[cgl_discussion] TIPC issues

Guo, Min min.guo at intel.com
Wed Apr 2 00:03:05 PST 2003


Hi:

	Here are some issues that I encountered in the test suites development, 
could you please take a look at them and give me a feedback?

Thanks
Guo Min
The content of this email message solely contains my own personal views,
and not those of my employer.



1. With the source codes cloned from http://tipc-linux.bkbits.net/, I try the subscritiption program,
 I find the TIPC does not support GET_LINK_EVENT function currently.

------------------------------------------------------------------------------------------------------
struct link_event link_ev = {0,0,0,0,0};
ex: assert(!ioctl(tipc_fd,GET_LINK_EVENT,&link_ev); //GET_LINK_EVENT=68u

System will tell us GET_LINK_EVENT is not supported now.
---------------------------------------------------------------------------------------------------------------
So, we think we can contribute codes to implement such function, how about your opinion?

2. When I insert the tipc.ko and run a normal TIPC server program and client program,
Here the TIPC server and client are in an infinite loop send/receive program, it is strange that
tipc.ko user number is still 0, and when I remove the tipc.ko ,and system will hang up. 
------------------------------------------------------------------------------
lsmod tipc
------------------------------------------
Module                  Size  Used by
tipc                  631832  0
-----------------------------------
rmmod tipc.ko
schedule in atomic ..
<Call Trace>.....
------------------------------------------
I think it is a bug of the TIPC module. I think when there are still  TIPC services,the tipc.ko 
user number should not be 0.

3. Because there are some different include files in the TIPC include/net,include/linux and corresponding standard linux include dir,
so I suggest the full include files should be found in the BK development tree, which can bring us more convenience to build application
programme.






More information about the cgl_discussion mailing list