> Subject: Re: iSER Connection via LIO not working > > Hi, > > after a bit of digging and adding about 100 logs I found the following > commit which screws with the qib implementation. > > https://gitlab- > beta.engr.illinois.edu/ejclark2/linux/commit/df749cdc45d9f97cb0a5e6ceab80 > e2e00ee9bf85 > > If you check, qib never sets the max_fast_reg_page_list_len (I assume, > haven't found it in the code, it's there for mellanox) > > therefore in iser_calc_scsi_params, the scsi_max_sectors is always set > to 0. > > Maybe someone that has experience there could figure out what exactly is > going on? > > Adding > > sup_sg_tablesize = ISCSI_ISER_MAX_SG_TABLESIZE; > > directly under the sup_sg_tablesize calculation at least fixes the iser > connection for QLogic Adapters (~800 - 900 MB/s direct IO vs. ~200 MB/s > via iscsi and no more errors) > Good job tracking this down. Thanks! static void iser_calc_scsi_params(struct iser_conn *iser_conn, unsigned int max_sectors) { struct iser_device *device = iser_conn->ib_conn.device; unsigned short sg_tablesize, sup_sg_tablesize; sg_tablesize = DIV_ROUND_UP(max_sectors * 512, SIZE_4K); sup_sg_tablesize = min_t(unsigned, ISCSI_ISER_MAX_SG_TABLESIZE, device->ib_device->attrs.max_fast_reg_page_list_len); iser_conn->scsi_sg_tablesize = min(sg_tablesize, sup_sg_tablesize); } The bug is that device->ib_device->attrs.max_fast_reg_page_list_len should only be considered valid IFF IB_DEVICE_MEM_MGT_EXTENSIONS is set in device->ib_device->device_cap_flags. The assignment you suggest should be an else on an if test of the device_cap_flags bit. This appears to be the only storage ULP that reads this field exported from drivers. Mike -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
