David,
A suggestion (in case you hadn't already considered this)...
The sys_arch_protect() and unprotect() functions should be
able to deal with nesting...
pval = sys_arch_protect();
pval = sys_arch_protect();
sys_arch_unprotect(pval); <<--- should not cause an unprotect
sys_arch_unprotect(pval);
Agree? It's similar to uC/OS's critical section type 3.
Ed
David Haas wrote:
I have been meaning to integrate the select() code into the source tree,
but last week I was very busy on a different project :-( .
If it is OK with everybody, I will work on integrating all these socket
changes and hope to have it done in a day or two. I will do some
(limited) testing before I check it in.
Also, I had put in some lightweight protection into pbuf and memp code
(in a patch) which disabled interrupts. Several people commented on this
and thought that we should not be disabling interrupts. I agree that on
many systems this is undesirable. So how about a sys_arch_protect() and
sys_arch_unprotect(). This can then be implemented differently on
different systems.
Any comments???
David.
_______________________________________________
lwip-users mailing list
address@hidden
http://mail.nongnu.org/mailman/listinfo/lwip-users
_______________________________________________
lwip-users mailing list
address@hidden
http://mail.nongnu.org/mailman/listinfo/lwip-users