1. 03 2月, 2017 1 次提交
  2. 08 10月, 2016 2 次提交
    • S
      iw_cxgb4: add fast-path for small REG_MR operations · 49b53a93
      Steve Wise 提交于
      When processing a REG_MR work request, if fw supports the
      FW_RI_NSMR_TPTE_WR work request, and if the page list for this
      registration is <= 2 pages, and the current state of the mr is INVALID,
      then use FW_RI_NSMR_TPTE_WR to pass down a fully populated TPTE for FW
      to write.  This avoids FW having to do an async read of the TPTE blocking
      the SQ until the read completes.
      
      To know if the current MR state is INVALID or not, iw_cxgb4 must track the
      state of each fastreg MR.  The c4iw_mr struct state is updated as REG_MR
      and LOCAL_INV WRs are posted and completed, when a reg_mr is destroyed,
      and when RECV completions are processed that include a local invalidation.
      
      This optimization increases small IO IOPS for both iSER and NVMF.
      Signed-off-by: NSteve Wise <swise@opengridcomputing.com>
      Signed-off-by: NDoug Ledford <dledford@redhat.com>
      49b53a93
    • S
      cxgb4: advertise support for FR_NSMR_TPTE_WR · 086de575
      Steve Wise 提交于
      Query firmware for the FW_PARAMS_PARAM_DEV_RI_FR_NSMR_TPTE_WR parameter.
      If it exists and is 1, then advertise support for FR_NSMR_TPTE_WR to
      the ULDs.
      Signed-off-by: NSteve Wise <swise@opengridcomputing.com>
      Signed-off-by: NDoug Ledford <dledford@redhat.com>
      086de575
  3. 21 9月, 2016 1 次提交
  4. 23 8月, 2016 1 次提交
  5. 19 8月, 2016 2 次提交
  6. 26 7月, 2016 1 次提交
  7. 27 4月, 2016 1 次提交
  8. 22 3月, 2016 1 次提交
  9. 01 3月, 2016 1 次提交
    • H
      cxgb4/iw_cxgb4: TOS support · ac8e4c69
      Hariprasad S 提交于
      This series provides support for iWARP applications to specify a TOS
      value and have that map to a VLAN Priority for iw_cxgb4 iWARP connections.
      
      In iw_cxgb4, when allocating an L2T entry, pass the skb_priority based
      on the tos value in the cm_id. Also pass the correct tos value during
      connection setup so the passive side gets the client's desired tos.
      When sending the FLOWC work request to FW, if the egress device is
      in a vlan, then use the vlan priority bits as the scheduling class.
      This allows associating RDMA connections with scheduling classes to
      provide traffic shaping per flow.
      Signed-off-by: NSteve Wise <swise@opengridcomputing.com>
      Signed-off-by: NHariprasad Shenai <hariprasad@chelsio.com>
      Signed-off-by: NDoug Ledford <dledford@redhat.com>
      ac8e4c69
  10. 11 2月, 2016 1 次提交
    • H
      cxgb4/iw_cxgb4: TOS support · ba9cee6a
      Hariprasad Shenai 提交于
      This series provides support for iWARP applications to specify a TOS
      value and have that map to a VLAN Priority for iw_cxgb4 iWARP connections.
      
      In iw_cxgb4, when allocating an L2T entry, pass the skb_priority based
      on the tos value in the cm_id. Also pass the correct tos value during
      connection setup so the passive side gets the client's desired tos.
      When sending the FLOWC work request to FW, if the egress device is
      in a vlan, then use the vlan priority bits as the scheduling class.
      This allows associating RDMA connections with scheduling classes to
      provide traffic shaping per flow.
      Signed-off-by: NSteve Wise <swise@opengridcomputing.com>
      Signed-off-by: NHariprasad Shenai <hariprasad@chelsio.com>
      Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      ba9cee6a
  11. 10 9月, 2015 1 次提交
  12. 06 6月, 2015 1 次提交
  13. 02 6月, 2015 1 次提交
  14. 25 5月, 2015 1 次提交
  15. 13 5月, 2015 1 次提交
  16. 06 5月, 2015 2 次提交
  17. 02 4月, 2015 1 次提交
  18. 09 3月, 2015 1 次提交
  19. 10 2月, 2015 1 次提交
  20. 08 2月, 2015 1 次提交
  21. 25 1月, 2015 1 次提交
  22. 09 1月, 2015 1 次提交
  23. 19 12月, 2014 1 次提交
  24. 13 12月, 2014 2 次提交
  25. 23 11月, 2014 5 次提交
  26. 14 11月, 2014 1 次提交
  27. 11 11月, 2014 1 次提交
  28. 02 9月, 2014 1 次提交
  29. 22 8月, 2014 1 次提交
  30. 08 8月, 2014 1 次提交
    • A
      cxgb4: IEEE fixes for DCBx state machine · 10b00466
      Anish Bhatt 提交于
      * Changes required due to 16eecd9b ("dcbnl : Fix misleading
        dcb_app->priority explanation")
      * Driver was previously not aware of what DCBx version was negotiated by
        firmware, this could lead to DCB app table  in kernel or in firmware being
        populated wrong  since IEEE/CEE used different formats made clear by above
        mentioned commit
      * Driver was missing a couple of state transitions that could be caused
        by other drivers that use chelsio hardware, resulting in incorrect behaviour
        (the change that addresses this also flips the state machine to switch on
         state instead of transition, hope this is okay in current window)
      * Prio queue info & tsa is no longer thrown away
      
      v2: Print DCBx state transition messages only when debug is enabled
      Signed-off-by: NAnish Bhatt <anish@chelsio.com>
      Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      10b00466
  31. 16 7月, 2014 1 次提交
    • H
      cxgb4/iw_cxgb4: use firmware ord/ird resource limits · 4c2c5763
      Hariprasad Shenai 提交于
      Advertise a larger max read queue depth for qps, and gather the resource limits
      from fw and use them to avoid exhaustinq all the resources.
      
      Design:
      
      cxgb4:
      
      Obtain the max_ordird_qp and max_ird_adapter device params from FW
      at init time and pass them up to the ULDs when they attach.  If these
      parameters are not available, due to older firmware, then hard-code
      the values based on the known values for older firmware.
      iw_cxgb4:
      
      Fix the c4iw_query_device() to report these correct values based on
      adapter parameters.  ibv_query_device() will always return:
      
      max_qp_rd_atom = max_qp_init_rd_atom = min(module_max, max_ordird_qp)
      max_res_rd_atom = max_ird_adapter
      
      Bump up the per qp max module option to 32, allowing it to be increased
      by the user up to the device max of max_ordird_qp.  32 seems to be
      sufficient to maximize throughput for streaming read benchmarks.
      
      Fail connection setup if the negotiated IRD exhausts the available
      adapter ird resources.  So the driver will track the amount of ird
      resource in use and not send an RI_WR/INIT to FW that would reduce the
      available ird resources below zero.
      Signed-off-by: NSteve Wise <swise@opengridcomputing.com>
      Signed-off-by: NHariprasad Shenai <hariprasad@chelsio.com>
      Signed-off-by: NDavid S. Miller <davem@davemloft.net>
      4c2c5763
  32. 23 6月, 2014 1 次提交