From: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
To: amwang@redhat.com
Cc: opurdila@ixiacom.com, eric.dumazet@gmail.com,
netdev@vger.kernel.org, nhorman@tuxdriver.com,
davem@davemloft.net, ebiederm@xmission.com,
linux-kernel@vger.kernel.org
Subject: Re: [Patch 3/3] net: reserve ports for applications using fixed port numbers
Date: Tue, 13 Apr 2010 10:21:43 +0900 [thread overview]
Message-ID: <201004130121.o3D1Lhh7099571@www262.sakura.ne.jp> (raw)
In-Reply-To: <20100412100816.5302.74919.sendpatchset@localhost.localdomain>
Hello.
> --- linux-2.6.orig/drivers/infiniband/core/cma.c
> +++ linux-2.6/drivers/infiniband/core/cma.c
> @@ -1980,6 +1980,8 @@ retry:
> /* FIXME: add proper port randomization per like inet_csk_get_port */
> do {
> ret = idr_get_new_above(ps, bind_list, next_port, &port);
> + if (!ret && inet_is_reserved_local_port(port))
> + ret = -EAGAIN;
> } while ((ret == -EAGAIN) && idr_pre_get(ps, GFP_KERNEL));
>
> if (ret)
>
I think above part is wrong. Below program
--------------------
#include <linux/module.h>
#include <linux/sched.h>
#include <linux/idr.h>
static DEFINE_IDR(idr);
static int idr_demo_init(void)
{
int next_port = 65530;
int port = 0;
int ret = -EINTR;
while (!signal_pending(current)) {
msleep(1000);
ret = idr_get_new_above(&idr, NULL, next_port, &port);
printk(KERN_INFO "idr_get_new_above() = %d\n", ret);
if (!ret) {
/* Emulate inet_is_reserved_local_port(port) = true */
printk(KERN_INFO "Port %u is reserved.\n", port);
ret = -EAGAIN;
}
if (ret == -EAGAIN) {
if (idr_pre_get(&idr, GFP_KERNEL)) {
printk(KERN_INFO "idr_pre_get() succeeded.\n");
continue;
}
printk(KERN_INFO "idr_pre_get() failed.\n");
break;
} else {
printk(KERN_INFO "next_port=%u port=%u\n",
next_port, port);
break;
}
}
if (!ret)
idr_remove(&idr, port);
idr_destroy(&idr);
return -EINVAL;
}
module_init(idr_demo_init);
MODULE_LICENSE("GPL");
--------------------
generated below output.
idr_get_new_above() = -11
idr_pre_get() succeeded.
idr_get_new_above() = 0
Port 65530 is reserved.
idr_pre_get() succeeded.
idr_get_new_above() = 0
Port 65531 is reserved.
idr_pre_get() succeeded.
idr_get_new_above() = 0
Port 65532 is reserved.
idr_pre_get() succeeded.
idr_get_new_above() = 0
Port 65533 is reserved.
idr_pre_get() succeeded.
idr_get_new_above() = 0
Port 65534 is reserved.
idr_pre_get() succeeded.
idr_get_new_above() = 0
Port 65535 is reserved.
idr_pre_get() succeeded.
idr_get_new_above() = 0
Port 65536 is reserved.
idr_pre_get() succeeded.
idr_get_new_above() = 0
Port 65537 is reserved.
idr_pre_get() succeeded.
idr_get_new_above() = 0
(...snipped...)
This result suggests that above loop will continue until idr_pre_get() fails
due to out of memory if all ports were reserved.
Also, if idr_get_new_above() returned 0, bind_list (which is a kmalloc()ed
pointer) is already installed into a free slot (see comment on
idr_get_new_above_int()). Thus, simply calling idr_get_new_above() again will
install the same pointer into multiple slots. I guess it will malfunction later.
next prev parent reply other threads:[~2010-04-13 1:22 UTC|newest]
Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-04-12 10:03 [Patch v8 0/3] net: reserve ports for applications using fixed port numbers Amerigo Wang
2010-04-12 10:04 ` [Patch 1/3] sysctl: refactor integer handling proc code Amerigo Wang
2010-04-13 11:18 ` Alexey Dobriyan
2010-04-13 7:35 ` Cong Wang
2010-04-12 10:04 ` [Patch 2/3] sysctl: add proc_do_large_bitmap Amerigo Wang
2010-04-12 10:04 ` [Patch 3/3] net: reserve ports for applications using fixed port numbers Amerigo Wang
2010-04-13 1:21 ` Tetsuo Handa [this message]
2010-04-13 7:13 ` Cong Wang
2010-04-13 8:48 ` Cong Wang
2010-04-13 13:07 ` Tetsuo Handa
2010-04-13 16:32 ` Sean Hefty
2010-04-14 2:01 ` [PATCH] Infiniband: Randomize local port allocation penguin-kernel
2010-04-14 4:38 ` Cong Wang
2010-04-15 0:01 ` Sean Hefty
2010-04-15 2:29 ` Tetsuo Handa
2010-04-15 19:55 ` [PATCH] rdma/cm: " Sean Hefty
2010-04-16 2:22 ` Cong Wang
2010-04-16 13:54 ` Tetsuo Handa
2010-04-16 20:30 ` David Miller
2010-04-20 4:34 ` Cong Wang
2010-04-21 23:19 ` [PATCH] Infiniband: " Roland Dreier
2010-04-21 23:22 ` Roland Dreier
-- strict thread matches above, loose matches on Subject: below --
2010-05-05 10:26 [Patch v10 0/3] net: reserve ports for applications using fixed port numbers Amerigo Wang
2010-05-05 10:27 ` [Patch 3/3] " Amerigo Wang
2010-04-30 8:25 [Patch v9 0/3] " Amerigo Wang
2010-04-30 8:25 ` [Patch 3/3] " Amerigo Wang
2010-04-09 10:10 [Patch v7 0/3] " Amerigo Wang
2010-04-09 10:11 ` [Patch 3/3] " Amerigo Wang
2010-04-09 13:21 ` Tetsuo Handa
2010-04-12 6:52 ` Cong Wang
2010-04-12 7:06 ` Tetsuo Handa
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=201004130121.o3D1Lhh7099571@www262.sakura.ne.jp \
--to=penguin-kernel@i-love.sakura.ne.jp \
--cc=amwang@redhat.com \
--cc=davem@davemloft.net \
--cc=ebiederm@xmission.com \
--cc=eric.dumazet@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev@vger.kernel.org \
--cc=nhorman@tuxdriver.com \
--cc=opurdila@ixiacom.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 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).