[PATCH] Fix aarch64 decoding of arm struct stat64.

[enh]

On Sun, Apr 6, 2014 at 8:20 AM, Dmitry V. Levin <ldv at altlinux.org> wrote:
>> diff --git a/file.c b/file.c
>> index 7e0773d..b2790e0 100644
>> --- a/file.c
>> +++ b/file.c
>> @@ -181,10 +181,6 @@ struct __old_kernel_stat {
>>  #undef st_mtime
>>  #undef st_ctime
>>
>> -#ifdef AARCH64
>> -#define stat64 stat
>> -#endif /* AARCH64 */
>> -
>>  #include <fcntl.h>
>>  #ifdef HAVE_SYS_VFS_H
>>  # include <sys/vfs.h>
>
> This hunk doesn't apply - it conflicts with commit
> a07cfcc133f4adf5116a32b199c4230e022475da.

ah, sorry --- i'd already reverted that locally. (my patches are
relative to Android's strace tree where i can build and test.) what
would you like me to do?

>> @@ -1053,10 +1053,16 @@ struct stat64 {
>>   unsigned int st_ctime;
>>   unsigned int st_ctime_nsec;
>>   unsigned long long st_ino;
>> -} __attribute__((packed));
>> +}
>> +#if defined X86_64
>> +__attribute__((packed))
>> +#endif
>> +;
>>  # define HAVE_STAT64 1
>> +#if defined X86_64
>>  # define STAT64_SIZE 96
>>  #endif
>> +#endif
>
> Why do you disable STAT64_SIZE check for AARCH64?

because arm's struct stat64 is a different size than the x86 one, and
i thought that you were just checking that the packing attribute
worked. it's 104 bytes if you want to add that for arm, which i guess
would protect against someone accidentally turning on packing for
aarch64/arm.

> BTW, this hunk introduces wrong indentation.

sorry. i hadn't noticed that you indent preprocessor stuff.

do you want me to redo this patch or is it easier for you to fit it locally?

-- 
Elliott Hughes - http://who/enh - http://jessies.org/~enh/
Java i18n/JNI/NIO, or bionic questions? Mail me/drop by/add me as a reviewer.