netdev - Re: [PATCH net-next] net: ethernet: rmnet: Restructure if checks to avoid uninitialized warning

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] [day] [month] [year] [list]

Message-ID: <2145e27f-c8b3-ef4b-793a-841cb2f7e60f@kernel.org>
Date:   Fri, 4 Jun 2021 09:54:02 -0700
From:   Nathan Chancellor <nathan@...nel.org>
To:     subashab@...eaurora.org, patchwork-bot+netdevbpf@...nel.org
Cc:     stranche@...eaurora.org, davem@...emloft.net, kuba@...nel.org,
        ndesaulniers@...gle.com, sharathv@...eaurora.org,
        netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
        clang-built-linux@...glegroups.com
Subject: Re: [PATCH net-next] net: ethernet: rmnet: Restructure if checks to
 avoid uninitialized warning

Hi Subash,

On 6/3/2021 10:15 PM, subashab@...eaurora.org wrote:
> On 2021-06-03 16:40, patchwork-bot+netdevbpf@...nel.org wrote:
>> Hello:
>>
>> This patch was applied to netdev/net-next.git (refs/heads/master):
>>
>> On Thu,  3 Jun 2021 10:34:10 -0700 you wrote:
>>> Clang warns that proto in rmnet_map_v5_checksum_uplink_packet() might be
>>> used uninitialized:
>>>
>>> drivers/net/ethernet/qualcomm/rmnet/rmnet_map_data.c:283:14: warning:
>>> variable 'proto' is used uninitialized whenever 'if' condition is false
>>> [-Wsometimes-uninitialized]
>>>                 } else if (skb->protocol == htons(ETH_P_IPV6)) {
>>>                            ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>>> drivers/net/ethernet/qualcomm/rmnet/rmnet_map_data.c:295:36: note:
>>> uninitialized use occurs here
>>>                 check = rmnet_map_get_csum_field(proto, trans);
>>>                                                  ^~~~~
>>> drivers/net/ethernet/qualcomm/rmnet/rmnet_map_data.c:283:10: note:
>>> remove the 'if' if its condition is always true
>>>                 } else if (skb->protocol == htons(ETH_P_IPV6)) {
>>>                        ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>>> drivers/net/ethernet/qualcomm/rmnet/rmnet_map_data.c:270:11: note:
>>> initialize the variable 'proto' to silence this warning
>>>                 u8 proto;
>>>                         ^
>>>                          = '\0'
>>> 1 warning generated.
>>>
>>> [...]
>>
>> Here is the summary with links:
>>   - [net-next] net: ethernet: rmnet: Restructure if checks to avoid
>> uninitialized warning
>>     https://git.kernel.org/netdev/net-next/c/118de6106735
>>
>> You are awesome, thank you!
>> -- 
>> Deet-doot-dot, I am a bot.
>> https://korg.docs.kernel.org/patchwork/pwbot.html
> 
> Hi Nathan
> 
> Can you tell why CLANG detected this error.
> Does it require a bug fix.

As far as I understand it, clang does not remember the conditions of 
previous if statements when generating this warning. Basically:

void bar(int x)
{
}

int foo(int a, int b)
{
	int x;

	if (!a && !b)
		goto out;

	if (a)
		x = 1;
	else if (b)
		x = 2;

	bar(x);

out:
	return 0;
}

clang will warn that x is uninitialized when neither of the second if 
statement's conditions are true, even though we as humans know that is 
not possible due to the first if statement. I am guessing this has 
something to do with how clang generates its control flow graphs. While 
this is a false positive, I do not personally see this as a bug in the 
compiler. The code is more clear to both the compiler and humans if it 
is written as:

	if (a)
		x = 1;
	else if (b)
		x = 2;
	else
		goto out;

Cheers,
Nathan