Thread (22 messages) 22 messages, 7 authors, 2020-08-23

Re: [PATCH v2 1/2] mm: add GFP mask param to strndup_user

From: Andrew Morton <akpm@linux-foundation.org>
Date: 2020-08-22 03:51:59
Also in: linux-api, lkml

On Fri, 21 Aug 2020 20:28:26 -0700 Pascal Bouchareine [off-list ref] wrote:
Let caller specify allocation.
Preserve existing calls with GFP_USER.

 21 files changed, 65 insertions(+), 43 deletions(-)
Why change all existing callsites so that one callsite can pass in a
different gfp_t?
quoted hunk ↗ jump to hunk
diff --git a/drivers/dma-buf/dma-buf.c b/drivers/dma-buf/dma-buf.c
index 1ca609f66fdf..3d94ba811f4b 100644
--- a/drivers/dma-buf/dma-buf.c
+++ b/drivers/dma-buf/dma-buf.c
@@ -326,7 +326,7 @@ static __poll_t dma_buf_poll(struct file *file, poll_table *poll)
  */
 static long dma_buf_set_name(struct dma_buf *dmabuf, const char __user *buf)
 {
-	char *name = strndup_user(buf, DMA_BUF_NAME_LEN);
+	char *name = strndup_user(buf, DMA_BUF_NAME_LEN, GFP_USER);
 	long ret = 0;
Wouldn't

#include <linux/gfp.h>

char *__strndup_user(const char __user *s, long n, gfp_t gfp);

static inline char *strndup_user(const char __user *s, long n)
{
	return __strndup_user(s, n, GFP_USER);
}

be simpler?



Also...

why does strndup_user() use GFP_USER?  Nobody will be mapping the
resulting strings into user pagetables (will they?).  This was done by
Al's 6c2c97a24f096e32, which doesn't have a changelog :(


In [patch 2/2],

+	desc = strndup_user(user_desc, SK_MAX_DESC_SIZE, GFP_KERNEL_ACCOUNT);

if GFP_USER is legit then shouldn't this be GFP_USER_ACCOUNT (ie,
GFP_USER|__GFP_ACCOUNT)?
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help