If THP is failed to migrate due to -ENOSYS or -ENOMEM case, the THP will
be split, and the subpages of fail-to-migrate THP will be tried to migrate
again, so we should not account the retry counter in the second loop, since
we already accounted 'nr_thp_failed' in the first loop.
Moreover we also do not need retry 10 times for -EAGAIN case for the subpages
of fail-to-migrate THP in the second loop, since we already regarded the
THP as migration failure, and save some migration time (for the worst case,
will try 512 * 10 times) according to previous discussion [1].
[1] https://lore.kernel.org/linux-mm/[email protected]/
Signed-off-by: Baolin Wang <[email protected]>
---
Note, this patch is based on the patch set [1] from Huang Ying.
[1] https://lore.kernel.org/linux-mm/[email protected]/
---
mm/migrate.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/mm/migrate.c b/mm/migrate.c
index 8429206..e36a084 100644
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -1507,7 +1507,7 @@ int migrate_pages(struct list_head *from, new_page_t get_new_page,
case -EAGAIN:
if (is_thp)
thp_retry++;
- else
+ else if (!no_subpage_counting)
retry++;
nr_retry_pages += nr_subpages;
break;
@@ -1533,8 +1533,7 @@ int migrate_pages(struct list_head *from, new_page_t get_new_page,
}
}
}
- if (!no_subpage_counting)
- nr_failed += retry;
+ nr_failed += retry;
nr_thp_failed += thp_retry;
nr_failed_pages += nr_retry_pages;
/*
--
1.8.3.1
On 21 Jul 2022, at 8:05, Baolin Wang wrote:
> If THP is failed to migrate due to -ENOSYS or -ENOMEM case, the THP will
> be split, and the subpages of fail-to-migrate THP will be tried to migrate
> again, so we should not account the retry counter in the second loop, since
> we already accounted 'nr_thp_failed' in the first loop.
>
> Moreover we also do not need retry 10 times for -EAGAIN case for the subpages
> of fail-to-migrate THP in the second loop, since we already regarded the
> THP as migration failure, and save some migration time (for the worst case,
> will try 512 * 10 times) according to previous discussion [1].
>
> [1] https://lore.kernel.org/linux-mm/[email protected]/
>
> Signed-off-by: Baolin Wang <[email protected]>
> ---
> Note, this patch is based on the patch set [1] from Huang Ying.
> [1] https://lore.kernel.org/linux-mm/[email protected]/
> ---
> mm/migrate.c | 5 ++---
> 1 file changed, 2 insertions(+), 3 deletions(-)
>
> diff --git a/mm/migrate.c b/mm/migrate.c
> index 8429206..e36a084 100644
> --- a/mm/migrate.c
> +++ b/mm/migrate.c
> @@ -1507,7 +1507,7 @@ int migrate_pages(struct list_head *from, new_page_t get_new_page,
> case -EAGAIN:
> if (is_thp)
> thp_retry++;
> - else
> + else if (!no_subpage_counting)
> retry++;
> nr_retry_pages += nr_subpages;
> break;
> @@ -1533,8 +1533,7 @@ int migrate_pages(struct list_head *from, new_page_t get_new_page,
> }
> }
> }
> - if (!no_subpage_counting)
> - nr_failed += retry;
> + nr_failed += retry;
> nr_thp_failed += thp_retry;
> nr_failed_pages += nr_retry_pages;
> /*
> --
> 1.8.3.1
LGTM. Thanks. Reviewed-by: Zi Yan <[email protected]>
--
Best Regards,
Yan, Zi
Hi, Baolin,
Baolin Wang <[email protected]> writes:
> If THP is failed to migrate due to -ENOSYS or -ENOMEM case, the THP will
> be split, and the subpages of fail-to-migrate THP will be tried to migrate
> again, so we should not account the retry counter in the second loop, since
> we already accounted 'nr_thp_failed' in the first loop.
>
> Moreover we also do not need retry 10 times for -EAGAIN case for the subpages
> of fail-to-migrate THP in the second loop, since we already regarded the
> THP as migration failure, and save some migration time (for the worst case,
> will try 512 * 10 times) according to previous discussion [1].
>
> [1] https://lore.kernel.org/linux-mm/[email protected]/
>
> Signed-off-by: Baolin Wang <[email protected]>
I have tested this patch, and it works as expected.
Tested-by: "Huang, Ying" <[email protected]>
> ---
> Note, this patch is based on the patch set [1] from Huang Ying.
> [1] https://lore.kernel.org/linux-mm/[email protected]/
Please this patch is based on my patchset that hasn't been merged. I
can add it as the last patch of my migrate_pages failure path fixing
patchset if you don't object.
Best Regards,
Huang, Ying
> ---
> mm/migrate.c | 5 ++---
> 1 file changed, 2 insertions(+), 3 deletions(-)
>
> diff --git a/mm/migrate.c b/mm/migrate.c
> index 8429206..e36a084 100644
> --- a/mm/migrate.c
> +++ b/mm/migrate.c
> @@ -1507,7 +1507,7 @@ int migrate_pages(struct list_head *from, new_page_t get_new_page,
> case -EAGAIN:
> if (is_thp)
> thp_retry++;
> - else
> + else if (!no_subpage_counting)
> retry++;
> nr_retry_pages += nr_subpages;
> break;
> @@ -1533,8 +1533,7 @@ int migrate_pages(struct list_head *from, new_page_t get_new_page,
> }
> }
> }
> - if (!no_subpage_counting)
> - nr_failed += retry;
> + nr_failed += retry;
> nr_thp_failed += thp_retry;
> nr_failed_pages += nr_retry_pages;
> /*
On 7/28/2022 2:16 PM, Huang, Ying wrote:
> Hi, Baolin,
>
> Baolin Wang <[email protected]> writes:
>
>> If THP is failed to migrate due to -ENOSYS or -ENOMEM case, the THP will
>> be split, and the subpages of fail-to-migrate THP will be tried to migrate
>> again, so we should not account the retry counter in the second loop, since
>> we already accounted 'nr_thp_failed' in the first loop.
>>
>> Moreover we also do not need retry 10 times for -EAGAIN case for the subpages
>> of fail-to-migrate THP in the second loop, since we already regarded the
>> THP as migration failure, and save some migration time (for the worst case,
>> will try 512 * 10 times) according to previous discussion [1].
>>
>> [1] https://lore.kernel.org/linux-mm/[email protected]/
>>
>> Signed-off-by: Baolin Wang <[email protected]>
>
> I have tested this patch, and it works as expected.
>
> Tested-by: "Huang, Ying" <[email protected]>
Thanks.
>
>> ---
>> Note, this patch is based on the patch set [1] from Huang Ying.
>> [1] https://lore.kernel.org/linux-mm/[email protected]/
>
> Please this patch is based on my patchset that hasn't been merged. I
> can add it as the last patch of my migrate_pages failure path fixing
> patchset if you don't object.
Sure, please fold it into your series. Thanks for helping.