[PATCH net 1/1] net: sock: Use double send/recv buff value to compare with max value

Subsystems: networking [general], networking [sockets], the rest

STALE3476d

6 messages, 3 authors, 2017-02-15 · open the first message on its own page

[PATCH net 1/1] net: sock: Use double send/recv buff value to compare with max value

From: <hidden>
Date: 2017-02-08 13:17:14

From: Gao Feng <redacted>

Because the value of SO_SNDBUF and SO_RCVBUF is doubled before
assignment, so the real value of send and recv buffer could be more
than the max sysctl config sysctl_wmem_max and sysctl_rmem_max.

Now use doulbe send/recv buffer value to compare with sysctl_wmem_max
and sysctl_rmem_max, and it keeps consistence with SOCK_MIN_SNDBUF
and SOCK_MIN_RCVBUF.

Signed-off-by: Gao Feng <redacted>
---
 net/core/sock.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/net/core/sock.c b/net/core/sock.c
index 4eca27d..fa40dff 100644
--- a/net/core/sock.c
+++ b/net/core/sock.c

@@ -712,7 +712,7 @@ int sock_setsockopt(struct socket *sock, int level, int optname,
 		 * play 'guess the biggest size' games. RCVBUF/SNDBUF
 		 * are treated in BSD as hints
 		 */
-		val = min_t(u32, val, sysctl_wmem_max);
+		val = min_t(u32, val * 2, sysctl_wmem_max);
 set_sndbuf:
 		sk->sk_userlocks |= SOCK_SNDBUF_LOCK;
 		sk->sk_sndbuf = max_t(int, val * 2, SOCK_MIN_SNDBUF);

@@ -733,7 +733,7 @@ int sock_setsockopt(struct socket *sock, int level, int optname,
 		 * play 'guess the biggest size' games. RCVBUF/SNDBUF
 		 * are treated in BSD as hints
 		 */
-		val = min_t(u32, val, sysctl_rmem_max);
+		val = min_t(u32, val * 2, sysctl_rmem_max);
 set_rcvbuf:
 		sk->sk_userlocks |= SOCK_RCVBUF_LOCK;
 		/*

-- 
1.9.1

Re: [PATCH net 1/1] net: sock: Use double send/recv buff value to compare with max value

From: Eric Dumazet <hidden>
Date: 2017-02-08 16:00:32

On Wed, 2017-02-08 at 21:07 +0800, fgao@ikuai8.com wrote:

From: Gao Feng <redacted>

Because the value of SO_SNDBUF and SO_RCVBUF is doubled before
assignment, so the real value of send and recv buffer could be more
than the max sysctl config sysctl_wmem_max and sysctl_rmem_max.

Now use doulbe send/recv buffer value to compare with sysctl_wmem_max
and sysctl_rmem_max, and it keeps consistence with SOCK_MIN_SNDBUF
and SOCK_MIN_RCVBUF.

Signed-off-by: Gao Feng <redacted>
---

Looks completely bogus, based on your comprehension of this code.

If you need to, fix the doc, not this code.

Unless you give more details of course, why we should take your patch.

Re: [PATCH net 1/1] net: sock: Use double send/recv buff value to compare with max value

From: Gao Feng <hidden>
Date: 2017-02-09 09:10:38

On Thu, Feb 9, 2017 at 12:00 AM, Eric Dumazet [off-list ref] wrote:

On Wed, 2017-02-08 at 21:07 +0800, fgao@ikuai8.com wrote:

quoted

From: Gao Feng <redacted>

Because the value of SO_SNDBUF and SO_RCVBUF is doubled before
assignment, so the real value of send and recv buffer could be more
than the max sysctl config sysctl_wmem_max and sysctl_rmem_max.

Now use doulbe send/recv buffer value to compare with sysctl_wmem_max
and sysctl_rmem_max, and it keeps consistence with SOCK_MIN_SNDBUF
and SOCK_MIN_RCVBUF.

Signed-off-by: Gao Feng <redacted>
---

Looks completely bogus, based on your comprehension of this code.

It is a config param, user could config any value.
So why give it one bogus?
If need more, user could config it by himself.

If you need to, fix the doc, not this code.

The current codes mean the buffer could exceed the sysctl max value.
It seems inconsistent.

Regards
Feng

Unless you give more details of course, why we should take your patch.

Re: [PATCH net 1/1] net: sock: Use double send/recv buff value to compare with max value

From: Eric Dumazet <hidden>
Date: 2017-02-09 14:35:02

On Thu, 2017-02-09 at 17:08 +0800, Gao Feng wrote:

On Thu, Feb 9, 2017 at 12:00 AM, Eric Dumazet [off-list ref] wrote:

quoted

On Wed, 2017-02-08 at 21:07 +0800, fgao@ikuai8.com wrote:

quoted

From: Gao Feng <redacted>

Because the value of SO_SNDBUF and SO_RCVBUF is doubled before
assignment, so the real value of send and recv buffer could be more
than the max sysctl config sysctl_wmem_max and sysctl_rmem_max.

Now use doulbe send/recv buffer value to compare with sysctl_wmem_max
and sysctl_rmem_max, and it keeps consistence with SOCK_MIN_SNDBUF
and SOCK_MIN_RCVBUF.

Signed-off-by: Gao Feng <redacted>
---

Looks completely bogus, based on your comprehension of this code.

It is a config param, user could config any value.
So why give it one bogus?
If need more, user could config it by himself.

quoted

If you need to, fix the doc, not this code.

The current codes mean the buffer could exceed the sysctl max value.
It seems inconsistent.


Well, after your patch, we end up doubling what the user requested.

Say user used val = 20000

val = min_t(u32, val * 2, sysctl_wmem_max);

Now val is 40000.

sk->sk_sndbuf = max_t(int, val * 2, SOCK_MIN_SNDBUF);

Now sk_sndbuf is 80000, or am I missing something ?

Before your patch, sk_sndbuf was 40000. See the difference ?

Some applications carefully tuned this, some people care about
bufferbloat.

This is too late now to change this behavior, especially the way you did
it, breaking some applications.

Unless you can show me a real application that hit an issue, I will not
accept your patch.

[lkp-robot] [net] 0148239373: ltp.test_1_to_1_sockopt.fail

From: kernel test robot <hidden>
Date: 2017-02-15 01:47:04

FYI, we noticed the following commit:

commit: 0148239373801a9c0c63eefc481b710ae7ce2941 ("net: sock: Use double send/recv buff value to compare with max value")
url: https://github.com/0day-ci/linux/commits/fgao-ikuai8-com/net-sock-Use-double-send-recv-buff-value-to-compare-with-max-value/20170208-213745


in testcase: ltp
with following parameters:

	test: net.sctp

test-description: The LTP testsuite contains a collection of tools for testing the Linux kernel and related features.
test-url: http://linux-test-project.github.io/


on test machine: 88 threads Intel(R) Xeon(R) CPU E5-2699 v4 @ 2.20GHz with 128G memory

caused below changes (please refer to attached dmesg/kmsg for entire log/backtrace):



user  :notice: [   54.674772] tag=test_1_to_1_sockopt stime=1487093637

user  :notice: [   54.675078] cmdline="test_1_to_1_sockopt"

user  :notice: [   54.675194] contacts=""

user  :notice: [   54.675333] analysis=exit

user  :notice: [   54.675513] <<<test_output>>>

user  :notice: [   54.676497] test_1_to_1_sockopt.c    1  TPASS  :  setsockopt() with a bad socket descriptor - EBADF

user  :notice: [   54.677439] test_1_to_1_sockopt.c    2  TPASS  :  setsockopt() with an invalid socket - ENOTSOCK

user  :notice: [   54.678409] test_1_to_1_sockopt.c    3  TPASS  :  setsockopt() with an invalid level - ENOPROTOOPT

user  :notice: [   54.679378] test_1_to_1_sockopt.c    4  TPASS  :  setsockopt() with invalid option buffer - EFAULT

user  :notice: [   54.680375] test_1_to_1_sockopt.c    5  TPASS  :  setsockopt() with invalid option name - EOPNOTSUPP

user  :notice: [   54.681354] test_1_to_1_sockopt.c    6  TPASS  :  getsockopt() with a bad socket descriptor - EBADF

user  :notice: [   54.682297] test_1_to_1_sockopt.c    7  TPASS  :  getsockopt() with an invalid socket - ENOTSOCK

user  :notice: [   54.683265] test_1_to_1_sockopt.c    8  TPASS  :  getsockopt() with invalid option buffer - EFAULT

user  :notice: [   54.684267] test_1_to_1_sockopt.c    9  TPASS  :  getsockopt() with invalid option name - EOPNOTSUPP

user  :notice: [   54.685079] test_1_to_1_sockopt.c   10  TPASS  :  getsockopt() SCTP_INITMSG - SUCCESS

user  :notice: [   54.685885] test_1_to_1_sockopt.c   11  TPASS  :  setsockopt() SCTP_INITMSG - SUCCESS

user  :notice: [   54.686653] test_1_to_1_sockopt.c   12  TPASS  :  setsockopt() SO_LINGER - SUCCESS

user  :notice: [   54.687418] test_1_to_1_sockopt.c   13  TPASS  :  getsockopt() SO_LINGER - SUCCESS

user  :notice: [   54.688187] test_1_to_1_sockopt.c   14  TPASS  :  getsockopt() SO_RCVBUF - SUCCESS

user  :notice: [   54.688979] test_1_to_1_sockopt.c   15  TPASS  :  getsockopt() SCTP_STATUS - SUCCESS

user  :notice: [   54.689752] test_1_to_1_sockopt.c   16  TPASS  :  setsockopt() SO_RCVBUF - SUCCESS

user  :notice: [   54.691539] test_1_to_1_sockopt.c   17  TBROK  :  test_1_to_1_sockopt.c:341: Comparison failed:Set value and got value differs Set Value=4096 Get Value=8192

user  :notice: [   54.692520] test_1_to_1_sockopt.c   18  TBROK  :  test_1_to_1_sockopt.c:341: Remaining cases broken



To reproduce:

        git clone git://git.kernel.org/pub/scm/linux/kernel/git/wfg/lkp-tests.git
        cd lkp-tests
        bin/lkp install job.yaml  # job file is attached in this email
        bin/lkp run     job.yaml



Thanks,
Xiaolong

Re: [lkp-robot] [net] 0148239373: ltp.test_1_to_1_sockopt.fail

From: Gao Feng <hidden>
Date: 2017-02-15 02:03:04

Hi Xiaolong,

On Wed, Feb 15, 2017 at 9:46 AM, kernel test robot
[off-list ref] wrote:

FYI, we noticed the following commit:

commit: 0148239373801a9c0c63eefc481b710ae7ce2941 ("net: sock: Use double send/recv buff value to compare with max value")
url: https://github.com/0day-ci/linux/commits/fgao-ikuai8-com/net-sock-Use-double-send-recv-buff-value-to-compare-with-max-value/20170208-213745


in testcase: ltp
with following parameters:

        test: net.sctp

test-description: The LTP testsuite contains a collection of tools for testing the Linux kernel and related features.
test-url: http://linux-test-project.github.io/


on test machine: 88 threads Intel(R) Xeon(R) CPU E5-2699 v4 @ 2.20GHz with 128G memory

caused below changes (please refer to attached dmesg/kmsg for entire log/backtrace):



user  :notice: [   54.674772] tag=test_1_to_1_sockopt stime=1487093637

user  :notice: [   54.675078] cmdline="test_1_to_1_sockopt"

user  :notice: [   54.675194] contacts=""

user  :notice: [   54.675333] analysis=exit

user  :notice: [   54.675513] <<<test_output>>>

user  :notice: [   54.676497] test_1_to_1_sockopt.c    1  TPASS  :  setsockopt() with a bad socket descriptor - EBADF

user  :notice: [   54.677439] test_1_to_1_sockopt.c    2  TPASS  :  setsockopt() with an invalid socket - ENOTSOCK

user  :notice: [   54.678409] test_1_to_1_sockopt.c    3  TPASS  :  setsockopt() with an invalid level - ENOPROTOOPT

user  :notice: [   54.679378] test_1_to_1_sockopt.c    4  TPASS  :  setsockopt() with invalid option buffer - EFAULT

user  :notice: [   54.680375] test_1_to_1_sockopt.c    5  TPASS  :  setsockopt() with invalid option name - EOPNOTSUPP

user  :notice: [   54.681354] test_1_to_1_sockopt.c    6  TPASS  :  getsockopt() with a bad socket descriptor - EBADF

user  :notice: [   54.682297] test_1_to_1_sockopt.c    7  TPASS  :  getsockopt() with an invalid socket - ENOTSOCK

user  :notice: [   54.683265] test_1_to_1_sockopt.c    8  TPASS  :  getsockopt() with invalid option buffer - EFAULT

user  :notice: [   54.684267] test_1_to_1_sockopt.c    9  TPASS  :  getsockopt() with invalid option name - EOPNOTSUPP

user  :notice: [   54.685079] test_1_to_1_sockopt.c   10  TPASS  :  getsockopt() SCTP_INITMSG - SUCCESS

user  :notice: [   54.685885] test_1_to_1_sockopt.c   11  TPASS  :  setsockopt() SCTP_INITMSG - SUCCESS

user  :notice: [   54.686653] test_1_to_1_sockopt.c   12  TPASS  :  setsockopt() SO_LINGER - SUCCESS

user  :notice: [   54.687418] test_1_to_1_sockopt.c   13  TPASS  :  getsockopt() SO_LINGER - SUCCESS

user  :notice: [   54.688187] test_1_to_1_sockopt.c   14  TPASS  :  getsockopt() SO_RCVBUF - SUCCESS

user  :notice: [   54.688979] test_1_to_1_sockopt.c   15  TPASS  :  getsockopt() SCTP_STATUS - SUCCESS

user  :notice: [   54.689752] test_1_to_1_sockopt.c   16  TPASS  :  setsockopt() SO_RCVBUF - SUCCESS

user  :notice: [   54.691539] test_1_to_1_sockopt.c   17  TBROK  :  test_1_to_1_sockopt.c:341: Comparison failed:Set value and got value differs Set Value=4096 Get Value=8192

user  :notice: [   54.692520] test_1_to_1_sockopt.c   18  TBROK  :  test_1_to_1_sockopt.c:341: Remaining cases broken



To reproduce:

        git clone git://git.kernel.org/pub/scm/linux/kernel/git/wfg/lkp-tests.git
        cd lkp-tests
        bin/lkp install job.yaml  # job file is attached in this email
        bin/lkp run     job.yaml



Thanks,
Xiaolong

This commit is not accepted, thanks.
After discuss with Eric, I know it is too late to correct it.

Regards
Feng

`h`	back out one level
`j`	next message in thread
`k`	previous message in thread
`l`	drill in
`Esc`	close help / fold thread tree
`?`	toggle this help