Skip to contentSkip to conversation

rds: convert get_user_pages() --> pin_user_pages()

PATCH Applied John Hubbard 2 participants 2 messages May 17, 2020

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 rds: convert get_user_pages() --> pin_user_pages()

This code was using get_user_pages_fast(), in a "Case 2" scenario (DMA/RDMA), using the categorization from [1]. That means that it's time to convert the get_user_pages_fast() + put_page() calls to pin_user_pages_fast() + unpin_user_pages() calls.

There is some helpful background in [2]: basically, this is a small part of fixing a long-standing disconnect between pinning pages, and file systems' use of those pages.

[1] Documentation/core-api/pin_user_pages.rst

[2] "Explicit pinning of user-space pages":

    https://lwn.net/Articles/807108/
---
 net/rds/info.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)
net/rds/info.c +2 -4
index 03f6fd56d237..e1d63563e81c 100644
@@ -162,7 +162,6 @@ int rds_info_getsockopt(struct socket *sock, int optname, char __user *optval,
 	struct rds_info_lengths lens;
 	unsigned long nr_pages = 0;
 	unsigned long start;
-	unsigned long i;
 	rds_info_func func;
 	struct page **pages = NULL;
 	int ret;
@@ -193,7 +192,7 @@ int rds_info_getsockopt(struct socket *sock, int optname, char __user *optval,
 		ret = -ENOMEM;
 		goto out;
 	}
-	ret = get_user_pages_fast(start, nr_pages, FOLL_WRITE, pages);
+	ret = pin_user_pages_fast(start, nr_pages, FOLL_WRITE, pages);
 	if (ret != nr_pages) {
 		if (ret > 0)
 			nr_pages = ret;
@@ -235,8 +234,7 @@ int rds_info_getsockopt(struct socket *sock, int optname, char __user *optval,
 		ret = -EFAULT;
 
 out:
-	for (i = 0; pages && i < nr_pages; i++)
-		put_page(pages[i]);
+	unpin_user_pages(pages, nr_pages);
 	kfree(pages);
 
 	return ret;

base-commit: 3d1c1e5931ce45b3a3f309385bbc00c78e9951c6

Signature · 1 line
2.26.2
maintainer Applied
#1 lore raw

From: John Hubbard <[email protected]>
Date: Sat, 16 May 2020 18:23:36 -0700

Quoted 12 lines

This code was using get_user_pages_fast(), in a "Case 2" scenario (DMA/RDMA), using the categorization from [1]. That means that it's time to convert the get_user_pages_fast() + put_page() calls to pin_user_pages_fast() + unpin_user_pages() calls.

There is some helpful background in [2]: basically, this is a small part of fixing a long-standing disconnect between pinning pages, and file systems' use of those pages.

[1] Documentation/core-api/pin_user_pages.rst

[2] "Explicit pinning of user-space pages":

Quoted 8 lines
    https://lwn.net/Articles/807108/

Applied to net-next, thanks.