linux-kernel - Re: [PATCH] ipc/sem: ensure we left shift a ULL rather than a 32 bit integer

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: <1e0bacea-d3f1-a793-9e23-5e175c33346e@colorfullife.com>
Date:   Sun, 30 Oct 2016 16:33:23 +0100
From:   Manfred Spraul <manfred@...orfullife.com>
To:     Colin Ian King <colin.king@...onical.com>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Davidlohr Bueso <dave@...olabs.net>,
        Peter Zijlstra <peterz@...radead.org>,
        Ingo Molnar <mingo@...nel.org>,
        Nikolay Borisov <kernel@...p.com>
Cc:     linux-kernel@...r.kernel.org
Subject: Re: [PATCH] ipc/sem: ensure we left shift a ULL rather than a 32 bit
 integer

On 10/28/2016 09:29 PM, Colin Ian King wrote:
> On 28/10/16 20:21, Manfred Spraul wrote:
>> Hi Colin,
>>
>> On 10/28/2016 08:11 PM, Colin King wrote:
>> [...]
>>> --- a/ipc/sem.c
>>> +++ b/ipc/sem.c
>>> @@ -1839,7 +1839,7 @@ SYSCALL_DEFINE4(semtimedop, int, semid, struct
>>> sembuf __user *, tsops,
>>>          max = 0;
>>>        for (sop = sops; sop < sops + nsops; sop++) {
>>> -        unsigned long mask = 1 << ((sop->sem_num) % BITS_PER_LONG);
>>> +        unsigned long mask = 1ULL << ((sop->sem_num) % BITS_PER_LONG);
>>>    
>> Why 1ULL? Is 1UL not sufficient?
> For example, 1UL i386 is 32 bits, where as 1ULL is 64.
Exactly: on i386, 'unsigned long" is 32 bits. BITS_PER_LONG is 32.
Thus with 1UL, the code should be correct.
With 1ULL & -Wconversion, gcc would even report a warning:
> gcc -m32 -Wall -Wconversion -O1 test.c
> test.c: In function ‘main’:
> test.c:13:6: warning: conversion to ‘long unsigned int’ from ‘long 
> long unsigned int’ may alter its value [-Wconversion]
>    j= 1ULL << k;
>       ^~~~

test.c:
> #include <stdio.h>
> #include <stdlib.h>
>
> int main(int argc, char **argv)
> {
>     unsigned long j;
>     int i;
>
>     for (i=1;i<argc;i++) {
>         long k;
>
>         k=atoi(argv[i]);
>         j= 1ULL << k;
>         printf("%d: %lu %ld.\n", i, j, k);
>     }
>     return 0;
> }
>

--
     Manfred
(still thinks "1UL" is what is required)