On Sat, Apr 14, 2012 at 01:34:15PM +0200, Marek Lindner wrote:
On Saturday, April 14, 2012 13:29:15 Antonio Quartulli wrote:
On Sat, Apr 14, 2012 at 01:15:25PM +0200, Antonio Quartulli wrote:
The following two patches fix the two issues found by Al Viro and reported in his email.
Patch 1) simply convert the tt_crc field in the bat_priv structure from atomic_t to uint16_t. Actually there is no reason to declare it as atomic_t.
Patch 2) fixes a little bug we have when sending a tt_request message: now we don't convert the tt_crc that we are sending within the request to network order. OTOH we convert this field from NO to HO on the receiver side.
Therefore, nodes on the path of the request, which have HO different from NO, will fail to reply to the request and will forward it towards the final destination.
I think David'd want to see this fix in the next pull request or so...should we commit these patches to next directly?
Probably. The second patch could even go to stable, right ?
I'd say so
Cheers,