
- Specify time unit of the duration in the parameter description, (suggested by Maximilian Heyne)
- Change default aggressive shrinking duration from 1ms to 10ms
- Merge two patches into one single patch
SeongJae Park (2):
xenbus/backend: Add memory pressure handler callback xen/blkback: Squeeze page pools if a memory pressure is detected
drivers/block/xen-blkback/blkback.c | 23 +++++++++++++++-- drivers/block/xen-blkback/common.h | 1 + drivers/block/xen-blkback/xenbus.c | 3 ++- drivers/xen/xenbus/xenbus_probe_backend.c | 31 +++++++++++++++++++++++ include/xen/xenbus.h | 1 + 5 files changed, 56 insertions(+), 3 deletions(-)
Signature · 1 line
2.17.1
PATCH v7 1/3 xenbus/backend: Add memory pressure handler callback
Granting pages consumes backend system memory. In systems configured with insufficient spare memory for those pages, it can cause a memory pressure situation. However, finding the optimal amount of the spare memory is challenging for large systems having dynamic resource utilization patterns. Also, such a static configuration might lack flexibility.
To mitigate such problems, this commit adds a memory reclaim callback to 'xenbus_driver'. If a memory pressure is detected, 'xenbus' requests every backend driver to volunarily release its memory.
Note that it would be able to improve the callback facility for more sophisticated handlings of general pressures. For example, it would be possible to monitor the memory consumption of each device and issue the release requests to only devices which causing the pressure. Also, the callback could be extended to handle not only memory, but general resources. Nevertheless, this version of the implementation defers such sophisticated goals as a future work.
--- drivers/xen/xenbus/xenbus_probe_backend.c | 32 +++++++++++++++++++++++ include/xen/xenbus.h | 1 + 2 files changed, 33 insertions(+)
drivers/xen/xenbus/xenbus_probe_backend.c +32 -0
@@ -248,6 +248,35 @@ static int backend_probe_and_watch(struct notifier_block *notifier,return NOTIFY_DONE ; } +static int backend_reclaim_memory (struct device *dev,void *data) +{ +const struct xenbus_driver *drv; + +if (!dev->driver) +return 0 ; + drv =to_xenbus_driver (dev->driver); +if (drv && drv->reclaim_memory) + drv->reclaim_memory (to_xenbus_device (dev)); +return 0 ; +} + +/* +* Returns 0 always because we are using shrinker to only detect memory +* pressure. +*/ +static unsigned long backend_shrink_memory_count (struct shrinker *shrinker, +struct shrink_control *sc) +{ +bus_for_each_dev (&xenbus_backend.bus,NULL ,NULL , + backend_reclaim_memory); +return 0 ; +} + +static struct shrinker backend_memory_shrinker = { + .count_objects = backend_shrink_memory_count, + .seeks =DEFAULT_SEEKS , +}; +static int __init xenbus_probe_backend_init (void ) {static struct notifier_block xenstore_notifier = {
@@ -264,6 +293,9 @@ static int __init xenbus_probe_backend_init(void)register_xenstore_notifier (&xenstore_notifier); +if (register_shrinker (&backend_memory_shrinker)) +pr_warn ("shrinker registration failed\n" ); +return 0 ; }subsys_initcall (xenbus_probe_backend_init);
include/xen/xenbus.h +1 -0
@@ -104,6 +104,7 @@ struct xenbus_driver {struct device_driver driver;int (*read_otherend_details)(struct xenbus_device *dev);int (*is_ready)(struct xenbus_device *dev); +void (*reclaim_memory)(struct xenbus_device *dev); };static inline struct xenbus_driver *to_xenbus_driver (struct device_driver *drv)
Signature · 1 line
2.17.1
Re: [Xen-devel] [PATCH v7 1/3] xenbus/backend: Add memory pressure handler callback
Quoted 19 lines · SeongJae Park
Granting pages consumes backend system memory. In systems configured with insufficient spare memory for those pages, it can cause a memory pressure situation. However, finding the optimal amount of the spare memory is challenging for large systems having dynamic resource utilization patterns. Also, such a static configuration might lack flexibility.
To mitigate such problems, this commit adds a memory reclaim callback to '
xenbus_driver'. If a memory pressure is detected, 'xenbus' requests every backend driver to volunarily release its memory.Note that it would be able to improve the callback facility for more sophisticated handlings of general pressures. For example, it would be possible to monitor the memory consumption of each device and issue the release requests to only devices which causing the pressure. Also, the callback could be extended to handle not only memory, but general resources. Nevertheless, this version of the implementation defers such sophisticated goals as a future work.
Quoted 2 lines
Reviewed-by Juergen Gross <[email protected]>Signed-off-by SeongJae Park <[email protected]>
Thanks, Roger.
PATCH v7 2/3 xen/blkback: Squeeze page pools if a memory pressure is detected
In reply to SeongJae ParkEach blkif has a free pages pool for the grant mapping. The size of the pool starts from zero and be increased on demand while processing the I/O requests. If current I/O requests handling is finished or 100 milliseconds has passed since last I/O requests handling, it checks and shrinks the pool to not exceed the size limit, max_buffer_pages.
Therefore, host administrators can cause memory pressure in blkback by attaching a large number of block devices and inducing I/O. Such problematic situations can be avoided by limiting the maximum number of devices that can be attached, but finding the optimal limit is not so easy. Improper set of the limit can results in the memory pressure or a resource underutilization. This commit avoids such problematic situations by squeezing the pools (returns every free page in the pool to the system) for a while (users can set this duration via a module parameter) if a memory pressure is detected.
Discussions
===========
The blkback's original shrinking mechanism returns only pages in the pool, which are not currently be used by blkback, to the system. In other words, the pages that are not mapped with granted pages. Because this commit is changing only the shrink limit but still uses the same freeing mechanism it does not touch pages which are currently mapping grants.
Once a memory pressure is detected, this commit keeps the squeezing limit for a user-specified time duration. The duration should be neither too long nor too short. If it is too long, the squeezing incurring overhead can reduce the I/O performance. If it is too short, blkback will not free enough pages to reduce the memory pressure. This commit sets the value as 10 milliseconds by default because it is a short time in terms of I/O while it is a long time in terms of memory operations. Also, as the original shrinking mechanism works for at least every 100 milliseconds, this could be a somewhat reasonable choice. I also tested other durations (refer to the below section for more details) and confirmed that 10 milliseconds is the one that works best with the test. That said, the proper duration depends on actual configurations and workloads. That's why this commit allows users to set the duration as a module parameter.
Memory Pressure Test
====================
To show how this commit fixes the memory pressure situation well, I configured a test environment on a xen-running virtualization system. On the blkfront running guest instances, I attach a large number of network-backed volume devices and induce I/O to those. Meanwhile, I measure the number of pages that swapped in (pswpin) and out (pswpout) on the blkback running guest. The test ran twice, once for the blkback before this commit and once for that after this commit. As shown below, this commit has dramatically reduced the memory pressure:
pswpin pswpout
before 76,672 185,799
after 212 3,325Optimal Aggressive Shrinking Duration
-------------------------------------
To find a best squeezing duration, I repeated the test with three different durations (1ms, 10ms, and 100ms). The results are as below:
duration pswpin pswpout
1 852 6,424
10 212 3,325
100 203 3,340As expected, the memory pressure has decreased as the duration is increased, but the reduction stopped from the 10ms. Based on this results, I chose the default duration as 10ms.
Performance Overhead Test
=========================
This commit could incur I/O performance degradation under severe memory pressure because the squeezing will require more page allocations per I/O. To show the overhead, I artificially made a worst-case squeezing situation and measured the I/O performance of a blkfront running guest.
For the artificial squeezing, I set the blkback.max_buffer_pages using the /sys/module/xen_blkback/parameters/max_buffer_pages file. In this test, I set the value to 1024 and 0. The 1024 is the default value. Setting the value as 0 is same to a situation doing the squeezing always (worst-case).
For the I/O performance measurement, I run a simple dd command 5 times as below and collect the 'MB/s' results.
$ for i in {1..5}; do dd if=/dev/zero of=file \
bs=4k count=$((256*512)); sync; doneIf the underlying block device is slow enough, the squeezing overhead could be hidden. For the reason, I do this test for both a slow block device and a fast block device. I use a popular cloud block storage service, ebs[1] as a slow device and the ramdisk block device[2] for the fast device.
The results are as below. 'max_pgs' represents the value of the blkback.max_buffer_pages parameter.
On the slow block device
------------------------
max_pgs Min Max Median Avg Stddev
0 38.7 45.8 38.7 40.12 3.1752165
1024 38.7 45.8 38.7 40.12 3.1752165No difference proven at 95.0% confidence
On the fast block device
------------------------
max_pgs Min Max Median Avg Stddev
0 417 423 420 419.4 2.5099801
1024 414 425 416 417.8 4.4384682No difference proven at 95.0% confidence
In short, even worst case squeezing on ramdisk based fast block device makes no visible performance degradation. Please note that this is just a very simple and minimal test. On systems using super-fast block devices and a special I/O workload, the results might be different. If you have any doubt, test on your machine for your workload to find the optimal squeezing duration for you.
[1] https://aws.amazon.com/ebs/
[2] https://www.kernel.org/doc/html/latest/admin-guide/blockdev/ramdisk.html
--- drivers/block/xen-blkback/blkback.c | 22 ++++++++++++++++++++-- drivers/block/xen-blkback/common.h | 1 + drivers/block/xen-blkback/xenbus.c | 3 ++- 3 files changed, 23 insertions(+), 3 deletions(-)
drivers/block/xen-blkback/blkback.c +20 -2
@@ -142,6 +142,21 @@ static inline bool persistent_gnt_timeout(struct persistent_gnt *persistent_gnt) HZ * xen_blkif_pgrant_timeout); } +/* Once a memory pressure is detected, squeeze free page pools for a while. */ +static unsigned int buffer_squeeze_duration_ms =10 ; +module_param_named (buffer_squeeze_duration_ms, + buffer_squeeze_duration_ms, int,0644 ); +MODULE_PARM_DESC (buffer_squeeze_duration_ms, +"Duration in ms to squeeze pages buffer when a memory pressure is detected" ); + +static unsigned long buffer_squeeze_end; + +void xen_blkbk_reclaim_memory (struct xenbus_device *dev) +{ + buffer_squeeze_end = jiffies + +msecs_to_jiffies (buffer_squeeze_duration_ms); +} +static inline int get_free_page (struct xen_blkif_ring *ring,struct page **page) {unsigned long flags;
@@ -656,8 +671,11 @@ int xen_blkif_schedule(void *arg) ring->next_lru = jiffies +msecs_to_jiffies (LRU_INTERVAL ); } -/* Shrink if we have more than xen_blkif_max_buffer_pages */ -shrink_free_pagepool (ring, xen_blkif_max_buffer_pages); +/* Shrink the free pages pool if it is too large. */ +if (time_before (jiffies, buffer_squeeze_end)) +shrink_free_pagepool (ring,0 ); +else +shrink_free_pagepool (ring, xen_blkif_max_buffer_pages);if (log_stats &&time_after (jiffies, ring->st_print))print_stats (ring);
drivers/block/xen-blkback/common.h +1 -0
@@ -383,6 +383,7 @@ irqreturn_t xen_blkif_be_int(int irq, void *dev_id);int xen_blkif_schedule (void *arg);int xen_blkif_purge_persistent (void *arg);void xen_blkbk_free_caches (struct xen_blkif_ring *ring); +void xen_blkbk_reclaim_memory (struct xenbus_device *dev);int xen_blkbk_flush_diskcache (struct xenbus_transaction xbt,struct backend_info *be,int state);
drivers/block/xen-blkback/xenbus.c +2 -1
@@ -1115,7 +1115,8 @@ static struct xenbus_driver xen_blkbk_driver = { .ids = xen_blkbk_ids, .probe = xen_blkbk_probe, .remove = xen_blkbk_remove, - .otherend_changed = frontend_changed + .otherend_changed = frontend_changed, + .reclaim_memory = xen_blkbk_reclaim_memory, };int xen_blkif_xenbus_init (void )
Signature · 1 line
2.17.1
Re: [Xen-devel] [PATCH v7 2/3] xen/blkback: Squeeze page pools if a memory pressure is detected
Hello,
Please make sure you Cc me in blkback related patches.
Quoted 2 lines · SeongJae Park
Each
blkifhas a free pages pool for the grant mapping. The size of the pool starts from zero and be increased on demand while processing
^ is
Quoted 9 lines
the I/O requests. If current I/O requests handling is finished or 100 milliseconds has passed since last I/O requests handling, it checks and shrinks the pool to not exceed the size limit,
max_buffer_pages.Therefore, host administrators can cause memory pressure in blkback by attaching a large number of block devices and inducing I/O. Such problematic situations can be avoided by limiting the maximum number of devices that can be attached, but finding the optimal limit is not so easy. Improper set of the limit can results in the memory pressure or a
^ s/the//
Quoted 4 lines
resource underutilization. This commit avoids such problematic situations by squeezing the pools (returns every free page in the pool to the system) for a while (users can set this duration via a module parameter) if a memory pressure is detected.
^ s/a//
Quoted 6 lines
Discussions
===========The
blkback's original shrinking mechanism returns only pages in the pool, which are not currently be used byblkback, to the system. In
I think you can remove both comas in the above sentence.
Quoted 6 lines
other words, the pages that are not mapped with granted pages. Because this commit is changing only the shrink limit but still uses the same freeing mechanism it does not touch pages which are currently mapping grants.
Once a memory pressure is detected, this commit keeps the squeezing
^ s/a//
Quoted 60 lines
limit for a user-specified time duration. The duration should be neither too long nor too short. If it is too long, the squeezing incurring overhead can reduce the I/O performance. If it is too short,
blkbackwill not free enough pages to reduce the memory pressure. This commit sets the value as10 millisecondsby default because it is a short time in terms of I/O while it is a long time in terms of memory operations. Also, as the original shrinking mechanism works for at least every 100 milliseconds, this could be a somewhat reasonable choice. I also tested other durations (refer to the below section for more details) and confirmed that 10 milliseconds is the one that works best with the test. That said, the proper duration depends on actual configurations and workloads. That's why this commit allows users to set the duration as a module parameter.Memory Pressure Test
====================To show how this commit fixes the memory pressure situation well, I configured a test environment on a xen-running virtualization system. On the
blkfrontrunning guest instances, I attach a large number of network-backed volume devices and induce I/O to those. Meanwhile, I measure the number of pages that swapped in (pswpin) and out (pswpout) on theblkbackrunning guest. The test ran twice, once for theblkbackbefore this commit and once for that after this commit. As shown below, this commit has dramatically reduced the memory pressure:pswpin pswpout before 76,672 185,799 after 212 3,325Optimal Aggressive Shrinking Duration
-------------------------------------To find a best squeezing duration, I repeated the test with three different durations (1ms, 10ms, and 100ms). The results are as below:
duration pswpin pswpout 1 852 6,424 10 212 3,325 100 203 3,340As expected, the memory pressure has decreased as the duration is increased, but the reduction stopped from the
10ms. Based on this results, I chose the default duration as 10ms.Performance Overhead Test
=========================This commit could incur I/O performance degradation under severe memory pressure because the squeezing will require more page allocations per I/O. To show the overhead, I artificially made a worst-case squeezing situation and measured the I/O performance of a
blkfrontrunning guest.For the artificial squeezing, I set the
blkback.max_buffer_pagesusing the/sys/module/xen_blkback/parameters/max_buffer_pagesfile. In this test, I set the value to1024and0. The1024is the default value. Setting the value as0is same to a situation doing the squeezing always (worst-case).
Quoted 5 lines
For the I/O performance measurement, I run a simple
ddcommand 5 times as below and collect the 'MB/s' results.$ for i in {1..5}; do dd if=/dev/zero of=file \ bs=4k count=$((256*512)); sync; done
I think it would be better if you could skip the filesystem overhead by writing directly to a block device, ie:
Attach a null_blk based block device to the guest (on dom0):
# xl block-attach guest phy:/dev/null_blk0 xvdb w
Run a workload against the device (inside the guest):
# fio --filename=/dev/xvdb --direct=1 --rw=randrw --bs=4k --ioengine=libaio \
--iodepth=64 --runtime=120 --numjobs=4 --time_based --group_reporting \
--name=throughput-test-job --eta-newline=1You should run this on a multi-vcpu guest so that multiple queues are used, and adjust the numjobs to (at least) match the number of queues.
Quoted 18 lines
If the underlying block device is slow enough, the squeezing overhead could be hidden. For the reason, I do this test for both a slow block device and a fast block device. I use a popular cloud block storage service, ebs[1] as a slow device and the ramdisk block device[2] for the fast device.
The results are as below. '
max_pgs' represents the value of theblkback.max_buffer_pagesparameter.On the slow block device
------------------------max_pgs Min Max Median Avg Stddev 0 38.7 45.8 38.7 40.12 3.1752165 1024 38.7 45.8 38.7 40.12 3.1752165No difference proven at 95.0% confidence
Quoted 7 lines
On the fast block device
------------------------max_pgs Min Max Median Avg Stddev 0 417 423 420 419.4 2.5099801 1024 414 425 416 417.8 4.4384682No difference proven at 95.0% confidence
This is intriguing, as it seems to prove that the usage of a cache of free pages is irrelevant performance wise.
The pool of free pages was introduced long ago, and it's possible that recent improvements to the balloon driver had made such pool useless, at which point it could be removed instead of worked around.
Do you think you could perform some more tests (as pointed out above against the block device to skip the fs overhead) and report back the results?
Quoted 24 lines
In short, even worst case squeezing on ramdisk based fast block device makes no visible performance degradation. Please note that this is just a very simple and minimal test. On systems using super-fast block devices and a special I/O workload, the results might be different. If you have any doubt, test on your machine for your workload to find the optimal squeezing duration for you.
[1] https://aws.amazon.com/ebs/
[2] https://www.kernel.org/doc/html/latest/admin-guide/blockdev/ramdisk.htmlReviewed-by Juergen Gross <[email protected]>Signed-off-by SeongJae Park <[email protected]>--- drivers/block/xen-blkback/blkback.c | 22 ++++++++++++++++++++-- drivers/block/xen-blkback/common.h | 1 + drivers/block/xen-blkback/xenbus.c | 3 ++- 3 files changed, 23 insertions(+), 3 deletions(-)drivers/block/xen-blkback/blkback.c +0
-0
drivers/block/xen-blkback/blkback.c static inline bool persistent_gnt_timeout(struct persistent_gnt *persistent_gnt) · line 150
Quoted 9 lines
drivers/block/xen-blkback/blkback.c +6
-0@@ -142,6 +142,21 @@ static inline bool persistent_gnt_timeout(struct persistent_gnt *persistent_gnt) HZ * xen_blkif_pgrant_timeout); } +/* Once a memory pressure is detected, squeeze free page pools for a while. */ +static unsigned int buffer_squeeze_duration_ms =10 ; +module_param_named (buffer_squeeze_duration_ms, + buffer_squeeze_duration_ms, int,0644 ); +MODULE_PARM_DESC (buffer_squeeze_duration_ms, +"Duration in ms to squeeze pages buffer when a memory pressure is detected" );
You should add a description about this parameter to Documentation/ABI/testing/sysfs-driver-xen-blkback
Thanks, Roger.
Re: Re: [Xen-devel] [PATCH v7 2/3] xen/blkback: Squeeze page pools if a memory pressure is detected
Quoted 2 lines · On Thu, 12 Dec 2019 12:42:47 +0100 "Roger Pau Monné" <[email protected]>…
Please make sure you Cc me in blkback related patches.
Sorry for forgotting you! I will never forget again.
Quoted 29 lines · SeongJae Park
Quoted 2 lines
Each
blkifhas a free pages pool for the grant mapping. The size of the pool starts from zero and be increased on demand while processing^ is
Quoted 9 lines
the I/O requests. If current I/O requests handling is finished or 100 milliseconds has passed since last I/O requests handling, it checks and shrinks the pool to not exceed the size limit,
max_buffer_pages.Therefore, host administrators can cause memory pressure in blkback by attaching a large number of block devices and inducing I/O. Such problematic situations can be avoided by limiting the maximum number of devices that can be attached, but finding the optimal limit is not so easy. Improper set of the limit can results in the memory pressure or a
^ s/the//Quoted 4 lines
resource underutilization. This commit avoids such problematic situations by squeezing the pools (returns every free page in the pool to the system) for a while (users can set this duration via a module parameter) if a memory pressure is detected.
^ s/a//Quoted 6 lines
Discussions
===========The
blkback's original shrinking mechanism returns only pages in the pool, which are not currently be used byblkback, to the system. InI think you can remove both comas in the above sentence.
Quoted 7 lines
Quoted 6 lines
other words, the pages that are not mapped with granted pages. Because this commit is changing only the shrink limit but still uses the same freeing mechanism it does not touch pages which are currently mapping grants.
Once a memory pressure is detected, this commit keeps the squeezing
^ s/a//
Thank you for corrections, will apply!
Quoted 72 lines
Quoted 60 lines
limit for a user-specified time duration. The duration should be neither too long nor too short. If it is too long, the squeezing incurring overhead can reduce the I/O performance. If it is too short,
blkbackwill not free enough pages to reduce the memory pressure. This commit sets the value as10 millisecondsby default because it is a short time in terms of I/O while it is a long time in terms of memory operations. Also, as the original shrinking mechanism works for at least every 100 milliseconds, this could be a somewhat reasonable choice. I also tested other durations (refer to the below section for more details) and confirmed that 10 milliseconds is the one that works best with the test. That said, the proper duration depends on actual configurations and workloads. That's why this commit allows users to set the duration as a module parameter.Memory Pressure Test
====================To show how this commit fixes the memory pressure situation well, I configured a test environment on a xen-running virtualization system. On the
blkfrontrunning guest instances, I attach a large number of network-backed volume devices and induce I/O to those. Meanwhile, I measure the number of pages that swapped in (pswpin) and out (pswpout) on theblkbackrunning guest. The test ran twice, once for theblkbackbefore this commit and once for that after this commit. As shown below, this commit has dramatically reduced the memory pressure:pswpin pswpout before 76,672 185,799 after 212 3,325Optimal Aggressive Shrinking Duration
-------------------------------------To find a best squeezing duration, I repeated the test with three different durations (1ms, 10ms, and 100ms). The results are as below:
duration pswpin pswpout 1 852 6,424 10 212 3,325 100 203 3,340As expected, the memory pressure has decreased as the duration is increased, but the reduction stopped from the
10ms. Based on this results, I chose the default duration as 10ms.Performance Overhead Test
=========================This commit could incur I/O performance degradation under severe memory pressure because the squeezing will require more page allocations per I/O. To show the overhead, I artificially made a worst-case squeezing situation and measured the I/O performance of a
blkfrontrunning guest.For the artificial squeezing, I set the
blkback.max_buffer_pagesusing the/sys/module/xen_blkback/parameters/max_buffer_pagesfile. In this test, I set the value to1024and0. The1024is the default value. Setting the value as0is same to a situation doing the squeezing always (worst-case).Quoted 5 lines
For the I/O performance measurement, I run a simple
ddcommand 5 times as below and collect the 'MB/s' results.$ for i in {1..5}; do dd if=/dev/zero of=file \ bs=4k count=$((256*512)); sync; doneI think it would be better if you could skip the filesystem overhead by writing directly to a block device, ie:
Attach a
null_blkbased block device to the guest (on dom0):# xl block-attach guest phy:/dev/null_blk0 xvdb w
Quoted 7 lines
Run a workload against the device (inside the guest):
# fio --filename=/dev/xvdb --direct=1 --rw=randrw --bs=4k --ioengine=libaio \ --iodepth=64 --runtime=120 --numjobs=4 --time_based --group_reporting \ --name=throughput-test-job --eta-newline=1You should run this on a multi-vcpu guest so that multiple queues are used, and adjust the numjobs to (at least) match the number of queues.
I forgot to update the dd command. I used the command for the slow block device test, but directly induced the I/O towards the block device for the fast block device test as below:
# xl block-attach guest phy:/dev/ram0 xvdb w
$ for i in {1..5}; do dd if=/dev/zero of=file \
bs=4k count=$((256*512)); sync; doneNevertheless, I agree that you suggested test will provide much more accurate. As stated before, my test is only designed for a minimal proof-of-concept.
Quoted 28 lines
Quoted 20 lines
If the underlying block device is slow enough, the squeezing overhead could be hidden. For the reason, I do this test for both a slow block device and a fast block device. I use a popular cloud block storage service, ebs[1] as a slow device and the ramdisk block device[2] for the fast device.
The results are as below. '
max_pgs' represents the value of theblkback.max_buffer_pagesparameter.On the slow block device
------------------------max_pgs Min Max Median Avg Stddev 0 38.7 45.8 38.7 40.12 3.1752165 1024 38.7 45.8 38.7 40.12 3.1752165No difference proven at 95.0% confidence
Quoted 7 lines
On the fast block device
------------------------max_pgs Min Max Median Avg Stddev 0 417 423 420 419.4 2.5099801 1024 414 425 416 417.8 4.4384682No difference proven at 95.0% confidence
Quoted 6 lines
This is intriguing, as it seems to prove that the usage of a cache of free pages is irrelevant performance wise.
The pool of free pages was introduced long ago, and it's possible that recent improvements to the balloon driver had made such pool useless, at which point it could be removed instead of worked around.
I guess the grant page allocation overhead in this test scenario is really small. In an absence of memory pressure, fragmentation, and NUMA imbalance, the latency of the page allocation ('get_page()') is very short, as it will success in the fast path.
Few years ago, I once measured the page allocation latency on my machine. Roughly speaking, it was about 1us in best case, 100us in worst case, and 5us in average. Please keep in mind that the measurement was not designed and performed in serious way. Thus the results could have profile overhead in it, though. While keeping that in mind, let's simply believe the number and ignore the latency of the block layer, blkback itself (including the grant mapping), and anything else including context switch, cache miss, but the allocation. In other words, suppose that the grant page allocation is only one source of the overhead. It will be able to achieve 1 million IOPS (4KB * 1MIOPS = 4 GB/s) in the best case, 200 thousand IOPS (800 MB/s) in average, and 10 thousand IOPS (40 MB/s) in worst case. Based on this coarse calculation, I think the test results is reasonable.
This also means that the effect of the blkback's free pages pool might be visible under page allocation fast path failure situation. Nevertheless, it would be also hard to measure that in micro level unless the measurement is well designed and controlled.
Quoted 4 lines
Do you think you could perform some more tests (as pointed out above against the block device to skip the fs overhead) and report back the results?
To be honest, I'm not sure whether additional tests are really necessary, because I think the dd test and the results explanation already makes some sense and provide the minimal proof of the concept. Also, this change is a fallback for the memory pressure situation, which is an error path in some point of view. Such errorneous situation might not happen frequently and if the situation is not solved in short time, something much worse (e.g., OOM kill of the user space xen control processes) than temporal I/O performance degradation could happen. Thus, I'm not sure whether such detailed performance measurement is necessary for this rare error handling change. The comment of 'xen_blkbk_unmap()' also says, "This could accumulate ops up to the batch size to reduce the number of hypercalls, but since this is only used in error paths there's no real need.".
That said, if you still want me to do the test, I will gladly do it. Note that since I'm now traveling US without the power cable of my now discharged laptop which is the only way to connect to my test environments in my office, I would not be able to do the additional test quickly. I assume I could do the test and give you the result within a couple of week. If you want me to post next version with minor changes such as commit message update meanwhile, please just let me know. Again, if you want it, I will gladly do it :)
Quoted 28 lines
Quoted 28 lines
In short, even worst case squeezing on ramdisk based fast block device makes no visible performance degradation. Please note that this is just a very simple and minimal test. On systems using super-fast block devices and a special I/O workload, the results might be different. If you have any doubt, test on your machine for your workload to find the optimal squeezing duration for you.
[1] https://aws.amazon.com/ebs/
[2] https://www.kernel.org/doc/html/latest/admin-guide/blockdev/ramdisk.htmlReviewed-by Juergen Gross <[email protected]>Signed-off-by SeongJae Park <[email protected]>--- drivers/block/xen-blkback/blkback.c | 22 ++++++++++++++++++++-- drivers/block/xen-blkback/common.h | 1 + drivers/block/xen-blkback/xenbus.c | 3 ++- 3 files changed, 23 insertions(+), 3 deletions(-)drivers/block/xen-blkback/blkback.c +0
-0@@ -142,6 +142,21 @@ static inline bool persistent_gnt_timeout(struct persistent_gnt *persistent_gnt) HZ * xen_blkif_pgrant_timeout); }
Quoted 9 lines
drivers/block/xen-blkback/blkback.c static inline bool persistent_gnt_timeout(struct persistent_gnt *persistent_gnt) · line 150
Quoted 6 lines
drivers/block/xen-blkback/blkback.c +6
-0+/* Once a memory pressure is detected, squeeze free page pools for a while. */ +static unsigned int buffer_squeeze_duration_ms =10 ; +module_param_named (buffer_squeeze_duration_ms, + buffer_squeeze_duration_ms, int,0644 ); +MODULE_PARM_DESC (buffer_squeeze_duration_ms, +"Duration in ms to squeeze pages buffer when a memory pressure is detected" );You should add a description about this parameter to
Documentation/ABI/testing/sysfs-driver-xen-blkback
Good point, I will.
Thanks,
SeongJae Park
Quoted 3 lines
Thanks, Roger.
Re: Re: [Xen-devel] [PATCH v7 2/3] xen/blkback: Squeeze page pools if a memory pressure is detected
Quoted 20 lines · SeongJae Park
Quoted 19 lines · On Thu, 12 Dec 2019 12:42:47 +0100 "Roger Pau Monné" <[email protected]>…
Quoted 8 lines
On the slow block device
------------------------max_pgs Min Max Median Avg Stddev 0 38.7 45.8 38.7 40.12 3.1752165 1024 38.7 45.8 38.7 40.12 3.1752165No difference proven at 95.0% confidence
Quoted 7 lines
On the fast block device
------------------------max_pgs Min Max Median Avg Stddev 0 417 423 420 419.4 2.5099801 1024 414 425 416 417.8 4.4384682No difference proven at 95.0% confidence
This is intriguing, as it seems to prove that the usage of a cache of free pages is irrelevant performance wise.
Quoted 8 lines
Quoted 3 lines
The pool of free pages was introduced long ago, and it's possible that recent improvements to the balloon driver had made such pool useless, at which point it could be removed instead of worked around.
I guess the grant page allocation overhead in this test scenario is really small. In an absence of memory pressure, fragmentation, and NUMA imbalance, the latency of the page allocation ('
get_page()') is very short, as it will success in the fast path.
The allocation of the pool of free pages involves more than get_page, it uses gnttab_alloc_pages which in the worse case will allocate a page and balloon it out issuing one hypercall.
Quoted 23 lines
Few years ago, I once measured the page allocation latency on my machine. Roughly speaking, it was about 1us in best case, 100us in worst case, and 5us in average. Please keep in mind that the measurement was not designed and performed in serious way. Thus the results could have profile overhead in it, though. While keeping that in mind, let's simply believe the number and ignore the latency of the block layer, blkback itself (including the grant mapping), and anything else including context switch, cache miss, but the allocation. In other words, suppose that the grant page allocation is only one source of the overhead. It will be able to achieve 1 million IOPS (4KB * 1MIOPS = 4 GB/s) in the best case, 200 thousand IOPS (800 MB/s) in average, and 10 thousand IOPS (40 MB/s) in worst case. Based on this coarse calculation, I think the test results is reasonable.
This also means that the effect of the blkback's free pages pool might be visible under page allocation fast path failure situation. Nevertheless, it would be also hard to measure that in micro level unless the measurement is well designed and controlled.
Quoted 4 lines
Do you think you could perform some more tests (as pointed out above against the block device to skip the fs overhead) and report back the results?
Quoted 9 lines
To be honest, I'm not sure whether additional tests are really necessary, because I think the
ddtest and the results explanation already makes some sense and provide the minimal proof of the concept. Also, this change is a fallback for the memory pressure situation, which is an error path in some point of view. Such errorneous situation might not happen frequently and if the situation is not solved in short time, something much worse (e.g., OOM kill of the user space xen control processes) than temporal I/O performance degradation could happen. Thus, I'm not sure whether such detailed performance measurement is necessary for this rare error handling change.
Right, my main concern is that we seem to be adding duck tape so things don't fall apart, but if such cache is really not beneficial from a performance PoV I would rather see it go away than adding more stuff to it in order to workaround corner cases like memory starvation.
Anyway, I guess we can take such change, but long term we need to look into fixing grants to not use ballooned pages, and figure out if the blkback free page cache is really useful or not.
Thanks, Roger.
Re: Re: [Xen-devel] [PATCH v7 2/3] xen/blkback: Squeeze page pools if a memory pressure is detected
Quoted 66 lines · On Thu, 12 Dec 2019 16:23:17 +0100 "Roger Pau Monné" <[email protected]>…
Quoted 20 lines · On Thu, 12 Dec 2019 12:42:47 +0100 "Roger Pau Monné" <[email protected]>…
Quoted 19 lines
Quoted 8 lines
On the slow block device
------------------------max_pgs Min Max Median Avg Stddev 0 38.7 45.8 38.7 40.12 3.1752165 1024 38.7 45.8 38.7 40.12 3.1752165No difference proven at 95.0% confidence
Quoted 7 lines
On the fast block device
------------------------max_pgs Min Max Median Avg Stddev 0 417 423 420 419.4 2.5099801 1024 414 425 416 417.8 4.4384682No difference proven at 95.0% confidence
This is intriguing, as it seems to prove that the usage of a cache of free pages is irrelevant performance wise.
Quoted 8 lines
Quoted 3 lines
The pool of free pages was introduced long ago, and it's possible that recent improvements to the balloon driver had made such pool useless, at which point it could be removed instead of worked around.
I guess the grant page allocation overhead in this test scenario is really small. In an absence of memory pressure, fragmentation, and NUMA imbalance, the latency of the page allocation ('
get_page()') is very short, as it will success in the fast path.The allocation of the pool of free pages involves more than
get_page, it usesgnttab_alloc_pageswhich in the worse case will allocate a page and balloon it out issuing one hypercall.Quoted 23 lines
Few years ago, I once measured the page allocation latency on my machine. Roughly speaking, it was about 1us in best case, 100us in worst case, and 5us in average. Please keep in mind that the measurement was not designed and performed in serious way. Thus the results could have profile overhead in it, though. While keeping that in mind, let's simply believe the number and ignore the latency of the block layer, blkback itself (including the grant mapping), and anything else including context switch, cache miss, but the allocation. In other words, suppose that the grant page allocation is only one source of the overhead. It will be able to achieve 1 million IOPS (4KB * 1MIOPS = 4 GB/s) in the best case, 200 thousand IOPS (800 MB/s) in average, and 10 thousand IOPS (40 MB/s) in worst case. Based on this coarse calculation, I think the test results is reasonable.
This also means that the effect of the blkback's free pages pool might be visible under page allocation fast path failure situation. Nevertheless, it would be also hard to measure that in micro level unless the measurement is well designed and controlled.
Quoted 4 lines
Do you think you could perform some more tests (as pointed out above against the block device to skip the fs overhead) and report back the results?
Quoted 9 lines
To be honest, I'm not sure whether additional tests are really necessary, because I think the
ddtest and the results explanation already makes some sense and provide the minimal proof of the concept. Also, this change is a fallback for the memory pressure situation, which is an error path in some point of view. Such errorneous situation might not happen frequently and if the situation is not solved in short time, something much worse (e.g., OOM kill of the user space xen control processes) than temporal I/O performance degradation could happen. Thus, I'm not sure whether such detailed performance measurement is necessary for this rare error handling change.
Quoted 5 lines
Right, my main concern is that we seem to be adding duck tape so things don't fall apart, but if such cache is really not beneficial from a performance PoV I would rather see it go away than adding more stuff to it in order to workaround corner cases like memory starvation.
Right, if the cache is really giving no benefit, it would be much better to simply remove it. However, as mentioned before, I'm not sure whether it is useless at all. Maybe we could do some more detailed test to know that, but it would be an out of scope of this patch.
Quoted 4 lines
Anyway, I guess we can take such change, but long term we need to look into fixing grants to not use ballooned pages, and figure out if the blkback free page cache is really useful or not.
Totally agreed.
Thanks,
SeongJae Park
Quoted 3 lines
Thanks, Roger.
Re: [Xen-devel] [PATCH v7 2/3] xen/blkback: Squeeze page pools if a memory pressure is detected
In reply to SeongJae ParkQuoted 11 lines · SeongJae Park
drivers/block/xen-blkback/blkback.c +3
-0@@ -142,6 +142,21 @@ static inline bool persistent_gnt_timeout(struct persistent_gnt *persistent_gnt) HZ * xen_blkif_pgrant_timeout); } +/* Once a memory pressure is detected, squeeze free page pools for a while. */ +static unsigned int buffer_squeeze_duration_ms =10 ; +module_param_named(buffer_squeeze_duration_ms,
drivers/block/xen-blkback/blkback.c static inline bool persistent_gnt_timeout(struct persistent_gnt *persistent_gnt) · line 157
Quoted 10 lines
drivers/block/xen-blkback/blkback.c +10
-0+ buffer_squeeze_duration_ms, int, 0644); +MODULE_PARM_DESC (buffer_squeeze_duration_ms, +"Duration in ms to squeeze pages buffer when a memory pressure is detected" ); + +static unsigned long buffer_squeeze_end; + +void xen_blkbk_reclaim_memory (struct xenbus_device *dev) +{ + buffer_squeeze_end = jiffies + +msecs_to_jiffies (buffer_squeeze_duration_ms);
I'm not sure this is fully correct. This function will be called for each blkback instance, but the timeout is stored in a global variable that's shared between all blkback instances. Shouldn't this timeout be stored in xen_blkif so each instance has it's own local variable?
Or else in the case you have 1k blkback instances the timeout is certainly going to be longer than expected, because each call to xen_blkbk_reclaim_memory will move it forward.
Thanks, Roger.
Re: Re: [Xen-devel] [PATCH v7 2/3] xen/blkback: Squeeze page pools if a memory pressure is detected
Quoted 27 lines · On Thu, 12 Dec 2019 16:27:57 +0100 "Roger Pau Monné" <[email protected]>…
Quoted 11 lines
drivers/block/xen-blkback/blkback.c +3
-0@@ -142,6 +142,21 @@ static inline bool persistent_gnt_timeout(struct persistent_gnt *persistent_gnt) HZ * xen_blkif_pgrant_timeout); } +/* Once a memory pressure is detected, squeeze free page pools for a while. */ +static unsigned int buffer_squeeze_duration_ms =10 ; +module_param_named(buffer_squeeze_duration_ms,drivers/block/xen-blkback/blkback.c static inline bool persistent_gnt_timeout(struct persistent_gnt *persistent_gnt) · line 157
Quoted 10 lines
drivers/block/xen-blkback/blkback.c +10
-0+ buffer_squeeze_duration_ms, int, 0644); +MODULE_PARM_DESC (buffer_squeeze_duration_ms, +"Duration in ms to squeeze pages buffer when a memory pressure is detected" ); + +static unsigned long buffer_squeeze_end; + +void xen_blkbk_reclaim_memory (struct xenbus_device *dev) +{ + buffer_squeeze_end = jiffies + +msecs_to_jiffies (buffer_squeeze_duration_ms);I'm not sure this is fully correct. This function will be called for each blkback instance, but the timeout is stored in a global variable that's shared between all blkback instances. Shouldn't this timeout be stored in
xen_blkifso each instance has it's own local variable?
Quoted 3 lines
Or else in the case you have 1k blkback instances the timeout is certainly going to be longer than expected, because each call to
xen_blkbk_reclaim_memorywill move it forward.
Agreed that. I think the extended timeout would not make a visible performance, though, because the time that 1k-loop take would be short enough to be ignored compared to the millisecond-scope duration.
I took this way because I wanted to minimize such structural changes as far as I can, as this is just a point-fix rather than ultimate solution. That said, it is not fully correct and very confusing. My another colleague also pointed out it in internal review. Correct solution would be to adding a variable in the struct as you suggested or avoiding duplicated update of the variable by initializing the variable once the squeezing duration passes. I would prefer the later way, as it is more straightforward and still not introducing structural change. For example, it might be like below:
drivers/block/xen-blkback/blkback.c +8 -4
@@ -152,8 +152,9 @@ static unsigned long buffer_squeeze_end;void xen_blkbk_reclaim_memory (struct xenbus_device *dev) { - buffer_squeeze_end = jiffies + -msecs_to_jiffies (buffer_squeeze_duration_ms); +if (!buffer_squeeze_end) + buffer_squeeze_end = jiffies + +msecs_to_jiffies (buffer_squeeze_duration_ms); }static inline int get_free_page (struct xen_blkif_ring *ring,struct page **page)
@@ -669,10 +670,13 @@ int xen_blkif_schedule(void *arg) }/* Shrink the free pages pool if it is too large. */ -if (time_before (jiffies, buffer_squeeze_end)) +if (time_before (jiffies, buffer_squeeze_end)) {shrink_free_pagepool (ring,0 ); -else + }else { +if (unlikely (buffer_squeeze_end)) + buffer_squeeze_end =0 ;shrink_free_pagepool (ring, max_buffer_pages); + }if (log_stats &&time_after (jiffies, ring->st_print))print_stats (ring);
May I ask you what way would you prefer?
Thanks,
SeongJae Park
Quoted 3 lines
Thanks, Roger.
Re: Re: [Xen-devel] [PATCH v7 2/3] xen/blkback: Squeeze page pools if a memory pressure is detected
Quoted 79 lines · SeongJae Park
Quoted 27 lines · On Thu, 12 Dec 2019 16:27:57 +0100 "Roger Pau Monné" <[email protected]>…
Quoted 11 lines
drivers/block/xen-blkback/blkback.c +3
-0@@ -142,6 +142,21 @@ static inline bool persistent_gnt_timeout(struct persistent_gnt *persistent_gnt) HZ * xen_blkif_pgrant_timeout); } +/* Once a memory pressure is detected, squeeze free page pools for a while. */ +static unsigned int buffer_squeeze_duration_ms =10 ; +module_param_named(buffer_squeeze_duration_ms,drivers/block/xen-blkback/blkback.c static inline bool persistent_gnt_timeout(struct persistent_gnt *persistent_gnt) · line 157
Quoted 10 lines
drivers/block/xen-blkback/blkback.c +10
-0+ buffer_squeeze_duration_ms, int, 0644); +MODULE_PARM_DESC (buffer_squeeze_duration_ms, +"Duration in ms to squeeze pages buffer when a memory pressure is detected" ); + +static unsigned long buffer_squeeze_end; + +void xen_blkbk_reclaim_memory (struct xenbus_device *dev) +{ + buffer_squeeze_end = jiffies + +msecs_to_jiffies (buffer_squeeze_duration_ms);I'm not sure this is fully correct. This function will be called for each blkback instance, but the timeout is stored in a global variable that's shared between all blkback instances. Shouldn't this timeout be stored in
xen_blkifso each instance has it's own local variable?Quoted 3 lines
Or else in the case you have 1k blkback instances the timeout is certainly going to be longer than expected, because each call to
xen_blkbk_reclaim_memorywill move it forward.Agreed that. I think the extended timeout would not make a visible performance, though, because the time that 1k-loop take would be short enough to be ignored compared to the millisecond-scope duration.
I took this way because I wanted to minimize such structural changes as far as I can, as this is just a point-fix rather than ultimate solution. That said, it is not fully correct and very confusing. My another colleague also pointed out it in internal review. Correct solution would be to adding a variable in the struct as you suggested or avoiding duplicated update of the variable by initializing the variable once the squeezing duration passes. I would prefer the later way, as it is more straightforward and still not introducing structural change. For example, it might be like below:
drivers/block/xen-blkback/blkback.c +8
-4@@ -152,8 +152,9 @@ static unsigned long buffer_squeeze_end;void xen_blkbk_reclaim_memory (struct xenbus_device *dev) { - buffer_squeeze_end = jiffies + -msecs_to_jiffies (buffer_squeeze_duration_ms); +if (!buffer_squeeze_end) + buffer_squeeze_end = jiffies + +msecs_to_jiffies (buffer_squeeze_duration_ms); }static inline int get_free_page (struct xen_blkif_ring *ring,struct page **page)@@ -669,10 +670,13 @@ int xen_blkif_schedule(void *arg) }/* Shrink the free pages pool if it is too large. */ -if (time_before (jiffies, buffer_squeeze_end)) +if (time_before (jiffies, buffer_squeeze_end)) {shrink_free_pagepool (ring,0 ); -else + }else { +if (unlikely (buffer_squeeze_end)) + buffer_squeeze_end =0 ;shrink_free_pagepool (ring, max_buffer_pages); + }if (log_stats &&time_after (jiffies, ring->st_print))print_stats (ring);
Quoted 1 line
May I ask you what way would you prefer?
I'm not particularly found of this approach, as I think it's racy. Ie:
you would have to add some kind of lock to make sure the contents of buffer_squeeze_end stay unmodified during the read and set cycle, or else xen_blkif_schedule will race with xen_blkbk_reclaim_memory.
This is likely not a big deal ATM since the code will work as expected in most cases AFAICT, but I would still prefer to have a per-instance buffer_squeeze_end added to xen_blkif, given that the callback is per-instance. I wouldn't call it a structural change, it's just adding a variable to a struct instead of having a shared one, but the code is almost the same as the current version.
Thanks, Roger.
Jürgen Groß
maintainerRe: [Xen-devel] [PATCH v7 2/3] xen/blkback: Squeeze page pools if a memory pressure is detected
Quoted 87 lines · Roger Pau Monné
Quoted 79 lines · SeongJae Park
Quoted 27 lines · On Thu, 12 Dec 2019 16:27:57 +0100 "Roger Pau Monné" <[email protected]>…
Quoted 11 lines
drivers/block/xen-blkback/blkback.c +3
-0@@ -142,6 +142,21 @@ static inline bool persistent_gnt_timeout(struct persistent_gnt *persistent_gnt) HZ * xen_blkif_pgrant_timeout); } +/* Once a memory pressure is detected, squeeze free page pools for a while. */ +static unsigned int buffer_squeeze_duration_ms =10 ; +module_param_named(buffer_squeeze_duration_ms,drivers/block/xen-blkback/blkback.c static inline bool persistent_gnt_timeout(struct persistent_gnt *persistent_gnt) · line 157
Quoted 10 lines
drivers/block/xen-blkback/blkback.c +10
-0+ buffer_squeeze_duration_ms, int, 0644); +MODULE_PARM_DESC (buffer_squeeze_duration_ms, +"Duration in ms to squeeze pages buffer when a memory pressure is detected" ); + +static unsigned long buffer_squeeze_end; + +void xen_blkbk_reclaim_memory (struct xenbus_device *dev) +{ + buffer_squeeze_end = jiffies + +msecs_to_jiffies (buffer_squeeze_duration_ms);I'm not sure this is fully correct. This function will be called for each blkback instance, but the timeout is stored in a global variable that's shared between all blkback instances. Shouldn't this timeout be stored in
xen_blkifso each instance has it's own local variable?Quoted 3 lines
Or else in the case you have 1k blkback instances the timeout is certainly going to be longer than expected, because each call to
xen_blkbk_reclaim_memorywill move it forward.Agreed that. I think the extended timeout would not make a visible performance, though, because the time that 1k-loop take would be short enough to be ignored compared to the millisecond-scope duration.
I took this way because I wanted to minimize such structural changes as far as I can, as this is just a point-fix rather than ultimate solution. That said, it is not fully correct and very confusing. My another colleague also pointed out it in internal review. Correct solution would be to adding a variable in the struct as you suggested or avoiding duplicated update of the variable by initializing the variable once the squeezing duration passes. I would prefer the later way, as it is more straightforward and still not introducing structural change. For example, it might be like below:
drivers/block/xen-blkback/blkback.c +8
-4@@ -152,8 +152,9 @@ static unsigned long buffer_squeeze_end;void xen_blkbk_reclaim_memory (struct xenbus_device *dev) { - buffer_squeeze_end = jiffies + -msecs_to_jiffies (buffer_squeeze_duration_ms); +if (!buffer_squeeze_end) + buffer_squeeze_end = jiffies + +msecs_to_jiffies (buffer_squeeze_duration_ms); }static inline int get_free_page (struct xen_blkif_ring *ring,struct page **page)@@ -669,10 +670,13 @@ int xen_blkif_schedule(void *arg) }/* Shrink the free pages pool if it is too large. */ -if (time_before (jiffies, buffer_squeeze_end)) +if (time_before (jiffies, buffer_squeeze_end)) {shrink_free_pagepool (ring,0 ); -else + }else { +if (unlikely (buffer_squeeze_end)) + buffer_squeeze_end =0 ;shrink_free_pagepool (ring, max_buffer_pages); + }if (log_stats &&time_after (jiffies, ring->st_print))print_stats (ring);Quoted 1 line
May I ask you what way would you prefer?
I'm not particularly found of this approach, as I think it's racy. Ie:
you would have to add some kind of lock to make sure the contents ofbuffer_squeeze_endstay unmodified during the read and set cycle, or elsexen_blkif_schedulewill race withxen_blkbk_reclaim_memory.
Quoted 6 lines
This is likely not a big deal ATM since the code will work as expected in most cases AFAICT, but I would still prefer to have a per-instance
buffer_squeeze_endadded toxen_blkif, given that the callback is per-instance. I wouldn't call it a structural change, it's just adding a variable to a struct instead of having a shared one, but the code is almost the same as the current version.
FWIW, I agree.
Juergen
Re: [Xen-devel] [PATCH v7 2/3] xen/blkback: Squeeze page pools if a memory pressure is detected
Quoted 89 lines · Jürgen Groß
Quoted 87 lines · Roger Pau Monné
Quoted 79 lines · SeongJae Park
Quoted 27 lines · On Thu, 12 Dec 2019 16:27:57 +0100 "Roger Pau Monné" <[email protected]>…
Quoted 11 lines
drivers/block/xen-blkback/blkback.c +3
-0@@ -142,6 +142,21 @@ static inline bool persistent_gnt_timeout(struct persistent_gnt *persistent_gnt) HZ * xen_blkif_pgrant_timeout); } +/* Once a memory pressure is detected, squeeze free page pools for a while. */ +static unsigned int buffer_squeeze_duration_ms =10 ; +module_param_named(buffer_squeeze_duration_ms,drivers/block/xen-blkback/blkback.c static inline bool persistent_gnt_timeout(struct persistent_gnt *persistent_gnt) · line 157
Quoted 10 lines
drivers/block/xen-blkback/blkback.c +10
-0+ buffer_squeeze_duration_ms, int, 0644); +MODULE_PARM_DESC (buffer_squeeze_duration_ms, +"Duration in ms to squeeze pages buffer when a memory pressure is detected" ); + +static unsigned long buffer_squeeze_end; + +void xen_blkbk_reclaim_memory (struct xenbus_device *dev) +{ + buffer_squeeze_end = jiffies + +msecs_to_jiffies (buffer_squeeze_duration_ms);I'm not sure this is fully correct. This function will be called for each blkback instance, but the timeout is stored in a global variable that's shared between all blkback instances. Shouldn't this timeout be stored in
xen_blkifso each instance has it's own local variable?Quoted 3 lines
Or else in the case you have 1k blkback instances the timeout is certainly going to be longer than expected, because each call to
xen_blkbk_reclaim_memorywill move it forward.Agreed that. I think the extended timeout would not make a visible performance, though, because the time that 1k-loop take would be short enough to be ignored compared to the millisecond-scope duration.
I took this way because I wanted to minimize such structural changes as far as I can, as this is just a point-fix rather than ultimate solution. That said, it is not fully correct and very confusing. My another colleague also pointed out it in internal review. Correct solution would be to adding a variable in the struct as you suggested or avoiding duplicated update of the variable by initializing the variable once the squeezing duration passes. I would prefer the later way, as it is more straightforward and still not introducing structural change. For example, it might be like below:
drivers/block/xen-blkback/blkback.c +8
-4@@ -152,8 +152,9 @@ static unsigned long buffer_squeeze_end;void xen_blkbk_reclaim_memory (struct xenbus_device *dev) { - buffer_squeeze_end = jiffies + -msecs_to_jiffies (buffer_squeeze_duration_ms); +if (!buffer_squeeze_end) + buffer_squeeze_end = jiffies + +msecs_to_jiffies (buffer_squeeze_duration_ms); }static inline int get_free_page (struct xen_blkif_ring *ring,struct page **page)@@ -669,10 +670,13 @@ int xen_blkif_schedule(void *arg) }/* Shrink the free pages pool if it is too large. */ -if (time_before (jiffies, buffer_squeeze_end)) +if (time_before (jiffies, buffer_squeeze_end)) {shrink_free_pagepool (ring,0 ); -else + }else { +if (unlikely (buffer_squeeze_end)) + buffer_squeeze_end =0 ;shrink_free_pagepool (ring, max_buffer_pages); + }if (log_stats &&time_after (jiffies, ring->st_print))print_stats (ring);Quoted 1 line
May I ask you what way would you prefer?
I'm not particularly found of this approach, as I think it's racy. Ie:
you would have to add some kind of lock to make sure the contents ofbuffer_squeeze_endstay unmodified during the read and set cycle, or elsexen_blkif_schedulewill race withxen_blkbk_reclaim_memory.
Quoted 8 lines
Quoted 6 lines
This is likely not a big deal ATM since the code will work as expected in most cases AFAICT, but I would still prefer to have a per-instance
buffer_squeeze_endadded toxen_blkif, given that the callback is per-instance. I wouldn't call it a structural change, it's just adding a variable to a struct instead of having a shared one, but the code is almost the same as the current version.FWIW, I agree.
Agreed, will send v8 soon!
Thanks,
SeongJae Park
Quoted 3 lines
Juergen
PATCH v7 3/3 xen/blkback: Remove unnecessary static variable name prefixes
In reply to SeongJae ParkA few of static variables in blkback have 'xen_blkif_' prefix, though it is unnecessary for static variables. This commit removes such prefixes.
--- drivers/block/xen-blkback/blkback.c | 37 +++++++++++++---------------- 1 file changed, 17 insertions(+), 20 deletions(-)
drivers/block/xen-blkback/blkback.c +17 -20
@@ -62,8 +62,8 @@* IO workloads. */ -static int xen_blkif_max_buffer_pages =1024 ; -module_param_named (max_buffer_pages, xen_blkif_max_buffer_pages, int,0644 ); +static int max_buffer_pages =1024 ; +module_param_named (max_buffer_pages, max_buffer_pages, int,0644 );MODULE_PARM_DESC (max_buffer_pages,"Maximum number of free pages to keep in each block backend buffer" );
@@ -78,8 +78,8 @@ MODULE_PARM_DESC(max_buffer_pages,* algorithm. */ -static int xen_blkif_max_pgrants =1056 ; -module_param_named (max_persistent_grants, xen_blkif_max_pgrants, int,0644 ); +static int max_pgrants =1056 ; +module_param_named (max_persistent_grants, max_pgrants, int,0644 );MODULE_PARM_DESC (max_persistent_grants,"Maximum number of grants to map persistently" );
@@ -88,8 +88,8 @@ MODULE_PARM_DESC(max_persistent_grants,* use. The time is in seconds, 0 means indefinitely long. */ -static unsigned int xen_blkif_pgrant_timeout =60 ; -module_param_named (persistent_grant_unused_seconds, xen_blkif_pgrant_timeout, +static unsigned int pgrant_timeout =60 ; +module_param_named (persistent_grant_unused_seconds, pgrant_timeout, uint,0644 );MODULE_PARM_DESC (persistent_grant_unused_seconds, "Time in seconds an unused persistent grant is allowed to "
@@ -137,9 +137,8 @@ module_param(log_stats, int, 0644);static inline bool persistent_gnt_timeout (struct persistent_gnt *persistent_gnt) { -return xen_blkif_pgrant_timeout && - (jiffies - persistent_gnt->last_used >= -HZ * xen_blkif_pgrant_timeout); +return pgrant_timeout && (jiffies - persistent_gnt->last_used >= +HZ * pgrant_timeout); }/* Once a memory pressure is detected, squeeze free page pools for a while. */
@@ -249,7 +248,7 @@ static int add_persistent_gnt(struct xen_blkif_ring *ring,struct persistent_gnt *this;struct xen_blkif *blkif = ring->blkif; -if (ring->persistent_gnt_c >= xen_blkif_max_pgrants) { +if (ring->persistent_gnt_c >= max_pgrants) {if (!blkif->vbd.overflow_max_grants) blkif->vbd.overflow_max_grants =1 ;return -EBUSY ;
@@ -412,14 +411,13 @@ static void purge_persistent_gnt(struct xen_blkif_ring *ring) goto out; } -if (ring->persistent_gnt_c < xen_blkif_max_pgrants || - (ring->persistent_gnt_c == xen_blkif_max_pgrants && +if (ring->persistent_gnt_c < max_pgrants || + (ring->persistent_gnt_c == max_pgrants && !ring->blkif->vbd.overflow_max_grants)) { num_clean =0 ; }else { - num_clean = (xen_blkif_max_pgrants /100 ) *LRU_PERCENT_CLEAN ; - num_clean = ring->persistent_gnt_c - xen_blkif_max_pgrants + - num_clean; + num_clean = (max_pgrants /100 ) *LRU_PERCENT_CLEAN ; + num_clean = ring->persistent_gnt_c - max_pgrants + num_clean; num_clean =min (ring->persistent_gnt_c, num_clean);pr_debug ("Going to purge at least %u persistent grants\n" , num_clean);
@@ -614,8 +612,7 @@ static void print_stats(struct xen_blkif_ring *ring) current->comm, ring->st_oo_req, ring->st_rd_req, ring->st_wr_req, ring->st_f_req, ring->st_ds_req, - ring->persistent_gnt_c, - xen_blkif_max_pgrants); + ring->persistent_gnt_c, max_pgrants); ring->st_print = jiffies +msecs_to_jiffies (10 *1000 ); ring->st_rd_req =0 ; ring->st_wr_req =0 ;
@@ -675,7 +672,7 @@ int xen_blkif_schedule(void *arg)if (time_before (jiffies, buffer_squeeze_end))shrink_free_pagepool (ring,0 );else -shrink_free_pagepool (ring, xen_blkif_max_buffer_pages); +shrink_free_pagepool (ring, max_buffer_pages);if (log_stats &&time_after (jiffies, ring->st_print))print_stats (ring);
@@ -902,7 +899,7 @@ static int xen_blkbk_map(struct xen_blkif_ring *ring,continue ; }if (use_persistent_gnts && - ring->persistent_gnt_c < xen_blkif_max_pgrants) { + ring->persistent_gnt_c < max_pgrants) {/* * We are using persistent grants, the grant is * not mapped but we might have room for it.
@@ -929,7 +926,7 @@ static int xen_blkbk_map(struct xen_blkif_ring *ring, pages[seg_idx]->persistent_gnt = persistent_gnt;pr_debug ("grant %u added to the tree of persistent grants, using %u/%u\n" , persistent_gnt->gnt, ring->persistent_gnt_c, - xen_blkif_max_pgrants); + max_pgrants); goto next; }if (use_persistent_gnts && !blkif->vbd.overflow_max_grants) {
Signature · 1 line
2.17.1