Hi Jeff,
thanks for submitting the patches. Since we last spoke, there was also a parallel submission of SLES11 SP2 patches from another user about a week ago, as you are already aware. It looks like your patch series addresses a number of other issues beyond the patches that have already been submitted, so I hope two patch series can be worked out and landed.
For reference, the patch series are at:
master series: http://review.whamcloud.com/5113
b2_1 series: http://review.whamcloud.com/5057
I wasn't aware that you would be submitting patches for SP2 support on b2_1. Currently we do not have plans to update the b2_1 stable branch for kernels beyond 2.6.37 due to the changes needed in the dcache. Are all of the client-side patches in the b2_1 series also included in the master series as well? It looks like a number of the b2_1 patches are back-ports from newer Lustre branches, but some are new to me and I want to make sure that they are also targetted at the master branch.
Hi Andreas -
Yes, I've included the client-side patches in both the b2_1 and master branches.
The patch in http://review.whamcloud.com/#change,5121 recently landed to master seems to have broken the build for me on sles11sp2. I'm not exactly sure why. The section of the new code
#ifndef __REQ_WRITE /* pre-2.6.35 */ #define __REQ_WRITE BIO_RW #endif
appears to be going wrong. Even though __REQ_WRITE is defined in an enum in linux/blk_types.h, the build does the local #define and fails due to no BIO_RW being #define'd anywhere.
If I just comment out that #ifndef ... #endif section the build completes fine.
Oops. That's a typo. http://review.whamcloud.com/5503 fixes that.
Thanks, Jeff. That 1 line repair works for me.
Last patch to support SLES11 SP2 server side has landed. Only the kernel tuning and fail over simulation patch are left which belong under LU-1812. Peter you can close this ticket.
With pleasure 