Re: [PATCH] src/ext4_resize.c: set errno to 0 before the strtoull call

["xuyang2018.jy@fujitsu.com" <xuyang2018.jy@fujitsu.com>]

Hi Andrew

errno doesn't be initialized to 0 when c program link with lcap since
libcap-2.30 (commit f1f62a748d7c Refactor the way we do the psx linkage
in libcap introduced this bug.)

The c example code as below:
-------------------------------------------
#include <stdio.h>
#include <errno.h>

int main(int argc, char **argv)
{
        printf("errno %d\n", errno);
        return 0;
}
---------------------------------------

#gcc test.c -lcap -o test
#./test
errno 1

Best Regards
Yang Xu
> on 2022/1/18 22:02, Florian Weimer wrote:
>> * xuyang2018.jy:
>>
>>> on 2022/1/18 11:56, Theodore Ts'o wrote:
>>>> On Tue, Jan 18, 2022 at 02:43:26AM +0000, xuyang2018.jy@fujitsu.com wrote:
>>>>>> You're right of course, but out of curiosity, which C library are you
>>>>>> using?
>>>>> I use glibc-2.34.
>>>>
>>>> Hmm, ok.  I'm using glibc 2.31, and in this particular program, errno
>>>> shouldn't have been set by any prior system call.  I'm guessing maybe
>>>> it was something in crt0 which ended up setting errno?
>>> It maybe a glibc bug.
>>> I cc glibc mailing list and see whether they have met this problem.
>>>
>>> @Florian
>>>
>>> Now, I use glibc-2.34 and run the following program[1] but the errno is
>>> not 0 in the beginning. So is this a known bug on glibc-2.34(Theodore
>>> doesn't meet this problem on glicb-2.31)?
>>>
>>> [1]https://git.kernel.org/pub/scm/fs/xfs/xfstests-dev.git/tree/src/ext4_resize.c
>>
>> I'm not aware of this issue.  Could you run strace or strace -k to see
>> where the failing system call is coming from?  Thanks.
> Even before open call, errno is also 1.
> 
> The strace output see attachment.
> 
> I found the following difference(compare with glibc-2.28, search 'err'
> keywords)
> 
> bad(glibc-2.34):
> 
> map(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0)
> = 0x7fdf815a3000
>   >  /usr/lib64/ld-linux-x86-64.so.2(__mmap+0x26) [0x21c56]
>   >  /usr/lib64/ld-linux-x86-64.so.2(rtld_malloc+0xa0) [0x1dc50]
>   >  /usr/lib64/ld-linux-x86-64.so.2(_dl_check_map_versions+0x4ec) [0x121fc]
>   >  /usr/lib64/ld-linux-x86-64.so.2(_dl_check_all_versions+0x3e) [0x124de]
>   >  /usr/lib64/ld-linux-x86-64.so.2(version_check_doit+0x1b) [0x1cab]
>   >  /usr/lib64/ld-linux-x86-64.so.2(_dl_receive_error+0x31) [0x1e811]
>   >  /usr/lib64/ld-linux-x86-64.so.2(dl_main+0x1e46) [0x4416]
>   >  /usr/lib64/ld-linux-x86-64.so.2(_dl_sysdep_start+0x3f6) [0x1d696]
>   >  /usr/lib64/ld-linux-x86-64.so.2(_dl_start+0x217) [0x2097]
>   >  /usr/lib64/ld-linux-x86-64.so.2(_start+0x7) [0x1097]
>   >  no matching address range
> arch_prctl(ARCH_SET_FS, 0x7fdf815a4480) = 0
> 
> good(glibc-2.28):
> 
> map(NULL, 12288, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0)
> = 0x7febd628b000
>   >  /usr/lib64/ld-2.28.so(__mmap+0x47) [0x1c377]
>   >  /usr/lib64/ld-2.28.so(malloc+0x8a) [0x1a94a]
>   >  /usr/lib64/ld-2.28.so(_dl_allocate_tls_storage+0x23) [0x12ac3]
>   >  /usr/lib64/ld-2.28.so(init_tls+0xa5) [0x1da5]
>   >  /usr/lib64/ld-2.28.so(dl_main+0x2c6d) [0x51ed]
>   >  /usr/lib64/ld-2.28.so(_dl_sysdep_start+0x48e) [0x1a14e]
>   >  /usr/lib64/ld-2.28.so(_dl_start+0x275) [0x2135]
>   >  /usr/lib64/ld-2.28.so(_start+0x7) [0x1087]
>   >  No DWARF information found
> arch_prctl(ARCH_SET_FS, 0x7febd628b740) = 0
> 
> Also if use gcc to compile this c program directly instead of using make
> in xfstests testsuite,  the errno value is normal(0).
> 
> After tried, I found if using the following command, errno is still 0
> gcc ext4_resize.c -g -O2  -D_GNU_SOURCE   -D_FILE_OFFSET_BITS=64
> -funsigned-char -fno-strict-aliasing  -Wall -o ext4_resize
> 
> But if I used link with lcap, then errno becomes 1 in the beginning.
> 
> Even I use lastest upstream libcap, this problem still occurs.
> 
> I am still looking into it .
> 
> Best Regards
> Yang Xu
> 
>>
>> Florian
>>
>