linux-kernel - Re: [PATCH] selftests/mm: Fix compiler -Wmaybe-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] [thread-next>] [day] [month] [year] [list]

Message-ID: <dcf585d8-8293-42ea-9d31-23825cc2f658@redhat.com>
Date: Wed, 9 Apr 2025 12:21:48 +0200
From: David Hildenbrand <david@...hat.com>
To: Anshuman Khandual <anshuman.khandual@....com>, linux-mm@...ck.org
Cc: Andrew Morton <akpm@...ux-foundation.org>, Shuah Khan <shuah@...nel.org>,
 linux-kselftest@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] selftests/mm: Fix compiler -Wmaybe-uninitialized warning

On 09.04.25 12:09, Anshuman Khandual wrote:
> 
> 
> On 4/9/25 15:27, David Hildenbrand wrote:
>> On 09.04.25 11:50, Anshuman Khandual wrote:
>>> Following build warning comes up for cow test as 'transferred' variable has
>>> not been initialized. Fix the warning via zero init for the variable.
>>>
>>>     CC       cow
>>> cow.c: In function ‘do_test_vmsplice_in_parent’:
>>> cow.c:365:61: warning: ‘transferred’ may be used uninitialized [-Wmaybe-uninitialized]
>>>     365 |                 cur = read(fds[0], new + total, transferred - total);
>>>         |                                                 ~~~~~~~~~~~~^~~~~~~
>>> cow.c:296:29: note: ‘transferred’ was declared here
>>>     296 |         ssize_t cur, total, transferred;
>>>         |                             ^~~~~~~~~~~
>>>     CC       compaction_test
>>>     CC       gup_longterm
>>>
>>> Cc: Andrew Morton <akpm@...ux-foundation.org>
>>> Cc: Shuah Khan <shuah@...nel.org>
>>> Cc: linux-mm@...ck.org
>>> Cc: linux-kselftest@...r.kernel.org
>>> Cc: linux-kernel@...r.kernel.org
>>> Signed-off-by: Anshuman Khandual <anshuman.khandual@....com>
>>> ---
>>>    tools/testing/selftests/mm/cow.c | 2 +-
>>>    1 file changed, 1 insertion(+), 1 deletion(-)
>>>
>>> diff --git a/tools/testing/selftests/mm/cow.c b/tools/testing/selftests/mm/cow.c
>>> index f0cb14ea8608..b6cfe0a4b7df 100644
>>> --- a/tools/testing/selftests/mm/cow.c
>>> +++ b/tools/testing/selftests/mm/cow.c
>>> @@ -293,7 +293,7 @@ static void do_test_vmsplice_in_parent(char *mem, size_t size,
>>>            .iov_base = mem,
>>>            .iov_len = size,
>>>        };
>>> -    ssize_t cur, total, transferred;
>>> +    ssize_t cur, total, transferred = 0;
>>>        struct comm_pipes comm_pipes;
>>>        char *old, *new;
>>>        int ret, fds[2];
>>
>>
>> if (before_fork) {
>>      transferred = vmsplice(fds[1], &iov, 1, 0);
>> ...
>>
>> if (!before_fork) {
>>      transferred = vmsplice(fds[1], &iov, 1, 0);
>> ...
>>
>> for (total = 0; total < transferred; total += cur) {
>> ...
>>
>>
>> And I don't see any jump label that could jump to code that would ve using transferred.
>>
>> What am I missing?
> 
> Probably because both those conditional statements are not mutually
> exclusive above with an if-else construct. Hence compiler flags it
> rather as a false positive ? Initializing with 0 just works around
> that false positive.

This is something the compiler should clearly be able to verify. 
before_fork is never changed in that function.

We should not work around wrong compilers.

Which compiler are you using such that you run into this issue?

-- 
Cheers,

David / dhildenb