From: kingsley@aurema.com
To: Tom Zanussi <zanussi@us.ibm.com>, Karim Yaghmour <karim@opersys.com>
Cc: linux-kernel@vger.kernel.org
Subject: read() on relayfs channel returns premature 0
Date: Wed, 23 Mar 2005 20:02:54 +1100 [thread overview]
Message-ID: <20050323090254.GA10630@aurema.com> (raw)
[-- Attachment #1: Type: text/plain, Size: 909 bytes --]
Hi
I'm using relayfs to relay data from a kernel module to user space on
a SuSE 2.6.5 kernel. I'm not absolutely sure what version of relayfs
has been back ported to it.
While reading data from the channel I've been seeing read() return 0
prematurely. However, the 0 does not signify that the file is being
closed for there is still data available afterwards.
I've noticed that zeros occur when roughly one page of data has been
read. I suspect that they occur whenever there is a read across the
relayfs sub-buffers.
Now I understand that this is not the latest release of relayfs (there
are the redux patches, which I have yet to try). Nonetheless I'd like
to know whether this behaviour is deliberate. Is it?
Thanks,
--
Kingsley
P.S. I've been able to get around this by deliberately modifying
do_read() with the attached patch. I'm not absolutely sure its
correct but it seems to work.
[-- Attachment #2: relayfs-premature-zero.patch --]
[-- Type: text/plain, Size: 1340 bytes --]
Index: fs/relayfs/relay.c
===================================================================
RCS file: /export/cvsroot/SuSE-Kernel-2.4.21/fs/relayfs/Attic/relay.c,v
retrieving revision 1.1.2.1
diff -u -r1.1.2.1 relay.c
--- fs/relayfs/relay.c 18 Feb 2005 00:00:51 -0000 1.1.2.1
+++ fs/relayfs/relay.c 23 Mar 2005 09:00:42 -0000
@@ -1445,26 +1445,22 @@
avail_offset = cur_idx = relay_get_offset(rchan, &max_offset);
+ last_buf_byte_offset = (read_bufno + 1) * buf_size - 1;
if (cur_idx == read_offset) {
if (atomic_read(&rchan->suspended) == 1) {
- read_offset += 1;
+ read_offset = last_buf_byte_offset + 1;
if (read_offset >= max_offset)
read_offset = 0;
*actual_read_offset = read_offset;
} else {
- *new_offset = read_offset;
+ *new_offset = read_offset;
return 0;
}
- } else {
- last_buf_byte_offset = (read_bufno + 1) * buf_size - 1;
- if (read_offset == last_buf_byte_offset) {
- if (unused_bytes != 1) {
- read_offset += 1;
- if (read_offset >= max_offset)
- read_offset = 0;
- *actual_read_offset = read_offset;
- }
- }
+ } else if ((read_offset + unused_bytes) > last_buf_byte_offset) {
+ read_offset = last_buf_byte_offset + 1;
+ if (read_offset >= max_offset)
+ read_offset = 0;
+ *actual_read_offset = read_offset;
}
read_bufno = read_offset / buf_size;
next reply other threads:[~2005-03-23 9:12 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-03-23 9:02 kingsley [this message]
2005-03-23 15:29 ` read() on relayfs channel returns premature 0 Tom Zanussi
2005-03-24 1:29 ` Kingsley Cheung
2005-03-24 6:56 ` Jan Engelhardt
2005-03-24 19:11 ` Tom Zanussi
2005-03-24 19:45 ` Jan Engelhardt
2005-03-25 12:27 ` Karim Yaghmour
2005-03-28 23:43 ` Kingsley Cheung
2005-04-18 1:29 ` Relayfs Question: Use of relay_reset(). Potential race? Kingsley Cheung
2005-04-18 15:56 ` Tom Zanussi
2005-04-19 0:40 ` Kingsley Cheung
2005-05-04 9:04 ` kingsley
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20050323090254.GA10630@aurema.com \
--to=kingsley@aurema.com \
--cc=karim@opersys.com \
--cc=linux-kernel@vger.kernel.org \
--cc=zanussi@us.ibm.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.