diff mbox series

[v5,09/12] mm: dump_page(): better diagnostics for huge pinned pages

Message ID 20200207033735.308000-10-jhubbard@nvidia.com (mailing list archive)
State New
Headers show
Series mm/gup: track FOLL_PIN pages | expand

Commit Message

John Hubbard Feb. 7, 2020, 3:37 a.m. UTC
As part of pin_user_pages() and related API calls, pages are
"dma-pinned". For the case of compound pages of order > 1, the per-page
accounting of dma pins is accomplished via the 3rd struct page in the
compound page. In order to support debugging of any pin_user_pages()-
related problems, enhance dump_page() so as to report the pin count
in that case.

Documentation/core-api/pin_user_pages.rst is also updated accordingly.

Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: John Hubbard <jhubbard@nvidia.com>
---
 Documentation/core-api/pin_user_pages.rst |  7 +++++
 mm/debug.c                                | 34 +++++++++++++++++------
 2 files changed, 33 insertions(+), 8 deletions(-)
diff mbox series

Patch

diff --git a/Documentation/core-api/pin_user_pages.rst b/Documentation/core-api/pin_user_pages.rst
index 7e5dd8b1b3f2..24641f1a1eba 100644
--- a/Documentation/core-api/pin_user_pages.rst
+++ b/Documentation/core-api/pin_user_pages.rst
@@ -215,6 +215,13 @@  Those are both going to show zero, unless CONFIG_DEBUG_VM is set. This is
 because there is a noticeable performance drop in unpin_user_page(), when they
 are activated.
 
+Other diagnostics
+=================
+
+dump_page() has been enhanced slightly, to handle these new counting fields, and
+to better report on compound pages in general. Specifically, for compound pages
+with order > 1, the exact (hpage_pinned_refcount) pincount is reported.
+
 References
 ==========
 
diff --git a/mm/debug.c b/mm/debug.c
index f074077eee11..e82c878c27df 100644
--- a/mm/debug.c
+++ b/mm/debug.c
@@ -57,10 +57,20 @@  static void __dump_tail_page(struct page *page, int mapcount)
 			page, page_ref_count(page), mapcount, page->mapping,
 			page_to_pgoff(page));
 	} else {
-		pr_warn("page:%px compound refcount:%d mapcount:%d mapping:%px "
-			"index:%#lx compound_mapcount:%d\n",
-			page, page_ref_count(head), mapcount, head->mapping,
-			page_to_pgoff(head), compound_mapcount(page));
+		if (hpage_pincount_available(page))
+			pr_warn("page:%px compound refcount:%d mapcount:%d "
+				"mapping:%px index:%#lx compound_mapcount:%d "
+				"compound_pincount:%d\n",
+				page, page_ref_count(head), mapcount,
+				head->mapping, page_to_pgoff(head),
+				compound_mapcount(page),
+				compound_pincount(page));
+		else
+			pr_warn("page:%px compound refcount:%d mapcount:%d "
+				"mapping:%px index:%#lx compound_mapcount:%d\n",
+				page, page_ref_count(head), mapcount,
+				head->mapping, page_to_pgoff(head),
+				compound_mapcount(page));
 	}
 
 	if (page_ref_count(page) != 0) {
@@ -104,10 +114,18 @@  void __dump_page(struct page *page, const char *reason)
 
 	if (PageTail(page))
 		__dump_tail_page(page, mapcount);
-	else
-		pr_warn("page:%px refcount:%d mapcount:%d mapping:%px index:%#lx\n",
-			page, page_ref_count(page), mapcount,
-			page->mapping, page_to_pgoff(page));
+	else {
+		if (hpage_pincount_available(page))
+			pr_warn("page:%px refcount:%d mapcount:%d mapping:%px "
+				"index:%#lx compound pincount: %d\n",
+				page, page_ref_count(page), mapcount,
+				page->mapping, page_to_pgoff(page),
+				compound_pincount(page));
+		else
+			pr_warn("page:%px refcount:%d mapcount:%d mapping:%px "
+				"index:%#lx\n", page, page_ref_count(page),
+				mapcount, page->mapping, page_to_pgoff(page));
+	}
 	if (PageKsm(page))
 		type = "ksm ";
 	else if (PageAnon(page))