diff mbox series

[v14,6/8] mm: hugetlb: introduce nr_free_vmemmap_pages in the struct hstate

Message ID 20210204035043.36609-7-songmuchun@bytedance.com (mailing list archive)
State New, archived
Headers show
Series Free some vmemmap pages of HugeTLB page | expand

Commit Message

Muchun Song Feb. 4, 2021, 3:50 a.m. UTC 
All the infrastructure is ready, so we introduce nr_free_vmemmap_pages
field in the hstate to indicate how many vmemmap pages associated with
a HugeTLB page that can be freed to buddy allocator. And initialize it
in the hugetlb_vmemmap_init(). This patch is actual enablement of the
feature.

Signed-off-by: Muchun Song <songmuchun@bytedance.com>
Acked-by: Mike Kravetz <mike.kravetz@oracle.com>
Reviewed-by: Oscar Salvador <osalvador@suse.de>
---
 include/linux/hugetlb.h |  3 +++
 mm/hugetlb.c            |  1 +
 mm/hugetlb_vmemmap.c    | 30 ++++++++++++++++++++++++++----
 mm/hugetlb_vmemmap.h    |  5 +++++
 4 files changed, 35 insertions(+), 4 deletions(-)

Comments

Miaohe Lin Feb. 5, 2021, 7:29 a.m. UTC | #1 
On 2021/2/4 11:50, Muchun Song wrote:
> All the infrastructure is ready, so we introduce nr_free_vmemmap_pages
> field in the hstate to indicate how many vmemmap pages associated with
> a HugeTLB page that can be freed to buddy allocator. And initialize it
> in the hugetlb_vmemmap_init(). This patch is actual enablement of the
> feature.
> 
> Signed-off-by: Muchun Song <songmuchun@bytedance.com>
> Acked-by: Mike Kravetz <mike.kravetz@oracle.com>
> Reviewed-by: Oscar Salvador <osalvador@suse.de>
> ---
>  include/linux/hugetlb.h |  3 +++
>  mm/hugetlb.c            |  1 +
>  mm/hugetlb_vmemmap.c    | 30 ++++++++++++++++++++++++++----
>  mm/hugetlb_vmemmap.h    |  5 +++++
>  4 files changed, 35 insertions(+), 4 deletions(-)
> 
> diff --git a/include/linux/hugetlb.h b/include/linux/hugetlb.h
> index ad249e56ac49..775aea53669a 100644
> --- a/include/linux/hugetlb.h
> +++ b/include/linux/hugetlb.h
> @@ -560,6 +560,9 @@ struct hstate {
>  	unsigned int nr_huge_pages_node[MAX_NUMNODES];
>  	unsigned int free_huge_pages_node[MAX_NUMNODES];
>  	unsigned int surplus_huge_pages_node[MAX_NUMNODES];
> +#ifdef CONFIG_HUGETLB_PAGE_FREE_VMEMMAP
> +	unsigned int nr_free_vmemmap_pages;
> +#endif
>  #ifdef CONFIG_CGROUP_HUGETLB
>  	/* cgroup control files */
>  	struct cftype cgroup_files_dfl[7];
> diff --git a/mm/hugetlb.c b/mm/hugetlb.c
> index 5518283aa667..04dde2b71f3e 100644
> --- a/mm/hugetlb.c
> +++ b/mm/hugetlb.c
> @@ -3220,6 +3220,7 @@ void __init hugetlb_add_hstate(unsigned int order)
>  	h->next_nid_to_free = first_memory_node;
>  	snprintf(h->name, HSTATE_NAME_LEN, "hugepages-%lukB",
>  					huge_page_size(h)/1024);
> +	hugetlb_vmemmap_init(h);
>  
>  	parsed_hstate = h;
>  }
> diff --git a/mm/hugetlb_vmemmap.c b/mm/hugetlb_vmemmap.c
> index 224a3cb69bf9..36ebd677e606 100644
> --- a/mm/hugetlb_vmemmap.c
> +++ b/mm/hugetlb_vmemmap.c
> @@ -208,13 +208,10 @@ early_param("hugetlb_free_vmemmap", early_hugetlb_free_vmemmap_param);
>  /*
>   * How many vmemmap pages associated with a HugeTLB page that can be freed
>   * to the buddy allocator.
> - *
> - * Todo: Returns zero for now, which means the feature is disabled. We will
> - * enable it once all the infrastructure is there.
>   */
>  static inline unsigned int free_vmemmap_pages_per_hpage(struct hstate *h)
>  {
> -	return 0;
> +	return h->nr_free_vmemmap_pages;
>  }
>  
>  static inline unsigned long free_vmemmap_pages_size_per_hpage(struct hstate *h)
> @@ -269,3 +266,28 @@ void free_huge_page_vmemmap(struct hstate *h, struct page *head)
>  	 */
>  	vmemmap_remap_free(vmemmap_addr, vmemmap_end, vmemmap_reuse);
>  }
> +
> +void __init hugetlb_vmemmap_init(struct hstate *h)
> +{
> +	unsigned int nr_pages = pages_per_huge_page(h);
> +	unsigned int vmemmap_pages;
> +
> +	if (!hugetlb_free_vmemmap_enabled)
> +		return;
> +
> +	vmemmap_pages = (nr_pages * sizeof(struct page)) >> PAGE_SHIFT;
> +	/*
> +	 * The head page and the first tail page are not to be freed to buddy
> +	 * allocator, the other pages will map to the first tail page, so they
> +	 * can be freed.
> +	 *
> +	 * Could RESERVE_VMEMMAP_NR be greater than @vmemmap_pages? It is true
> +	 * on some architectures (e.g. aarch64). See Documentation/arm64/
> +	 * hugetlbpage.rst for more details.
> +	 */
> +	if (likely(vmemmap_pages > RESERVE_VMEMMAP_NR))
> +		h->nr_free_vmemmap_pages = vmemmap_pages - RESERVE_VMEMMAP_NR;

Not a problem. Should we set h->nr_free_vmemmap_pages to 0 in 'else' case explicitly ?

Anyway, looks good to me. Thanks.
Reviewed-by: Miaohe Lin <linmiaohe@huawei.com>

> +
> +	pr_info("can free %d vmemmap pages for %s\n", h->nr_free_vmemmap_pages,
> +		h->name);
> +}
> diff --git a/mm/hugetlb_vmemmap.h b/mm/hugetlb_vmemmap.h
> index 6f89a9eed02c..02a21604ef1d 100644
> --- a/mm/hugetlb_vmemmap.h
> +++ b/mm/hugetlb_vmemmap.h
> @@ -14,6 +14,7 @@
>  int alloc_huge_page_vmemmap(struct hstate *h, struct page *head,
>  			    gfp_t gfp_mask);
>  void free_huge_page_vmemmap(struct hstate *h, struct page *head);
> +void hugetlb_vmemmap_init(struct hstate *h);
>  #else
>  static inline int alloc_huge_page_vmemmap(struct hstate *h, struct page *head,
>  					  gfp_t gfp_mask)
> @@ -24,5 +25,9 @@ static inline int alloc_huge_page_vmemmap(struct hstate *h, struct page *head,
>  static inline void free_huge_page_vmemmap(struct hstate *h, struct page *head)
>  {
>  }
> +
> +static inline void hugetlb_vmemmap_init(struct hstate *h)
> +{
> +}
>  #endif /* CONFIG_HUGETLB_PAGE_FREE_VMEMMAP */
>  #endif /* _LINUX_HUGETLB_VMEMMAP_H */
>
Oscar Salvador Feb. 5, 2021, 8:22 a.m. UTC | #2 
On Fri, Feb 05, 2021 at 03:29:43PM +0800, Miaohe Lin wrote:
> > +	if (likely(vmemmap_pages > RESERVE_VMEMMAP_NR))
> > +		h->nr_free_vmemmap_pages = vmemmap_pages - RESERVE_VMEMMAP_NR;
> 
> Not a problem. Should we set h->nr_free_vmemmap_pages to 0 in 'else' case explicitly ?

No, hstate fields are already zeroed.
Miaohe Lin Feb. 5, 2021, 8:39 a.m. UTC | #3 
Hi:
On 2021/2/5 16:22, Oscar Salvador wrote:
> On Fri, Feb 05, 2021 at 03:29:43PM +0800, Miaohe Lin wrote:
>>> +	if (likely(vmemmap_pages > RESERVE_VMEMMAP_NR))
>>> +		h->nr_free_vmemmap_pages = vmemmap_pages - RESERVE_VMEMMAP_NR;
>>
>> Not a problem. Should we set h->nr_free_vmemmap_pages to 0 in 'else' case explicitly ?
> 
> No, hstate fields are already zeroed.

I know hstate fields are already zeroed. What I mean is should we set nr_free_vmemmap_pages
to 0 _explicitly_ like nr_huge_pages and free_huge_pages in hugetlb_add_hstate() ?
But this is really trival.

Many thanks for reply.

> 
>
Oscar Salvador Feb. 5, 2021, 8:56 a.m. UTC | #4 
On Fri, Feb 05, 2021 at 04:39:26PM +0800, Miaohe Lin wrote:
> Hi:
> On 2021/2/5 16:22, Oscar Salvador wrote:
> > On Fri, Feb 05, 2021 at 03:29:43PM +0800, Miaohe Lin wrote:
> >>> +	if (likely(vmemmap_pages > RESERVE_VMEMMAP_NR))
> >>> +		h->nr_free_vmemmap_pages = vmemmap_pages - RESERVE_VMEMMAP_NR;
> >>
> >> Not a problem. Should we set h->nr_free_vmemmap_pages to 0 in 'else' case explicitly ?
> > 
> > No, hstate fields are already zeroed.
> 
> I know hstate fields are already zeroed. What I mean is should we set nr_free_vmemmap_pages
> to 0 _explicitly_ like nr_huge_pages and free_huge_pages in hugetlb_add_hstate() ?
> But this is really trival.

We do not anny more [1]

[1] https://patchwork.kernel.org/project/linux-mm/patch/20201119112141.6452-1-osalvador@suse.de/
Miaohe Lin Feb. 5, 2021, 9:12 a.m. UTC | #5 
On 2021/2/5 16:56, Oscar Salvador wrote:
> On Fri, Feb 05, 2021 at 04:39:26PM +0800, Miaohe Lin wrote:
>> Hi:
>> On 2021/2/5 16:22, Oscar Salvador wrote:
>>> On Fri, Feb 05, 2021 at 03:29:43PM +0800, Miaohe Lin wrote:
>>>>> +	if (likely(vmemmap_pages > RESERVE_VMEMMAP_NR))
>>>>> +		h->nr_free_vmemmap_pages = vmemmap_pages - RESERVE_VMEMMAP_NR;
>>>>
>>>> Not a problem. Should we set h->nr_free_vmemmap_pages to 0 in 'else' case explicitly ?
>>>
>>> No, hstate fields are already zeroed.
>>
>> I know hstate fields are already zeroed. What I mean is should we set nr_free_vmemmap_pages
>> to 0 _explicitly_ like nr_huge_pages and free_huge_pages in hugetlb_add_hstate() ?
>> But this is really trival.
> 
> We do not anny more [1]
> 
> [1] https://patchwork.kernel.org/project/linux-mm/patch/20201119112141.6452-1-osalvador@suse.de/
> 
> 

Oh, I see. I thought we do it deliberately. Many thanks for your patient answer. :)
diff mbox series

Patch

diff --git a/include/linux/hugetlb.h b/include/linux/hugetlb.h
index ad249e56ac49..775aea53669a 100644
--- a/include/linux/hugetlb.h
+++ b/include/linux/hugetlb.h
@@ -560,6 +560,9 @@  struct hstate {
 	unsigned int nr_huge_pages_node[MAX_NUMNODES];
 	unsigned int free_huge_pages_node[MAX_NUMNODES];
 	unsigned int surplus_huge_pages_node[MAX_NUMNODES];
+#ifdef CONFIG_HUGETLB_PAGE_FREE_VMEMMAP
+	unsigned int nr_free_vmemmap_pages;
+#endif
 #ifdef CONFIG_CGROUP_HUGETLB
 	/* cgroup control files */
 	struct cftype cgroup_files_dfl[7];
diff --git a/mm/hugetlb.c b/mm/hugetlb.c
index 5518283aa667..04dde2b71f3e 100644
--- a/mm/hugetlb.c
+++ b/mm/hugetlb.c
@@ -3220,6 +3220,7 @@  void __init hugetlb_add_hstate(unsigned int order)
 	h->next_nid_to_free = first_memory_node;
 	snprintf(h->name, HSTATE_NAME_LEN, "hugepages-%lukB",
 					huge_page_size(h)/1024);
+	hugetlb_vmemmap_init(h);
 
 	parsed_hstate = h;
 }
diff --git a/mm/hugetlb_vmemmap.c b/mm/hugetlb_vmemmap.c
index 224a3cb69bf9..36ebd677e606 100644
--- a/mm/hugetlb_vmemmap.c
+++ b/mm/hugetlb_vmemmap.c
@@ -208,13 +208,10 @@  early_param("hugetlb_free_vmemmap", early_hugetlb_free_vmemmap_param);
 /*
  * How many vmemmap pages associated with a HugeTLB page that can be freed
  * to the buddy allocator.
- *
- * Todo: Returns zero for now, which means the feature is disabled. We will
- * enable it once all the infrastructure is there.
  */
 static inline unsigned int free_vmemmap_pages_per_hpage(struct hstate *h)
 {
-	return 0;
+	return h->nr_free_vmemmap_pages;
 }
 
 static inline unsigned long free_vmemmap_pages_size_per_hpage(struct hstate *h)
@@ -269,3 +266,28 @@  void free_huge_page_vmemmap(struct hstate *h, struct page *head)
 	 */
 	vmemmap_remap_free(vmemmap_addr, vmemmap_end, vmemmap_reuse);
 }
+
+void __init hugetlb_vmemmap_init(struct hstate *h)
+{
+	unsigned int nr_pages = pages_per_huge_page(h);
+	unsigned int vmemmap_pages;
+
+	if (!hugetlb_free_vmemmap_enabled)
+		return;
+
+	vmemmap_pages = (nr_pages * sizeof(struct page)) >> PAGE_SHIFT;
+	/*
+	 * The head page and the first tail page are not to be freed to buddy
+	 * allocator, the other pages will map to the first tail page, so they
+	 * can be freed.
+	 *
+	 * Could RESERVE_VMEMMAP_NR be greater than @vmemmap_pages? It is true
+	 * on some architectures (e.g. aarch64). See Documentation/arm64/
+	 * hugetlbpage.rst for more details.
+	 */
+	if (likely(vmemmap_pages > RESERVE_VMEMMAP_NR))
+		h->nr_free_vmemmap_pages = vmemmap_pages - RESERVE_VMEMMAP_NR;
+
+	pr_info("can free %d vmemmap pages for %s\n", h->nr_free_vmemmap_pages,
+		h->name);
+}
diff --git a/mm/hugetlb_vmemmap.h b/mm/hugetlb_vmemmap.h
index 6f89a9eed02c..02a21604ef1d 100644
--- a/mm/hugetlb_vmemmap.h
+++ b/mm/hugetlb_vmemmap.h
@@ -14,6 +14,7 @@ 
 int alloc_huge_page_vmemmap(struct hstate *h, struct page *head,
 			    gfp_t gfp_mask);
 void free_huge_page_vmemmap(struct hstate *h, struct page *head);
+void hugetlb_vmemmap_init(struct hstate *h);
 #else
 static inline int alloc_huge_page_vmemmap(struct hstate *h, struct page *head,
 					  gfp_t gfp_mask)
@@ -24,5 +25,9 @@  static inline int alloc_huge_page_vmemmap(struct hstate *h, struct page *head,
 static inline void free_huge_page_vmemmap(struct hstate *h, struct page *head)
 {
 }
+
+static inline void hugetlb_vmemmap_init(struct hstate *h)
+{
+}
 #endif /* CONFIG_HUGETLB_PAGE_FREE_VMEMMAP */
 #endif /* _LINUX_HUGETLB_VMEMMAP_H */