From mboxrd@z Thu Jan 1 00:00:00 1970 From: ebiederm@xmission.com (Eric W. Biederman) Subject: Re: [PATCH net] net:sysctl fix the confusing corner of tcp_mem Date: Wed, 09 Dec 2015 10:38:01 -0600 Message-ID: <877fknbms6.fsf@x220.int.ebiederm.org> References: <1449626509-3736-1-git-send-email-wangyufen@huawei.com> Mime-Version: 1.0 Content-Type: text/plain Cc: , , , , , , Eric Dumazet To: Wang Yufen Return-path: Received: from out03.mta.xmission.com ([166.70.13.233]:50719 "EHLO out03.mta.xmission.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751332AbbLIQq2 (ORCPT ); Wed, 9 Dec 2015 11:46:28 -0500 In-Reply-To: <1449626509-3736-1-git-send-email-wangyufen@huawei.com> (Wang Yufen's message of "Wed, 9 Dec 2015 10:01:49 +0800") Sender: netdev-owner@vger.kernel.org List-ID: Wang Yufen writes: > From: Yufen Wang > > I tried on linux-4.1: > linux:~# cat /proc/sys/net/ipv4/tcp_mem > 8388608 12582912 16777216 > linux:~# echo 1234 >/proc/sys/net/ipv4/tcp_mem > -bash: echo: write error: Invalid argument > linux:~# cat /proc/sys/net/ipv4/tcp_mem > 1234 12582912 16777216 > > the echo operation got error, but value already written to tcp_mem. > If a write() returns an error like EINVAL, we expect no change occurred. > This patch fix the confusing corner and makes __do_proc_doulongvec_minmax > works the same as __do_proc_dointvec Nacked-by: "Eric W. Biederman" Except for possibly breaking your muscle memory this does not explain why this is a problem. Further you are changing a whole lot more than tcp_mem, without a word of justification in your description. I do not think changing every integer use of sysctl for some unknown reason. Is justified in this case. Eric > Signed-off-by: Yufen Wang > Cc: Eric Dumazet > Cc: "Eric W. Biederman" > --- > kernel/sysctl.c | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/kernel/sysctl.c b/kernel/sysctl.c > index c3eee4c..e3ee4be 100644 > --- a/kernel/sysctl.c > +++ b/kernel/sysctl.c > @@ -2318,6 +2318,8 @@ static int __do_proc_doulongvec_minmax(void *data, struct ctl_table *table, int > bool neg; > > left -= proc_skip_spaces(&kbuf); > + if (!left) > + break; > > err = proc_get_long(&kbuf, &left, &val, &neg, > proc_wspace_sep,