[rds-devel] RDS: updates for net-next

[Or Gerlitz]

Andy Grover wrote:
> Too big a change for getting into 2.6.31 -- the 2.6.32 net-next window could be open as soon as 2-3 weeks (right after 2.6.31-rc1 comes out) so I was hoping to shoot for that, as well as ofed 1.5 of course
Hi Andy,

I'm not sure what you mean by too big for 2.6.31 - as the .31 merge 
window hasn't opened yet and even when this starts, it lasts for two 
weeks. Did you mean that adding them on top of what you already pushed 
to .31 is too much? so what's makes it ready for ofed? Please take into 
account, that any patch that would come to fix a regression introduced 
by patches merged for .31-rc1 would be accepted for .31-rc2,3,etc. As 
for the patchset itself, I took a look on them and it seems that 5, 6 
and 7 are related to debugging and readability of the code, 1 and 2 and 
to large extend  3, 4 and 9 as well are small/local changes.

Looking forward, lets max the effort to enable distros which are to be 
released in the next months/quarters (e.g FC11, RHEL6, OEL6, SLES11 SP1, 
etc) to base their RDS code on mainline bits and not on other venues, 
thoughts?

Or.
> 1 2009-05-13     Andy Grover     RDS/IB: Drop connection when a fatal 
> QP event is received  
> 2 2009-05-13     Andy Grover     RDS/IB: Disable flow control in 
> sysctl and explain why     
> 3 2009-05-13     Andy Grover     RDS/IB: Move tx/rx ring init and 
> refill to later     
> 4 2009-05-13     Andy Grover     RDS: Don't set c_version in 
> __rds_conn_create()     
> 5 2009-05-13     Andy Grover     RDS/IB: Rename byte_len to data_len 
> to enhance readability
> 6 2009-05-13     Andy Grover     RDS/RDMA: Fix cut-n-paste errors in 
> printks in rdma ...     
> 7 2009-05-13     Andy Grover     RDS/IB: Fix printk to indicate remote 
> IP, not local     
> 8 2009-05-13     Andy Grover     RDS/IB: Handle connections using RDS 
> 3.0 wire protocol     
> 9 2009-05-13     Andy Grover     RDS/IB: Improve RDS protocol version 
> checking