staging/rdma/hfi1: Fix port bounce issues with 0.22 DC firmware
authorEaswar Hariharan <easwar.hariharan@intel.com>
Mon, 26 Oct 2015 14:28:36 +0000 (10:28 -0400)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Tue, 27 Oct 2015 08:19:22 +0000 (17:19 +0900)
The DC firmware overwrites the enable_lane_tx register and does not update it
on a host request to go to Poll. This causes an infinite loop through the LNI
state machine if a link width downgrade occurs. This patch re-sets the
enable_lane_tx register to all 4 lanes.

Reviewed-by: Dean Luick <dean.luick@intel.com>
Signed-off-by: Easwar Hariharan <easwar.hariharan@intel.com>
Signed-off-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/staging/rdma/hfi1/chip.c

index 52a75a6111487316faa41e8b9f12db11b8353be4..e47420f4786bb3146b426d3cd44f01b2ffb64175 100644 (file)
@@ -5415,6 +5415,8 @@ static int set_local_link_attributes(struct hfi1_pportdata *ppd)
                if (ppd->link_speed_enabled & OPA_LINK_SPEED_12_5G)
                        ppd->local_tx_rate |= 1;
        }
+
+       enable_lane_tx = 0xF; /* enable all four lanes */
        ret = write_tx_settings(dd, enable_lane_tx, tx_polarity_inversion,
                     rx_polarity_inversion, ppd->local_tx_rate);
        if (ret != HCMD_SUCCESS)