aboutsummaryrefslogtreecommitdiffstats
path: root/man2/open.2
diff options
context:
space:
mode:
authorMichael Kerrisk <mtk.manpages@gmail.com>2006-01-15 04:39:23 +0000
committerMichael Kerrisk <mtk.manpages@gmail.com>2006-01-15 04:39:23 +0000
commitb50582ebfa66a369df4686281b9f9c1c68cf1fa2 (patch)
tree2fc779bc48857117a12a0e0308f8de7c7fe05e1a /man2/open.2
parenta3a11667835380145d7d6f87c22ce940c54074af (diff)
downloadman-pages-b50582ebfa66a369df4686281b9f9c1c68cf1fa2.tar.gz
Added BUG noting that O_ASYNC can't be enabled via
open(): fcntl() must be used for this purpose.
Diffstat (limited to 'man2/open.2')
-rw-r--r--man2/open.215
1 files changed, 12 insertions, 3 deletions
diff --git a/man2/open.2 b/man2/open.2
index 22e25e98de..c46ae56651 100644
--- a/man2/open.2
+++ b/man2/open.2
@@ -114,9 +114,8 @@ appending to a file, so the client kernel has to simulate it, which
can't be done without a race condition.
.TP
.B O_ASYNC
-.\" FIXME -- as far as I can tell O_ASYNC doesn't work for open(2),
-.\" only when set via fcntl(2) -- MTK, Dec 04
-Generate a signal (SIGIO by default, but this can be changed via
+Enable signal-driven I/O:
+generate a signal (SIGIO by default, but this can be changed via
.BR fcntl (2))
when input or output becomes possible on this file descriptor.
This feature is only available for terminals, pseudo-terminals,
@@ -500,6 +499,16 @@ macro to get its definition.
"The thing that has always disturbed me about O_DIRECT is that the whole
interface is just stupid, and was probably designed by a deranged monkey
on some serious mind-controlling substances." \(em Linus
+
+Currently, it is not possible to enable signal-driven
+I/O by specifying
+.B O_ASYNC
+when calling
+.BR open ();
+use
+.BR fcntl (2)
+to enable this flag.
+.\" FIXME Raise a bug report about this O_ASYNC behaviour!?
.SH RESTRICTIONS
There are many infelicities in the protocol underlying NFS, affecting
amongst others