Le Saturday 04 September 2004 à 00:45:13 +0200, Loic Le Loarer a écrit: > Le Friday 03 September 2004 à 09:47:59 -0700, David S. Miller a écrit: > > Why not just decrement the counter in *_kill_one_slave()? > > Sure, it would be better, but num_slaves is a member of struct > slave_queue which is not given in the arguments of eql_kill_one_slave > and the decrement already appears after one of the eql_kill_one_slave > calls, so I correct the situation the same way, which is also the > fastest way. > > I agree it should be better for futur to decrement the counter in > eql_kill_one_slave, this require to add the queue or the counter address > as an argument to this function, it may seem strange, but it is less > error prone and futur proof. > > So here is a patch which use this method, chose the one you prefer. I confirm that the second patch is the best, in the first version, I have just discovered that I forget to decrement num_slaves after one of the call to eql_kill_one_slave, so decremnting in eql_kill_one_slave is much better ! Please consider my second patch. Regards -- Loïc "heaven is not a place, it's a feeling"
Attachment:
signature.asc
Description: Digital signature
