On 2/20/2014 1:14 AM, Nicholas A. Bellinger wrote:
On Wed, 2014-02-19 at 17:50 +0200, Sagi Grimberg wrote:Hey Nic, I addressed your comments in the this set. I'll try to send the initiator code for review in the next couple of days. This patchset introduces target side T10-PI offload support over RDMA. Currently the implementation is for iSER transport but can be easily extended to SRP (or FCoE in the future). Should mention that this patchset depends on RDMA signature verbs making it for inclusion which will hopefully happen in near future. This code was tested against iSER legacy initiator, and also initiator that uses T10-PI offload as well. I'll clean up the initiator code in the following days and submit that as well. This code works under medium loads using backstores: - FileIO with DIF emulation. - RD with DIF emulation. - iBLOCK (scsi_debug with DIF support). Chnages from v1: - Rebased rdma_dif (3.14-rc2) - Target core: - Pass zero_flag=true to alloc_sgl for protection buffers - Removed Unneeded inline function rwprotect - Pass is_write bool to sbc_set_prot_op_checks - Send failure response for absence of protection buffers when needed. - iSER: - Removed unneeded assignments in isert_set_sig_attrs Chnages from v0: - Rebased from for-next - Target core: - Don't minor fixes for check_prot - to support transports that doesn't use submit_map_sgls. - file format - use escape values. - Removed redundant prot_handover. - Added protection checks and operation set. - iSER: - Added preperation routines for mapping/unmapping buffers to ease the amount of code in isert_reg_rdma. - Fixed print of DIF error (sector instead of offset). - Fix RDMA length for protection on wire domain. - Refactored reg_sig_mr to use helper routines.Hey Sagi, FYI, this series did not compile: drivers/infiniband/ulp/isert/ib_isert.c: In function ‘isert_reg_sig_mr’: drivers/infiniband/ulp/isert/ib_isert.c:2531:2: error: expected ‘;’ before ‘ret’ drivers/infiniband/ulp/isert/ib_isert.c: In function ‘isert_reg_rdma’: drivers/infiniband/ulp/isert/ib_isert.c:2611:24: warning: comparison between ‘enum target_prot_type’ and ‘enum target_prot_op’ [-Wenum-compare] drivers/infiniband/ulp/isert/ib_isert.c:2625:24: warning: comparison between ‘enum target_prot_type’ and ‘enum target_prot_op’ [-Wenum-compare] drivers/infiniband/ulp/isert/ib_isert.c:2667:43: warning: comparison between ‘enum target_prot_type’ and ‘enum target_prot_op’ [-Wenum-compare] drivers/infiniband/ulp/isert/ib_isert.c: In function ‘isert_put_datain’: drivers/infiniband/ulp/isert/ib_isert.c:2711:24: warning: comparison between ‘enum target_prot_type’ and ‘enum target_prot_op’ [-Wenum-compare] drivers/infiniband/ulp/isert/ib_isert.c:2732:24: warning: comparison between ‘enum target_prot_type’ and ‘enum target_prot_op’ [-Wenum-compare] drivers/infiniband/ulp/isert/ib_isert.c: At top level: drivers/infiniband/ulp/isert/ib_isert.c:2468:1: warning: ‘isert_set_sig_attrs’ defined but not used [-Wunused-function] drivers/target/target_core_transport.c: In function ‘transport_generic_new_cmd’: drivers/target/target_core_transport.c:2221:22: warning: comparison between ‘enum target_prot_type’ and ‘enum target_prot_op’ [-Wenum-compare] Squashing the following patch(es) into your original series, and applied to target-pending/rdma-dif.
Ohhh, last second cosmetics... Thanks for fixing it, won't happen again...
Thank you,
Thank you...
--nab
-- To unsubscribe from this list: send the line "unsubscribe target-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
