David Miller wrote:
From: Sven Eckelmann sven.eckelmann@gmx.de Date: Sun, 7 Nov 2010 14:26:17 +0100
if (seq_bits[word_num] & 1 << word_offset)...
- seq_bits[word_num] |= 1 << word_offset; /* turn the position on */
...
+#define TYPE_OF_WORD unsigned long
"1" is an 'int' and won't get promoted to unsigned long which means this code won't work on 64-bit, you need to explicitly say "1UL".
Thanks, that's correct. Will be replaced by the kernel facilities.
[...]
The badman-adv code is full of duplicated functionality, and until all of these cases are cured I refuse to integrate this code. I already complained about the hashing stuff, and now there's this stuff too.
Every time I review the batman-adv code I find a bug, and the bug is often in facilities which the kernel has already and are being duplicated. That is by definition a waste of my and everyone else's time.
Probably your submission will be almost half the size that it is currently once you take care of this issue. :-)
Do you have any more possible duplication you have in your mind right now?
thanks, Sven