From: Mike Kravetz <mike.kravetz@oracle.com>
To: linux-mm@kvack.org, linux-kernel@vger.kernel.org
Cc: Muchun Song <songmuchun@bytedance.com>,
Joao Martins <joao.m.martins@oracle.com>,
Oscar Salvador <osalvador@suse.de>,
David Hildenbrand <david@redhat.com>,
Miaohe Lin <linmiaohe@huawei.com>,
David Rientjes <rientjes@google.com>,
Anshuman Khandual <anshuman.khandual@arm.com>,
Naoya Horiguchi <naoya.horiguchi@linux.dev>,
Barry Song <song.bao.hua@hisilicon.com>,
Michal Hocko <mhocko@suse.com>,
Matthew Wilcox <willy@infradead.org>,
Xiongchun Duan <duanxiongchun@bytedance.com>,
Andrew Morton <akpm@linux-foundation.org>,
Mike Kravetz <mike.kravetz@oracle.com>
Subject: [PATCH 08/12] hugetlb: batch freeing of vmemmap pages
Date: Fri, 25 Aug 2023 12:04:28 -0700 [thread overview]
Message-ID: <20230825190436.55045-9-mike.kravetz@oracle.com> (raw)
In-Reply-To: <20230825190436.55045-1-mike.kravetz@oracle.com>
Now that batching of hugetlb vmemmap optimization processing is possible,
batch the freeing of vmemmap pages. When freeing vmemmap pages for a
hugetlb page, we add them to a list that is freed after the entire batch
has been processed.
This enhances the ability to return contiguous ranges of memory to the
low level allocators.
Signed-off-by: Mike Kravetz <mike.kravetz@oracle.com>
---
mm/hugetlb_vmemmap.c | 56 ++++++++++++++++++++++++++++++++------------
1 file changed, 41 insertions(+), 15 deletions(-)
diff --git a/mm/hugetlb_vmemmap.c b/mm/hugetlb_vmemmap.c
index d5e6b6c76dce..e390170c0887 100644
--- a/mm/hugetlb_vmemmap.c
+++ b/mm/hugetlb_vmemmap.c
@@ -305,11 +305,14 @@ static void vmemmap_restore_pte(pte_t *pte, unsigned long addr,
* @end: end address of the vmemmap virtual address range that we want to
* remap.
* @reuse: reuse address.
+ * @bulk_pages: list to deposit vmemmap pages to be freed in bulk operations
+ * or NULL in non-bulk case;
*
* Return: %0 on success, negative error code otherwise.
*/
static int vmemmap_remap_free(unsigned long start, unsigned long end,
- unsigned long reuse)
+ unsigned long reuse,
+ struct list_head *bulk_pages)
{
int ret;
LIST_HEAD(vmemmap_pages);
@@ -372,7 +375,14 @@ static int vmemmap_remap_free(unsigned long start, unsigned long end,
}
mmap_read_unlock(&init_mm);
- free_vmemmap_page_list(&vmemmap_pages);
+ /*
+ * if performing bulk operation, do not free pages here.
+ * rather add them to the bulk list
+ */
+ if (!bulk_pages)
+ free_vmemmap_page_list(&vmemmap_pages);
+ else
+ list_splice(&vmemmap_pages, bulk_pages);
return ret;
}
@@ -546,17 +556,9 @@ static bool vmemmap_should_optimize(const struct hstate *h, const struct page *h
return true;
}
-/**
- * hugetlb_vmemmap_optimize - optimize @head page's vmemmap pages.
- * @h: struct hstate.
- * @head: the head page whose vmemmap pages will be optimized.
- *
- * This function only tries to optimize @head's vmemmap pages and does not
- * guarantee that the optimization will succeed after it returns. The caller
- * can use HPageVmemmapOptimized(@head) to detect if @head's vmemmap pages
- * have been optimized.
- */
-void hugetlb_vmemmap_optimize(const struct hstate *h, struct page *head)
+static void __hugetlb_vmemmap_optimize(const struct hstate *h,
+ struct page *head,
+ struct list_head *bulk_pages)
{
unsigned long vmemmap_start = (unsigned long)head, vmemmap_end;
unsigned long vmemmap_reuse;
@@ -575,18 +577,42 @@ void hugetlb_vmemmap_optimize(const struct hstate *h, struct page *head)
* to the page which @vmemmap_reuse is mapped to, then free the pages
* which the range [@vmemmap_start, @vmemmap_end] is mapped to.
*/
- if (vmemmap_remap_free(vmemmap_start, vmemmap_end, vmemmap_reuse))
+ if (vmemmap_remap_free(vmemmap_start, vmemmap_end, vmemmap_reuse, bulk_pages))
static_branch_dec(&hugetlb_optimize_vmemmap_key);
else
SetHPageVmemmapOptimized(head);
}
+/**
+ * hugetlb_vmemmap_optimize - optimize @head page's vmemmap pages.
+ * @h: struct hstate.
+ * @head: the head page whose vmemmap pages will be optimized.
+ *
+ * This function only tries to optimize @head's vmemmap pages and does not
+ * guarantee that the optimization will succeed after it returns. The caller
+ * can use HPageVmemmapOptimized(@head) to detect if @head's vmemmap pages
+ * have been optimized.
+ */
+void hugetlb_vmemmap_optimize(const struct hstate *h, struct page *head)
+{
+ __hugetlb_vmemmap_optimize(h, head, NULL);
+}
+
+void hugetlb_vmemmap_optimize_bulk(const struct hstate *h, struct page *head,
+ struct list_head *bulk_pages)
+{
+ __hugetlb_vmemmap_optimize(h, head, bulk_pages);
+}
+
void hugetlb_vmemmap_optimize_folios(struct hstate *h, struct list_head *folio_list)
{
struct folio *folio;
+ LIST_HEAD(vmemmap_pages);
list_for_each_entry(folio, folio_list, lru)
- hugetlb_vmemmap_optimize(h, &folio->page);
+ hugetlb_vmemmap_optimize_bulk(h, &folio->page, &vmemmap_pages);
+
+ free_vmemmap_page_list(&vmemmap_pages);
}
static struct ctl_table hugetlb_vmemmap_sysctls[] = {
--
2.41.0
next prev parent reply other threads:[~2023-08-25 19:05 UTC|newest]
Thread overview: 36+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-08-25 19:04 [PATCH 00/12] Batch hugetlb vmemmap modification operations Mike Kravetz
2023-08-25 19:04 ` [PATCH 01/12] hugetlb: clear flags in tail pages that will be freed individually Mike Kravetz
2023-08-25 19:04 ` [PATCH 02/12] hugetlb: Use a folio in free_hpage_workfn() Mike Kravetz
2023-08-25 19:04 ` [PATCH 03/12] hugetlb: Remove a few calls to page_folio() Mike Kravetz
2023-08-25 19:04 ` [PATCH 04/12] hugetlb: Convert remove_pool_huge_page() to remove_pool_hugetlb_folio() Mike Kravetz
2023-08-25 19:04 ` [PATCH 05/12] hugetlb: restructure pool allocations Mike Kravetz
2023-08-25 19:04 ` [PATCH 06/12] hugetlb: perform vmemmap optimization on a list of pages Mike Kravetz
2023-08-25 19:04 ` [PATCH 07/12] hugetlb: perform vmemmap restoration " Mike Kravetz
2023-08-26 6:58 ` kernel test robot
2023-08-30 8:33 ` Muchun Song
2023-08-30 17:53 ` Mike Kravetz
2023-08-25 19:04 ` Mike Kravetz [this message]
2023-08-26 4:00 ` [PATCH 08/12] hugetlb: batch freeing of vmemmap pages kernel test robot
2023-08-30 7:20 ` Muchun Song
2023-08-30 18:36 ` Mike Kravetz
2023-08-25 19:04 ` [PATCH 09/12] hugetlb_vmemmap: Optimistically set Optimized flag Mike Kravetz
2023-08-30 7:26 ` Muchun Song
2023-08-30 22:47 ` Mike Kravetz
2023-08-31 3:27 ` Muchun Song
2023-08-25 19:04 ` [PATCH 10/12] hugetlb: batch PMD split for bulk vmemmap dedup Mike Kravetz
2023-08-26 5:56 ` kernel test robot
2023-08-28 9:42 ` Joao Martins
2023-08-28 16:44 ` Mike Kravetz
2023-08-29 3:47 ` Muchun Song
2023-08-26 18:14 ` kernel test robot
2023-08-30 8:09 ` Muchun Song
2023-08-30 11:13 ` Joao Martins
2023-08-30 16:03 ` Joao Martins
2023-08-31 3:54 ` Muchun Song
2023-08-31 9:26 ` Joao Martins
2023-08-25 19:04 ` [PATCH 11/12] hugetlb: batch TLB flushes when freeing vmemmap Mike Kravetz
2023-08-30 8:23 ` Muchun Song
2023-08-30 11:17 ` Joao Martins
2023-08-25 19:04 ` [PATCH 12/12] hugetlb: batch TLB flushes when restoring vmemmap Mike Kravetz
2023-08-26 8:01 ` kernel test robot
2023-08-30 8:47 ` Muchun Song
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20230825190436.55045-9-mike.kravetz@oracle.com \
--to=mike.kravetz@oracle.com \
--cc=akpm@linux-foundation.org \
--cc=anshuman.khandual@arm.com \
--cc=david@redhat.com \
--cc=duanxiongchun@bytedance.com \
--cc=joao.m.martins@oracle.com \
--cc=linmiaohe@huawei.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mhocko@suse.com \
--cc=naoya.horiguchi@linux.dev \
--cc=osalvador@suse.de \
--cc=rientjes@google.com \
--cc=song.bao.hua@hisilicon.com \
--cc=songmuchun@bytedance.com \
--cc=willy@infradead.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).