[PATCH] io_uring_enter(2): clarify OP_READ and OP_WRITE

[PATCH] io_uring_enter(2): clarify OP_READ and OP_WRITE

[Drew DeVault]

These do not advance the internal file offset, making them behave more
like pread/pwrite than read/write.
---
 man/io_uring_enter.2 | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/man/io_uring_enter.2 b/man/io_uring_enter.2
index f898ffd..81044c1 100644
--- a/man/io_uring_enter.2
+++ b/man/io_uring_enter.2
@@ -576,16 +576,18 @@ for the general description of the related system call. Available since 5.6.
 .TP
 .B IORING_OP_WRITE
 Issue the equivalent of a
-.BR read(2)
+.BR pread(2)
 or
-.BR write(2)
+.BR pwrite(2)
 system call.
 .I fd
 is the file descriptor to be operated on,
 .I addr
-contains the buffer in question, and
+contains the buffer in question,
 .I len
-contains the length of the IO operation. These are non-vectored versions of the
+contains the length of the IO operation, and
+.I offs
+contains the read or write offset. These are non-vectored versions of the
 .B IORING_OP_READV
 and
 .B IORING_OP_WRITEV
-- 
2.31.1

Re: [PATCH] io_uring_enter(2): clarify OP_READ and OP_WRITE

[Jens Axboe]

On 5/23/21 10:20 AM, Drew DeVault wrote:
> These do not advance the internal file offset, making them behave more
> like pread/pwrite than read/write.

They do, if -1 is used as the offset. Care to update the patch and include
that clarification?

-- 
Jens Axboe

Re: [PATCH] io_uring_enter(2): clarify OP_READ and OP_WRITE

[Drew DeVault]

On Sun May 23, 2021 at 12:41 PM EDT, Jens Axboe wrote:
> They do, if -1 is used as the offset. Care to update the patch and
> include that clarification?

Certainly.