Hi, There is a description of the locking policy in tipc_net.c. Briefly, the node_lock doesn't only protect the node structure, but also all the 1-4 subordinate link structures which are pointed to by that node. I found this model much easier to handle, and still acceptable from performance viewpoint, than having individual locks per link.
Regards ///jon -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Florian Westphal Sent: February 18, 2008 5:58 PM To: [email protected] Subject: Re: [tipc-discussion] [PATCH] Allow configuration ofdefault bearerpriority/window size/tolerance Jon Maloy <[EMAIL PROTECTED]> wrote: > As for your earlier question about the node_lock, (sorry for not > responding earlier); It is there to protect against parallel access to > the link object when doing tipc_link_send_proto_msg(). Ah. I'll check the code wether this method is used consistently. This is rather unintuitive -- so thank you for pointing this out. I'll put the node_lock back (unless i can prove you wrong 8-). > Otherwise I think your patch looks ok. Thanks for reviewing this. I'll look into the node_lock issue tomorrow and will then send and updated patch. Regards, Florian ------------------------------------------------------------------------ - This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ tipc-discussion mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/tipc-discussion ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ tipc-discussion mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/tipc-discussion
