Hi PrzemysÅaw,
Thank you for comment :)
Would you review this patch again?
Regards,
Akira Fujita
---
--- man-pages-3.32-a/man2/sendfile.2 2010-12-03 16:01:59.000000000 +0900
+++ man-pages-3.32-b/man2/sendfile.2 2011-02-14 09:02:57.000000000 +0900
@@ -87,15 +87,17 @@ and the file offset will be updated by t
.I count
is the number of bytes to copy between the file descriptors.
-Presently (Linux 2.6.9):
-.IR in_fd ,
+.IR in_fd
must correspond to a file which supports
.BR mmap (2)-like
operations
-(i.e., it cannot be a socket);
-and
-.I out_fd
-must refer to a socket.
+(i.e., it cannot be a socket).
+
+.IR out_fd
+was required to be a socket, but since Linux 2.6.33 it can be
+any file. If it is a regular file, then
+.BR sendfile()
+changes its offset appropriately.
Applications may wish to fall back to
.BR read (2)/ write (2)
@@ -165,7 +167,7 @@ to minimize the number of packets and to
In Linux 2.4 and earlier,
.I out_fd
-could refer to a regular file, and
+could also refer to a regular file, and
.BR sendfile ()
changed the current offset of that file.
(2011/02/13 5:11), PrzemysÅaw PaweÅczyk wrote:
Hi,
2011/1/27 Akira Fujita<a-fujita@xxxxxxxxxxxxx>:
commit cc56f7de7f00d188c7c4da1e9861581853b9e92f made
sendfile(2) can work with any output file.
Therefore the out_fd of sendfile(2) can refer to any file,
but current manual (man-pages-3.32) has not been changed so far.
And I thought I was the only one discovering it recently (beside a few
obvious people year ago). :)
diff -Nrup man-pages-3.32-a/man2/sendfile.2 man-pages-3.32-b/man2/sendfile.2
--- man-pages-3.32-a/man2/sendfile.2 2010-12-03 16:01:59.000000000 +0900
+++ man-pages-3.32-b/man2/sendfile.2 2011-01-27 16:03:48.000000000 +0900
@@ -87,15 +87,11 @@ and the file offset will be updated by t
.I count
is the number of bytes to copy between the file descriptors.
-Presently (Linux 2.6.9):
-.IR in_fd ,
+.IR in_fd
must correspond to a file which supports
.BR mmap (2)-like
operations
-(i.e., it cannot be a socket);
-and
-.I out_fd
-must refer to a socket.
+(i.e., it cannot be a socket).
I don't think that completely removing out_fd socket restriction part
is a proper way of fixing this page. Such information should not be
lost.
I would add here sth like:
out_fd was required to be a socket, but since Linux 2.6.33 it can be
any file. If it is a regular file, then sendfile() changes its offset
appropriately.
Applications may wish to fall back to
.BR read (2)/ write (2)
@@ -168,6 +164,9 @@ In Linux 2.4 and earlier,
could refer to a regular file, and
s/could/could also/
.BR sendfile ()
changed the current offset of that file.
+Since 2.6.33,
+.I out_fd
+can refer to any file.
IMO this change shouldn't be mentioned in NOTES section.
The original Linux
.BR sendfile ()
Regards.
--
Akira Fujita <a-fujita@xxxxxxxxxxxxx>
The First Fundamental Software Development Group,
Software Development Division,
NEC Software Tohoku, Ltd.
--
To unsubscribe from this list: send the line "unsubscribe linux-man" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
