linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH 1/2] splice: don't merge into linked buffers
@ 2018-10-15 15:04 Jann Horn
  2018-10-15 15:04 ` [PATCH 2/2] pipe: remove ->can_merge from pipe_buf_operations Jann Horn
  2018-10-15 21:00 ` [PATCH 1/2] splice: don't merge into linked buffers Eric Biggers
  0 siblings, 2 replies; 4+ messages in thread
From: Jann Horn @ 2018-10-15 15:04 UTC (permalink / raw)
  To: Al Viro, Miklos Szeredi, Jens Axboe, Jens Axboe, jannh
  Cc: linux-kernel, linux-fsdevel, Kees Cook, Eric Biggers

Before this patch, it was possible for two pipes to affect each other after
data had been transferred between them with tee():

============
$ cat tee_test.c

int main(void) {
  int pipe_a[2];
  if (pipe(pipe_a)) err(1, "pipe");
  int pipe_b[2];
  if (pipe(pipe_b)) err(1, "pipe");
  if (write(pipe_a[1], "abcd", 4) != 4) err(1, "write");
  if (tee(pipe_a[0], pipe_b[1], 2, 0) != 2) err(1, "tee");
  if (write(pipe_b[1], "xx", 2) != 2) err(1, "write");

  char buf[5];
  if (read(pipe_a[0], buf, 4) != 4) err(1, "read");
  buf[4] = 0;
  printf("got back: '%s'\n", buf);
}
$ gcc -o tee_test tee_test.c
$ ./tee_test
got back: 'abxx'
$
============

Fix it by explicitly marking buffers as mergeable and clearing that flag in
splice_pipe_to_pipe() and link_pipe().

Cc: <stable@vger.kernel.org>
Fixes: 7c77f0b3f920 ("splice: implement pipe to pipe splicing")
Fixes: 70524490ee2e ("[PATCH] splice: add support for sys_tee()")
Signed-off-by: Jann Horn <jannh@google.com>
---
Cleanup in the next patch, to simplify backporting.

 fs/pipe.c                 | 5 +++--
 fs/splice.c               | 6 ++++++
 include/linux/pipe_fs_i.h | 8 ++++++++
 3 files changed, 17 insertions(+), 2 deletions(-)

diff --git a/fs/pipe.c b/fs/pipe.c
index bdc5d3c0977d..4e2eee77f855 100644
--- a/fs/pipe.c
+++ b/fs/pipe.c
@@ -379,7 +379,8 @@ pipe_write(struct kiocb *iocb, struct iov_iter *from)
 		struct pipe_buffer *buf = pipe->bufs + lastbuf;
 		int offset = buf->offset + buf->len;
 
-		if (buf->ops->can_merge && offset + chars <= PAGE_SIZE) {
+		if (buf->ops->can_merge && offset + chars <= PAGE_SIZE &&
+		    (buf->flags & PIPE_BUF_FLAG_MAYMERGE)) {
 			ret = pipe_buf_confirm(pipe, buf);
 			if (ret)
 				goto out;
@@ -439,7 +440,7 @@ pipe_write(struct kiocb *iocb, struct iov_iter *from)
 			buf->ops = &anon_pipe_buf_ops;
 			buf->offset = 0;
 			buf->len = copied;
-			buf->flags = 0;
+			buf->flags = PIPE_BUF_FLAG_MAYMERGE;
 			if (is_packetized(filp)) {
 				buf->ops = &packet_pipe_buf_ops;
 				buf->flags = PIPE_BUF_FLAG_PACKET;
diff --git a/fs/splice.c b/fs/splice.c
index b3daa971f597..111977c80dfd 100644
--- a/fs/splice.c
+++ b/fs/splice.c
@@ -1593,6 +1593,9 @@ static int splice_pipe_to_pipe(struct pipe_inode_info *ipipe,
 			 */
 			obuf->flags &= ~PIPE_BUF_FLAG_GIFT;
 
+			/* We can't merge data into a buffer we don't own. */
+			obuf->flags &= ~PIPE_BUF_FLAG_MAYMERGE;
+
 			obuf->len = len;
 			opipe->nrbufs++;
 			ibuf->offset += obuf->len;
@@ -1667,6 +1670,9 @@ static int link_pipe(struct pipe_inode_info *ipipe,
 		 */
 		obuf->flags &= ~PIPE_BUF_FLAG_GIFT;
 
+		/* We can't merge data into a buffer we don't own. */
+		obuf->flags &= ~PIPE_BUF_FLAG_MAYMERGE;
+
 		if (obuf->len > len)
 			obuf->len = len;
 
diff --git a/include/linux/pipe_fs_i.h b/include/linux/pipe_fs_i.h
index 5a3bb3b7c9ad..8893711f9171 100644
--- a/include/linux/pipe_fs_i.h
+++ b/include/linux/pipe_fs_i.h
@@ -8,6 +8,14 @@
 #define PIPE_BUF_FLAG_ATOMIC	0x02	/* was atomically mapped */
 #define PIPE_BUF_FLAG_GIFT	0x04	/* page is a gift */
 #define PIPE_BUF_FLAG_PACKET	0x08	/* read() as a packet */
+/*
+ * Set this flag if the generic pipe read/write may coalesce data into an
+ * existing buffer. If this is not set, a new pipe page segment is always used
+ * for new data.
+ * When pipe data is copied by reference (as in the tee() syscall), this flag
+ * must be cleared on the copy.
+ */
+#define PIPE_BUF_FLAG_MAYMERGE	0x10
 
 /**
  *	struct pipe_buffer - a linux kernel pipe buffer
-- 
2.19.0.605.g01d371f741-goog


^ permalink raw reply related	[flat|nested] 4+ messages in thread

* [PATCH 2/2] pipe: remove ->can_merge from pipe_buf_operations
  2018-10-15 15:04 [PATCH 1/2] splice: don't merge into linked buffers Jann Horn
@ 2018-10-15 15:04 ` Jann Horn
  2018-10-15 21:00 ` [PATCH 1/2] splice: don't merge into linked buffers Eric Biggers
  1 sibling, 0 replies; 4+ messages in thread
From: Jann Horn @ 2018-10-15 15:04 UTC (permalink / raw)
  To: Al Viro, Miklos Szeredi, Jens Axboe, Jens Axboe, jannh
  Cc: linux-kernel, linux-fsdevel, Kees Cook, Eric Biggers

My last commit "splice: don't merge into linked buffers" made the
"can_merge" field in struct pipe_buf_operations redundant: Whether a buffer
is mergeable is now a property of the buffer, not of the operations struct.
Get rid of it.

Signed-off-by: Jann Horn <jannh@google.com>
---
 fs/pipe.c                 | 4 +---
 fs/splice.c               | 4 ----
 include/linux/pipe_fs_i.h | 7 -------
 kernel/relay.c            | 1 -
 kernel/trace/trace.c      | 2 --
 net/smc/smc_rx.c          | 1 -
 6 files changed, 1 insertion(+), 18 deletions(-)

diff --git a/fs/pipe.c b/fs/pipe.c
index 4e2eee77f855..b1ad3271c175 100644
--- a/fs/pipe.c
+++ b/fs/pipe.c
@@ -227,7 +227,6 @@ void generic_pipe_buf_release(struct pipe_inode_info *pipe,
 EXPORT_SYMBOL(generic_pipe_buf_release);
 
 static const struct pipe_buf_operations anon_pipe_buf_ops = {
-	.can_merge = 1,
 	.confirm = generic_pipe_buf_confirm,
 	.release = anon_pipe_buf_release,
 	.steal = anon_pipe_buf_steal,
@@ -235,7 +234,6 @@ static const struct pipe_buf_operations anon_pipe_buf_ops = {
 };
 
 static const struct pipe_buf_operations packet_pipe_buf_ops = {
-	.can_merge = 0,
 	.confirm = generic_pipe_buf_confirm,
 	.release = anon_pipe_buf_release,
 	.steal = anon_pipe_buf_steal,
@@ -379,7 +377,7 @@ pipe_write(struct kiocb *iocb, struct iov_iter *from)
 		struct pipe_buffer *buf = pipe->bufs + lastbuf;
 		int offset = buf->offset + buf->len;
 
-		if (buf->ops->can_merge && offset + chars <= PAGE_SIZE &&
+		if (offset + chars <= PAGE_SIZE &&
 		    (buf->flags & PIPE_BUF_FLAG_MAYMERGE)) {
 			ret = pipe_buf_confirm(pipe, buf);
 			if (ret)
diff --git a/fs/splice.c b/fs/splice.c
index 111977c80dfd..fb7412b2eadb 100644
--- a/fs/splice.c
+++ b/fs/splice.c
@@ -138,7 +138,6 @@ static int page_cache_pipe_buf_confirm(struct pipe_inode_info *pipe,
 }
 
 const struct pipe_buf_operations page_cache_pipe_buf_ops = {
-	.can_merge = 0,
 	.confirm = page_cache_pipe_buf_confirm,
 	.release = page_cache_pipe_buf_release,
 	.steal = page_cache_pipe_buf_steal,
@@ -156,7 +155,6 @@ static int user_page_pipe_buf_steal(struct pipe_inode_info *pipe,
 }
 
 static const struct pipe_buf_operations user_page_pipe_buf_ops = {
-	.can_merge = 0,
 	.confirm = generic_pipe_buf_confirm,
 	.release = page_cache_pipe_buf_release,
 	.steal = user_page_pipe_buf_steal,
@@ -326,7 +324,6 @@ ssize_t generic_file_splice_read(struct file *in, loff_t *ppos,
 EXPORT_SYMBOL(generic_file_splice_read);
 
 const struct pipe_buf_operations default_pipe_buf_ops = {
-	.can_merge = 0,
 	.confirm = generic_pipe_buf_confirm,
 	.release = generic_pipe_buf_release,
 	.steal = generic_pipe_buf_steal,
@@ -341,7 +338,6 @@ static int generic_pipe_buf_nosteal(struct pipe_inode_info *pipe,
 
 /* Pipe buffer operations for a socket and similar. */
 const struct pipe_buf_operations nosteal_pipe_buf_ops = {
-	.can_merge = 0,
 	.confirm = generic_pipe_buf_confirm,
 	.release = generic_pipe_buf_release,
 	.steal = generic_pipe_buf_nosteal,
diff --git a/include/linux/pipe_fs_i.h b/include/linux/pipe_fs_i.h
index 8893711f9171..03d3529c7654 100644
--- a/include/linux/pipe_fs_i.h
+++ b/include/linux/pipe_fs_i.h
@@ -81,13 +81,6 @@ struct pipe_inode_info {
  * in fs/pipe.c for the pipe and generic variants of these hooks.
  */
 struct pipe_buf_operations {
-	/*
-	 * This is set to 1, if the generic pipe read/write may coalesce
-	 * data into an existing buffer. If this is set to 0, a new pipe
-	 * page segment is always used for new data.
-	 */
-	int can_merge;
-
 	/*
 	 * ->confirm() verifies that the data in the pipe buffer is there
 	 * and that the contents are good. If the pages in the pipe belong
diff --git a/kernel/relay.c b/kernel/relay.c
index 04f248644e06..db3e419c25a6 100644
--- a/kernel/relay.c
+++ b/kernel/relay.c
@@ -1175,7 +1175,6 @@ static void relay_pipe_buf_release(struct pipe_inode_info *pipe,
 }
 
 static const struct pipe_buf_operations relay_pipe_buf_ops = {
-	.can_merge = 0,
 	.confirm = generic_pipe_buf_confirm,
 	.release = relay_pipe_buf_release,
 	.steal = generic_pipe_buf_steal,
diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index bf6f1d70484d..820eff45e907 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -5803,7 +5803,6 @@ static void tracing_spd_release_pipe(struct splice_pipe_desc *spd,
 }
 
 static const struct pipe_buf_operations tracing_pipe_buf_ops = {
-	.can_merge		= 0,
 	.confirm		= generic_pipe_buf_confirm,
 	.release		= generic_pipe_buf_release,
 	.steal			= generic_pipe_buf_steal,
@@ -6823,7 +6822,6 @@ static void buffer_pipe_buf_get(struct pipe_inode_info *pipe,
 
 /* Pipe buffer operations for a buffer. */
 static const struct pipe_buf_operations buffer_pipe_buf_ops = {
-	.can_merge		= 0,
 	.confirm		= generic_pipe_buf_confirm,
 	.release		= buffer_pipe_buf_release,
 	.steal			= generic_pipe_buf_steal,
diff --git a/net/smc/smc_rx.c b/net/smc/smc_rx.c
index bbcf0fe4ae10..413a6abf227e 100644
--- a/net/smc/smc_rx.c
+++ b/net/smc/smc_rx.c
@@ -136,7 +136,6 @@ static int smc_rx_pipe_buf_nosteal(struct pipe_inode_info *pipe,
 }
 
 static const struct pipe_buf_operations smc_pipe_ops = {
-	.can_merge = 0,
 	.confirm = generic_pipe_buf_confirm,
 	.release = smc_rx_pipe_buf_release,
 	.steal = smc_rx_pipe_buf_nosteal,
-- 
2.19.0.605.g01d371f741-goog


^ permalink raw reply related	[flat|nested] 4+ messages in thread

* Re: [PATCH 1/2] splice: don't merge into linked buffers
  2018-10-15 15:04 [PATCH 1/2] splice: don't merge into linked buffers Jann Horn
  2018-10-15 15:04 ` [PATCH 2/2] pipe: remove ->can_merge from pipe_buf_operations Jann Horn
@ 2018-10-15 21:00 ` Eric Biggers
  2018-10-15 21:13   ` Jann Horn
  1 sibling, 1 reply; 4+ messages in thread
From: Eric Biggers @ 2018-10-15 21:00 UTC (permalink / raw)
  To: Jann Horn
  Cc: Al Viro, Miklos Szeredi, Jens Axboe, Jens Axboe, linux-kernel,
	linux-fsdevel, Kees Cook

On Mon, Oct 15, 2018 at 05:04:18PM +0200, Jann Horn wrote:
> Before this patch, it was possible for two pipes to affect each other after
> data had been transferred between them with tee():
> 
> ============
> $ cat tee_test.c
> 
> int main(void) {
>   int pipe_a[2];
>   if (pipe(pipe_a)) err(1, "pipe");
>   int pipe_b[2];
>   if (pipe(pipe_b)) err(1, "pipe");
>   if (write(pipe_a[1], "abcd", 4) != 4) err(1, "write");
>   if (tee(pipe_a[0], pipe_b[1], 2, 0) != 2) err(1, "tee");
>   if (write(pipe_b[1], "xx", 2) != 2) err(1, "write");
> 
>   char buf[5];
>   if (read(pipe_a[0], buf, 4) != 4) err(1, "read");
>   buf[4] = 0;
>   printf("got back: '%s'\n", buf);
> }
> $ gcc -o tee_test tee_test.c
> $ ./tee_test
> got back: 'abxx'
> $
> ============
> 
> Fix it by explicitly marking buffers as mergeable and clearing that flag in
> splice_pipe_to_pipe() and link_pipe().
> 
> Cc: <stable@vger.kernel.org>
> Fixes: 7c77f0b3f920 ("splice: implement pipe to pipe splicing")
> Fixes: 70524490ee2e ("[PATCH] splice: add support for sys_tee()")
> Signed-off-by: Jann Horn <jannh@google.com>
> ---
> Cleanup in the next patch, to simplify backporting.
> 
>  fs/pipe.c                 | 5 +++--
>  fs/splice.c               | 6 ++++++
>  include/linux/pipe_fs_i.h | 8 ++++++++
>  3 files changed, 17 insertions(+), 2 deletions(-)
> 
> diff --git a/fs/pipe.c b/fs/pipe.c
> index bdc5d3c0977d..4e2eee77f855 100644
> --- a/fs/pipe.c
> +++ b/fs/pipe.c
> @@ -379,7 +379,8 @@ pipe_write(struct kiocb *iocb, struct iov_iter *from)
>  		struct pipe_buffer *buf = pipe->bufs + lastbuf;
>  		int offset = buf->offset + buf->len;
>  
> -		if (buf->ops->can_merge && offset + chars <= PAGE_SIZE) {
> +		if (buf->ops->can_merge && offset + chars <= PAGE_SIZE &&
> +		    (buf->flags & PIPE_BUF_FLAG_MAYMERGE)) {
>  			ret = pipe_buf_confirm(pipe, buf);
>  			if (ret)
>  				goto out;
> @@ -439,7 +440,7 @@ pipe_write(struct kiocb *iocb, struct iov_iter *from)
>  			buf->ops = &anon_pipe_buf_ops;
>  			buf->offset = 0;
>  			buf->len = copied;
> -			buf->flags = 0;
> +			buf->flags = PIPE_BUF_FLAG_MAYMERGE;
>  			if (is_packetized(filp)) {
>  				buf->ops = &packet_pipe_buf_ops;
>  				buf->flags = PIPE_BUF_FLAG_PACKET;
> diff --git a/fs/splice.c b/fs/splice.c
> index b3daa971f597..111977c80dfd 100644
> --- a/fs/splice.c
> +++ b/fs/splice.c
> @@ -1593,6 +1593,9 @@ static int splice_pipe_to_pipe(struct pipe_inode_info *ipipe,
>  			 */
>  			obuf->flags &= ~PIPE_BUF_FLAG_GIFT;
>  
> +			/* We can't merge data into a buffer we don't own. */
> +			obuf->flags &= ~PIPE_BUF_FLAG_MAYMERGE;
> +
>  			obuf->len = len;
>  			opipe->nrbufs++;
>  			ibuf->offset += obuf->len;
> @@ -1667,6 +1670,9 @@ static int link_pipe(struct pipe_inode_info *ipipe,
>  		 */
>  		obuf->flags &= ~PIPE_BUF_FLAG_GIFT;
>  
> +		/* We can't merge data into a buffer we don't own. */
> +		obuf->flags &= ~PIPE_BUF_FLAG_MAYMERGE;
> +
>  		if (obuf->len > len)
>  			obuf->len = len;
>  
> diff --git a/include/linux/pipe_fs_i.h b/include/linux/pipe_fs_i.h
> index 5a3bb3b7c9ad..8893711f9171 100644
> --- a/include/linux/pipe_fs_i.h
> +++ b/include/linux/pipe_fs_i.h
> @@ -8,6 +8,14 @@
>  #define PIPE_BUF_FLAG_ATOMIC	0x02	/* was atomically mapped */
>  #define PIPE_BUF_FLAG_GIFT	0x04	/* page is a gift */
>  #define PIPE_BUF_FLAG_PACKET	0x08	/* read() as a packet */
> +/*
> + * Set this flag if the generic pipe read/write may coalesce data into an
> + * existing buffer. If this is not set, a new pipe page segment is always used
> + * for new data.
> + * When pipe data is copied by reference (as in the tee() syscall), this flag
> + * must be cleared on the copy.
> + */
> +#define PIPE_BUF_FLAG_MAYMERGE	0x10
>  
>  /**
>   *	struct pipe_buffer - a linux kernel pipe buffer
> -- 
> 2.19.0.605.g01d371f741-goog
> 

Deja vu... https://marc.info/?l=linux-fsdevel&m=149003133809192

Thanks for fixing this; I think this is the right fix.
I verified it works for my reproducer too.

Did you check whether fuse_dev_splice_write() needs to clear
PIPE_BUF_FLAG_MAYMERGE?

- Eric

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH 1/2] splice: don't merge into linked buffers
  2018-10-15 21:00 ` [PATCH 1/2] splice: don't merge into linked buffers Eric Biggers
@ 2018-10-15 21:13   ` Jann Horn
  0 siblings, 0 replies; 4+ messages in thread
From: Jann Horn @ 2018-10-15 21:13 UTC (permalink / raw)
  To: ebiggers
  Cc: Al Viro, Miklos Szeredi, Jens Axboe, Jens Axboe, kernel list,
	linux-fsdevel, Kees Cook

On Mon, Oct 15, 2018 at 11:00 PM Eric Biggers <ebiggers@kernel.org> wrote:
> On Mon, Oct 15, 2018 at 05:04:18PM +0200, Jann Horn wrote:
> > Before this patch, it was possible for two pipes to affect each other after
> > data had been transferred between them with tee():
> >
> > ============
> > $ cat tee_test.c
> >
> > int main(void) {
> >   int pipe_a[2];
> >   if (pipe(pipe_a)) err(1, "pipe");
> >   int pipe_b[2];
> >   if (pipe(pipe_b)) err(1, "pipe");
> >   if (write(pipe_a[1], "abcd", 4) != 4) err(1, "write");
> >   if (tee(pipe_a[0], pipe_b[1], 2, 0) != 2) err(1, "tee");
> >   if (write(pipe_b[1], "xx", 2) != 2) err(1, "write");
> >
> >   char buf[5];
> >   if (read(pipe_a[0], buf, 4) != 4) err(1, "read");
> >   buf[4] = 0;
> >   printf("got back: '%s'\n", buf);
> > }
> > $ gcc -o tee_test tee_test.c
> > $ ./tee_test
> > got back: 'abxx'
> > $
> > ============
> >
> > Fix it by explicitly marking buffers as mergeable and clearing that flag in
> > splice_pipe_to_pipe() and link_pipe().
> >
> > Cc: <stable@vger.kernel.org>
> > Fixes: 7c77f0b3f920 ("splice: implement pipe to pipe splicing")
> > Fixes: 70524490ee2e ("[PATCH] splice: add support for sys_tee()")
> > Signed-off-by: Jann Horn <jannh@google.com>
> > ---
> > Cleanup in the next patch, to simplify backporting.
> >
> >  fs/pipe.c                 | 5 +++--
> >  fs/splice.c               | 6 ++++++
> >  include/linux/pipe_fs_i.h | 8 ++++++++
> >  3 files changed, 17 insertions(+), 2 deletions(-)
> >
> > diff --git a/fs/pipe.c b/fs/pipe.c
> > index bdc5d3c0977d..4e2eee77f855 100644
> > --- a/fs/pipe.c
> > +++ b/fs/pipe.c
> > @@ -379,7 +379,8 @@ pipe_write(struct kiocb *iocb, struct iov_iter *from)
> >               struct pipe_buffer *buf = pipe->bufs + lastbuf;
> >               int offset = buf->offset + buf->len;
> >
> > -             if (buf->ops->can_merge && offset + chars <= PAGE_SIZE) {
> > +             if (buf->ops->can_merge && offset + chars <= PAGE_SIZE &&
> > +                 (buf->flags & PIPE_BUF_FLAG_MAYMERGE)) {
> >                       ret = pipe_buf_confirm(pipe, buf);
> >                       if (ret)
> >                               goto out;
> > @@ -439,7 +440,7 @@ pipe_write(struct kiocb *iocb, struct iov_iter *from)
> >                       buf->ops = &anon_pipe_buf_ops;
> >                       buf->offset = 0;
> >                       buf->len = copied;
> > -                     buf->flags = 0;
> > +                     buf->flags = PIPE_BUF_FLAG_MAYMERGE;
> >                       if (is_packetized(filp)) {
> >                               buf->ops = &packet_pipe_buf_ops;
> >                               buf->flags = PIPE_BUF_FLAG_PACKET;
> > diff --git a/fs/splice.c b/fs/splice.c
> > index b3daa971f597..111977c80dfd 100644
> > --- a/fs/splice.c
> > +++ b/fs/splice.c
> > @@ -1593,6 +1593,9 @@ static int splice_pipe_to_pipe(struct pipe_inode_info *ipipe,
> >                        */
> >                       obuf->flags &= ~PIPE_BUF_FLAG_GIFT;
> >
> > +                     /* We can't merge data into a buffer we don't own. */
> > +                     obuf->flags &= ~PIPE_BUF_FLAG_MAYMERGE;
> > +
> >                       obuf->len = len;
> >                       opipe->nrbufs++;
> >                       ibuf->offset += obuf->len;
> > @@ -1667,6 +1670,9 @@ static int link_pipe(struct pipe_inode_info *ipipe,
> >                */
> >               obuf->flags &= ~PIPE_BUF_FLAG_GIFT;
> >
> > +             /* We can't merge data into a buffer we don't own. */
> > +             obuf->flags &= ~PIPE_BUF_FLAG_MAYMERGE;
> > +
> >               if (obuf->len > len)
> >                       obuf->len = len;
> >
> > diff --git a/include/linux/pipe_fs_i.h b/include/linux/pipe_fs_i.h
> > index 5a3bb3b7c9ad..8893711f9171 100644
> > --- a/include/linux/pipe_fs_i.h
> > +++ b/include/linux/pipe_fs_i.h
> > @@ -8,6 +8,14 @@
> >  #define PIPE_BUF_FLAG_ATOMIC 0x02    /* was atomically mapped */
> >  #define PIPE_BUF_FLAG_GIFT   0x04    /* page is a gift */
> >  #define PIPE_BUF_FLAG_PACKET 0x08    /* read() as a packet */
> > +/*
> > + * Set this flag if the generic pipe read/write may coalesce data into an
> > + * existing buffer. If this is not set, a new pipe page segment is always used
> > + * for new data.
> > + * When pipe data is copied by reference (as in the tee() syscall), this flag
> > + * must be cleared on the copy.
> > + */
> > +#define PIPE_BUF_FLAG_MAYMERGE       0x10
> >
> >  /**
> >   *   struct pipe_buffer - a linux kernel pipe buffer
> > --
> > 2.19.0.605.g01d371f741-goog
> >
>
> Deja vu... https://marc.info/?l=linux-fsdevel&m=149003133809192

Oh, hah, even the reproducer looks almost the same. I hadn't seen that...

> Thanks for fixing this; I think this is the right fix.
> I verified it works for my reproducer too.
>
> Did you check whether fuse_dev_splice_write() needs to clear
> PIPE_BUF_FLAG_MAYMERGE?

No, I hadn't looked at that. But from what I can tell,
fuse_dev_splice_write() shouldn't write into the provided buffers,
only read from them.

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2018-10-15 21:13 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-10-15 15:04 [PATCH 1/2] splice: don't merge into linked buffers Jann Horn
2018-10-15 15:04 ` [PATCH 2/2] pipe: remove ->can_merge from pipe_buf_operations Jann Horn
2018-10-15 21:00 ` [PATCH 1/2] splice: don't merge into linked buffers Eric Biggers
2018-10-15 21:13   ` Jann Horn

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).