netdev - Re: [PATCH net-next] net: remove redundant input checks in SIOCSIFTXQLEN case of dev

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20180723.140025.949946162106654052.davem@davemloft.net>
Date:   Mon, 23 Jul 2018 14:00:25 -0700 (PDT)
From:   David Miller <davem@...emloft.net>
To:     xiyou.wangcong@...il.com
Cc:     tariqt@...lanox.com, netdev@...r.kernel.org, eranbe@...lanox.com
Subject: Re: [PATCH net-next] net: remove redundant input checks in
 SIOCSIFTXQLEN case of dev_ifsioc

From: Cong Wang <xiyou.wangcong@...il.com>
Date: Mon, 23 Jul 2018 13:37:22 -0700

> On Sun, Jul 22, 2018 at 12:29 AM Tariq Toukan <tariqt@...lanox.com> wrote:
>>
>>
>>
>> On 19/07/2018 8:21 PM, Cong Wang wrote:
>> > On Thu, Jul 19, 2018 at 7:50 AM Tariq Toukan <tariqt@...lanox.com> wrote:
>> >> --- a/net/core/dev_ioctl.c
>> >> +++ b/net/core/dev_ioctl.c
>> >> @@ -282,14 +282,7 @@ static int dev_ifsioc(struct net *net, struct ifreq *ifr, unsigned int cmd)
>> >>                  return dev_mc_del_global(dev, ifr->ifr_hwaddr.sa_data);
>> >>
>> >>          case SIOCSIFTXQLEN:
>> >> -               if (ifr->ifr_qlen < 0)
>> >> -                       return -EINVAL;
>> >
>> > Are you sure we can remove this if check too?
>> >
>> > The other one is safe to remove.
>> >
>>
>> Hmm, let's see:
>> dev_change_tx_queue_len gets unsigned long new_len, any negative value
>> passed is interpreted as a very large number, then we test:
>> if (new_len != (unsigned int)new_len)
>>
>> This test returns true if range of unsigned long is larger than range of
>> unsigned int. AFAIK these ranges are Arch dependent and there is no
>> guarantee this holds.
> 
> I am not sure either, you probably have to give it a test.
> And at least, explain it in changelog if you still want to remove it.

On 64-bit we will fail with -ERANGE.  The 32-bit int ifr_qlen will be sign
extended to 64-bits when it is passed into dev_change_tx_queue_len(). And
then for negative values this test triggers:

	if (new_len != (unsigned int)new_len)
		return -ERANGE;

because:
	if (0xffffffffWHATEVER != 0x00000000WHATEVER)

On 32-bit the signed value will be accepted, changing behavior.

I think, therefore, that the < 0 check should be retained.

Thank you.