Skip to contentSkip to conversation

Documentation/unaligned-memory-access.txt: fix incorrect comparison operator

PATCH Applied Cihangir Akturk 3 participants 5 messages Dec 17, 2016 – Dec 27, 2016 Documentation

lore
Share & tools
lore
#0 lore raw

In the actual implementation ether_addr_equal function tests for equality to 0 when returning. It seems in commit 0d74c4 it is somehow overlooked to change this operator to reflect the actual function.

---
 Documentation/unaligned-memory-access.txt | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Documentation/unaligned-memory-access.txt +1 -1
index a445da0..3f76c0c 100644
@@ -151,7 +151,7 @@ bool ether_addr_equal(const u8 *addr1, const u8 *addr2)
 #else
 	const u16 *a = (const u16 *)addr1;
 	const u16 *b = (const u16 *)addr2;
-	return ((a[0] ^ b[0]) | (a[1] ^ b[1]) | (a[2] ^ b[2])) != 0;
+	return ((a[0] ^ b[0]) | (a[1] ^ b[1]) | (a[2] ^ b[2])) == 0;
 #endif
 }
 
Signature · 1 line
2.1.4
#1 lore raw

17.12.2016, 19:43, "Cihangir Akturk" <[email protected]>:

Quoted 3 lines

In the actual implementation ether_addr_equal function tests for equality to 0 when returning. It seems in commit 0d74c4 it is somehow overlooked to change this operator to reflect the actual function.

why this "return" function need to be ==0? I think, u16 functions read memory but "0" is should not be equalty. This way, -for the code to work- memory should be everytime unaligned !=0.

Quoted 20 lines

---
 Documentation/unaligned-memory-access.txt | 2 +-

 1 file changed, 1 insertion(+), 1 deletion(-)
Documentation/unaligned-memory-access.txt +1 -2
index a445da0..3f76c0c 100644
@@ -151,7 +151,7 @@ bool ether_addr_equal(const u8 *addr1, const u8 *addr2)
  #else
          const u16 *a = (const u16 *)addr1;
          const u16 *b = (const u16 *)addr2;
- return ((a[0] ^ b[0]) | (a[1] ^ b[1]) | (a[2] ^ b[2])) != 0;
+ return ((a[0] ^ b[0]) | (a[1] ^ b[1]) | (a[2] ^ b[2])) == 0;
  #endif
  }
 
--

2.1.4

#2 lore raw
Quoted 6 lines · Ozgur Karatas

17.12.2016, 19:43, "Cihangir Akturk" <[email protected]>:

Quoted 3 lines

In the actual implementation ether_addr_equal function tests for equality to 0 when returning. It seems in commit 0d74c4 it is somehow overlooked to change this operator to reflect the actual function.

why this "return" function need to be ==0? I think, u16 functions read memory but "0" is should not be equalty.

XOR is true only when inputs differ. That means if inputs are the same, then it outputs false (0) or whatever you call it. Then we perform OR operation between those outputs. So if the result is 0 then addr1 and addr2 is equal.

Quoted 1 line

This way, -for the code to work- memory should be everytime unaligned !=0.

Sorry I didn't quite get the point.

#3 lore raw

20.12.2016, 02:22, "Cihangir Akturk" <[email protected]>:

Quoted 6 lines · Ozgur Karatas
Quoted 5 lines

 17.12.2016, 19:43, "Cihangir Akturk" <[email protected]>:
 > In the actual implementation ether_addr_equal function tests for equality to 0 > when returning. It seems in commit 0d74c4 it is somehow overlooked to change > this operator to reflect the actual function.

Quoted 6 lines
Quoted 1 line

 why this "return" function need to be ==0? I think, u16 functions read memory but "0" is should not be equalty.

XOR is true only when inputs differ. That means if inputs are the same, then it outputs false (0) or whatever you call it. Then we perform OR operation between those outputs. So if the result is 0 then addr1 and addr2 is equal.

Thanks for this explanation to your patch. In this case the patch mentioned is valid. I checked, turned to "!=0" errors.

un-mem.c:29:18: error: ‘!’ (first use in this function)

         return ((a[0] ^ b[0]) | (a[1] ^ b[1]) | (a[2] ^ b[2])) != 0;
                                                                                    ^

Also, don't need to send patch it secon time :)

Quoted 3 lines
Quoted 1 line

 This way, -for the code to work- memory should be everytime unaligned !=0.

Sorry I didn't quite get the point.

Regards,

~Ozgur

maintainer Applied
#4 lore raw
In reply to Cihangir Akturk
Quoted 3 lines · On Sat, 17 Dec 2016 19:42:17 +0200 Cihangir Akturk <[email protected]> wrote:

In the actual implementation ether_addr_equal function tests for equality to 0 when returning. It seems in commit 0d74c4 it is somehow overlooked to change this operator to reflect the actual function.

I've applied this to the docs tree, thanks.

jon