Re: [PATCH 01/11] rxrpc: Add a common object cache

2016-03-08 Thread David Miller
From: David Howells Date: Tue, 08 Mar 2016 21:11:09 + > I can put in a limit per peer, where a 'peer' is either a particular remote > UDP port or particulat remote host. TCP has this by virtue of having a > limited number of ports available per IP address. But if I have 10 IP > addresses av

Re: [PATCH 01/11] rxrpc: Add a common object cache

2016-03-08 Thread David Howells
David Miller wrote: > > One could argue, I suppose, that things should've been arranged that the > > RxRPC client would manage the lifetime of each connection it sets up, > > rather than both ends letting it lapse by mutual loss of interest. But > > you *still* have to have a timeout, lest the c

Re: [PATCH 01/11] rxrpc: Add a common object cache

2016-03-08 Thread David Miller
From: David Howells Date: Tue, 08 Mar 2016 13:02:28 + > David Howells wrote: > >> Does it make sense to maintain a FIFO list of connections (though this would >> mean potentially taking a spinlock every time I get a packet)? > > It occurs to me that only inactive connections would need to

Re: [PATCH 01/11] rxrpc: Add a common object cache

2016-03-08 Thread David Miller
From: David Howells Date: Tue, 08 Mar 2016 11:39:57 + > One could argue, I suppose, that things should've been arranged that the RxRPC > client would manage the lifetime of each connection it sets up, rather than > both ends letting it lapse by mutual loss of interest. But you *still* have

Re: [PATCH 01/11] rxrpc: Add a common object cache

2016-03-08 Thread David Howells
David Howells wrote: > Does it make sense to maintain a FIFO list of connections (though this would > mean potentially taking a spinlock every time I get a packet)? It occurs to me that only inactive connections would need to be on an LRU list. Any connection with packets or active calls to dea

Re: [PATCH 01/11] rxrpc: Add a common object cache

2016-03-08 Thread David Howells
David Miller wrote: > >> I know you put a lot of time and effort into this, but I want to strongly > >> recommend against a garbage collected hash table for anything whatsoever. > >> > >> Especially if the given objects are in some way created/destroyed/etc. by > >> operations triggerable remote

Re: [PATCH 01/11] rxrpc: Add a common object cache

2016-03-07 Thread David Miller
From: David Howells Date: Mon, 07 Mar 2016 22:45:14 + > David Miller wrote: > >> I know you put a lot of time and effort into this, but I want to strongly >> recommend against a garbage collected hash table for anything whatsoever. >> >> Especially if the given objects are in some way crea

Re: [PATCH 01/11] rxrpc: Add a common object cache

2016-03-07 Thread David Howells
David Miller wrote: > I know you put a lot of time and effort into this, but I want to strongly > recommend against a garbage collected hash table for anything whatsoever. > > Especially if the given objects are in some way created/destroyed/etc. by > operations triggerable remotely. > > This c

Re: [PATCH 01/11] rxrpc: Add a common object cache

2016-03-07 Thread David Miller
From: David Howells Date: Mon, 07 Mar 2016 14:38:06 + > Add a common object cache implementation for RxRPC. This will be used to > cache objects of various types (calls, connections, local and remote > endpoint records). Each object that would be cached must contain an > obj_node struct for