lkml.org 
[lkml]   [2015]   [Feb]   [2]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
SubjectRe: Can this be a invalid memory access? (was: Re: [PATCH] spi/xilinx: Cast ioread32/iowrite32 function pointers)
From
Hi Ricardo,

On Mon, Feb 2, 2015 at 3:05 PM, Ricardo Ribalda Delgado
<ricardo.ribalda@gmail.com> wrote:
> On Mon, Feb 2, 2015 at 2:04 PM, Geert Uytterhoeven <geert@linux-m68k.org> wrote:
>>> void * pvar=varB;
>>
>> ... = &varB;
>>
>>> *pvar = ioread8(valid_memory);
>>>
>>> Depending if ioread8 returns a u8 or a unsigned int, aren't we also
>>> accessing varC?
>>>
>>> Could not this be a problem?
>>
>> Please try to compile the above.
>> The compiler will tell you you cannot dereference a void pointer.
>>
>> Now, replace "void *" by "unsigned int *".
>> After that, varC will indeed be overwritten. But the compiler will still have
>> warned you that you assigned the address of an u8 variable to an
>> unsigned int pointer.
>
> I am still a bit worried, that the same function has different
> prototypes and implementations depending on the arch. But
> "unfortunately" I cannot make a counter-example that can cause an
> error.
>
> So far, the closest I have get is this:
>
> ricardo@neopili:/tmp$ cat arch.c
> #include <stdint.h>

You should include io.h here...

>
> unsigned int my_ioread8(){

... then the compiler will complain here.

> return 0xdeadbeef;
> }
>
> void test_read();
> int main(int argc, char *argv[]){
>
> test_read();
> return 0;
> }
>
>
> ricardo@neopili:/tmp$ cat io.h
> #ifndef IO_H
> #define IO_H
> #include <stdint.h>
>
> uint8_t my_ioread8();
>
> #endif
>
>
> ricardo@neopili:/tmp$ cat driver.c
> #include "io.h"
> #include <stdint.h>
> #include <stdio.h>
>
> void test_read(){
> uint8_t varA=0x1;
> uint8_t varB=0x2;
> uint8_t varC=0x3;
>
> varB = my_ioread8();
>
> fprintf(stdout, "varA=%d varB=%d varC=%d\n",
> varA, varB, varC);
> }
>
> It does not produce any error at build time:
> ricardo@neopili:/tmp$ make
> cc -Wall -c -o arch.o arch.c
> cc -Wall -c -o driver.o driver.c
> cc -Wall arch.o driver.o -o test
>
> and it works fine on amd86
>
> ricardo@neopili:/tmp$ ./test
> varA=1 varB=239 varC=3
>
> But in the hypothetical case where the calling convention will be
> different for unsigned int and u8, there will be an issue....

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds


\
 
 \ /
  Last update: 2015-02-02 15:21    [W:1.211 / U:0.012 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site