From: Nishanth Aravamudan <nacc@us.ibm.com>
To: akpm@osdl.org
Cc: dwmw2@infradead.org, bunk@stusta.de, johnstul@us.ibm.com,
drepper@redhat.com, Franz.Fischer@goyellow.de,
LKML <linux-kernel@vger.kernel.org>
Subject: [PATCH][Bug 5132] fix sys_poll() large timeout handling
Date: Wed, 31 Aug 2005 13:01:09 -0700 [thread overview]
Message-ID: <20050831200109.GB3017@us.ibm.com> (raw)
Sorry everybody, forgot the most important Cc: :)
-Nish
Hi Andrew,
In looking at Bug 5132 and sys_poll(), I think there is a flaw in the
current code.
The @timeout parameter to sys_poll() is in milliseconds but we compare
it to (MAX_SCHEDULE_TIMEOUT / HZ), which is jiffies/jiffies-per-sec or
seconds. That seems blatantly broken. Also, I think we are better served
by converting to jiffies first then comparing, as opposed to converting
our maximum to milliseconds (or seconds, incorrectly) and comparing.
Comments, suggestions for improvement?
Description: The current sys_poll() implementation does not seem to
handle large timeouts correctly. Any value in milliseconds (@timeout)
which exceeds the maximum representable jiffy value
(MAX_SCHEDULE_TIMEOUT) should result in a MAX_SCHEDULE_TIMEOUT
schedule_timeout() call. To achieve this, convert @timeout to jiffies
first, then compare to MAX_SCHEDULE_TIMEOUT.
Signed-off-by: Nishanth Aravamudan <nacc@us.ibm.com>
---
fs/select.c | 14 +++++++++-----
1 files changed, 9 insertions(+), 5 deletions(-)
diff -urpN 2.6.13/fs/select.c 2.6.13-dev/fs/select.c
--- 2.6.13/fs/select.c 2005-08-28 17:46:14.000000000 -0700
+++ 2.6.13-dev/fs/select.c 2005-08-31 12:43:52.000000000 -0700
@@ -470,12 +470,16 @@ asmlinkage long sys_poll(struct pollfd _
return -EINVAL;
if (timeout) {
- /* Careful about overflow in the intermediate values */
- if ((unsigned long) timeout < MAX_SCHEDULE_TIMEOUT / HZ)
- timeout = (unsigned long)(timeout*HZ+999)/1000+1;
- else /* Negative or overflow */
- timeout = MAX_SCHEDULE_TIMEOUT;
+ /*
+ * Convert the value from msecs to jiffies - if overflow
+ * occurs we get a negative value, which gets handled by
+ * the next block
+ */
+ timeout = msecs_to_jiffies(timeout) + 1;
}
+ if (timeout < 0) /* Negative requests result in infinite timeouts */
+ timeout = MAX_SCHEDULE_TIMEOUT;
+ /* 0 case falls through */
poll_initwait(&table);
next reply other threads:[~2005-08-31 20:01 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-08-31 20:01 Nishanth Aravamudan [this message]
2005-09-06 21:25 ` [PATCH][Bug 5132] fix sys_poll() large timeout handling Nishanth Aravamudan
2005-09-10 0:35 ` [UPDATE PATCH][Bug " Nishanth Aravamudan
2005-09-10 1:16 ` Andrew Morton
2005-09-10 2:23 ` Nishanth Aravamudan
2005-09-10 2:36 ` Andrew Morton
2005-09-10 2:55 ` Nishanth Aravamudan
2005-09-12 14:30 ` Peter Staubach
2005-09-12 15:05 ` Nishanth Aravamudan
2005-09-12 15:19 ` Peter Staubach
2005-09-12 16:06 ` Nishanth Aravamudan
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=20050831200109.GB3017@us.ibm.com \
--to=nacc@us.ibm.com \
--cc=Franz.Fischer@goyellow.de \
--cc=akpm@osdl.org \
--cc=bunk@stusta.de \
--cc=drepper@redhat.com \
--cc=dwmw2@infradead.org \
--cc=johnstul@us.ibm.com \
--cc=linux-kernel@vger.kernel.org \
/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 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).