[ofa-general] [PATCH] perftest Add rdma_cm retries

Or Gerlitz ogerlitz at voltaire.com
Tue Jul 28 02:03:28 PDT 2009


David McMillen wrote:
> I would be happy to rework this patch again to make it a proper 
> example.  Please let me know what to change.  I read through the 
> threads referenced below and did not see anything that relates to this 
> patch.
Over these threads (along with others on which I didn't provide you with 
a pointer), people were attempting to discuss the root cause of the 
problem and few proposed approaches/solutions. Your patch simply add 
retries to a synthetic test program and as I wrote you, I don't have any 
comments on it, expect for that this may not an end-in-mind, that's it.
> In all versions of rdma_cm, [...] and assuming a properly functioning 
> fabric, both rdma_resolve_addr and rdma_resolve_route can fail due to 
> ETIMEDOUT, and both can be retried with success.  Is there an example 
> that I missed somewhere that shows how I should be doing things?
no. In many cases the retry is done in a higher level such as the SCSI 
stack or the file system core trying to mount, etc. As I wrote, the 
retry for address and route resolution may be implemented in the rdma-cm 
and layers below it, but if you feel this is beyond your scope, no 
problem, at some point the next generations will have to deal with it.

Or.




More information about the general mailing list