Re: [PATCH elfutils 2/2] [tests] parse inode in /proc/pid/maps correctly in run-backtrace-data.sh

[Yonghong Song <yhs at fb dot com>]

On 1/29/19 12:50 PM, Mark Wielaard wrote:
> On Fri, Jan 25, 2019 at 01:20:09PM -0800, Yonghong Song wrote:
>> The backtrace-data.c parsed the inode in /proc/pid/maps with
>> format "%*x".
>> This caused failure if inode is big. For example,
>>    7f269223d000-7f269226b000 r-xp 00000000 00:50 10224326387095067468       /home/...
> 
> I have a bit of trouble replicating this (with a simple sscanf).
> How exactly does it fail?

The error message looks like:

-bash-4.4$ cat run-backtrace-data.sh.log
backtrace-data: 
/home/engshare/elfutils/0.174/src/elfutils-0.174/tests/backtrace-data.c:110: 
maps_lookup: Assertion `errno == 0' failed.
/home/engshare/elfutils/0.174/src/elfutils-0.174/tests/test-subr.sh: 
line 84: 3123578 Aborted                 (core dumped) 
LD_LIBRARY_PATH="${built_library_path}${LD_LIBRARY_PATH:+:}$LD_LIBRARY_PATH" 
$VALGRIND_CMD "$@"
data: no main
-bash-4.4$

The reason is errno is ERANGE.

> 
>> The correct format should be "%*lu" to reflect inode "unsigned long" type.
>> But that caused the following compilation error.
>>    acktrace-data.c: In function ‘maps_lookup’:
>>    backtrace-data.c:109:22: error: use of assignment suppression and length modifier
>>    together in gnu_scanf format [-Werror=format=]
>>         i = fscanf (f, "%lx-%lx %*s %lx %*x:%*x %*lu", &start, &end, &offset);
> 
> Not that it matters much, since we are really ignoring the rest of the
> line and this is just a test. But I do wonder why %*u doesn't work.
> The warning says you cannot combine a length specifier with
> a ignored format specifier. Which kind of makes sense given that the
> length is for the variable to assign the value for, not the format.
> So it seems $*u should do the trick. But since I haven't been able
> to make the original fail, I might be wrong.

"%u" works as well. Let me submit another patch for this.
Thanks!

> 
> Cheers,
> 
> Mark
>