From: Paul Jackson <pj@sgi.com>
To: Simon Derr <simon.derr@bull.net>
Cc: akpm@osdl.org, linux-kernel@vger.kernel.org
Subject: [PATCH] cpusets: alternative fix for possible race in cpuset_tasks_read()
Date: Sat, 11 Sep 2004 01:01:20 -0700 [thread overview]
Message-ID: <20040911010120.572595e3.pj@sgi.com> (raw)
In-Reply-To: <Pine.LNX.4.58.0409101632100.2891@daphne.frec.bull.fr>
Here's an alternative fix for the race condition on read that Simon
reports.
Andrew,
Don't apply this one yet, until Simon Derr gets a chance to
compare with his alternative patch, and render his analysis.
Move the code that sets up the character buffer of text to read out
when reading a "tasks" file from the read routine to the open routine.
Multiple cloned threads could be doing the first read on a shared
file descriptor open on a "tasks" file, resulting in confused
or leaked kernel memory as multiple threads initialized the same
file private_data at the same time. Rather than add locks to the
initialization code, move it into the open(), where it belongs anyway.
Signed-off-by: Paul Jackson
Index: 2.6.9-rc1-mm4/kernel/cpuset.c
===================================================================
--- 2.6.9-rc1-mm4.orig/kernel/cpuset.c 2004-09-10 15:27:32.000000000 -0700
+++ 2.6.9-rc1-mm4/kernel/cpuset.c 2004-09-10 21:20:02.000000000 -0700
@@ -1034,7 +1034,7 @@ static int pid_array_to_buf(char *buf, i
return cnt;
}
-static inline struct ctr_struct *cpuset_tasks_mkctr(struct file *file)
+static int cpuset_tasks_open(struct inode *unused, struct file *file)
{
struct cpuset *cs = __d_cs(file->f_dentry->d_parent);
struct ctr_struct *ctr;
@@ -1069,14 +1069,14 @@ static inline struct ctr_struct *cpuset_
kfree(pidarray);
file->private_data = (void *)ctr;
- return ctr;
+ return 0;
err2:
kfree(pidarray);
err1:
kfree(ctr);
err0:
- return NULL;
+ return -ENOMEM;
}
static ssize_t cpuset_tasks_read(struct file *file, char __user *buf,
@@ -1084,13 +1084,6 @@ static ssize_t cpuset_tasks_read(struct
{
struct ctr_struct *ctr = (struct ctr_struct *)file->private_data;
- /* allocate buffer and fill it on first call to read() */
- if (!ctr) {
- ctr = cpuset_tasks_mkctr(file);
- if (!ctr)
- return -ENOMEM;
- }
-
if (*ppos + nbytes > ctr->bufsz)
nbytes = ctr->bufsz - *ppos;
if (copy_to_user(buf, ctr->buf + *ppos, nbytes))
@@ -1121,6 +1114,7 @@ static int cpuset_tasks_release(struct i
static struct cftype cft_tasks = {
.name = "tasks",
+ .open = cpuset_tasks_open,
.read = cpuset_tasks_read,
.release = cpuset_tasks_release,
.private = FILE_TASKLIST,
--
I won't rest till it's the best ...
Programmer, Linux Scalability
Paul Jackson <pj@sgi.com> 1.650.933.1373
next prev parent reply other threads:[~2004-09-11 8:02 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2004-09-10 14:38 [PATCH] cpusets: fix possible race in cpuset_tasks_read() Simon Derr
2004-09-11 8:01 ` Paul Jackson [this message]
2004-09-13 8:24 ` [PATCH] cpusets: alternative fix for " Simon.Derr
2004-09-13 9:25 ` Paul Jackson
2004-09-13 9:39 ` Paul Jackson
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=20040911010120.572595e3.pj@sgi.com \
--to=pj@sgi.com \
--cc=akpm@osdl.org \
--cc=linux-kernel@vger.kernel.org \
--cc=simon.derr@bull.net \
/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.