[apparmor] [patch 18/26] Convert to htoleXX fns instead of ifdef on endian

[John Johansen]

On 03/28/2014 08:52 AM, Tyler Hicks wrote:
> On 2014-03-27 08:45:31, john.johansen at canonical.com wrote:
>> This patch makes use of the htoleXX() functions (see endian(3))
>> defined as part of endian.h (already included in parser_interface.c),
>> instead of defining a function differently based on the detection of
>> endian related macros.
>>
>> This fixes a build failure experienced on powerpc with John's patch
>> set applied. This patch has been updated with John's feedback to use
>> letoh16() in the le16_to_cpu() macro.
>>
>>
>> Signed-off-by: Steve Beattie <steve at nxnw.org>
>> Acked-by: John Johansen <john.johansen at canonical.com>
>> ---
>>  parser/parser.h           |   15 ++++++---------
>>  parser/parser_interface.c |    3 ---
>>  parser/parser_main.c      |    6 +-----
>>  3 files changed, 7 insertions(+), 17 deletions(-)
>>
>> --- 2.9-test.orig/parser/parser.h
>> +++ 2.9-test/parser/parser.h
>> @@ -23,9 +23,11 @@
>>  #define __AA_PARSER_H
>>  
>>  
>> +#include <endian.h>
>>  #include <string.h>
>>  #include <netinet/in.h>
>>  #include <sys/resource.h>
>> +
>>  #include "immunix.h"
>>  #include "libapparmor_re/apparmor_re.h"
>>  #include "libapparmor_re/aare_rules.h"
>> @@ -258,15 +260,10 @@
>>  #define u16 uint16_t
>>  #define u32 uint32_t
>>  #define u64 uint64_t
>> -#if __BYTE_ORDER == __BIG_ENDIAN
>> -#  define cpu_to_le16(x) ((u16)(bswap_16 ((u16) x)))
>> -#  define cpu_to_le32(x) ((u32)(bswap_32 ((u32) x)))
>> -#  define cpu_to_le64(x) ((u64)(bswap_64 ((u64) x)))
>> -#else
>> -#  define cpu_to_le16(x) ((u16)(x))
>> -#  define cpu_to_le32(x) ((u32)(x))
>> -#  define cpu_to_le64(x) ((u64)(x))
>> -#endif
>> +
>> +#define cpu_to_le16(x) ((u16)(htole16 ((u16) x)))
>> +#define cpu_to_le32(x) ((u32)(htole32 ((u32) x)))
>> +#define cpu_to_le64(x) ((u64)(htole64 ((u64) x)))
>>  
>>  /* The encoding for kernal abi > 5 is
>>   * 28-31: reserved
>> --- 2.9-test.orig/parser/parser_interface.c
>> +++ 2.9-test/parser/parser_interface.c
>> @@ -37,9 +37,6 @@
>>  #include <unistd.h>
>>  #include <linux/unistd.h>
>>  
>> -/* only for x86 at the moment */
>> -#include <endian.h>
>> -#include <byteswap.h>
>>  #include <libintl.h>
>>  #define _(s) gettext(s)
>>  
>> --- 2.9-test.orig/parser/parser_main.c
>> +++ 2.9-test/parser/parser_main.c
>> @@ -946,11 +946,7 @@
>>  	return rc;
>>  }
>>  
>> -#if __BYTE_ORDER == __BIG_ENDIAN
>> -#  define le16_to_cpu(x) ((uint16_t)(bswap_16 (*(uint16_t *) x)))
>> -#else
>> -#  define le16_to_cpu(x) (*(uint16_t *)(x))
>> -#endif
>> +#define le16_to_cpu(x) ((uint16_t)(letoh16 (*(uint16_t *) x)))
> 
> You didn't pick up your feedback from the original thread. This should
> be le16toh().
> 
err? it is.

Quadruple checked in my queue and then here. Steve's patch used to use
htole16

> https://lists.ubuntu.com/archives/apparmor/2014-March/005411.html
> 
> On a related note, we need to start getting some of this patch set acked
> and committed upstream or this is going to happen more often.
> 
yes, I have pushed any of the earlier ones that have been acked up, but
I haven't tried bubbling up the later ones that have been acked, even
though several of them probably could be.

The big problem is that first patch in the queue is a bear (or so I hear,
/me wonders what is going to happen when the real bears hit), and people
are going to need some time to tackle it.

> Tyler
> 
>>  
>>  const char header_string[] = "\004\010\000version\000\002";
>>  #define HEADER_STRING_SIZE 12
>>
>>
>> -- 
>> AppArmor mailing list
>> AppArmor at lists.ubuntu.com
>> Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/apparmor