[PATCH] staging: gdm724: adding kernel endianness header

[PATCH] staging: gdm724: adding kernel endianness header

[Jaime Arrocha]

>From TODO list: remove test for host endian
Included header to gather information about host endianness.
Please let me know if the code addition requires corrections
to meet standards.

Signed-off-by: Jaime Arrocha <jarr@kerneldev.net>
---
 drivers/staging/gdm724x/gdm_endian.c |   22 ++++++++++------------
 1 file changed, 10 insertions(+), 12 deletions(-)

diff --git a/drivers/staging/gdm724x/gdm_endian.c b/drivers/staging/gdm724x/gdm_endian.c
index f6cc90a..5dfd9d3 100644
--- a/drivers/staging/gdm724x/gdm_endian.c
+++ b/drivers/staging/gdm724x/gdm_endian.c
@@ -11,27 +11,25 @@
  * GNU General Public License for more details.
  */
 
-#include <linux/slab.h>
+#include<asm/byteorder.h>
+#ifdef __LITTLE_ENDIAN
+#include<linux/byteorder/little_endian.h>
+#define H_ENDIAN ENDIANNESS_LITTLE
+#else
+#include<linux/byteorder/big_endian.h>
+#define H_ENDIAN ENDIANNESS_BIG
+#endif
+
 #include "gdm_endian.h"
 
 void gdm_set_endian(struct gdm_endian *ed, u8 dev_endian)
 {
-	u8 a[2] = {0x12, 0x34};
-	u8 b[2] = {0, };
-	u16 c = 0x1234;
-
 	if (dev_endian == ENDIANNESS_BIG)
 		ed->dev_ed = ENDIANNESS_BIG;
 	else
 		ed->dev_ed = ENDIANNESS_LITTLE;
 
-	memcpy(b, &c, 2);
-
-	if (a[0] != b[0])
-		ed->host_ed = ENDIANNESS_LITTLE;
-	else
-		ed->host_ed = ENDIANNESS_BIG;
-
+	ed->host_ed = H_ENDIAN;
 }
 
 u16 gdm_cpu_to_dev16(struct gdm_endian *ed, u16 x)
-- 
1.7.10.4

Re: [PATCH] staging: gdm724: adding kernel endianness header

[Dan Carpenter]

On Fri, May 08, 2015 at 06:29:24PM -0500, Jaime Arrocha wrote:
> >From TODO list: remove test for host endian
> Included header to gather information about host endianness.
> Please let me know if the code addition requires corrections
> to meet standards.
> 

No.  This isn't really an improvement.  We need to get rid of ->host_ed
entirely.  Which is actually easy.  Basically change gdm_cpu_to_dev16()
and friends to do so something:

u16 gdm_cpu_to_dev16(struct gdm_endian *ed, u16 x)
{
	if (ed->dev_ed == ENDIANNESS_LITTLE)
		return cpu_to_le16(x);
	else
		return cpu_to_be16(x);
}

regards,
dan carpenter

Re: Re: [PATCH] staging: gdm724: adding kernel endianness header

[Jaime Arrocha]

---- On Sat, 09 May 2015 05:59:40 -0500 Dan Carpenter <dan.carpenter@oracle.com> wrote ---- 

 > On Fri, May 08, 2015 at 06:29:24PM -0500, Jaime Arrocha wrote: 
 > > >From TODO list: remove test for host endian 
 > > Included header to gather information about host endianness. 
 > > Please let me know if the code addition requires corrections 
 > > to meet standards. 
 > >  
 >  
 > No.  This isn't really an improvement.  We need to get rid of ->host_ed 
 > entirely.  Which is actually easy.  Basically change gdm_cpu_to_dev16() 
 > and friends to do so something: 
 >  
 > u16 gdm_cpu_to_dev16(struct gdm_endian *ed, u16 x) 
 > { 
 >     if (ed->dev_ed == ENDIANNESS_LITTLE) 
 >         return cpu_to_le16(x); 
 >     else 
 >         return cpu_to_be16(x); 
 > } 
 >  
 > regards, 
 > dan carpenter 
 >  
 > 

You are right, I will dive a little further to find out the use for host_ed and eliminate it. I did
read an email from a while ago where one author mentioned that he needed to keep this function.

Also, the kernel already has functions to swap numbers for endianness. Will you rather use those?

Thanks for your feedback,
Jaime