Re: [PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Vlastimil Babka]

On 12/23/2015 07:57 AM, Joonsoo Kim wrote:

What are the cases where pageblock_pfn_to_page() is used for a subset of
the pageblock and the result would be problematic for compaction?  I.e.,
do we actually care to use pageblocks that are not contiguous at all?


The problematic pageblocks are those that have pages from more than one zone in
them, so we just skip them. Supposedly that can only happen by switching once
between two zones somewhere in the middle of the pageblock, so it's sufficient
to check first and last pfn and compare their zones. So using
pageblock_pfn_to_page() on a subset from compaction would be wrong. Holes (==no
pages) within pageblock is a different thing checked by pfn_valid_within()
(#defined out on archs where such holes cannot happen) when scanning the block.

That's why I'm not entirely happy with how the patch conflates both the
first/last pfn's zone checks and pfn_valid_within() checks. Yes, a fully
contiguous zone does *imply* that pageblock_pfn_to_page() doesn't have to check
first/last pfn for a matching zone. But it's not *equality*. And any (now just
*potential*) user of pageblock_pfn_to_page() with pfn's different than
first/last pfn of a pageblock is likely wrong.


Now, I understand your concern. What makes me mislead is that
3 of 4 callers to pageblock_pfn_to_page() in compaction.c could call it with
non-pageblock boundary pfn.


Oh, I thought you were talking about potential new callers, now that the 
function was exported. So let's see about the existing callers:


isolate_migratepages() - first pfn can be non-boundary when restarting 
from a middle of pageblock, that's true. But it means the pageblock has 
already passed the check in previous call where it was boundary, so it's 
safe. Worst can happen that the restarting pfn will be in a 
intra-pageblock hole so pageblock will be falsely skipped over.


isolate_freepages() - always boundary AFAICS?

isolate_migratepages_range() and isolate_freepages_range() - yeah the 
CMA parts say it doesn't have to be aligned, I don't know about actual users



Maybe, they should be fixed first.


It would be probably best, even for isolate_migratepages() for 
consistency and less-surprisibility.



Then, yes. I can
separate first/last pfn's zone checks and pfn_valid_within() checks.
If then, would you be entirely happy? :)


Maybe, if the patch also made me a coffee :P


Thanks.


Thanks!

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Re: [PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Vlastimil Babka]

On 12/23/2015 07:57 AM, Joonsoo Kim wrote:

What are the cases where pageblock_pfn_to_page() is used for a subset of
the pageblock and the result would be problematic for compaction?  I.e.,
do we actually care to use pageblocks that are not contiguous at all?


The problematic pageblocks are those that have pages from more than one zone in
them, so we just skip them. Supposedly that can only happen by switching once
between two zones somewhere in the middle of the pageblock, so it's sufficient
to check first and last pfn and compare their zones. So using
pageblock_pfn_to_page() on a subset from compaction would be wrong. Holes (==no
pages) within pageblock is a different thing checked by pfn_valid_within()
(#defined out on archs where such holes cannot happen) when scanning the block.

That's why I'm not entirely happy with how the patch conflates both the
first/last pfn's zone checks and pfn_valid_within() checks. Yes, a fully
contiguous zone does *imply* that pageblock_pfn_to_page() doesn't have to check
first/last pfn for a matching zone. But it's not *equality*. And any (now just
*potential*) user of pageblock_pfn_to_page() with pfn's different than
first/last pfn of a pageblock is likely wrong.


Now, I understand your concern. What makes me mislead is that
3 of 4 callers to pageblock_pfn_to_page() in compaction.c could call it with
non-pageblock boundary pfn.


Oh, I thought you were talking about potential new callers, now that the 
function was exported. So let's see about the existing callers:


isolate_migratepages() - first pfn can be non-boundary when restarting 
from a middle of pageblock, that's true. But it means the pageblock has 
already passed the check in previous call where it was boundary, so it's 
safe. Worst can happen that the restarting pfn will be in a 
intra-pageblock hole so pageblock will be falsely skipped over.


isolate_freepages() - always boundary AFAICS?

isolate_migratepages_range() and isolate_freepages_range() - yeah the 
CMA parts say it doesn't have to be aligned, I don't know about actual users



Maybe, they should be fixed first.


It would be probably best, even for isolate_migratepages() for 
consistency and less-surprisibility.



Then, yes. I can
separate first/last pfn's zone checks and pfn_valid_within() checks.
If then, would you be entirely happy? :)


Maybe, if the patch also made me a coffee :P


Thanks.


Thanks!

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Re: [PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Joonsoo Kim]

On Wed, Dec 23, 2015 at 07:14:21AM +0100, Vlastimil Babka wrote:
> On 22.12.2015 23:17, David Rientjes wrote:
> > On Mon, 21 Dec 2015, Joonsoo Kim wrote:
> > 
> >> Before vs After
> >> Max: 1096 MB/s vs 1325 MB/s
> >> Min: 635 MB/s 1015 MB/s
> >> Avg: 899 MB/s 1194 MB/s
> >>
> >> Avg is improved by roughly 30% [2].
> >>
> > 
> > Wow, ok!
> > 
> > I'm wondering if it would be better to maintain this as a characteristic 
> > of each pageblock rather than each zone.  Have you tried to introduce a 
> > couple new bits to pageblock_bits that would track (1) if a cached value 
> > makes sense and (2) if the pageblock is contiguous?  On the first call to 
> > pageblock_pfn_to_page(), set the first bit, PB_cached, and set the second 
> > bit, PB_contiguous, iff it is contiguous.  On subsequent calls, if 
> > PB_cached is true, then return PB_contiguous.  On memory hot-add or 
> > remove (or init), clear PB_cached.
> 
> I can imagine these bitmap operation to be as expensive as what
> __pageblock_pfn_to_page() does (or close)? But if not, we could also just be a
> bit smarter about PG_skip and check that before doing pfn_to_page.

Although I don't think carefully, to get PB_xxx, we need to check pfn's zone
and it requires pfn_valid() and pfn_to_page(). So, I guess that cost would be
same or half compared to cost of __pageblock_pfn_to_page().

> 
> > What are the cases where pageblock_pfn_to_page() is used for a subset of 
> > the pageblock and the result would be problematic for compaction?  I.e., 
> > do we actually care to use pageblocks that are not contiguous at all?
> 
> The problematic pageblocks are those that have pages from more than one zone 
> in
> them, so we just skip them. Supposedly that can only happen by switching once
> between two zones somewhere in the middle of the pageblock, so it's sufficient
> to check first and last pfn and compare their zones. So using
> pageblock_pfn_to_page() on a subset from compaction would be wrong. Holes 
> (==no
> pages) within pageblock is a different thing checked by pfn_valid_within()
> (#defined out on archs where such holes cannot happen) when scanning the 
> block.
> 
> That's why I'm not entirely happy with how the patch conflates both the
> first/last pfn's zone checks and pfn_valid_within() checks. Yes, a fully
> contiguous zone does *imply* that pageblock_pfn_to_page() doesn't have to 
> check
> first/last pfn for a matching zone. But it's not *equality*. And any (now just
> *potential*) user of pageblock_pfn_to_page() with pfn's different than
> first/last pfn of a pageblock is likely wrong.

Now, I understand your concern. What makes me mislead is that
3 of 4 callers to pageblock_pfn_to_page() in compaction.c could call it with
non-pageblock boundary pfn. Maybe, they should be fixed first. Then, yes. I can
separate first/last pfn's zone checks and pfn_valid_within() checks.
If then, would you be entirely happy? :)

Thanks.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Re: [PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Vlastimil Babka]

On 22.12.2015 23:17, David Rientjes wrote:
> On Mon, 21 Dec 2015, Joonsoo Kim wrote:
> 
>> Before vs After
>> Max: 1096 MB/s vs 1325 MB/s
>> Min: 635 MB/s 1015 MB/s
>> Avg: 899 MB/s 1194 MB/s
>>
>> Avg is improved by roughly 30% [2].
>>
> 
> Wow, ok!
> 
> I'm wondering if it would be better to maintain this as a characteristic 
> of each pageblock rather than each zone.  Have you tried to introduce a 
> couple new bits to pageblock_bits that would track (1) if a cached value 
> makes sense and (2) if the pageblock is contiguous?  On the first call to 
> pageblock_pfn_to_page(), set the first bit, PB_cached, and set the second 
> bit, PB_contiguous, iff it is contiguous.  On subsequent calls, if 
> PB_cached is true, then return PB_contiguous.  On memory hot-add or 
> remove (or init), clear PB_cached.

I can imagine these bitmap operation to be as expensive as what
__pageblock_pfn_to_page() does (or close)? But if not, we could also just be a
bit smarter about PG_skip and check that before doing pfn_to_page.

> What are the cases where pageblock_pfn_to_page() is used for a subset of 
> the pageblock and the result would be problematic for compaction?  I.e., 
> do we actually care to use pageblocks that are not contiguous at all?

The problematic pageblocks are those that have pages from more than one zone in
them, so we just skip them. Supposedly that can only happen by switching once
between two zones somewhere in the middle of the pageblock, so it's sufficient
to check first and last pfn and compare their zones. So using
pageblock_pfn_to_page() on a subset from compaction would be wrong. Holes (==no
pages) within pageblock is a different thing checked by pfn_valid_within()
(#defined out on archs where such holes cannot happen) when scanning the block.

That's why I'm not entirely happy with how the patch conflates both the
first/last pfn's zone checks and pfn_valid_within() checks. Yes, a fully
contiguous zone does *imply* that pageblock_pfn_to_page() doesn't have to check
first/last pfn for a matching zone. But it's not *equality*. And any (now just
*potential*) user of pageblock_pfn_to_page() with pfn's different than
first/last pfn of a pageblock is likely wrong.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Re: [PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[David Rientjes]

On Mon, 21 Dec 2015, Joonsoo Kim wrote:

> There is a performance drop report due to hugepage allocation and in there
> half of cpu time are spent on pageblock_pfn_to_page() in compaction [1].
> In that workload, compaction is triggered to make hugepage but most of
> pageblocks are un-available for compaction due to pageblock type and
> skip bit so compaction usually fails. Most costly operations in this case
> is to find valid pageblock while scanning whole zone range. To check
> if pageblock is valid to compact, valid pfn within pageblock is required
> and we can obtain it by calling pageblock_pfn_to_page(). This function
> checks whether pageblock is in a single zone and return valid pfn
> if possible. Problem is that we need to check it every time before
> scanning pageblock even if we re-visit it and this turns out to
> be very expensive in this workload.
> 
> Although we have no way to skip this pageblock check in the system
> where hole exists at arbitrary position, we can use cached value for
> zone continuity and just do pfn_to_page() in the system where hole doesn't
> exist. This optimization considerably speeds up in above workload.
> 
> Before vs After
> Max: 1096 MB/s vs 1325 MB/s
> Min: 635 MB/s 1015 MB/s
> Avg: 899 MB/s 1194 MB/s
> 
> Avg is improved by roughly 30% [2].
> 

Wow, ok!

I'm wondering if it would be better to maintain this as a characteristic 
of each pageblock rather than each zone.  Have you tried to introduce a 
couple new bits to pageblock_bits that would track (1) if a cached value 
makes sense and (2) if the pageblock is contiguous?  On the first call to 
pageblock_pfn_to_page(), set the first bit, PB_cached, and set the second 
bit, PB_contiguous, iff it is contiguous.  On subsequent calls, if 
PB_cached is true, then return PB_contiguous.  On memory hot-add or 
remove (or init), clear PB_cached.

What are the cases where pageblock_pfn_to_page() is used for a subset of 
the pageblock and the result would be problematic for compaction?  I.e., 
do we actually care to use pageblocks that are not contiguous at all?
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Re: [PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[David Rientjes]

On Mon, 21 Dec 2015, Joonsoo Kim wrote:

> There is a performance drop report due to hugepage allocation and in there
> half of cpu time are spent on pageblock_pfn_to_page() in compaction [1].
> In that workload, compaction is triggered to make hugepage but most of
> pageblocks are un-available for compaction due to pageblock type and
> skip bit so compaction usually fails. Most costly operations in this case
> is to find valid pageblock while scanning whole zone range. To check
> if pageblock is valid to compact, valid pfn within pageblock is required
> and we can obtain it by calling pageblock_pfn_to_page(). This function
> checks whether pageblock is in a single zone and return valid pfn
> if possible. Problem is that we need to check it every time before
> scanning pageblock even if we re-visit it and this turns out to
> be very expensive in this workload.
> 
> Although we have no way to skip this pageblock check in the system
> where hole exists at arbitrary position, we can use cached value for
> zone continuity and just do pfn_to_page() in the system where hole doesn't
> exist. This optimization considerably speeds up in above workload.
> 
> Before vs After
> Max: 1096 MB/s vs 1325 MB/s
> Min: 635 MB/s 1015 MB/s
> Avg: 899 MB/s 1194 MB/s
> 
> Avg is improved by roughly 30% [2].
> 

Wow, ok!

I'm wondering if it would be better to maintain this as a characteristic 
of each pageblock rather than each zone.  Have you tried to introduce a 
couple new bits to pageblock_bits that would track (1) if a cached value 
makes sense and (2) if the pageblock is contiguous?  On the first call to 
pageblock_pfn_to_page(), set the first bit, PB_cached, and set the second 
bit, PB_contiguous, iff it is contiguous.  On subsequent calls, if 
PB_cached is true, then return PB_contiguous.  On memory hot-add or 
remove (or init), clear PB_cached.

What are the cases where pageblock_pfn_to_page() is used for a subset of 
the pageblock and the result would be problematic for compaction?  I.e., 
do we actually care to use pageblocks that are not contiguous at all?
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Re: [PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Vlastimil Babka]

On 22.12.2015 23:17, David Rientjes wrote:
> On Mon, 21 Dec 2015, Joonsoo Kim wrote:
> 
>> Before vs After
>> Max: 1096 MB/s vs 1325 MB/s
>> Min: 635 MB/s 1015 MB/s
>> Avg: 899 MB/s 1194 MB/s
>>
>> Avg is improved by roughly 30% [2].
>>
> 
> Wow, ok!
> 
> I'm wondering if it would be better to maintain this as a characteristic 
> of each pageblock rather than each zone.  Have you tried to introduce a 
> couple new bits to pageblock_bits that would track (1) if a cached value 
> makes sense and (2) if the pageblock is contiguous?  On the first call to 
> pageblock_pfn_to_page(), set the first bit, PB_cached, and set the second 
> bit, PB_contiguous, iff it is contiguous.  On subsequent calls, if 
> PB_cached is true, then return PB_contiguous.  On memory hot-add or 
> remove (or init), clear PB_cached.

I can imagine these bitmap operation to be as expensive as what
__pageblock_pfn_to_page() does (or close)? But if not, we could also just be a
bit smarter about PG_skip and check that before doing pfn_to_page.

> What are the cases where pageblock_pfn_to_page() is used for a subset of 
> the pageblock and the result would be problematic for compaction?  I.e., 
> do we actually care to use pageblocks that are not contiguous at all?

The problematic pageblocks are those that have pages from more than one zone in
them, so we just skip them. Supposedly that can only happen by switching once
between two zones somewhere in the middle of the pageblock, so it's sufficient
to check first and last pfn and compare their zones. So using
pageblock_pfn_to_page() on a subset from compaction would be wrong. Holes (==no
pages) within pageblock is a different thing checked by pfn_valid_within()
(#defined out on archs where such holes cannot happen) when scanning the block.

That's why I'm not entirely happy with how the patch conflates both the
first/last pfn's zone checks and pfn_valid_within() checks. Yes, a fully
contiguous zone does *imply* that pageblock_pfn_to_page() doesn't have to check
first/last pfn for a matching zone. But it's not *equality*. And any (now just
*potential*) user of pageblock_pfn_to_page() with pfn's different than
first/last pfn of a pageblock is likely wrong.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Re: [PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Joonsoo Kim]

On Wed, Dec 23, 2015 at 07:14:21AM +0100, Vlastimil Babka wrote:
> On 22.12.2015 23:17, David Rientjes wrote:
> > On Mon, 21 Dec 2015, Joonsoo Kim wrote:
> > 
> >> Before vs After
> >> Max: 1096 MB/s vs 1325 MB/s
> >> Min: 635 MB/s 1015 MB/s
> >> Avg: 899 MB/s 1194 MB/s
> >>
> >> Avg is improved by roughly 30% [2].
> >>
> > 
> > Wow, ok!
> > 
> > I'm wondering if it would be better to maintain this as a characteristic 
> > of each pageblock rather than each zone.  Have you tried to introduce a 
> > couple new bits to pageblock_bits that would track (1) if a cached value 
> > makes sense and (2) if the pageblock is contiguous?  On the first call to 
> > pageblock_pfn_to_page(), set the first bit, PB_cached, and set the second 
> > bit, PB_contiguous, iff it is contiguous.  On subsequent calls, if 
> > PB_cached is true, then return PB_contiguous.  On memory hot-add or 
> > remove (or init), clear PB_cached.
> 
> I can imagine these bitmap operation to be as expensive as what
> __pageblock_pfn_to_page() does (or close)? But if not, we could also just be a
> bit smarter about PG_skip and check that before doing pfn_to_page.

Although I don't think carefully, to get PB_xxx, we need to check pfn's zone
and it requires pfn_valid() and pfn_to_page(). So, I guess that cost would be
same or half compared to cost of __pageblock_pfn_to_page().

> 
> > What are the cases where pageblock_pfn_to_page() is used for a subset of 
> > the pageblock and the result would be problematic for compaction?  I.e., 
> > do we actually care to use pageblocks that are not contiguous at all?
> 
> The problematic pageblocks are those that have pages from more than one zone 
> in
> them, so we just skip them. Supposedly that can only happen by switching once
> between two zones somewhere in the middle of the pageblock, so it's sufficient
> to check first and last pfn and compare their zones. So using
> pageblock_pfn_to_page() on a subset from compaction would be wrong. Holes 
> (==no
> pages) within pageblock is a different thing checked by pfn_valid_within()
> (#defined out on archs where such holes cannot happen) when scanning the 
> block.
> 
> That's why I'm not entirely happy with how the patch conflates both the
> first/last pfn's zone checks and pfn_valid_within() checks. Yes, a fully
> contiguous zone does *imply* that pageblock_pfn_to_page() doesn't have to 
> check
> first/last pfn for a matching zone. But it's not *equality*. And any (now just
> *potential*) user of pageblock_pfn_to_page() with pfn's different than
> first/last pfn of a pageblock is likely wrong.

Now, I understand your concern. What makes me mislead is that
3 of 4 callers to pageblock_pfn_to_page() in compaction.c could call it with
non-pageblock boundary pfn. Maybe, they should be fixed first. Then, yes. I can
separate first/last pfn's zone checks and pfn_valid_within() checks.
If then, would you be entirely happy? :)

Thanks.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Re: [PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Joonsoo Kim]

2015-12-21 21:18 GMT+09:00 Joonsoo Kim :
> 2015-12-21 19:46 GMT+09:00 Vlastimil Babka :
>> On 12/21/2015 07:13 AM, Joonsoo Kim wrote:
>>>
>>> There is a performance drop report due to hugepage allocation and in there
>>> half of cpu time are spent on pageblock_pfn_to_page() in compaction [1].
>>> In that workload, compaction is triggered to make hugepage but most of
>>> pageblocks are un-available for compaction due to pageblock type and
>>> skip bit so compaction usually fails. Most costly operations in this case
>>> is to find valid pageblock while scanning whole zone range. To check
>>> if pageblock is valid to compact, valid pfn within pageblock is required
>>> and we can obtain it by calling pageblock_pfn_to_page(). This function
>>> checks whether pageblock is in a single zone and return valid pfn
>>> if possible. Problem is that we need to check it every time before
>>> scanning pageblock even if we re-visit it and this turns out to
>>> be very expensive in this workload.
>>>
>>> Although we have no way to skip this pageblock check in the system
>>> where hole exists at arbitrary position, we can use cached value for
>>> zone continuity and just do pfn_to_page() in the system where hole doesn't
>>> exist. This optimization considerably speeds up in above workload.
>>>
>>> Before vs After
>>> Max: 1096 MB/s vs 1325 MB/s
>>> Min: 635 MB/s 1015 MB/s
>>> Avg: 899 MB/s 1194 MB/s
>>>
>>> Avg is improved by roughly 30% [2].
>>>
>>> [1]: http://www.spinics.net/lists/linux-mm/msg97378.html
>>> [2]: https://lkml.org/lkml/2015/12/9/23
>>>
>>> v2
>>> o checking zone continuity after initialization
>>> o handle memory-hotplug case
>>>
>>> Reported and Tested-by: Aaron Lu 
>>> Signed-off-by: Joonsoo Kim 
>>
>> [...]
>>
>>> --- a/mm/memory_hotplug.c
>>> +++ b/mm/memory_hotplug.c
>>> @@ -505,6 +505,9 @@ int __ref __add_pages(int nid, struct zone *zone,
>>> unsigned long phys_start_pfn,
>>> unsigned long i;
>>> int err = 0;
>>> int start_sec, end_sec;
>>> +
>>> +   clear_zone_contiguous(zone);
>>> +
>>> /* during initialize mem_map, align hot-added range to section */
>>> start_sec = pfn_to_section_nr(phys_start_pfn);
>>> end_sec = pfn_to_section_nr(phys_start_pfn + nr_pages - 1);
>>> @@ -523,6 +526,8 @@ int __ref __add_pages(int nid, struct zone *zone,
>>> unsigned long phys_start_pfn,
>>> }
>>> vmemmap_populate_print_last();
>>>
>>> +   set_zone_contiguous(zone);
>>> +
>>> return err;
>>>   }
>>>   EXPORT_SYMBOL_GPL(__add_pages);
>>> @@ -770,6 +775,8 @@ int __remove_pages(struct zone *zone, unsigned long
>>> phys_start_pfn,
>>> resource_size_t start, size;
>>> int ret = 0;
>>>
>>> +   clear_zone_contiguous(zone);
>>> +
>>> /*
>>>  * We can only remove entire sections
>>>  */
>>> @@ -796,6 +803,9 @@ int __remove_pages(struct zone *zone, unsigned long
>>> phys_start_pfn,
>>> if (ret)
>>> break;
>>> }
>>> +
>>> +   set_zone_contiguous(zone);
>>> +
>>> return ret;
>>
>>
>> Hm I wonder how many __add_ or __remove_pages calls there might be per a
>> major hotplug event (e.g. whole node). IIRC there may be many subranges that
>> are onlined/offlined separately?
>
> __add_ or __removed_pages are called whenever memory device is added
> or removed, that is, 1 call per 1 device add/remove. If device is 1GB,
> it is called
> for 1GB range and if device is 2GB, it is called for 2GB range. I think they 
> are
> reasonable place to check continuity.
>
>> Doing a full zone rescan on each
>> sub-operation could be quite costly, no?
>
> Doing full zone rescan makes code much simpler. It can be optimized further 
> but
> at this point there is no reason to handle complexity.
>
> You should have added
>> mm/hotplug_memory.c people to CC to comment, as you did in the [RFC]
>> theoretical race... mail. Doing that now.
>
> Okay. Thanks.
>
>> If the hotplug people confirm it might be an issue, I guess one solution is
>> to call set_zone_contiguous() lazily on-demand as you did in the v1 (but not
>> relying on cached pfn initialization to determine whether contiguous was
>> already evaluated). Add another variable like zone->contiguous_evaluated and
>> make hotplug code just set it to false.
>
> It could be, but, I guess that this is not too expensive for memory
> hotplug. They can
> tolerate 120 sec for offline memory and rescanning full zone to check
> continuity is
> very cheap than it.
>
> Anyway, it's better to confirm these things by memory hotplug people.
>
> Thanks.
>
>>
>>>   }
>>>   EXPORT_SYMBOL_GPL(__remove_pages);
>>> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
>>> index bac8842..4f5ad2b 100644
>>> --- a/mm/page_alloc.c
>>> +++ b/mm/page_alloc.c
>>> @@ -1271,9 +1271,13 @@ free_range:
>>> pgdat_init_report_one_done();
>>> return 0;
>>>   }
>>> +#endif /* CONFIG_DEFERRED_STRUCT_PAGE_INIT */
>>>
>>>   void __init 

Re: [PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Joonsoo Kim]

2015-12-21 19:46 GMT+09:00 Vlastimil Babka :
> On 12/21/2015 07:13 AM, Joonsoo Kim wrote:
>>
>> There is a performance drop report due to hugepage allocation and in there
>> half of cpu time are spent on pageblock_pfn_to_page() in compaction [1].
>> In that workload, compaction is triggered to make hugepage but most of
>> pageblocks are un-available for compaction due to pageblock type and
>> skip bit so compaction usually fails. Most costly operations in this case
>> is to find valid pageblock while scanning whole zone range. To check
>> if pageblock is valid to compact, valid pfn within pageblock is required
>> and we can obtain it by calling pageblock_pfn_to_page(). This function
>> checks whether pageblock is in a single zone and return valid pfn
>> if possible. Problem is that we need to check it every time before
>> scanning pageblock even if we re-visit it and this turns out to
>> be very expensive in this workload.
>>
>> Although we have no way to skip this pageblock check in the system
>> where hole exists at arbitrary position, we can use cached value for
>> zone continuity and just do pfn_to_page() in the system where hole doesn't
>> exist. This optimization considerably speeds up in above workload.
>>
>> Before vs After
>> Max: 1096 MB/s vs 1325 MB/s
>> Min: 635 MB/s 1015 MB/s
>> Avg: 899 MB/s 1194 MB/s
>>
>> Avg is improved by roughly 30% [2].
>>
>> [1]: http://www.spinics.net/lists/linux-mm/msg97378.html
>> [2]: https://lkml.org/lkml/2015/12/9/23
>>
>> v2
>> o checking zone continuity after initialization
>> o handle memory-hotplug case
>>
>> Reported and Tested-by: Aaron Lu 
>> Signed-off-by: Joonsoo Kim 
>
> [...]
>
>> --- a/mm/memory_hotplug.c
>> +++ b/mm/memory_hotplug.c
>> @@ -505,6 +505,9 @@ int __ref __add_pages(int nid, struct zone *zone,
>> unsigned long phys_start_pfn,
>> unsigned long i;
>> int err = 0;
>> int start_sec, end_sec;
>> +
>> +   clear_zone_contiguous(zone);
>> +
>> /* during initialize mem_map, align hot-added range to section */
>> start_sec = pfn_to_section_nr(phys_start_pfn);
>> end_sec = pfn_to_section_nr(phys_start_pfn + nr_pages - 1);
>> @@ -523,6 +526,8 @@ int __ref __add_pages(int nid, struct zone *zone,
>> unsigned long phys_start_pfn,
>> }
>> vmemmap_populate_print_last();
>>
>> +   set_zone_contiguous(zone);
>> +
>> return err;
>>   }
>>   EXPORT_SYMBOL_GPL(__add_pages);
>> @@ -770,6 +775,8 @@ int __remove_pages(struct zone *zone, unsigned long
>> phys_start_pfn,
>> resource_size_t start, size;
>> int ret = 0;
>>
>> +   clear_zone_contiguous(zone);
>> +
>> /*
>>  * We can only remove entire sections
>>  */
>> @@ -796,6 +803,9 @@ int __remove_pages(struct zone *zone, unsigned long
>> phys_start_pfn,
>> if (ret)
>> break;
>> }
>> +
>> +   set_zone_contiguous(zone);
>> +
>> return ret;
>
>
> Hm I wonder how many __add_ or __remove_pages calls there might be per a
> major hotplug event (e.g. whole node). IIRC there may be many subranges that
> are onlined/offlined separately?

__add_ or __removed_pages are called whenever memory device is added
or removed, that is, 1 call per 1 device add/remove. If device is 1GB,
it is called
for 1GB range and if device is 2GB, it is called for 2GB range. I think they are
reasonable place to check continuity.

> Doing a full zone rescan on each
> sub-operation could be quite costly, no?

Doing full zone rescan makes code much simpler. It can be optimized further but
at this point there is no reason to handle complexity.

You should have added
> mm/hotplug_memory.c people to CC to comment, as you did in the [RFC]
> theoretical race... mail. Doing that now.

Okay. Thanks.

> If the hotplug people confirm it might be an issue, I guess one solution is
> to call set_zone_contiguous() lazily on-demand as you did in the v1 (but not
> relying on cached pfn initialization to determine whether contiguous was
> already evaluated). Add another variable like zone->contiguous_evaluated and
> make hotplug code just set it to false.

It could be, but, I guess that this is not too expensive for memory
hotplug. They can
tolerate 120 sec for offline memory and rescanning full zone to check
continuity is
very cheap than it.

Anyway, it's better to confirm these things by memory hotplug people.

Thanks.

>
>>   }
>>   EXPORT_SYMBOL_GPL(__remove_pages);
>> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
>> index bac8842..4f5ad2b 100644
>> --- a/mm/page_alloc.c
>> +++ b/mm/page_alloc.c
>> @@ -1271,9 +1271,13 @@ free_range:
>> pgdat_init_report_one_done();
>> return 0;
>>   }
>> +#endif /* CONFIG_DEFERRED_STRUCT_PAGE_INIT */
>>
>>   void __init page_alloc_init_late(void)
>>   {
>> +   struct zone *zone;
>> +
>> +#ifdef CONFIG_DEFERRED_STRUCT_PAGE_INIT
>> int nid;
>>
>> /* There will be num_node_state(N_MEMORY) threads */
>> @@ -1287,8 

Re: [PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Vlastimil Babka]

On 12/21/2015 07:13 AM, Joonsoo Kim wrote:

There is a performance drop report due to hugepage allocation and in there
half of cpu time are spent on pageblock_pfn_to_page() in compaction [1].
In that workload, compaction is triggered to make hugepage but most of
pageblocks are un-available for compaction due to pageblock type and
skip bit so compaction usually fails. Most costly operations in this case
is to find valid pageblock while scanning whole zone range. To check
if pageblock is valid to compact, valid pfn within pageblock is required
and we can obtain it by calling pageblock_pfn_to_page(). This function
checks whether pageblock is in a single zone and return valid pfn
if possible. Problem is that we need to check it every time before
scanning pageblock even if we re-visit it and this turns out to
be very expensive in this workload.

Although we have no way to skip this pageblock check in the system
where hole exists at arbitrary position, we can use cached value for
zone continuity and just do pfn_to_page() in the system where hole doesn't
exist. This optimization considerably speeds up in above workload.

Before vs After
Max: 1096 MB/s vs 1325 MB/s
Min: 635 MB/s 1015 MB/s
Avg: 899 MB/s 1194 MB/s

Avg is improved by roughly 30% [2].

[1]: http://www.spinics.net/lists/linux-mm/msg97378.html
[2]: https://lkml.org/lkml/2015/12/9/23

v2
o checking zone continuity after initialization
o handle memory-hotplug case

Reported and Tested-by: Aaron Lu 
Signed-off-by: Joonsoo Kim 

[...]

--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -505,6 +505,9 @@ int __ref __add_pages(int nid, struct zone *zone, unsigned 
long phys_start_pfn,
unsigned long i;
int err = 0;
int start_sec, end_sec;
+
+   clear_zone_contiguous(zone);
+
/* during initialize mem_map, align hot-added range to section */
start_sec = pfn_to_section_nr(phys_start_pfn);
end_sec = pfn_to_section_nr(phys_start_pfn + nr_pages - 1);
@@ -523,6 +526,8 @@ int __ref __add_pages(int nid, struct zone *zone, unsigned 
long phys_start_pfn,
}
vmemmap_populate_print_last();

+   set_zone_contiguous(zone);
+
return err;
  }
  EXPORT_SYMBOL_GPL(__add_pages);
@@ -770,6 +775,8 @@ int __remove_pages(struct zone *zone, unsigned long 
phys_start_pfn,
resource_size_t start, size;
int ret = 0;

+   clear_zone_contiguous(zone);
+
/*
 * We can only remove entire sections
 */
@@ -796,6 +803,9 @@ int __remove_pages(struct zone *zone, unsigned long 
phys_start_pfn,
if (ret)
break;
}
+
+   set_zone_contiguous(zone);
+
return ret;


Hm I wonder how many __add_ or __remove_pages calls there might be per a 
major hotplug event (e.g. whole node). IIRC there may be many subranges 
that are onlined/offlined separately? Doing a full zone rescan on each 
sub-operation could be quite costly, no? You should have added 
mm/hotplug_memory.c people to CC to comment, as you did in the [RFC] 
theoretical race... mail. Doing that now.
If the hotplug people confirm it might be an issue, I guess one solution 
is to call set_zone_contiguous() lazily on-demand as you did in the v1 
(but not relying on cached pfn initialization to determine whether 
contiguous was already evaluated). Add another variable like 
zone->contiguous_evaluated and make hotplug code just set it to false.



  }
  EXPORT_SYMBOL_GPL(__remove_pages);
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index bac8842..4f5ad2b 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -1271,9 +1271,13 @@ free_range:
pgdat_init_report_one_done();
return 0;
  }
+#endif /* CONFIG_DEFERRED_STRUCT_PAGE_INIT */

  void __init page_alloc_init_late(void)
  {
+   struct zone *zone;
+
+#ifdef CONFIG_DEFERRED_STRUCT_PAGE_INIT
int nid;

/* There will be num_node_state(N_MEMORY) threads */
@@ -1287,8 +1291,87 @@ void __init page_alloc_init_late(void)

/* Reinit limits that are based on free pages after the kernel is up */
files_maxfiles_init();
+#endif
+
+   for_each_populated_zone(zone)
+   set_zone_contiguous(zone);
+}
+
+/*
+ * Check that the whole (or subset of) a pageblock given by the interval of
+ * [start_pfn, end_pfn) is valid and within the same zone, before scanning it
+ * with the migration of free compaction scanner. The scanners then need to
+ * use only pfn_valid_within() check for arches that allow holes within
+ * pageblocks.
+ *
+ * Return struct page pointer of start_pfn, or NULL if checks were not passed.
+ *
+ * It's possible on some configurations to have a setup like node0 node1 node0
+ * i.e. it's possible that all pages within a zones range of pages do not
+ * belong to a single zone. We assume that a border between node0 and node1
+ * can occur within a single pageblock, but not a node0 node1 node0
+ * interleaving within a single pageblock. It is therefore sufficient 

Re: [PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Vlastimil Babka]

On 12/21/2015 07:13 AM, Joonsoo Kim wrote:

There is a performance drop report due to hugepage allocation and in there
half of cpu time are spent on pageblock_pfn_to_page() in compaction [1].
In that workload, compaction is triggered to make hugepage but most of
pageblocks are un-available for compaction due to pageblock type and
skip bit so compaction usually fails. Most costly operations in this case
is to find valid pageblock while scanning whole zone range. To check
if pageblock is valid to compact, valid pfn within pageblock is required
and we can obtain it by calling pageblock_pfn_to_page(). This function
checks whether pageblock is in a single zone and return valid pfn
if possible. Problem is that we need to check it every time before
scanning pageblock even if we re-visit it and this turns out to
be very expensive in this workload.

Although we have no way to skip this pageblock check in the system
where hole exists at arbitrary position, we can use cached value for
zone continuity and just do pfn_to_page() in the system where hole doesn't
exist. This optimization considerably speeds up in above workload.

Before vs After
Max: 1096 MB/s vs 1325 MB/s
Min: 635 MB/s 1015 MB/s
Avg: 899 MB/s 1194 MB/s

Avg is improved by roughly 30% [2].

[1]: http://www.spinics.net/lists/linux-mm/msg97378.html
[2]: https://lkml.org/lkml/2015/12/9/23

v2
o checking zone continuity after initialization
o handle memory-hotplug case

Reported and Tested-by: Aaron Lu 
Signed-off-by: Joonsoo Kim 

[...]

--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -505,6 +505,9 @@ int __ref __add_pages(int nid, struct zone *zone, unsigned 
long phys_start_pfn,
unsigned long i;
int err = 0;
int start_sec, end_sec;
+
+   clear_zone_contiguous(zone);
+
/* during initialize mem_map, align hot-added range to section */
start_sec = pfn_to_section_nr(phys_start_pfn);
end_sec = pfn_to_section_nr(phys_start_pfn + nr_pages - 1);
@@ -523,6 +526,8 @@ int __ref __add_pages(int nid, struct zone *zone, unsigned 
long phys_start_pfn,
}
vmemmap_populate_print_last();

+   set_zone_contiguous(zone);
+
return err;
  }
  EXPORT_SYMBOL_GPL(__add_pages);
@@ -770,6 +775,8 @@ int __remove_pages(struct zone *zone, unsigned long 
phys_start_pfn,
resource_size_t start, size;
int ret = 0;

+   clear_zone_contiguous(zone);
+
/*
 * We can only remove entire sections
 */
@@ -796,6 +803,9 @@ int __remove_pages(struct zone *zone, unsigned long 
phys_start_pfn,
if (ret)
break;
}
+
+   set_zone_contiguous(zone);
+
return ret;


Hm I wonder how many __add_ or __remove_pages calls there might be per a 
major hotplug event (e.g. whole node). IIRC there may be many subranges 
that are onlined/offlined separately? Doing a full zone rescan on each 
sub-operation could be quite costly, no? You should have added 
mm/hotplug_memory.c people to CC to comment, as you did in the [RFC] 
theoretical race... mail. Doing that now.
If the hotplug people confirm it might be an issue, I guess one solution 
is to call set_zone_contiguous() lazily on-demand as you did in the v1 
(but not relying on cached pfn initialization to determine whether 
contiguous was already evaluated). Add another variable like 
zone->contiguous_evaluated and make hotplug code just set it to false.



  }
  EXPORT_SYMBOL_GPL(__remove_pages);
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index bac8842..4f5ad2b 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -1271,9 +1271,13 @@ free_range:
pgdat_init_report_one_done();
return 0;
  }
+#endif /* CONFIG_DEFERRED_STRUCT_PAGE_INIT */

  void __init page_alloc_init_late(void)
  {
+   struct zone *zone;
+
+#ifdef CONFIG_DEFERRED_STRUCT_PAGE_INIT
int nid;

/* There will be num_node_state(N_MEMORY) threads */
@@ -1287,8 +1291,87 @@ void __init page_alloc_init_late(void)

/* Reinit limits that are based on free pages after the kernel is up */
files_maxfiles_init();
+#endif
+
+   for_each_populated_zone(zone)
+   set_zone_contiguous(zone);
+}
+
+/*
+ * Check that the whole (or subset of) a pageblock given by the interval of
+ * [start_pfn, end_pfn) is valid and within the same zone, before scanning it
+ * with the migration of free compaction scanner. The scanners then need to
+ * use only pfn_valid_within() check for arches that allow holes within
+ * pageblocks.
+ *
+ * Return struct page pointer of start_pfn, or NULL if checks were not passed.
+ *
+ * It's possible on some configurations to have a setup like node0 node1 node0
+ * i.e. it's possible that all pages within a zones range of pages do not
+ * belong to a single zone. We assume that a border between node0 and node1
+ * can occur within a single pageblock, but not a node0 node1 node0
+ * interleaving within a 

Re: [PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Joonsoo Kim]

2015-12-21 19:46 GMT+09:00 Vlastimil Babka :
> On 12/21/2015 07:13 AM, Joonsoo Kim wrote:
>>
>> There is a performance drop report due to hugepage allocation and in there
>> half of cpu time are spent on pageblock_pfn_to_page() in compaction [1].
>> In that workload, compaction is triggered to make hugepage but most of
>> pageblocks are un-available for compaction due to pageblock type and
>> skip bit so compaction usually fails. Most costly operations in this case
>> is to find valid pageblock while scanning whole zone range. To check
>> if pageblock is valid to compact, valid pfn within pageblock is required
>> and we can obtain it by calling pageblock_pfn_to_page(). This function
>> checks whether pageblock is in a single zone and return valid pfn
>> if possible. Problem is that we need to check it every time before
>> scanning pageblock even if we re-visit it and this turns out to
>> be very expensive in this workload.
>>
>> Although we have no way to skip this pageblock check in the system
>> where hole exists at arbitrary position, we can use cached value for
>> zone continuity and just do pfn_to_page() in the system where hole doesn't
>> exist. This optimization considerably speeds up in above workload.
>>
>> Before vs After
>> Max: 1096 MB/s vs 1325 MB/s
>> Min: 635 MB/s 1015 MB/s
>> Avg: 899 MB/s 1194 MB/s
>>
>> Avg is improved by roughly 30% [2].
>>
>> [1]: http://www.spinics.net/lists/linux-mm/msg97378.html
>> [2]: https://lkml.org/lkml/2015/12/9/23
>>
>> v2
>> o checking zone continuity after initialization
>> o handle memory-hotplug case
>>
>> Reported and Tested-by: Aaron Lu 
>> Signed-off-by: Joonsoo Kim 
>
> [...]
>
>> --- a/mm/memory_hotplug.c
>> +++ b/mm/memory_hotplug.c
>> @@ -505,6 +505,9 @@ int __ref __add_pages(int nid, struct zone *zone,
>> unsigned long phys_start_pfn,
>> unsigned long i;
>> int err = 0;
>> int start_sec, end_sec;
>> +
>> +   clear_zone_contiguous(zone);
>> +
>> /* during initialize mem_map, align hot-added range to section */
>> start_sec = pfn_to_section_nr(phys_start_pfn);
>> end_sec = pfn_to_section_nr(phys_start_pfn + nr_pages - 1);
>> @@ -523,6 +526,8 @@ int __ref __add_pages(int nid, struct zone *zone,
>> unsigned long phys_start_pfn,
>> }
>> vmemmap_populate_print_last();
>>
>> +   set_zone_contiguous(zone);
>> +
>> return err;
>>   }
>>   EXPORT_SYMBOL_GPL(__add_pages);
>> @@ -770,6 +775,8 @@ int __remove_pages(struct zone *zone, unsigned long
>> phys_start_pfn,
>> resource_size_t start, size;
>> int ret = 0;
>>
>> +   clear_zone_contiguous(zone);
>> +
>> /*
>>  * We can only remove entire sections
>>  */
>> @@ -796,6 +803,9 @@ int __remove_pages(struct zone *zone, unsigned long
>> phys_start_pfn,
>> if (ret)
>> break;
>> }
>> +
>> +   set_zone_contiguous(zone);
>> +
>> return ret;
>
>
> Hm I wonder how many __add_ or __remove_pages calls there might be per a
> major hotplug event (e.g. whole node). IIRC there may be many subranges that
> are onlined/offlined separately?

__add_ or __removed_pages are called whenever memory device is added
or removed, that is, 1 call per 1 device add/remove. If device is 1GB,
it is called
for 1GB range and if device is 2GB, it is called for 2GB range. I think they are
reasonable place to check continuity.

> Doing a full zone rescan on each
> sub-operation could be quite costly, no?

Doing full zone rescan makes code much simpler. It can be optimized further but
at this point there is no reason to handle complexity.

You should have added
> mm/hotplug_memory.c people to CC to comment, as you did in the [RFC]
> theoretical race... mail. Doing that now.

Okay. Thanks.

> If the hotplug people confirm it might be an issue, I guess one solution is
> to call set_zone_contiguous() lazily on-demand as you did in the v1 (but not
> relying on cached pfn initialization to determine whether contiguous was
> already evaluated). Add another variable like zone->contiguous_evaluated and
> make hotplug code just set it to false.

It could be, but, I guess that this is not too expensive for memory
hotplug. They can
tolerate 120 sec for offline memory and rescanning full zone to check
continuity is
very cheap than it.

Anyway, it's better to confirm these things by memory hotplug people.

Thanks.

>
>>   }
>>   EXPORT_SYMBOL_GPL(__remove_pages);
>> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
>> index bac8842..4f5ad2b 100644
>> --- a/mm/page_alloc.c
>> +++ b/mm/page_alloc.c
>> @@ -1271,9 +1271,13 @@ free_range:
>> pgdat_init_report_one_done();
>> return 0;
>>   }
>> +#endif /* CONFIG_DEFERRED_STRUCT_PAGE_INIT */
>>
>>   void __init page_alloc_init_late(void)
>>   {
>> +   struct zone *zone;
>> +
>> +#ifdef CONFIG_DEFERRED_STRUCT_PAGE_INIT
>> int nid;
>>
>> /* 

Re: [PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Joonsoo Kim]

2015-12-21 21:18 GMT+09:00 Joonsoo Kim :
> 2015-12-21 19:46 GMT+09:00 Vlastimil Babka :
>> On 12/21/2015 07:13 AM, Joonsoo Kim wrote:
>>>
>>> There is a performance drop report due to hugepage allocation and in there
>>> half of cpu time are spent on pageblock_pfn_to_page() in compaction [1].
>>> In that workload, compaction is triggered to make hugepage but most of
>>> pageblocks are un-available for compaction due to pageblock type and
>>> skip bit so compaction usually fails. Most costly operations in this case
>>> is to find valid pageblock while scanning whole zone range. To check
>>> if pageblock is valid to compact, valid pfn within pageblock is required
>>> and we can obtain it by calling pageblock_pfn_to_page(). This function
>>> checks whether pageblock is in a single zone and return valid pfn
>>> if possible. Problem is that we need to check it every time before
>>> scanning pageblock even if we re-visit it and this turns out to
>>> be very expensive in this workload.
>>>
>>> Although we have no way to skip this pageblock check in the system
>>> where hole exists at arbitrary position, we can use cached value for
>>> zone continuity and just do pfn_to_page() in the system where hole doesn't
>>> exist. This optimization considerably speeds up in above workload.
>>>
>>> Before vs After
>>> Max: 1096 MB/s vs 1325 MB/s
>>> Min: 635 MB/s 1015 MB/s
>>> Avg: 899 MB/s 1194 MB/s
>>>
>>> Avg is improved by roughly 30% [2].
>>>
>>> [1]: http://www.spinics.net/lists/linux-mm/msg97378.html
>>> [2]: https://lkml.org/lkml/2015/12/9/23
>>>
>>> v2
>>> o checking zone continuity after initialization
>>> o handle memory-hotplug case
>>>
>>> Reported and Tested-by: Aaron Lu 
>>> Signed-off-by: Joonsoo Kim 
>>
>> [...]
>>
>>> --- a/mm/memory_hotplug.c
>>> +++ b/mm/memory_hotplug.c
>>> @@ -505,6 +505,9 @@ int __ref __add_pages(int nid, struct zone *zone,
>>> unsigned long phys_start_pfn,
>>> unsigned long i;
>>> int err = 0;
>>> int start_sec, end_sec;
>>> +
>>> +   clear_zone_contiguous(zone);
>>> +
>>> /* during initialize mem_map, align hot-added range to section */
>>> start_sec = pfn_to_section_nr(phys_start_pfn);
>>> end_sec = pfn_to_section_nr(phys_start_pfn + nr_pages - 1);
>>> @@ -523,6 +526,8 @@ int __ref __add_pages(int nid, struct zone *zone,
>>> unsigned long phys_start_pfn,
>>> }
>>> vmemmap_populate_print_last();
>>>
>>> +   set_zone_contiguous(zone);
>>> +
>>> return err;
>>>   }
>>>   EXPORT_SYMBOL_GPL(__add_pages);
>>> @@ -770,6 +775,8 @@ int __remove_pages(struct zone *zone, unsigned long
>>> phys_start_pfn,
>>> resource_size_t start, size;
>>> int ret = 0;
>>>
>>> +   clear_zone_contiguous(zone);
>>> +
>>> /*
>>>  * We can only remove entire sections
>>>  */
>>> @@ -796,6 +803,9 @@ int __remove_pages(struct zone *zone, unsigned long
>>> phys_start_pfn,
>>> if (ret)
>>> break;
>>> }
>>> +
>>> +   set_zone_contiguous(zone);
>>> +
>>> return ret;
>>
>>
>> Hm I wonder how many __add_ or __remove_pages calls there might be per a
>> major hotplug event (e.g. whole node). IIRC there may be many subranges that
>> are onlined/offlined separately?
>
> __add_ or __removed_pages are called whenever memory device is added
> or removed, that is, 1 call per 1 device add/remove. If device is 1GB,
> it is called
> for 1GB range and if device is 2GB, it is called for 2GB range. I think they 
> are
> reasonable place to check continuity.
>
>> Doing a full zone rescan on each
>> sub-operation could be quite costly, no?
>
> Doing full zone rescan makes code much simpler. It can be optimized further 
> but
> at this point there is no reason to handle complexity.
>
> You should have added
>> mm/hotplug_memory.c people to CC to comment, as you did in the [RFC]
>> theoretical race... mail. Doing that now.
>
> Okay. Thanks.
>
>> If the hotplug people confirm it might be an issue, I guess one solution is
>> to call set_zone_contiguous() lazily on-demand as you did in the v1 (but not
>> relying on cached pfn initialization to determine whether contiguous was
>> already evaluated). Add another variable like zone->contiguous_evaluated and
>> make hotplug code just set it to false.
>
> It could be, but, I guess that this is not too expensive for memory
> hotplug. They can
> tolerate 120 sec for offline memory and rescanning full zone to check
> continuity is
> very cheap than it.
>
> Anyway, it's better to confirm these things by memory hotplug people.
>
> Thanks.
>
>>
>>>   }
>>>   EXPORT_SYMBOL_GPL(__remove_pages);
>>> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
>>> index bac8842..4f5ad2b 100644
>>> --- a/mm/page_alloc.c
>>> +++ b/mm/page_alloc.c
>>> @@ -1271,9 +1271,13 @@ free_range:
>>> pgdat_init_report_one_done();
>>> return 0;
>>>   }
>>> 

[PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Joonsoo Kim]

There is a performance drop report due to hugepage allocation and in there
half of cpu time are spent on pageblock_pfn_to_page() in compaction [1].
In that workload, compaction is triggered to make hugepage but most of
pageblocks are un-available for compaction due to pageblock type and
skip bit so compaction usually fails. Most costly operations in this case
is to find valid pageblock while scanning whole zone range. To check
if pageblock is valid to compact, valid pfn within pageblock is required
and we can obtain it by calling pageblock_pfn_to_page(). This function
checks whether pageblock is in a single zone and return valid pfn
if possible. Problem is that we need to check it every time before
scanning pageblock even if we re-visit it and this turns out to
be very expensive in this workload.

Although we have no way to skip this pageblock check in the system
where hole exists at arbitrary position, we can use cached value for
zone continuity and just do pfn_to_page() in the system where hole doesn't
exist. This optimization considerably speeds up in above workload.

Before vs After
Max: 1096 MB/s vs 1325 MB/s
Min: 635 MB/s 1015 MB/s
Avg: 899 MB/s 1194 MB/s

Avg is improved by roughly 30% [2].

[1]: http://www.spinics.net/lists/linux-mm/msg97378.html
[2]: https://lkml.org/lkml/2015/12/9/23

v2
o checking zone continuity after initialization
o handle memory-hotplug case

Reported and Tested-by: Aaron Lu 
Signed-off-by: Joonsoo Kim 
---
 include/linux/gfp.h|  6 ---
 include/linux/memory_hotplug.h |  3 ++
 include/linux/mmzone.h |  2 +
 mm/compaction.c| 43 -
 mm/internal.h  | 12 ++
 mm/memory_hotplug.c| 10 +
 mm/page_alloc.c| 85 +-
 7 files changed, 111 insertions(+), 50 deletions(-)

diff --git a/include/linux/gfp.h b/include/linux/gfp.h
index 91f74e7..6eb3eca 100644
--- a/include/linux/gfp.h
+++ b/include/linux/gfp.h
@@ -515,13 +515,7 @@ void drain_zone_pages(struct zone *zone, struct 
per_cpu_pages *pcp);
 void drain_all_pages(struct zone *zone);
 void drain_local_pages(struct zone *zone);
 
-#ifdef CONFIG_DEFERRED_STRUCT_PAGE_INIT
 void page_alloc_init_late(void);
-#else
-static inline void page_alloc_init_late(void)
-{
-}
-#endif
 
 /*
  * gfp_allowed_mask is set to GFP_BOOT_MASK during early boot to restrict what
diff --git a/include/linux/memory_hotplug.h b/include/linux/memory_hotplug.h
index 2ea574f..18c2676 100644
--- a/include/linux/memory_hotplug.h
+++ b/include/linux/memory_hotplug.h
@@ -196,6 +196,9 @@ void put_online_mems(void);
 void mem_hotplug_begin(void);
 void mem_hotplug_done(void);
 
+extern void set_zone_contiguous(struct zone *zone);
+extern void clear_zone_contiguous(struct zone *zone);
+
 #else /* ! CONFIG_MEMORY_HOTPLUG */
 /*
  * Stub functions for when hotplug is off
diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index 68cc063..eb5d88e 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -523,6 +523,8 @@ struct zone {
boolcompact_blockskip_flush;
 #endif
 
+   boolcontiguous;
+
ZONE_PADDING(_pad3_)
/* Zone statistics */
atomic_long_t   vm_stat[NR_VM_ZONE_STAT_ITEMS];
diff --git a/mm/compaction.c b/mm/compaction.c
index 56fa321..9c89d46 100644
--- a/mm/compaction.c
+++ b/mm/compaction.c
@@ -71,49 +71,6 @@ static inline bool migrate_async_suitable(int migratetype)
return is_migrate_cma(migratetype) || migratetype == MIGRATE_MOVABLE;
 }
 
-/*
- * Check that the whole (or subset of) a pageblock given by the interval of
- * [start_pfn, end_pfn) is valid and within the same zone, before scanning it
- * with the migration of free compaction scanner. The scanners then need to
- * use only pfn_valid_within() check for arches that allow holes within
- * pageblocks.
- *
- * Return struct page pointer of start_pfn, or NULL if checks were not passed.
- *
- * It's possible on some configurations to have a setup like node0 node1 node0
- * i.e. it's possible that all pages within a zones range of pages do not
- * belong to a single zone. We assume that a border between node0 and node1
- * can occur within a single pageblock, but not a node0 node1 node0
- * interleaving within a single pageblock. It is therefore sufficient to check
- * the first and last page of a pageblock and avoid checking each individual
- * page in a pageblock.
- */
-static struct page *pageblock_pfn_to_page(unsigned long start_pfn,
-   unsigned long end_pfn, struct zone *zone)
-{
-   struct page *start_page;
-   struct page *end_page;
-
-   /* end_pfn is one past the range we are checking */
-   end_pfn--;
-
-   if (!pfn_valid(start_pfn) || !pfn_valid(end_pfn))
-   return NULL;
-
-   start_page = pfn_to_page(start_pfn);
-
-   if (page_zone(start_page) != zone)
-   return NULL;

[PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Joonsoo Kim]

There is a performance drop report due to hugepage allocation and in there
half of cpu time are spent on pageblock_pfn_to_page() in compaction [1].
In that workload, compaction is triggered to make hugepage but most of
pageblocks are un-available for compaction due to pageblock type and
skip bit so compaction usually fails. Most costly operations in this case
is to find valid pageblock while scanning whole zone range. To check
if pageblock is valid to compact, valid pfn within pageblock is required
and we can obtain it by calling pageblock_pfn_to_page(). This function
checks whether pageblock is in a single zone and return valid pfn
if possible. Problem is that we need to check it every time before
scanning pageblock even if we re-visit it and this turns out to
be very expensive in this workload.

Although we have no way to skip this pageblock check in the system
where hole exists at arbitrary position, we can use cached value for
zone continuity and just do pfn_to_page() in the system where hole doesn't
exist. This optimization considerably speeds up in above workload.

Before vs After
Max: 1096 MB/s vs 1325 MB/s
Min: 635 MB/s 1015 MB/s
Avg: 899 MB/s 1194 MB/s

Avg is improved by roughly 30% [2].

[1]: http://www.spinics.net/lists/linux-mm/msg97378.html
[2]: https://lkml.org/lkml/2015/12/9/23

v2
o checking zone continuity after initialization
o handle memory-hotplug case

Reported and Tested-by: Aaron Lu 
Signed-off-by: Joonsoo Kim 
---
 include/linux/gfp.h|  6 ---
 include/linux/memory_hotplug.h |  3 ++
 include/linux/mmzone.h |  2 +
 mm/compaction.c| 43 -
 mm/internal.h  | 12 ++
 mm/memory_hotplug.c| 10 +
 mm/page_alloc.c| 85 +-
 7 files changed, 111 insertions(+), 50 deletions(-)

diff --git a/include/linux/gfp.h b/include/linux/gfp.h
index 91f74e7..6eb3eca 100644
--- a/include/linux/gfp.h
+++ b/include/linux/gfp.h
@@ -515,13 +515,7 @@ void drain_zone_pages(struct zone *zone, struct 
per_cpu_pages *pcp);
 void drain_all_pages(struct zone *zone);
 void drain_local_pages(struct zone *zone);
 
-#ifdef CONFIG_DEFERRED_STRUCT_PAGE_INIT
 void page_alloc_init_late(void);
-#else
-static inline void page_alloc_init_late(void)
-{
-}
-#endif
 
 /*
  * gfp_allowed_mask is set to GFP_BOOT_MASK during early boot to restrict what
diff --git a/include/linux/memory_hotplug.h b/include/linux/memory_hotplug.h
index 2ea574f..18c2676 100644
--- a/include/linux/memory_hotplug.h
+++ b/include/linux/memory_hotplug.h
@@ -196,6 +196,9 @@ void put_online_mems(void);
 void mem_hotplug_begin(void);
 void mem_hotplug_done(void);
 
+extern void set_zone_contiguous(struct zone *zone);
+extern void clear_zone_contiguous(struct zone *zone);
+
 #else /* ! CONFIG_MEMORY_HOTPLUG */
 /*
  * Stub functions for when hotplug is off
diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index 68cc063..eb5d88e 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -523,6 +523,8 @@ struct zone {
boolcompact_blockskip_flush;
 #endif
 
+   boolcontiguous;
+
ZONE_PADDING(_pad3_)
/* Zone statistics */
atomic_long_t   vm_stat[NR_VM_ZONE_STAT_ITEMS];
diff --git a/mm/compaction.c b/mm/compaction.c
index 56fa321..9c89d46 100644
--- a/mm/compaction.c
+++ b/mm/compaction.c
@@ -71,49 +71,6 @@ static inline bool migrate_async_suitable(int migratetype)
return is_migrate_cma(migratetype) || migratetype == MIGRATE_MOVABLE;
 }
 
-/*
- * Check that the whole (or subset of) a pageblock given by the interval of
- * [start_pfn, end_pfn) is valid and within the same zone, before scanning it
- * with the migration of free compaction scanner. The scanners then need to
- * use only pfn_valid_within() check for arches that allow holes within
- * pageblocks.
- *
- * Return struct page pointer of start_pfn, or NULL if checks were not passed.
- *
- * It's possible on some configurations to have a setup like node0 node1 node0
- * i.e. it's possible that all pages within a zones range of pages do not
- * belong to a single zone. We assume that a border between node0 and node1
- * can occur within a single pageblock, but not a node0 node1 node0
- * interleaving within a single pageblock. It is therefore sufficient to check
- * the first and last page of a pageblock and avoid checking each individual
- * page in a pageblock.
- */
-static struct page *pageblock_pfn_to_page(unsigned long start_pfn,
-   unsigned long end_pfn, struct zone *zone)
-{
-   struct page *start_page;
-   struct page *end_page;
-
-   /* end_pfn is one past the range we are checking */
-   end_pfn--;
-
-   if (!pfn_valid(start_pfn) || !pfn_valid(end_pfn))
-   return NULL;
-
-   start_page = pfn_to_page(start_pfn);
-
-   if 

Re: [PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Vlastimil Babka]

On 12/15/2015 02:06 AM, Aaron Lu wrote:

On 12/14/2015 11:25 PM, Joonsoo Kim wrote:

2015-12-14 19:29 GMT+09:00 Vlastimil Babka :

Unless I'm mistaken, these results also include my RFC series (Aaron can you
clarify?). These patches should better be tested standalone on top of base,
as being simpler they will probably be included sooner (the RFC series needs
reviews at the very least :) - although the memory hotplug concerns might
make the "sooner" here relative too.


AFAIK, these patches are tested standalone on top of base. When I sent it,
I asked to Aaron to test it on top of base.


Right, it is tested standalone on top of base.


Thanks, sorry about the noise then.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Re: [PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Vlastimil Babka]

On 12/15/2015 02:06 AM, Aaron Lu wrote:

On 12/14/2015 11:25 PM, Joonsoo Kim wrote:

2015-12-14 19:29 GMT+09:00 Vlastimil Babka :

Unless I'm mistaken, these results also include my RFC series (Aaron can you
clarify?). These patches should better be tested standalone on top of base,
as being simpler they will probably be included sooner (the RFC series needs
reviews at the very least :) - although the memory hotplug concerns might
make the "sooner" here relative too.


AFAIK, these patches are tested standalone on top of base. When I sent it,
I asked to Aaron to test it on top of base.


Right, it is tested standalone on top of base.


Thanks, sorry about the noise then.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Re: [PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Aaron Lu]

On 12/14/2015 11:25 PM, Joonsoo Kim wrote:
> 2015-12-14 19:29 GMT+09:00 Vlastimil Babka :
>> On 12/14/2015 06:02 AM, Joonsoo Kim wrote:
>>> Before vs After
>>> Max: 1096 MB/s vs 1325 MB/s
>>> Min: 635 MB/s 1015 MB/s
>>> Avg: 899 MB/s 1194 MB/s
>>>
>>> Avg is improved by roughly 30% [2].
>>
>>
>> Unless I'm mistaken, these results also include my RFC series (Aaron can you
>> clarify?). These patches should better be tested standalone on top of base,
>> as being simpler they will probably be included sooner (the RFC series needs
>> reviews at the very least :) - although the memory hotplug concerns might
>> make the "sooner" here relative too.
> 
> AFAIK, these patches are tested standalone on top of base. When I sent it,
> I asked to Aaron to test it on top of base.

Right, it is tested standalone on top of base.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Re: [PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Joonsoo Kim]

2015-12-14 19:29 GMT+09:00 Vlastimil Babka :
> On 12/14/2015 06:02 AM, Joonsoo Kim wrote:
>>
>> There is a performance drop report due to hugepage allocation and in there
>> half of cpu time are spent on pageblock_pfn_to_page() in compaction [1].
>> In that workload, compaction is triggered to make hugepage but most of
>> pageblocks are un-available for compaction due to pageblock type and
>> skip bit so compaction usually fails. Most costly operations in this case
>> is to find valid pageblock while scanning whole zone range. To check
>> if pageblock is valid to compact, valid pfn within pageblock is required
>> and we can obtain it by calling pageblock_pfn_to_page(). This function
>> checks whether pageblock is in a single zone and return valid pfn
>> if possible. Problem is that we need to check it every time before
>> scanning pageblock even if we re-visit it and this turns out to
>> be very expensive in this workload.
>
>
> Hm I wonder if this is safe wrt memory hotplug? Shouldn't there be a
> rechecking plugged into the appropriate hotplug add/remove callbacks? Which
> would make the whole thing generic too, zone->contiguous information doesn't
> have to be limited to compaction. And it would remove the rather ugly part
> where cached pfn info is used as an indication of zone->contiguous being
> already set...

Will check it.

>> Although we have no way to skip this pageblock check in the system
>> where hole exists at arbitrary position, we can use cached value for
>> zone continuity and just do pfn_to_page() in the system where hole doesn't
>> exist. This optimization considerably speeds up in above workload.
>>
>> Before vs After
>> Max: 1096 MB/s vs 1325 MB/s
>> Min: 635 MB/s 1015 MB/s
>> Avg: 899 MB/s 1194 MB/s
>>
>> Avg is improved by roughly 30% [2].
>
>
> Unless I'm mistaken, these results also include my RFC series (Aaron can you
> clarify?). These patches should better be tested standalone on top of base,
> as being simpler they will probably be included sooner (the RFC series needs
> reviews at the very least :) - although the memory hotplug concerns might
> make the "sooner" here relative too.

AFAIK, these patches are tested standalone on top of base. When I sent it,
I asked to Aaron to test it on top of base.

Btw, I missed adding Reported/Tested-by tag for Aaron. I will add it
on next spin.

> Anyway it's interesting that this patch improved "Min", and variance in
> general (on top of my RFC) so much. I would expect the overhead of
> pageblock_pfn_to_page() to be quite stable, hmm.

Perhaps, pageblock_pfn_to_page() would be stable. Combination of
slow scanning and kswapd's skip bit flushing would result in unstable result.

Thanks.

>
>> Not to disturb the system where compaction isn't triggered, checking will
>> be done at first compaction invocation.
>>
>> [1]: http://www.spinics.net/lists/linux-mm/msg97378.html
>> [2]: https://lkml.org/lkml/2015/12/9/23
>>
>> Signed-off-by: Joonsoo Kim 
>> ---
>>   include/linux/mmzone.h |  1 +
>>   mm/compaction.c| 49
>> -
>>   2 files changed, 49 insertions(+), 1 deletion(-)
>>
>> diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
>> index 68cc063..cd3736e 100644
>> --- a/include/linux/mmzone.h
>> +++ b/include/linux/mmzone.h
>> @@ -521,6 +521,7 @@ struct zone {
>>   #if defined CONFIG_COMPACTION || defined CONFIG_CMA
>> /* Set to true when the PG_migrate_skip bits should be cleared */
>> boolcompact_blockskip_flush;
>> +   boolcontiguous;
>>   #endif
>>
>> ZONE_PADDING(_pad3_)
>> diff --git a/mm/compaction.c b/mm/compaction.c
>> index 56fa321..ce60b38 100644
>> --- a/mm/compaction.c
>> +++ b/mm/compaction.c
>> @@ -88,7 +88,7 @@ static inline bool migrate_async_suitable(int
>> migratetype)
>>* the first and last page of a pageblock and avoid checking each
>> individual
>>* page in a pageblock.
>>*/
>> -static struct page *pageblock_pfn_to_page(unsigned long start_pfn,
>> +static struct page *__pageblock_pfn_to_page(unsigned long start_pfn,
>> unsigned long end_pfn, struct zone *zone)
>>   {
>> struct page *start_page;
>> @@ -114,6 +114,51 @@ static struct page *pageblock_pfn_to_page(unsigned
>> long start_pfn,
>> return start_page;
>>   }
>>
>> +static inline struct page *pageblock_pfn_to_page(unsigned long start_pfn,
>> +   unsigned long end_pfn, struct zone *zone)
>> +{
>> +   if (zone->contiguous)
>> +   return pfn_to_page(start_pfn);
>> +
>> +   return __pageblock_pfn_to_page(start_pfn, end_pfn, zone);
>> +}
>> +
>> +static void check_zone_contiguous(struct zone *zone)
>> +{
>> +   unsigned long block_start_pfn = zone->zone_start_pfn;
>> +   unsigned long block_end_pfn;
>> +   unsigned long pfn;
>> +
>> +   /* Already initialized if cached pfn is non-zero */
>> +   if 

Re: [PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Vlastimil Babka]

On 12/14/2015 06:02 AM, Joonsoo Kim wrote:

There is a performance drop report due to hugepage allocation and in there
half of cpu time are spent on pageblock_pfn_to_page() in compaction [1].
In that workload, compaction is triggered to make hugepage but most of
pageblocks are un-available for compaction due to pageblock type and
skip bit so compaction usually fails. Most costly operations in this case
is to find valid pageblock while scanning whole zone range. To check
if pageblock is valid to compact, valid pfn within pageblock is required
and we can obtain it by calling pageblock_pfn_to_page(). This function
checks whether pageblock is in a single zone and return valid pfn
if possible. Problem is that we need to check it every time before
scanning pageblock even if we re-visit it and this turns out to
be very expensive in this workload.


Hm I wonder if this is safe wrt memory hotplug? Shouldn't there be a 
rechecking plugged into the appropriate hotplug add/remove callbacks? 
Which would make the whole thing generic too, zone->contiguous 
information doesn't have to be limited to compaction. And it would 
remove the rather ugly part where cached pfn info is used as an 
indication of zone->contiguous being already set...



Although we have no way to skip this pageblock check in the system
where hole exists at arbitrary position, we can use cached value for
zone continuity and just do pfn_to_page() in the system where hole doesn't
exist. This optimization considerably speeds up in above workload.

Before vs After
Max: 1096 MB/s vs 1325 MB/s
Min: 635 MB/s 1015 MB/s
Avg: 899 MB/s 1194 MB/s

Avg is improved by roughly 30% [2].


Unless I'm mistaken, these results also include my RFC series (Aaron can 
you clarify?). These patches should better be tested standalone on top 
of base, as being simpler they will probably be included sooner (the RFC 
series needs reviews at the very least :) - although the memory hotplug 
concerns might make the "sooner" here relative too.


Anyway it's interesting that this patch improved "Min", and variance in 
general (on top of my RFC) so much. I would expect the overhead of 
pageblock_pfn_to_page() to be quite stable, hmm.



Not to disturb the system where compaction isn't triggered, checking will
be done at first compaction invocation.

[1]: http://www.spinics.net/lists/linux-mm/msg97378.html
[2]: https://lkml.org/lkml/2015/12/9/23

Signed-off-by: Joonsoo Kim 
---
  include/linux/mmzone.h |  1 +
  mm/compaction.c| 49 -
  2 files changed, 49 insertions(+), 1 deletion(-)

diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index 68cc063..cd3736e 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -521,6 +521,7 @@ struct zone {
  #if defined CONFIG_COMPACTION || defined CONFIG_CMA
/* Set to true when the PG_migrate_skip bits should be cleared */
boolcompact_blockskip_flush;
+   boolcontiguous;
  #endif

ZONE_PADDING(_pad3_)
diff --git a/mm/compaction.c b/mm/compaction.c
index 56fa321..ce60b38 100644
--- a/mm/compaction.c
+++ b/mm/compaction.c
@@ -88,7 +88,7 @@ static inline bool migrate_async_suitable(int migratetype)
   * the first and last page of a pageblock and avoid checking each individual
   * page in a pageblock.
   */
-static struct page *pageblock_pfn_to_page(unsigned long start_pfn,
+static struct page *__pageblock_pfn_to_page(unsigned long start_pfn,
unsigned long end_pfn, struct zone *zone)
  {
struct page *start_page;
@@ -114,6 +114,51 @@ static struct page *pageblock_pfn_to_page(unsigned long 
start_pfn,
return start_page;
  }

+static inline struct page *pageblock_pfn_to_page(unsigned long start_pfn,
+   unsigned long end_pfn, struct zone *zone)
+{
+   if (zone->contiguous)
+   return pfn_to_page(start_pfn);
+
+   return __pageblock_pfn_to_page(start_pfn, end_pfn, zone);
+}
+
+static void check_zone_contiguous(struct zone *zone)
+{
+   unsigned long block_start_pfn = zone->zone_start_pfn;
+   unsigned long block_end_pfn;
+   unsigned long pfn;
+
+   /* Already initialized if cached pfn is non-zero */
+   if (zone->compact_cached_migrate_pfn[0] ||
+   zone->compact_cached_free_pfn)
+   return;
+
+   /* Mark that checking is in progress */
+   zone->compact_cached_free_pfn = ULONG_MAX;
+
+   block_end_pfn = ALIGN(block_start_pfn + 1, pageblock_nr_pages);
+   for (; block_start_pfn < zone_end_pfn(zone);
+   block_start_pfn = block_end_pfn,
+   block_end_pfn += pageblock_nr_pages) {
+
+   block_end_pfn = min(block_end_pfn, zone_end_pfn(zone));
+
+   if (!__pageblock_pfn_to_page(block_start_pfn,
+   block_end_pfn, zone))
+   return;
+
+   /* Check 

Re: [PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Aaron Lu]

On 12/14/2015 11:25 PM, Joonsoo Kim wrote:
> 2015-12-14 19:29 GMT+09:00 Vlastimil Babka :
>> On 12/14/2015 06:02 AM, Joonsoo Kim wrote:
>>> Before vs After
>>> Max: 1096 MB/s vs 1325 MB/s
>>> Min: 635 MB/s 1015 MB/s
>>> Avg: 899 MB/s 1194 MB/s
>>>
>>> Avg is improved by roughly 30% [2].
>>
>>
>> Unless I'm mistaken, these results also include my RFC series (Aaron can you
>> clarify?). These patches should better be tested standalone on top of base,
>> as being simpler they will probably be included sooner (the RFC series needs
>> reviews at the very least :) - although the memory hotplug concerns might
>> make the "sooner" here relative too.
> 
> AFAIK, these patches are tested standalone on top of base. When I sent it,
> I asked to Aaron to test it on top of base.

Right, it is tested standalone on top of base.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Re: [PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Vlastimil Babka]

On 12/14/2015 06:02 AM, Joonsoo Kim wrote:

There is a performance drop report due to hugepage allocation and in there
half of cpu time are spent on pageblock_pfn_to_page() in compaction [1].
In that workload, compaction is triggered to make hugepage but most of
pageblocks are un-available for compaction due to pageblock type and
skip bit so compaction usually fails. Most costly operations in this case
is to find valid pageblock while scanning whole zone range. To check
if pageblock is valid to compact, valid pfn within pageblock is required
and we can obtain it by calling pageblock_pfn_to_page(). This function
checks whether pageblock is in a single zone and return valid pfn
if possible. Problem is that we need to check it every time before
scanning pageblock even if we re-visit it and this turns out to
be very expensive in this workload.


Hm I wonder if this is safe wrt memory hotplug? Shouldn't there be a 
rechecking plugged into the appropriate hotplug add/remove callbacks? 
Which would make the whole thing generic too, zone->contiguous 
information doesn't have to be limited to compaction. And it would 
remove the rather ugly part where cached pfn info is used as an 
indication of zone->contiguous being already set...



Although we have no way to skip this pageblock check in the system
where hole exists at arbitrary position, we can use cached value for
zone continuity and just do pfn_to_page() in the system where hole doesn't
exist. This optimization considerably speeds up in above workload.

Before vs After
Max: 1096 MB/s vs 1325 MB/s
Min: 635 MB/s 1015 MB/s
Avg: 899 MB/s 1194 MB/s

Avg is improved by roughly 30% [2].


Unless I'm mistaken, these results also include my RFC series (Aaron can 
you clarify?). These patches should better be tested standalone on top 
of base, as being simpler they will probably be included sooner (the RFC 
series needs reviews at the very least :) - although the memory hotplug 
concerns might make the "sooner" here relative too.


Anyway it's interesting that this patch improved "Min", and variance in 
general (on top of my RFC) so much. I would expect the overhead of 
pageblock_pfn_to_page() to be quite stable, hmm.



Not to disturb the system where compaction isn't triggered, checking will
be done at first compaction invocation.

[1]: http://www.spinics.net/lists/linux-mm/msg97378.html
[2]: https://lkml.org/lkml/2015/12/9/23

Signed-off-by: Joonsoo Kim 
---
  include/linux/mmzone.h |  1 +
  mm/compaction.c| 49 -
  2 files changed, 49 insertions(+), 1 deletion(-)

diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index 68cc063..cd3736e 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -521,6 +521,7 @@ struct zone {
  #if defined CONFIG_COMPACTION || defined CONFIG_CMA
/* Set to true when the PG_migrate_skip bits should be cleared */
boolcompact_blockskip_flush;
+   boolcontiguous;
  #endif

ZONE_PADDING(_pad3_)
diff --git a/mm/compaction.c b/mm/compaction.c
index 56fa321..ce60b38 100644
--- a/mm/compaction.c
+++ b/mm/compaction.c
@@ -88,7 +88,7 @@ static inline bool migrate_async_suitable(int migratetype)
   * the first and last page of a pageblock and avoid checking each individual
   * page in a pageblock.
   */
-static struct page *pageblock_pfn_to_page(unsigned long start_pfn,
+static struct page *__pageblock_pfn_to_page(unsigned long start_pfn,
unsigned long end_pfn, struct zone *zone)
  {
struct page *start_page;
@@ -114,6 +114,51 @@ static struct page *pageblock_pfn_to_page(unsigned long 
start_pfn,
return start_page;
  }

+static inline struct page *pageblock_pfn_to_page(unsigned long start_pfn,
+   unsigned long end_pfn, struct zone *zone)
+{
+   if (zone->contiguous)
+   return pfn_to_page(start_pfn);
+
+   return __pageblock_pfn_to_page(start_pfn, end_pfn, zone);
+}
+
+static void check_zone_contiguous(struct zone *zone)
+{
+   unsigned long block_start_pfn = zone->zone_start_pfn;
+   unsigned long block_end_pfn;
+   unsigned long pfn;
+
+   /* Already initialized if cached pfn is non-zero */
+   if (zone->compact_cached_migrate_pfn[0] ||
+   zone->compact_cached_free_pfn)
+   return;
+
+   /* Mark that checking is in progress */
+   zone->compact_cached_free_pfn = ULONG_MAX;
+
+   block_end_pfn = ALIGN(block_start_pfn + 1, pageblock_nr_pages);
+   for (; block_start_pfn < zone_end_pfn(zone);
+   block_start_pfn = block_end_pfn,
+   block_end_pfn += pageblock_nr_pages) {
+
+   block_end_pfn = min(block_end_pfn, zone_end_pfn(zone));
+
+   if (!__pageblock_pfn_to_page(block_start_pfn,
+   block_end_pfn, zone))
+   return;
+
+  

Re: [PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Joonsoo Kim]

2015-12-14 19:29 GMT+09:00 Vlastimil Babka :
> On 12/14/2015 06:02 AM, Joonsoo Kim wrote:
>>
>> There is a performance drop report due to hugepage allocation and in there
>> half of cpu time are spent on pageblock_pfn_to_page() in compaction [1].
>> In that workload, compaction is triggered to make hugepage but most of
>> pageblocks are un-available for compaction due to pageblock type and
>> skip bit so compaction usually fails. Most costly operations in this case
>> is to find valid pageblock while scanning whole zone range. To check
>> if pageblock is valid to compact, valid pfn within pageblock is required
>> and we can obtain it by calling pageblock_pfn_to_page(). This function
>> checks whether pageblock is in a single zone and return valid pfn
>> if possible. Problem is that we need to check it every time before
>> scanning pageblock even if we re-visit it and this turns out to
>> be very expensive in this workload.
>
>
> Hm I wonder if this is safe wrt memory hotplug? Shouldn't there be a
> rechecking plugged into the appropriate hotplug add/remove callbacks? Which
> would make the whole thing generic too, zone->contiguous information doesn't
> have to be limited to compaction. And it would remove the rather ugly part
> where cached pfn info is used as an indication of zone->contiguous being
> already set...

Will check it.

>> Although we have no way to skip this pageblock check in the system
>> where hole exists at arbitrary position, we can use cached value for
>> zone continuity and just do pfn_to_page() in the system where hole doesn't
>> exist. This optimization considerably speeds up in above workload.
>>
>> Before vs After
>> Max: 1096 MB/s vs 1325 MB/s
>> Min: 635 MB/s 1015 MB/s
>> Avg: 899 MB/s 1194 MB/s
>>
>> Avg is improved by roughly 30% [2].
>
>
> Unless I'm mistaken, these results also include my RFC series (Aaron can you
> clarify?). These patches should better be tested standalone on top of base,
> as being simpler they will probably be included sooner (the RFC series needs
> reviews at the very least :) - although the memory hotplug concerns might
> make the "sooner" here relative too.

AFAIK, these patches are tested standalone on top of base. When I sent it,
I asked to Aaron to test it on top of base.

Btw, I missed adding Reported/Tested-by tag for Aaron. I will add it
on next spin.

> Anyway it's interesting that this patch improved "Min", and variance in
> general (on top of my RFC) so much. I would expect the overhead of
> pageblock_pfn_to_page() to be quite stable, hmm.

Perhaps, pageblock_pfn_to_page() would be stable. Combination of
slow scanning and kswapd's skip bit flushing would result in unstable result.

Thanks.

>
>> Not to disturb the system where compaction isn't triggered, checking will
>> be done at first compaction invocation.
>>
>> [1]: http://www.spinics.net/lists/linux-mm/msg97378.html
>> [2]: https://lkml.org/lkml/2015/12/9/23
>>
>> Signed-off-by: Joonsoo Kim 
>> ---
>>   include/linux/mmzone.h |  1 +
>>   mm/compaction.c| 49
>> -
>>   2 files changed, 49 insertions(+), 1 deletion(-)
>>
>> diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
>> index 68cc063..cd3736e 100644
>> --- a/include/linux/mmzone.h
>> +++ b/include/linux/mmzone.h
>> @@ -521,6 +521,7 @@ struct zone {
>>   #if defined CONFIG_COMPACTION || defined CONFIG_CMA
>> /* Set to true when the PG_migrate_skip bits should be cleared */
>> boolcompact_blockskip_flush;
>> +   boolcontiguous;
>>   #endif
>>
>> ZONE_PADDING(_pad3_)
>> diff --git a/mm/compaction.c b/mm/compaction.c
>> index 56fa321..ce60b38 100644
>> --- a/mm/compaction.c
>> +++ b/mm/compaction.c
>> @@ -88,7 +88,7 @@ static inline bool migrate_async_suitable(int
>> migratetype)
>>* the first and last page of a pageblock and avoid checking each
>> individual
>>* page in a pageblock.
>>*/
>> -static struct page *pageblock_pfn_to_page(unsigned long start_pfn,
>> +static struct page *__pageblock_pfn_to_page(unsigned long start_pfn,
>> unsigned long end_pfn, struct zone *zone)
>>   {
>> struct page *start_page;
>> @@ -114,6 +114,51 @@ static struct page *pageblock_pfn_to_page(unsigned
>> long start_pfn,
>> return start_page;
>>   }
>>
>> +static inline struct page *pageblock_pfn_to_page(unsigned long start_pfn,
>> +   unsigned long end_pfn, struct zone *zone)
>> +{
>> +   if (zone->contiguous)
>> +   return pfn_to_page(start_pfn);
>> +
>> +   return __pageblock_pfn_to_page(start_pfn, end_pfn, zone);
>> +}
>> +
>> +static void check_zone_contiguous(struct zone *zone)
>> +{
>> +   unsigned long block_start_pfn = zone->zone_start_pfn;
>> +   unsigned long block_end_pfn;
>> +   unsigned long pfn;
>> +
>> +   /* Already initialized if cached pfn is 

[PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Joonsoo Kim]

There is a performance drop report due to hugepage allocation and in there
half of cpu time are spent on pageblock_pfn_to_page() in compaction [1].
In that workload, compaction is triggered to make hugepage but most of
pageblocks are un-available for compaction due to pageblock type and
skip bit so compaction usually fails. Most costly operations in this case
is to find valid pageblock while scanning whole zone range. To check
if pageblock is valid to compact, valid pfn within pageblock is required
and we can obtain it by calling pageblock_pfn_to_page(). This function
checks whether pageblock is in a single zone and return valid pfn
if possible. Problem is that we need to check it every time before
scanning pageblock even if we re-visit it and this turns out to
be very expensive in this workload.

Although we have no way to skip this pageblock check in the system
where hole exists at arbitrary position, we can use cached value for
zone continuity and just do pfn_to_page() in the system where hole doesn't
exist. This optimization considerably speeds up in above workload.

Before vs After
Max: 1096 MB/s vs 1325 MB/s
Min: 635 MB/s 1015 MB/s
Avg: 899 MB/s 1194 MB/s

Avg is improved by roughly 30% [2].

Not to disturb the system where compaction isn't triggered, checking will
be done at first compaction invocation.

[1]: http://www.spinics.net/lists/linux-mm/msg97378.html
[2]: https://lkml.org/lkml/2015/12/9/23

Signed-off-by: Joonsoo Kim 
---
 include/linux/mmzone.h |  1 +
 mm/compaction.c| 49 -
 2 files changed, 49 insertions(+), 1 deletion(-)

diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index 68cc063..cd3736e 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -521,6 +521,7 @@ struct zone {
 #if defined CONFIG_COMPACTION || defined CONFIG_CMA
/* Set to true when the PG_migrate_skip bits should be cleared */
boolcompact_blockskip_flush;
+   boolcontiguous;
 #endif
 
ZONE_PADDING(_pad3_)
diff --git a/mm/compaction.c b/mm/compaction.c
index 56fa321..ce60b38 100644
--- a/mm/compaction.c
+++ b/mm/compaction.c
@@ -88,7 +88,7 @@ static inline bool migrate_async_suitable(int migratetype)
  * the first and last page of a pageblock and avoid checking each individual
  * page in a pageblock.
  */
-static struct page *pageblock_pfn_to_page(unsigned long start_pfn,
+static struct page *__pageblock_pfn_to_page(unsigned long start_pfn,
unsigned long end_pfn, struct zone *zone)
 {
struct page *start_page;
@@ -114,6 +114,51 @@ static struct page *pageblock_pfn_to_page(unsigned long 
start_pfn,
return start_page;
 }
 
+static inline struct page *pageblock_pfn_to_page(unsigned long start_pfn,
+   unsigned long end_pfn, struct zone *zone)
+{
+   if (zone->contiguous)
+   return pfn_to_page(start_pfn);
+
+   return __pageblock_pfn_to_page(start_pfn, end_pfn, zone);
+}
+
+static void check_zone_contiguous(struct zone *zone)
+{
+   unsigned long block_start_pfn = zone->zone_start_pfn;
+   unsigned long block_end_pfn;
+   unsigned long pfn;
+
+   /* Already initialized if cached pfn is non-zero */
+   if (zone->compact_cached_migrate_pfn[0] ||
+   zone->compact_cached_free_pfn)
+   return;
+
+   /* Mark that checking is in progress */
+   zone->compact_cached_free_pfn = ULONG_MAX;
+
+   block_end_pfn = ALIGN(block_start_pfn + 1, pageblock_nr_pages);
+   for (; block_start_pfn < zone_end_pfn(zone);
+   block_start_pfn = block_end_pfn,
+   block_end_pfn += pageblock_nr_pages) {
+
+   block_end_pfn = min(block_end_pfn, zone_end_pfn(zone));
+
+   if (!__pageblock_pfn_to_page(block_start_pfn,
+   block_end_pfn, zone))
+   return;
+
+   /* Check validity of pfn within pageblock */
+   for (pfn = block_start_pfn; pfn < block_end_pfn; pfn++) {
+   if (!pfn_valid_within(pfn))
+   return;
+   }
+   }
+
+   /* We confirm that there is no hole */
+   zone->contiguous = true;
+}
+
 #ifdef CONFIG_COMPACTION
 
 /* Do not skip compaction more than 64 times */
@@ -1357,6 +1402,8 @@ static int compact_zone(struct zone *zone, struct 
compact_control *cc)
;
}
 
+   check_zone_contiguous(zone);
+
/*
 * Clear pageblock skip if there were failures recently and compaction
 * is about to be retried after being deferred. kswapd does not do
-- 
1.9.1

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

[PATCH 2/2] mm/compaction: speed up pageblock_pfn_to_page() when zone is contiguous

[Joonsoo Kim]

There is a performance drop report due to hugepage allocation and in there
half of cpu time are spent on pageblock_pfn_to_page() in compaction [1].
In that workload, compaction is triggered to make hugepage but most of
pageblocks are un-available for compaction due to pageblock type and
skip bit so compaction usually fails. Most costly operations in this case
is to find valid pageblock while scanning whole zone range. To check
if pageblock is valid to compact, valid pfn within pageblock is required
and we can obtain it by calling pageblock_pfn_to_page(). This function
checks whether pageblock is in a single zone and return valid pfn
if possible. Problem is that we need to check it every time before
scanning pageblock even if we re-visit it and this turns out to
be very expensive in this workload.

Although we have no way to skip this pageblock check in the system
where hole exists at arbitrary position, we can use cached value for
zone continuity and just do pfn_to_page() in the system where hole doesn't
exist. This optimization considerably speeds up in above workload.

Before vs After
Max: 1096 MB/s vs 1325 MB/s
Min: 635 MB/s 1015 MB/s
Avg: 899 MB/s 1194 MB/s

Avg is improved by roughly 30% [2].

Not to disturb the system where compaction isn't triggered, checking will
be done at first compaction invocation.

[1]: http://www.spinics.net/lists/linux-mm/msg97378.html
[2]: https://lkml.org/lkml/2015/12/9/23

Signed-off-by: Joonsoo Kim 
---
 include/linux/mmzone.h |  1 +
 mm/compaction.c| 49 -
 2 files changed, 49 insertions(+), 1 deletion(-)

diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h
index 68cc063..cd3736e 100644
--- a/include/linux/mmzone.h
+++ b/include/linux/mmzone.h
@@ -521,6 +521,7 @@ struct zone {
 #if defined CONFIG_COMPACTION || defined CONFIG_CMA
/* Set to true when the PG_migrate_skip bits should be cleared */
boolcompact_blockskip_flush;
+   boolcontiguous;
 #endif
 
ZONE_PADDING(_pad3_)
diff --git a/mm/compaction.c b/mm/compaction.c
index 56fa321..ce60b38 100644
--- a/mm/compaction.c
+++ b/mm/compaction.c
@@ -88,7 +88,7 @@ static inline bool migrate_async_suitable(int migratetype)
  * the first and last page of a pageblock and avoid checking each individual
  * page in a pageblock.
  */
-static struct page *pageblock_pfn_to_page(unsigned long start_pfn,
+static struct page *__pageblock_pfn_to_page(unsigned long start_pfn,
unsigned long end_pfn, struct zone *zone)
 {
struct page *start_page;
@@ -114,6 +114,51 @@ static struct page *pageblock_pfn_to_page(unsigned long 
start_pfn,
return start_page;
 }
 
+static inline struct page *pageblock_pfn_to_page(unsigned long start_pfn,
+   unsigned long end_pfn, struct zone *zone)
+{
+   if (zone->contiguous)
+   return pfn_to_page(start_pfn);
+
+   return __pageblock_pfn_to_page(start_pfn, end_pfn, zone);
+}
+
+static void check_zone_contiguous(struct zone *zone)
+{
+   unsigned long block_start_pfn = zone->zone_start_pfn;
+   unsigned long block_end_pfn;
+   unsigned long pfn;
+
+   /* Already initialized if cached pfn is non-zero */
+   if (zone->compact_cached_migrate_pfn[0] ||
+   zone->compact_cached_free_pfn)
+   return;
+
+   /* Mark that checking is in progress */
+   zone->compact_cached_free_pfn = ULONG_MAX;
+
+   block_end_pfn = ALIGN(block_start_pfn + 1, pageblock_nr_pages);
+   for (; block_start_pfn < zone_end_pfn(zone);
+   block_start_pfn = block_end_pfn,
+   block_end_pfn += pageblock_nr_pages) {
+
+   block_end_pfn = min(block_end_pfn, zone_end_pfn(zone));
+
+   if (!__pageblock_pfn_to_page(block_start_pfn,
+   block_end_pfn, zone))
+   return;
+
+   /* Check validity of pfn within pageblock */
+   for (pfn = block_start_pfn; pfn < block_end_pfn; pfn++) {
+   if (!pfn_valid_within(pfn))
+   return;
+   }
+   }
+
+   /* We confirm that there is no hole */
+   zone->contiguous = true;
+}
+
 #ifdef CONFIG_COMPACTION
 
 /* Do not skip compaction more than 64 times */
@@ -1357,6 +1402,8 @@ static int compact_zone(struct zone *zone, struct 
compact_control *cc)
;
}
 
+   check_zone_contiguous(zone);
+
/*
 * Clear pageblock skip if there were failures recently and compaction
 * is about to be retried after being deferred. kswapd does not do
-- 
1.9.1

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at