lwip-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[lwip-users] CORE_LOCKING


From: timmy brolin
Subject: [lwip-users] CORE_LOCKING
Date: Thu, 14 Jul 2011 00:23:42 -0700

To bring up the discussion about CORE_LOCKING again..
I would be interested in starting to use CORE_LOCKING to improve performance.
However, the implementation status of this differs in different parts of lwIP.
As I understand it, CORE_LOCKING is used on four interfaces to lwIP:

tcpip_input: Here it is fully implemented and should be pretty safe to use.
netifapi: Fully implemented. Should be pretty safe to use.
netconn: Fully implemented. Reasonably safe?
socket: Partial/broken implementation.

Perhaps it would make sense to enable CORE_LOCKING individually on those for interfaces?
That way CORE_LOCKING can be enabled for lets say tcpip_input and netifapi, while still keeping the message passing in netconn and socket.

So perhaps five defines:
LWIP_TCPIP_CORE_LOCKING (Enables the core lock)
LWIP_TCPIP_CORE_LOCKING_INPUT (Use core lock instead of message passing for input packets)
LWIP_TCPIP_CORE_LOCKING_NETIFAPI (Use core lock instead of message passing for netifapi)
LWIP_TCPIP_CORE_LOCKING_NETCONN (Use core lock instead of message passing for netconn)
LWIP_TCPIP_CORE_LOCKING_SOCKET (Use core lock instead of message passing for socket)
The four last defines are all dependent on the first define.

With this setup, people should be able to start using CORE_LOCKING safely.

Any thoughts?

Regards,
Timmy Brolin

_______________________________________________________________
Hitta ditt drömjobb - Turism Försäljning Vårdjobb Ekonomi Design Bank Media


reply via email to

[Prev in Thread] Current Thread [Next in Thread]