* Pull request: sfc-next 2013-12-12
@ 2013-12-12 22:56 Ben Hutchings
2013-12-12 22:59 ` [PATCH net-next 01/16] sfc: Demote "MC Scheduler error" messages Ben Hutchings
` (16 more replies)
0 siblings, 17 replies; 18+ messages in thread
From: Ben Hutchings @ 2013-12-12 22:56 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
The following changes since commit f54b311142a92ea2e42598e347b84e1655caf8e3:
tcp: auto corking (2013-12-06 12:51:41 -0500)
are available in the git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/bwh/sfc-next.git for-davem
for you to fetch changes up to ac36baf817c39fc9b53eff190f5901610c5dc5b7:
sfc: Remove dependency of PTP on having a dedicated channel (2013-12-12 22:07:02 +0000)
An assortment of changes for Linux 3.14:
1. Merge the sfc fixes that you have already merged into net.git.
(The branch point for those was such that this does not bring in any
other changes.)
2. Reduce log level for a generally useless warning message, from
Robert Stonehouse.
3. Include BISTs in ethtool offline self-test for EF10 and recover from
BISTs initiated through other functions, from Jon Cooper.
4. Improve a sanity check on RX completions.
5. Avoid incrementing RX dropped count while the interface is down, from
Jon Cooper.
6. Improve hardware sensor naming and log messages, from Edward Cree.
7. Log all unexpected errors returned by firmware, from Edward Cree.
8. Expose another NVRAM partition to userland.
9. Some refactoring of the PTP code in preparation for EF10 support.
10. Various minor cleanups.
Ben.
----------------------------------------------------------------
Alexandre Rames (1):
sfc: Stop/re-start PTP when stopping/starting the datapath.
Andrew Rybchenko (1):
sfc: RX buffer allocation takes prefix size into account in IP header alignment
Ben Hutchings (13):
sfc: Add length checks to efx_xmit_with_hwtstamp() and efx_ptp_is_ptp_tx()
sfc: Rate-limit log message for PTP packets without a matching timestamp event
sfc: Maintain current frequency adjustment when applying a time offset
Merge branch 'sfc-3.13' into master
sfc: Update MCDI protocol definitions
sfc: Tighten the check for RX merged completion events
sfc: Add new sensor names
sfc: Map MCDI error MC_CMD_ERR_ENOTSUP to Linux EOPNOTSUPP
sfc: Fold efx_flush_all() into efx_stop_port() and update comments
sfc: Expose NVRAM_PARTITION_TYPE_LICENSE on EF10
sfc: Return EBUSY for filter insertion on EF10, matching Falcon/Siena
sfc: Split PTP multicast filter insertion/removal out of efx_ptp_{start,stop}()
sfc: Remove dependency of PTP on having a dedicated channel
Edward Cree (3):
sfc: Report units in sensor warnings
sfc: Revise sensor names to be more understandable and consistent
sfc: Log all unexpected MCDI errors
Jon Cooper (3):
sfc: Add MC BISTs to ethtool offline self test on EF10
sfc: Make initial fill of RX descriptors synchronous
sfc: Correct RX dropped count for drops while interface is down
Laurence Evans (1):
sfc: PTP: Moderate log message on event queue overflow
Robert Stonehouse (3):
sfc: Refactor efx_mcdi_poll() by introducing efx_mcdi_poll_once()
sfc: Poll for MCDI completion once before timeout occurs
sfc: Demote "MC Scheduler error" messages
drivers/net/ethernet/sfc/ef10.c | 143 +++++--
drivers/net/ethernet/sfc/efx.c | 84 +++-
drivers/net/ethernet/sfc/efx.h | 2 +-
drivers/net/ethernet/sfc/enum.h | 1 +
drivers/net/ethernet/sfc/ethtool.c | 2 +
drivers/net/ethernet/sfc/falcon.c | 10 +
drivers/net/ethernet/sfc/farch.c | 2 +-
drivers/net/ethernet/sfc/mcdi.c | 424 ++++++++++++--------
drivers/net/ethernet/sfc/mcdi.h | 19 +
drivers/net/ethernet/sfc/mcdi_mon.c | 76 ++--
drivers/net/ethernet/sfc/mcdi_pcol.h | 733 ++++++++++++++++++++++++++++++----
drivers/net/ethernet/sfc/mcdi_port.c | 89 ++---
drivers/net/ethernet/sfc/net_driver.h | 9 +
drivers/net/ethernet/sfc/nic.c | 12 +
drivers/net/ethernet/sfc/nic.h | 7 +-
drivers/net/ethernet/sfc/ptp.c | 203 +++++++---
drivers/net/ethernet/sfc/rx.c | 15 +-
drivers/net/ethernet/sfc/selftest.c | 2 +-
drivers/net/ethernet/sfc/selftest.h | 1 +
drivers/net/ethernet/sfc/siena.c | 5 +-
20 files changed, 1418 insertions(+), 421 deletions(-)
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply [flat|nested] 18+ messages in thread
* [PATCH net-next 01/16] sfc: Demote "MC Scheduler error" messages
2013-12-12 22:56 Pull request: sfc-next 2013-12-12 Ben Hutchings
@ 2013-12-12 22:59 ` Ben Hutchings
2013-12-12 22:59 ` [PATCH net-next 02/16] sfc: Update MCDI protocol definitions Ben Hutchings
` (15 subsequent siblings)
16 siblings, 0 replies; 18+ messages in thread
From: Ben Hutchings @ 2013-12-12 22:59 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
From: Robert Stonehouse <rstonehouse@solarflare.com>
The MC firmware is cooperatively multitasking and its scheduler will
send an event when a task yields after running for more than the
expected maximum time. This can be useful for firmware development
but does not usually indicate a serious error and does not help to
detect a lockup (there is a hardware watchdog that does that).
Change the message and reduce log level accordingly.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/mcdi.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/sfc/mcdi.c b/drivers/net/ethernet/sfc/mcdi.c
index 4b0bd8a1514d..25f91c0ca6a6 100644
--- a/drivers/net/ethernet/sfc/mcdi.c
+++ b/drivers/net/ethernet/sfc/mcdi.c
@@ -867,8 +867,8 @@ void efx_mcdi_process_event(struct efx_channel *channel,
efx_mcdi_sensor_event(efx, event);
break;
case MCDI_EVENT_CODE_SCHEDERR:
- netif_info(efx, hw, efx->net_dev,
- "MC Scheduler error address=0x%x\n", data);
+ netif_dbg(efx, hw, efx->net_dev,
+ "MC Scheduler alert (0x%x)\n", data);
break;
case MCDI_EVENT_CODE_REBOOT:
case MCDI_EVENT_CODE_MC_REBOOT:
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH net-next 02/16] sfc: Update MCDI protocol definitions
2013-12-12 22:56 Pull request: sfc-next 2013-12-12 Ben Hutchings
2013-12-12 22:59 ` [PATCH net-next 01/16] sfc: Demote "MC Scheduler error" messages Ben Hutchings
@ 2013-12-12 22:59 ` Ben Hutchings
2013-12-12 22:59 ` [PATCH net-next 03/16] sfc: Add MC BISTs to ethtool offline self test on EF10 Ben Hutchings
` (14 subsequent siblings)
16 siblings, 0 replies; 18+ messages in thread
From: Ben Hutchings @ 2013-12-12 22:59 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/mcdi_pcol.h | 733 +++++++++++++++++++++++++++++++----
1 file changed, 663 insertions(+), 70 deletions(-)
diff --git a/drivers/net/ethernet/sfc/mcdi_pcol.h b/drivers/net/ethernet/sfc/mcdi_pcol.h
index e0a63ddb7a6c..a707fb5ef14c 100644
--- a/drivers/net/ethernet/sfc/mcdi_pcol.h
+++ b/drivers/net/ethernet/sfc/mcdi_pcol.h
@@ -224,6 +224,8 @@
#define MC_CMD_ERR_MAC_EXIST 0x1009
/* Slave core not present */
#define MC_CMD_ERR_SLAVE_NOT_PRESENT 0x100a
+/* The datapath is disabled. */
+#define MC_CMD_ERR_DATAPATH_DISABLED 0x100b
#define MC_CMD_ERR_CODE_OFST 0
@@ -390,6 +392,8 @@
* AOE_ERR_DATA)
*/
#define MCDI_EVENT_AOE_BYTEBLASTER 0x9
+/* enum: DDR ECC status update */
+#define MCDI_EVENT_AOE_DDR_ECC_STATUS 0xa
#define MCDI_EVENT_AOE_ERR_DATA_LBN 8
#define MCDI_EVENT_AOE_ERR_DATA_WIDTH 8
#define MCDI_EVENT_RX_ERR_RXQ_LBN 0
@@ -462,6 +466,10 @@
#define MCDI_EVENT_CODE_ECC_CORR_ERR 0x17
/* enum: the MC has detected an uncorrectable error */
#define MCDI_EVENT_CODE_ECC_FATAL_ERR 0x18
+/* enum: The MC has entered offline BIST mode */
+#define MCDI_EVENT_CODE_MC_BIST 0x19
+/* enum: PTP tick event providing current NIC time */
+#define MCDI_EVENT_CODE_PTP_TIME 0x1a
/* enum: Artificial event generated by host and posted via MC for test
* purposes.
*/
@@ -481,15 +489,32 @@
#define MCDI_EVENT_TX_ERR_DATA_OFST 0
#define MCDI_EVENT_TX_ERR_DATA_LBN 0
#define MCDI_EVENT_TX_ERR_DATA_WIDTH 32
-/* Seconds field of timestamp */
+/* For CODE_PTP_RX, CODE_PTP_PPS and CODE_HW_PPS events the seconds field of
+ * timestamp
+ */
#define MCDI_EVENT_PTP_SECONDS_OFST 0
#define MCDI_EVENT_PTP_SECONDS_LBN 0
#define MCDI_EVENT_PTP_SECONDS_WIDTH 32
-/* Nanoseconds field of timestamp */
+/* For CODE_PTP_RX, CODE_PTP_PPS and CODE_HW_PPS events the major field of
+ * timestamp
+ */
+#define MCDI_EVENT_PTP_MAJOR_OFST 0
+#define MCDI_EVENT_PTP_MAJOR_LBN 0
+#define MCDI_EVENT_PTP_MAJOR_WIDTH 32
+/* For CODE_PTP_RX, CODE_PTP_PPS and CODE_HW_PPS events the nanoseconds field
+ * of timestamp
+ */
#define MCDI_EVENT_PTP_NANOSECONDS_OFST 0
#define MCDI_EVENT_PTP_NANOSECONDS_LBN 0
#define MCDI_EVENT_PTP_NANOSECONDS_WIDTH 32
-/* Lowest four bytes of sourceUUID from PTP packet */
+/* For CODE_PTP_RX, CODE_PTP_PPS and CODE_HW_PPS events the minor field of
+ * timestamp
+ */
+#define MCDI_EVENT_PTP_MINOR_OFST 0
+#define MCDI_EVENT_PTP_MINOR_LBN 0
+#define MCDI_EVENT_PTP_MINOR_WIDTH 32
+/* For CODE_PTP_RX events, the lowest four bytes of sourceUUID from PTP packet
+ */
#define MCDI_EVENT_PTP_UUID_OFST 0
#define MCDI_EVENT_PTP_UUID_LBN 0
#define MCDI_EVENT_PTP_UUID_WIDTH 32
@@ -505,6 +530,13 @@
#define MCDI_EVENT_ECC_FATAL_ERR_DATA_OFST 0
#define MCDI_EVENT_ECC_FATAL_ERR_DATA_LBN 0
#define MCDI_EVENT_ECC_FATAL_ERR_DATA_WIDTH 32
+/* For CODE_PTP_TIME events, the major value of the PTP clock */
+#define MCDI_EVENT_PTP_TIME_MAJOR_OFST 0
+#define MCDI_EVENT_PTP_TIME_MAJOR_LBN 0
+#define MCDI_EVENT_PTP_TIME_MAJOR_WIDTH 32
+/* For CODE_PTP_TIME events, bits 19-26 of the minor value of the PTP clock */
+#define MCDI_EVENT_PTP_TIME_MINOR_26_19_LBN 36
+#define MCDI_EVENT_PTP_TIME_MINOR_26_19_WIDTH 8
/* FCDI_EVENT structuredef */
#define FCDI_EVENT_LEN 8
@@ -545,8 +577,10 @@
#define FCDI_EVENT_CODE_TIMED_READ 0x5
/* enum: One or more PPS IN events */
#define FCDI_EVENT_CODE_PPS_IN 0x6
-/* enum: One or more PPS OUT events */
-#define FCDI_EVENT_CODE_PPS_OUT 0x7
+/* enum: Tick event from PTP clock */
+#define FCDI_EVENT_CODE_PTP_TICK 0x7
+/* enum: ECC error counters */
+#define FCDI_EVENT_CODE_DDR_ECC_STATUS 0x8
#define FCDI_EVENT_ASSERT_INSTR_ADDRESS_OFST 0
#define FCDI_EVENT_ASSERT_INSTR_ADDRESS_LBN 0
#define FCDI_EVENT_ASSERT_INSTR_ADDRESS_WIDTH 32
@@ -560,14 +594,21 @@
#define FCDI_EVENT_LINK_STATE_DATA_OFST 0
#define FCDI_EVENT_LINK_STATE_DATA_LBN 0
#define FCDI_EVENT_LINK_STATE_DATA_WIDTH 32
-#define FCDI_EVENT_PPS_COUNT_OFST 0
-#define FCDI_EVENT_PPS_COUNT_LBN 0
-#define FCDI_EVENT_PPS_COUNT_WIDTH 32
-
-/* FCDI_EXTENDED_EVENT structuredef */
-#define FCDI_EXTENDED_EVENT_LENMIN 16
-#define FCDI_EXTENDED_EVENT_LENMAX 248
-#define FCDI_EXTENDED_EVENT_LEN(num) (8+8*(num))
+#define FCDI_EVENT_DDR_ECC_STATUS_BANK_ID_LBN 36
+#define FCDI_EVENT_DDR_ECC_STATUS_BANK_ID_WIDTH 8
+#define FCDI_EVENT_DDR_ECC_STATUS_STATUS_OFST 0
+#define FCDI_EVENT_DDR_ECC_STATUS_STATUS_LBN 0
+#define FCDI_EVENT_DDR_ECC_STATUS_STATUS_WIDTH 32
+
+/* FCDI_EXTENDED_EVENT_PPS structuredef: Extended FCDI event to send PPS events
+ * to the MC. Note that this structure | is overlayed over a normal FCDI event
+ * such that bits 32-63 containing | event code, level, source etc remain the
+ * same. In this case the data | field of the header is defined to be the
+ * number of timestamps
+ */
+#define FCDI_EXTENDED_EVENT_PPS_LENMIN 16
+#define FCDI_EXTENDED_EVENT_PPS_LENMAX 248
+#define FCDI_EXTENDED_EVENT_PPS_LEN(num) (8+8*(num))
/* Number of timestamps following */
#define FCDI_EXTENDED_EVENT_PPS_COUNT_OFST 0
#define FCDI_EXTENDED_EVENT_PPS_COUNT_LBN 0
@@ -581,14 +622,14 @@
#define FCDI_EXTENDED_EVENT_PPS_NANOSECONDS_LBN 96
#define FCDI_EXTENDED_EVENT_PPS_NANOSECONDS_WIDTH 32
/* Timestamp records comprising the event */
-#define FCDI_EXTENDED_EVENT_PPS_TIME_OFST 8
-#define FCDI_EXTENDED_EVENT_PPS_TIME_LEN 8
-#define FCDI_EXTENDED_EVENT_PPS_TIME_LO_OFST 8
-#define FCDI_EXTENDED_EVENT_PPS_TIME_HI_OFST 12
-#define FCDI_EXTENDED_EVENT_PPS_TIME_MINNUM 1
-#define FCDI_EXTENDED_EVENT_PPS_TIME_MAXNUM 30
-#define FCDI_EXTENDED_EVENT_PPS_TIME_LBN 64
-#define FCDI_EXTENDED_EVENT_PPS_TIME_WIDTH 64
+#define FCDI_EXTENDED_EVENT_PPS_TIMESTAMPS_OFST 8
+#define FCDI_EXTENDED_EVENT_PPS_TIMESTAMPS_LEN 8
+#define FCDI_EXTENDED_EVENT_PPS_TIMESTAMPS_LO_OFST 8
+#define FCDI_EXTENDED_EVENT_PPS_TIMESTAMPS_HI_OFST 12
+#define FCDI_EXTENDED_EVENT_PPS_TIMESTAMPS_MINNUM 1
+#define FCDI_EXTENDED_EVENT_PPS_TIMESTAMPS_MAXNUM 30
+#define FCDI_EXTENDED_EVENT_PPS_TIMESTAMPS_LBN 64
+#define FCDI_EXTENDED_EVENT_PPS_TIMESTAMPS_WIDTH 64
/***********************************/
@@ -642,6 +683,10 @@
#define MC_CMD_COPYCODE_IN_LEN 16
/* Source address */
#define MC_CMD_COPYCODE_IN_SRC_ADDR_OFST 0
+/* enum: The main image should be entered via a copy of a single word from and
+ * to this address when none of the other magic behaviours are required.
+ */
+#define MC_CMD_COPYCODE_HUNT_NO_MAGIC_ADDR 0x10000
/* enum: Entering the main image via a copy of a single word from and to this
* address indicates that it should not attempt to start the datapath CPUs.
* This is useful for certain soft rebooting scenarios. (Huntington only)
@@ -872,8 +917,28 @@
#define MC_CMD_PTP_OP_RST_CLK 0x14
/* enum: Enable the forwarding of PPS events to the host */
#define MC_CMD_PTP_OP_PPS_ENABLE 0x15
+/* enum: Get the time format used by this NIC for PTP operations */
+#define MC_CMD_PTP_OP_GET_TIME_FORMAT 0x16
+/* enum: Get the clock attributes. NOTE- extended version of
+ * MC_CMD_PTP_OP_GET_TIME_FORMAT
+ */
+#define MC_CMD_PTP_OP_GET_ATTRIBUTES 0x16
+/* enum: Get corrections that should be applied to the various different
+ * timestamps
+ */
+#define MC_CMD_PTP_OP_GET_TIMESTAMP_CORRECTIONS 0x17
+/* enum: Subscribe to receive periodic time events indicating the current NIC
+ * time
+ */
+#define MC_CMD_PTP_OP_TIME_EVENT_SUBSCRIBE 0x18
+/* enum: Unsubscribe to stop receiving time events */
+#define MC_CMD_PTP_OP_TIME_EVENT_UNSUBSCRIBE 0x19
+/* enum: PPS based manfacturing tests. Requires PPS output to be looped to PPS
+ * input on the same NIC.
+ */
+#define MC_CMD_PTP_OP_MANFTEST_PPS 0x1a
/* enum: Above this for future use. */
-#define MC_CMD_PTP_OP_MAX 0x16
+#define MC_CMD_PTP_OP_MAX 0x1b
/* MC_CMD_PTP_IN_ENABLE msgrequest */
#define MC_CMD_PTP_IN_ENABLE_LEN 16
@@ -938,8 +1003,12 @@
#define MC_CMD_PTP_IN_ADJUST_BITS 0x28
/* Time adjustment in seconds */
#define MC_CMD_PTP_IN_ADJUST_SECONDS_OFST 16
+/* Time adjustment major value */
+#define MC_CMD_PTP_IN_ADJUST_MAJOR_OFST 16
/* Time adjustment in nanoseconds */
#define MC_CMD_PTP_IN_ADJUST_NANOSECONDS_OFST 20
+/* Time adjustment minor value */
+#define MC_CMD_PTP_IN_ADJUST_MINOR_OFST 20
/* MC_CMD_PTP_IN_SYNCHRONIZE msgrequest */
#define MC_CMD_PTP_IN_SYNCHRONIZE_LEN 20
@@ -1005,8 +1074,12 @@
/* MC_CMD_PTP_IN_PERIPH_ID_OFST 4 */
/* Time adjustment in seconds */
#define MC_CMD_PTP_IN_CLOCK_OFFSET_ADJUST_SECONDS_OFST 8
+/* Time adjustment major value */
+#define MC_CMD_PTP_IN_CLOCK_OFFSET_ADJUST_MAJOR_OFST 8
/* Time adjustment in nanoseconds */
#define MC_CMD_PTP_IN_CLOCK_OFFSET_ADJUST_NANOSECONDS_OFST 12
+/* Time adjustment minor value */
+#define MC_CMD_PTP_IN_CLOCK_OFFSET_ADJUST_MINOR_OFST 12
/* MC_CMD_PTP_IN_CLOCK_FREQ_ADJUST msgrequest */
#define MC_CMD_PTP_IN_CLOCK_FREQ_ADJUST_LEN 16
@@ -1078,9 +1151,51 @@
#define MC_CMD_PTP_ENABLE_PPS 0x0
/* enum: Disable */
#define MC_CMD_PTP_DISABLE_PPS 0x1
-/* Queueid to send events back */
+/* Queue id to send events back */
#define MC_CMD_PTP_IN_PPS_ENABLE_QUEUE_ID_OFST 8
+/* MC_CMD_PTP_IN_GET_TIME_FORMAT msgrequest */
+#define MC_CMD_PTP_IN_GET_TIME_FORMAT_LEN 8
+/* MC_CMD_PTP_IN_CMD_OFST 0 */
+/* MC_CMD_PTP_IN_PERIPH_ID_OFST 4 */
+
+/* MC_CMD_PTP_IN_GET_ATTRIBUTES msgrequest */
+#define MC_CMD_PTP_IN_GET_ATTRIBUTES_LEN 8
+/* MC_CMD_PTP_IN_CMD_OFST 0 */
+/* MC_CMD_PTP_IN_PERIPH_ID_OFST 4 */
+
+/* MC_CMD_PTP_IN_GET_TIMESTAMP_CORRECTIONS msgrequest */
+#define MC_CMD_PTP_IN_GET_TIMESTAMP_CORRECTIONS_LEN 8
+/* MC_CMD_PTP_IN_CMD_OFST 0 */
+/* MC_CMD_PTP_IN_PERIPH_ID_OFST 4 */
+
+/* MC_CMD_PTP_IN_TIME_EVENT_SUBSCRIBE msgrequest */
+#define MC_CMD_PTP_IN_TIME_EVENT_SUBSCRIBE_LEN 12
+/* MC_CMD_PTP_IN_CMD_OFST 0 */
+/* MC_CMD_PTP_IN_PERIPH_ID_OFST 4 */
+/* Event queue to send PTP time events to */
+#define MC_CMD_PTP_IN_TIME_EVENT_SUBSCRIBE_QUEUE_OFST 8
+
+/* MC_CMD_PTP_IN_TIME_EVENT_UNSUBSCRIBE msgrequest */
+#define MC_CMD_PTP_IN_TIME_EVENT_UNSUBSCRIBE_LEN 16
+/* MC_CMD_PTP_IN_CMD_OFST 0 */
+/* MC_CMD_PTP_IN_PERIPH_ID_OFST 4 */
+/* Unsubscribe options */
+#define MC_CMD_PTP_IN_TIME_EVENT_UNSUBSCRIBE_CONTROL_OFST 8
+/* enum: Unsubscribe a single queue */
+#define MC_CMD_PTP_IN_TIME_EVENT_UNSUBSCRIBE_SINGLE 0x0
+/* enum: Unsubscribe all queues */
+#define MC_CMD_PTP_IN_TIME_EVENT_UNSUBSCRIBE_ALL 0x1
+/* Event queue ID */
+#define MC_CMD_PTP_IN_TIME_EVENT_UNSUBSCRIBE_QUEUE_OFST 12
+
+/* MC_CMD_PTP_IN_MANFTEST_PPS msgrequest */
+#define MC_CMD_PTP_IN_MANFTEST_PPS_LEN 12
+/* MC_CMD_PTP_IN_CMD_OFST 0 */
+/* MC_CMD_PTP_IN_PERIPH_ID_OFST 4 */
+/* 1 to enable PPS test mode, 0 to disable and return result. */
+#define MC_CMD_PTP_IN_MANFTEST_PPS_TEST_ENABLE_OFST 8
+
/* MC_CMD_PTP_OUT msgresponse */
#define MC_CMD_PTP_OUT_LEN 0
@@ -1088,15 +1203,29 @@
#define MC_CMD_PTP_OUT_TRANSMIT_LEN 8
/* Value of seconds timestamp */
#define MC_CMD_PTP_OUT_TRANSMIT_SECONDS_OFST 0
+/* Timestamp major value */
+#define MC_CMD_PTP_OUT_TRANSMIT_MAJOR_OFST 0
/* Value of nanoseconds timestamp */
#define MC_CMD_PTP_OUT_TRANSMIT_NANOSECONDS_OFST 4
+/* Timestamp minor value */
+#define MC_CMD_PTP_OUT_TRANSMIT_MINOR_OFST 4
+
+/* MC_CMD_PTP_OUT_TIME_EVENT_SUBSCRIBE msgresponse */
+#define MC_CMD_PTP_OUT_TIME_EVENT_SUBSCRIBE_LEN 0
+
+/* MC_CMD_PTP_OUT_TIME_EVENT_UNSUBSCRIBE msgresponse */
+#define MC_CMD_PTP_OUT_TIME_EVENT_UNSUBSCRIBE_LEN 0
/* MC_CMD_PTP_OUT_READ_NIC_TIME msgresponse */
#define MC_CMD_PTP_OUT_READ_NIC_TIME_LEN 8
/* Value of seconds timestamp */
#define MC_CMD_PTP_OUT_READ_NIC_TIME_SECONDS_OFST 0
+/* Timestamp major value */
+#define MC_CMD_PTP_OUT_READ_NIC_TIME_MAJOR_OFST 0
/* Value of nanoseconds timestamp */
#define MC_CMD_PTP_OUT_READ_NIC_TIME_NANOSECONDS_OFST 4
+/* Timestamp minor value */
+#define MC_CMD_PTP_OUT_READ_NIC_TIME_MINOR_OFST 4
/* MC_CMD_PTP_OUT_STATUS msgresponse */
#define MC_CMD_PTP_OUT_STATUS_LEN 64
@@ -1116,21 +1245,21 @@
#define MC_CMD_PTP_OUT_STATUS_STATS_PPS_OFLOW_OFST 24
/* Number of PPS bad periods */
#define MC_CMD_PTP_OUT_STATUS_STATS_PPS_BAD_OFST 28
-/* Minimum period of PPS pulse */
+/* Minimum period of PPS pulse in nanoseconds */
#define MC_CMD_PTP_OUT_STATUS_STATS_PPS_PER_MIN_OFST 32
-/* Maximum period of PPS pulse */
+/* Maximum period of PPS pulse in nanoseconds */
#define MC_CMD_PTP_OUT_STATUS_STATS_PPS_PER_MAX_OFST 36
-/* Last period of PPS pulse */
+/* Last period of PPS pulse in nanoseconds */
#define MC_CMD_PTP_OUT_STATUS_STATS_PPS_PER_LAST_OFST 40
-/* Mean period of PPS pulse */
+/* Mean period of PPS pulse in nanoseconds */
#define MC_CMD_PTP_OUT_STATUS_STATS_PPS_PER_MEAN_OFST 44
-/* Minimum offset of PPS pulse (signed) */
+/* Minimum offset of PPS pulse in nanoseconds (signed) */
#define MC_CMD_PTP_OUT_STATUS_STATS_PPS_OFF_MIN_OFST 48
-/* Maximum offset of PPS pulse (signed) */
+/* Maximum offset of PPS pulse in nanoseconds (signed) */
#define MC_CMD_PTP_OUT_STATUS_STATS_PPS_OFF_MAX_OFST 52
-/* Last offset of PPS pulse (signed) */
+/* Last offset of PPS pulse in nanoseconds (signed) */
#define MC_CMD_PTP_OUT_STATUS_STATS_PPS_OFF_LAST_OFST 56
-/* Mean offset of PPS pulse (signed) */
+/* Mean offset of PPS pulse in nanoseconds (signed) */
#define MC_CMD_PTP_OUT_STATUS_STATS_PPS_OFF_MEAN_OFST 60
/* MC_CMD_PTP_OUT_SYNCHRONIZE msgresponse */
@@ -1146,8 +1275,12 @@
#define MC_CMD_PTP_OUT_SYNCHRONIZE_HOSTSTART_OFST 0
/* Value of seconds timestamp */
#define MC_CMD_PTP_OUT_SYNCHRONIZE_SECONDS_OFST 4
+/* Timestamp major value */
+#define MC_CMD_PTP_OUT_SYNCHRONIZE_MAJOR_OFST 4
/* Value of nanoseconds timestamp */
#define MC_CMD_PTP_OUT_SYNCHRONIZE_NANOSECONDS_OFST 8
+/* Timestamp minor value */
+#define MC_CMD_PTP_OUT_SYNCHRONIZE_MINOR_OFST 8
/* Host time immediately after NIC's hardware clock read */
#define MC_CMD_PTP_OUT_SYNCHRONIZE_HOSTEND_OFST 12
/* Number of nanoseconds waited after reading NIC's hardware clock */
@@ -1177,6 +1310,16 @@
#define MC_CMD_PTP_MANF_PACKET_ENOUGH 0x8
/* enum: Timestamp trigger GPIO not working */
#define MC_CMD_PTP_MANF_GPIO_TRIGGER 0x9
+/* enum: Insufficient PPS events to perform checks */
+#define MC_CMD_PTP_MANF_PPS_ENOUGH 0xa
+/* enum: PPS time event period not sufficiently close to 1s. */
+#define MC_CMD_PTP_MANF_PPS_PERIOD 0xb
+/* enum: PPS time event nS reading not sufficiently close to zero. */
+#define MC_CMD_PTP_MANF_PPS_NS 0xc
+/* enum: PTP peripheral registers incorrect */
+#define MC_CMD_PTP_MANF_REGISTERS 0xd
+/* enum: Failed to read time from PTP peripheral */
+#define MC_CMD_PTP_MANF_CLOCK_READ 0xe
/* Presence of external oscillator */
#define MC_CMD_PTP_OUT_MANFTEST_BASIC_TEST_EXTOSC_OFST 4
@@ -1198,6 +1341,62 @@
#define MC_CMD_PTP_OUT_FPGAREAD_BUFFER_MINNUM 1
#define MC_CMD_PTP_OUT_FPGAREAD_BUFFER_MAXNUM 252
+/* MC_CMD_PTP_OUT_GET_TIME_FORMAT msgresponse */
+#define MC_CMD_PTP_OUT_GET_TIME_FORMAT_LEN 4
+/* Time format required/used by for this NIC. Applies to all PTP MCDI
+ * operations that pass times between the host and firmware. If this operation
+ * is not supported (older firmware) a format of seconds and nanoseconds should
+ * be assumed.
+ */
+#define MC_CMD_PTP_OUT_GET_TIME_FORMAT_FORMAT_OFST 0
+/* enum: Times are in seconds and nanoseconds */
+#define MC_CMD_PTP_OUT_GET_TIME_FORMAT_SECONDS_NANOSECONDS 0x0
+/* enum: Major register has units of 16 second per tick, minor 8 ns per tick */
+#define MC_CMD_PTP_OUT_GET_TIME_FORMAT_16SECONDS_8NANOSECONDS 0x1
+/* enum: Major register has units of seconds, minor 2^-27s per tick */
+#define MC_CMD_PTP_OUT_GET_TIME_FORMAT_SECONDS_27FRACTION 0x2
+
+/* MC_CMD_PTP_OUT_GET_ATTRIBUTES msgresponse */
+#define MC_CMD_PTP_OUT_GET_ATTRIBUTES_LEN 8
+/* Time format required/used by for this NIC. Applies to all PTP MCDI
+ * operations that pass times between the host and firmware. If this operation
+ * is not supported (older firmware) a format of seconds and nanoseconds should
+ * be assumed.
+ */
+#define MC_CMD_PTP_OUT_GET_ATTRIBUTES_TIME_FORMAT_OFST 0
+/* enum: Times are in seconds and nanoseconds */
+#define MC_CMD_PTP_OUT_GET_ATTRIBUTES_SECONDS_NANOSECONDS 0x0
+/* enum: Major register has units of 16 second per tick, minor 8 ns per tick */
+#define MC_CMD_PTP_OUT_GET_ATTRIBUTES_16SECONDS_8NANOSECONDS 0x1
+/* enum: Major register has units of seconds, minor 2^-27s per tick */
+#define MC_CMD_PTP_OUT_GET_ATTRIBUTES_SECONDS_27FRACTION 0x2
+/* Minimum acceptable value for a corrected synchronization timeset. When
+ * comparing host and NIC clock times, the MC returns a set of samples that
+ * contain the host start and end time, the MC time when the host start was
+ * detected and the time the MC waited between reading the time and detecting
+ * the host end. The corrected sync window is the difference between the host
+ * end and start times minus the time that the MC waited for host end.
+ */
+#define MC_CMD_PTP_OUT_GET_ATTRIBUTES_SYNC_WINDOW_MIN_OFST 4
+
+/* MC_CMD_PTP_OUT_GET_TIMESTAMP_CORRECTIONS msgresponse */
+#define MC_CMD_PTP_OUT_GET_TIMESTAMP_CORRECTIONS_LEN 16
+/* Uncorrected error on transmit timestamps in NIC clock format */
+#define MC_CMD_PTP_OUT_GET_TIMESTAMP_CORRECTIONS_TRANSMIT_OFST 0
+/* Uncorrected error on receive timestamps in NIC clock format */
+#define MC_CMD_PTP_OUT_GET_TIMESTAMP_CORRECTIONS_RECEIVE_OFST 4
+/* Uncorrected error on PPS output in NIC clock format */
+#define MC_CMD_PTP_OUT_GET_TIMESTAMP_CORRECTIONS_PPS_OUT_OFST 8
+/* Uncorrected error on PPS input in NIC clock format */
+#define MC_CMD_PTP_OUT_GET_TIMESTAMP_CORRECTIONS_PPS_IN_OFST 12
+
+/* MC_CMD_PTP_OUT_MANFTEST_PPS msgresponse */
+#define MC_CMD_PTP_OUT_MANFTEST_PPS_LEN 4
+/* Results of testing */
+#define MC_CMD_PTP_OUT_MANFTEST_PPS_TEST_RESULT_OFST 0
+/* Enum values, see field(s): */
+/* MC_CMD_PTP_OUT_MANFTEST_BASIC/TEST_RESULT */
+
/***********************************/
/* MC_CMD_CSR_READ32
@@ -1923,6 +2122,8 @@
#define MC_CMD_MEDIA_SFP_PLUS 0x5
/* enum: 10GBaseT. */
#define MC_CMD_MEDIA_BASE_T 0x6
+/* enum: QSFP+. */
+#define MC_CMD_MEDIA_QSFP_PLUS 0x7
#define MC_CMD_GET_PHY_CFG_OUT_MMD_MASK_OFST 48
/* enum: Native clause 22 */
#define MC_CMD_MMD_CLAUSE22 0x0
@@ -2223,6 +2424,8 @@
#define MC_CMD_LOOPBACK_SD_FEP_WS 0x21
/* enum: KR Serdes Serial Wireside. */
#define MC_CMD_LOOPBACK_SD_FES_WS 0x22
+/* enum: Near side of AOE Siena side port */
+#define MC_CMD_LOOPBACK_AOE_INT_NEAR 0x23
/* Supported loopbacks. */
#define MC_CMD_GET_LOOPBACK_MODES_OUT_1G_OFST 8
#define MC_CMD_GET_LOOPBACK_MODES_OUT_1G_LEN 8
@@ -2286,6 +2489,10 @@
#define MC_CMD_GET_LINK_OUT_BPX_LINK_WIDTH 1
#define MC_CMD_GET_LINK_OUT_PHY_LINK_LBN 3
#define MC_CMD_GET_LINK_OUT_PHY_LINK_WIDTH 1
+#define MC_CMD_GET_LINK_OUT_LINK_FAULT_RX_LBN 6
+#define MC_CMD_GET_LINK_OUT_LINK_FAULT_RX_WIDTH 1
+#define MC_CMD_GET_LINK_OUT_LINK_FAULT_TX_LBN 7
+#define MC_CMD_GET_LINK_OUT_LINK_FAULT_TX_WIDTH 1
/* This returns the negotiated flow control value. */
#define MC_CMD_GET_LINK_OUT_FCNTL_OFST 20
/* enum: Flow control is off. */
@@ -3175,7 +3382,7 @@
#define MC_CMD_SENSOR_INFO_EXT_IN_PAGE_OFST 0
/* MC_CMD_SENSOR_INFO_OUT msgresponse */
-#define MC_CMD_SENSOR_INFO_OUT_LENMIN 12
+#define MC_CMD_SENSOR_INFO_OUT_LENMIN 4
#define MC_CMD_SENSOR_INFO_OUT_LENMAX 252
#define MC_CMD_SENSOR_INFO_OUT_LEN(num) (4+8*(num))
#define MC_CMD_SENSOR_INFO_OUT_MASK_OFST 0
@@ -3269,16 +3476,18 @@
#define MC_CMD_SENSOR_VDD08D_VSS08D_CSR 0x2b
/* enum: voltage between VSS08D and VSS08D at CSR (external ADC): mV */
#define MC_CMD_SENSOR_VDD08D_VSS08D_CSR_EXTADC 0x2c
+/* enum: Hotpoint temperature: degC */
+#define MC_CMD_SENSOR_HOTPOINT_TEMP 0x2d
/* MC_CMD_SENSOR_INFO_ENTRY_TYPEDEF */
#define MC_CMD_SENSOR_ENTRY_OFST 4
#define MC_CMD_SENSOR_ENTRY_LEN 8
#define MC_CMD_SENSOR_ENTRY_LO_OFST 4
#define MC_CMD_SENSOR_ENTRY_HI_OFST 8
-#define MC_CMD_SENSOR_ENTRY_MINNUM 1
+#define MC_CMD_SENSOR_ENTRY_MINNUM 0
#define MC_CMD_SENSOR_ENTRY_MAXNUM 31
/* MC_CMD_SENSOR_INFO_EXT_OUT msgresponse */
-#define MC_CMD_SENSOR_INFO_EXT_OUT_LENMIN 12
+#define MC_CMD_SENSOR_INFO_EXT_OUT_LENMIN 4
#define MC_CMD_SENSOR_INFO_EXT_OUT_LENMAX 252
#define MC_CMD_SENSOR_INFO_EXT_OUT_LEN(num) (4+8*(num))
#define MC_CMD_SENSOR_INFO_EXT_OUT_MASK_OFST 0
@@ -3291,7 +3500,7 @@
/* MC_CMD_SENSOR_ENTRY_LEN 8 */
/* MC_CMD_SENSOR_ENTRY_LO_OFST 4 */
/* MC_CMD_SENSOR_ENTRY_HI_OFST 8 */
-/* MC_CMD_SENSOR_ENTRY_MINNUM 1 */
+/* MC_CMD_SENSOR_ENTRY_MINNUM 0 */
/* MC_CMD_SENSOR_ENTRY_MAXNUM 31 */
/* MC_CMD_SENSOR_INFO_ENTRY_TYPEDEF structuredef */
@@ -3864,6 +4073,18 @@
#define NVRAM_PARTITION_TYPE_ID_LBN 0
#define NVRAM_PARTITION_TYPE_ID_WIDTH 16
+/* LICENSED_APP_ID structuredef */
+#define LICENSED_APP_ID_LEN 4
+#define LICENSED_APP_ID_ID_OFST 0
+/* enum: OpenOnload */
+#define LICENSED_APP_ID_ONLOAD 0x1
+/* enum: PTP timestamping */
+#define LICENSED_APP_ID_PTP 0x2
+/* enum: SolarCapture Pro */
+#define LICENSED_APP_ID_SOLARCAPTURE_PRO 0x4
+#define LICENSED_APP_ID_ID_LBN 0
+#define LICENSED_APP_ID_ID_WIDTH 32
+
/***********************************/
/* MC_CMD_READ_REGS
@@ -4021,6 +4242,8 @@
#define MC_CMD_INIT_RXQ_IN_FLAG_CHAIN_WIDTH 1
#define MC_CMD_INIT_RXQ_IN_FLAG_PREFIX_LBN 8
#define MC_CMD_INIT_RXQ_IN_FLAG_PREFIX_WIDTH 1
+#define MC_CMD_INIT_RXQ_IN_FLAG_DISABLE_SCATTER_LBN 9
+#define MC_CMD_INIT_RXQ_IN_FLAG_DISABLE_SCATTER_WIDTH 1
/* Owner ID to use if in buffer mode (zero if physical) */
#define MC_CMD_INIT_RXQ_IN_OWNER_ID_OFST 20
/* The port ID associated with the v-adaptor which should contain this DMAQ. */
@@ -4179,6 +4402,9 @@
#define MC_CMD_PROXY_CMD_IN_TARGET_VF_WIDTH 16
#define MC_CMD_PROXY_CMD_IN_VF_NULL 0xffff /* enum */
+/* MC_CMD_PROXY_CMD_OUT msgresponse */
+#define MC_CMD_PROXY_CMD_OUT_LEN 0
+
/***********************************/
/* MC_CMD_ALLOC_BUFTBL_CHUNK
@@ -4213,7 +4439,7 @@
/* MC_CMD_PROGRAM_BUFTBL_ENTRIES_IN msgrequest */
#define MC_CMD_PROGRAM_BUFTBL_ENTRIES_IN_LENMIN 20
-#define MC_CMD_PROGRAM_BUFTBL_ENTRIES_IN_LENMAX 252
+#define MC_CMD_PROGRAM_BUFTBL_ENTRIES_IN_LENMAX 268
#define MC_CMD_PROGRAM_BUFTBL_ENTRIES_IN_LEN(num) (12+8*(num))
#define MC_CMD_PROGRAM_BUFTBL_ENTRIES_IN_HANDLE_OFST 0
/* ID */
@@ -4226,7 +4452,7 @@
#define MC_CMD_PROGRAM_BUFTBL_ENTRIES_IN_ENTRY_LO_OFST 12
#define MC_CMD_PROGRAM_BUFTBL_ENTRIES_IN_ENTRY_HI_OFST 16
#define MC_CMD_PROGRAM_BUFTBL_ENTRIES_IN_ENTRY_MINNUM 1
-#define MC_CMD_PROGRAM_BUFTBL_ENTRIES_IN_ENTRY_MAXNUM 30
+#define MC_CMD_PROGRAM_BUFTBL_ENTRIES_IN_ENTRY_MAXNUM 32
/* MC_CMD_PROGRAM_BUFTBL_ENTRIES_OUT msgresponse */
#define MC_CMD_PROGRAM_BUFTBL_ENTRIES_OUT_LEN 0
@@ -6800,6 +7026,30 @@
/***********************************/
+/* MC_CMD_CAP_BLK_READ
+ * Read multiple 64bit words from capture block memory
+ */
+#define MC_CMD_CAP_BLK_READ 0xe7
+
+/* MC_CMD_CAP_BLK_READ_IN msgrequest */
+#define MC_CMD_CAP_BLK_READ_IN_LEN 12
+#define MC_CMD_CAP_BLK_READ_IN_CAP_REG_OFST 0
+#define MC_CMD_CAP_BLK_READ_IN_ADDR_OFST 4
+#define MC_CMD_CAP_BLK_READ_IN_COUNT_OFST 8
+
+/* MC_CMD_CAP_BLK_READ_OUT msgresponse */
+#define MC_CMD_CAP_BLK_READ_OUT_LENMIN 8
+#define MC_CMD_CAP_BLK_READ_OUT_LENMAX 248
+#define MC_CMD_CAP_BLK_READ_OUT_LEN(num) (0+8*(num))
+#define MC_CMD_CAP_BLK_READ_OUT_BUFFER_OFST 0
+#define MC_CMD_CAP_BLK_READ_OUT_BUFFER_LEN 8
+#define MC_CMD_CAP_BLK_READ_OUT_BUFFER_LO_OFST 0
+#define MC_CMD_CAP_BLK_READ_OUT_BUFFER_HI_OFST 4
+#define MC_CMD_CAP_BLK_READ_OUT_BUFFER_MINNUM 1
+#define MC_CMD_CAP_BLK_READ_OUT_BUFFER_MAXNUM 31
+
+
+/***********************************/
/* MC_CMD_DUMP_DO
* Take a dump of the DUT state
*/
@@ -6826,6 +7076,10 @@
#define MC_CMD_DUMP_DO_IN_DUMPSPEC_SRC_CUSTOM_HOST_MEMORY_MLI_DEPTH_OFST 20
#define MC_CMD_DUMP_DO_IN_HOST_MEMORY_MLI_MAX_DEPTH 0x2 /* enum */
#define MC_CMD_DUMP_DO_IN_DUMPSPEC_SRC_CUSTOM_UART_PORT_OFST 12
+/* enum: The uart port this command was received over (if using a uart
+ * transport)
+ */
+#define MC_CMD_DUMP_DO_IN_UART_PORT_SRC 0xff
#define MC_CMD_DUMP_DO_IN_DUMPSPEC_SRC_CUSTOM_SIZE_OFST 24
#define MC_CMD_DUMP_DO_IN_DUMPFILE_DST_OFST 28
#define MC_CMD_DUMP_DO_IN_DUMPFILE_DST_CUSTOM 0x0 /* enum */
@@ -6942,39 +7196,68 @@
/***********************************/
-/* MC_CMD_START_KR_EYE_PLOT
- * Start KR Serdes Eye diagram plot on a given lane. Lane must have valid
- * signal.
- */
-#define MC_CMD_START_KR_EYE_PLOT 0xee
-
-/* MC_CMD_START_KR_EYE_PLOT_IN msgrequest */
-#define MC_CMD_START_KR_EYE_PLOT_IN_LEN 4
-#define MC_CMD_START_KR_EYE_PLOT_IN_LANE_OFST 0
-
-/* MC_CMD_START_KR_EYE_PLOT_OUT msgresponse */
-#define MC_CMD_START_KR_EYE_PLOT_OUT_LEN 0
-
-
-/***********************************/
-/* MC_CMD_POLL_KR_EYE_PLOT
- * Poll KR Serdes Eye diagram plot. Returns one row of BER data. The caller
- * should call this command repeatedly after starting eye plot, until no more
- * data is returned.
- */
-#define MC_CMD_POLL_KR_EYE_PLOT 0xef
-
-/* MC_CMD_POLL_KR_EYE_PLOT_IN msgrequest */
-#define MC_CMD_POLL_KR_EYE_PLOT_IN_LEN 0
-
-/* MC_CMD_POLL_KR_EYE_PLOT_OUT msgresponse */
-#define MC_CMD_POLL_KR_EYE_PLOT_OUT_LENMIN 0
-#define MC_CMD_POLL_KR_EYE_PLOT_OUT_LENMAX 252
-#define MC_CMD_POLL_KR_EYE_PLOT_OUT_LEN(num) (0+2*(num))
-#define MC_CMD_POLL_KR_EYE_PLOT_OUT_SAMPLES_OFST 0
-#define MC_CMD_POLL_KR_EYE_PLOT_OUT_SAMPLES_LEN 2
-#define MC_CMD_POLL_KR_EYE_PLOT_OUT_SAMPLES_MINNUM 0
-#define MC_CMD_POLL_KR_EYE_PLOT_OUT_SAMPLES_MAXNUM 126
+/* MC_CMD_UART_SEND_DATA
+ * Send checksummed[sic] block of data over the uart. Response is a placeholder
+ * should we wish to make this reliable; currently requests are fire-and-
+ * forget.
+ */
+#define MC_CMD_UART_SEND_DATA 0xee
+
+/* MC_CMD_UART_SEND_DATA_OUT msgrequest */
+#define MC_CMD_UART_SEND_DATA_OUT_LENMIN 16
+#define MC_CMD_UART_SEND_DATA_OUT_LENMAX 252
+#define MC_CMD_UART_SEND_DATA_OUT_LEN(num) (16+1*(num))
+/* CRC32 over OFFSET, LENGTH, RESERVED, DATA */
+#define MC_CMD_UART_SEND_DATA_OUT_CHECKSUM_OFST 0
+/* Offset at which to write the data */
+#define MC_CMD_UART_SEND_DATA_OUT_OFFSET_OFST 4
+/* Length of data */
+#define MC_CMD_UART_SEND_DATA_OUT_LENGTH_OFST 8
+/* Reserved for future use */
+#define MC_CMD_UART_SEND_DATA_OUT_RESERVED_OFST 12
+#define MC_CMD_UART_SEND_DATA_OUT_DATA_OFST 16
+#define MC_CMD_UART_SEND_DATA_OUT_DATA_LEN 1
+#define MC_CMD_UART_SEND_DATA_OUT_DATA_MINNUM 0
+#define MC_CMD_UART_SEND_DATA_OUT_DATA_MAXNUM 236
+
+/* MC_CMD_UART_SEND_DATA_IN msgresponse */
+#define MC_CMD_UART_SEND_DATA_IN_LEN 0
+
+
+/***********************************/
+/* MC_CMD_UART_RECV_DATA
+ * Request checksummed[sic] block of data over the uart. Only a placeholder,
+ * subject to change and not currently implemented.
+ */
+#define MC_CMD_UART_RECV_DATA 0xef
+
+/* MC_CMD_UART_RECV_DATA_OUT msgrequest */
+#define MC_CMD_UART_RECV_DATA_OUT_LEN 16
+/* CRC32 over OFFSET, LENGTH, RESERVED */
+#define MC_CMD_UART_RECV_DATA_OUT_CHECKSUM_OFST 0
+/* Offset from which to read the data */
+#define MC_CMD_UART_RECV_DATA_OUT_OFFSET_OFST 4
+/* Length of data */
+#define MC_CMD_UART_RECV_DATA_OUT_LENGTH_OFST 8
+/* Reserved for future use */
+#define MC_CMD_UART_RECV_DATA_OUT_RESERVED_OFST 12
+
+/* MC_CMD_UART_RECV_DATA_IN msgresponse */
+#define MC_CMD_UART_RECV_DATA_IN_LENMIN 16
+#define MC_CMD_UART_RECV_DATA_IN_LENMAX 252
+#define MC_CMD_UART_RECV_DATA_IN_LEN(num) (16+1*(num))
+/* CRC32 over RESERVED1, RESERVED2, RESERVED3, DATA */
+#define MC_CMD_UART_RECV_DATA_IN_CHECKSUM_OFST 0
+/* Offset at which to write the data */
+#define MC_CMD_UART_RECV_DATA_IN_RESERVED1_OFST 4
+/* Length of data */
+#define MC_CMD_UART_RECV_DATA_IN_RESERVED2_OFST 8
+/* Reserved for future use */
+#define MC_CMD_UART_RECV_DATA_IN_RESERVED3_OFST 12
+#define MC_CMD_UART_RECV_DATA_IN_DATA_OFST 16
+#define MC_CMD_UART_RECV_DATA_IN_DATA_LEN 1
+#define MC_CMD_UART_RECV_DATA_IN_DATA_MINNUM 0
+#define MC_CMD_UART_RECV_DATA_IN_DATA_MAXNUM 236
/***********************************/
@@ -7026,6 +7309,15 @@
#define MC_CMD_KR_TUNE_IN_TXEQ_SET 0x3
/* enum: Force KR Serdes reset / recalibration */
#define MC_CMD_KR_TUNE_IN_RECAL 0x4
+/* enum: Start KR Serdes Eye diagram plot on a given lane. Lane must have valid
+ * signal.
+ */
+#define MC_CMD_KR_TUNE_IN_START_EYE_PLOT 0x5
+/* enum: Poll KR Serdes Eye diagram plot. Returns one row of BER data. The
+ * caller should call this command repeatedly after starting eye plot, until no
+ * more data is returned.
+ */
+#define MC_CMD_KR_TUNE_IN_POLL_EYE_PLOT 0x6
/* Align the arguments to 32 bits */
#define MC_CMD_KR_TUNE_IN_KR_TUNE_RSVD_OFST 1
#define MC_CMD_KR_TUNE_IN_KR_TUNE_RSVD_LEN 3
@@ -7123,6 +7415,91 @@
/* MC_CMD_KR_TUNE_RXEQ_SET_OUT msgresponse */
#define MC_CMD_KR_TUNE_RXEQ_SET_OUT_LEN 0
+/* MC_CMD_KR_TUNE_TXEQ_GET_IN msgrequest */
+#define MC_CMD_KR_TUNE_TXEQ_GET_IN_LEN 4
+/* Requested operation */
+#define MC_CMD_KR_TUNE_TXEQ_GET_IN_KR_TUNE_OP_OFST 0
+#define MC_CMD_KR_TUNE_TXEQ_GET_IN_KR_TUNE_OP_LEN 1
+/* Align the arguments to 32 bits */
+#define MC_CMD_KR_TUNE_TXEQ_GET_IN_KR_TUNE_RSVD_OFST 1
+#define MC_CMD_KR_TUNE_TXEQ_GET_IN_KR_TUNE_RSVD_LEN 3
+
+/* MC_CMD_KR_TUNE_TXEQ_GET_OUT msgresponse */
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_LENMIN 4
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_LENMAX 252
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_LEN(num) (0+4*(num))
+/* TXEQ Parameter */
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_PARAM_OFST 0
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_PARAM_LEN 4
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_PARAM_MINNUM 1
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_PARAM_MAXNUM 63
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_PARAM_ID_LBN 0
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_PARAM_ID_WIDTH 8
+/* enum: TX Amplitude */
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_TX_LEV 0x0
+/* enum: De-Emphasis Tap1 Magnitude (0-7) */
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_TX_MODE 0x1
+/* enum: De-Emphasis Tap1 Fine */
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_TX_DTLEV 0x2
+/* enum: De-Emphasis Tap2 Magnitude (0-6) */
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_TX_D2 0x3
+/* enum: De-Emphasis Tap2 Fine */
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_TX_D2TLEV 0x4
+/* enum: Pre-Emphasis Magnitude */
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_TX_E 0x5
+/* enum: Pre-Emphasis Fine */
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_TX_ETLEV 0x6
+/* enum: TX Slew Rate Coarse control */
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_TX_PREDRV_DLY 0x7
+/* enum: TX Slew Rate Fine control */
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_TX_SR_SET 0x8
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_PARAM_LANE_LBN 8
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_PARAM_LANE_WIDTH 3
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_LANE_0 0x0 /* enum */
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_LANE_1 0x1 /* enum */
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_LANE_2 0x2 /* enum */
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_LANE_3 0x3 /* enum */
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_LANE_ALL 0x4 /* enum */
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_RESERVED_LBN 11
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_RESERVED_WIDTH 5
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_PARAM_INITIAL_LBN 16
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_PARAM_INITIAL_WIDTH 8
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_RESERVED2_LBN 24
+#define MC_CMD_KR_TUNE_TXEQ_GET_OUT_RESERVED2_WIDTH 8
+
+/* MC_CMD_KR_TUNE_TXEQ_SET_IN msgrequest */
+#define MC_CMD_KR_TUNE_TXEQ_SET_IN_LENMIN 8
+#define MC_CMD_KR_TUNE_TXEQ_SET_IN_LENMAX 252
+#define MC_CMD_KR_TUNE_TXEQ_SET_IN_LEN(num) (4+4*(num))
+/* Requested operation */
+#define MC_CMD_KR_TUNE_TXEQ_SET_IN_KR_TUNE_OP_OFST 0
+#define MC_CMD_KR_TUNE_TXEQ_SET_IN_KR_TUNE_OP_LEN 1
+/* Align the arguments to 32 bits */
+#define MC_CMD_KR_TUNE_TXEQ_SET_IN_KR_TUNE_RSVD_OFST 1
+#define MC_CMD_KR_TUNE_TXEQ_SET_IN_KR_TUNE_RSVD_LEN 3
+/* TXEQ Parameter */
+#define MC_CMD_KR_TUNE_TXEQ_SET_IN_PARAM_OFST 4
+#define MC_CMD_KR_TUNE_TXEQ_SET_IN_PARAM_LEN 4
+#define MC_CMD_KR_TUNE_TXEQ_SET_IN_PARAM_MINNUM 1
+#define MC_CMD_KR_TUNE_TXEQ_SET_IN_PARAM_MAXNUM 62
+#define MC_CMD_KR_TUNE_TXEQ_SET_IN_PARAM_ID_LBN 0
+#define MC_CMD_KR_TUNE_TXEQ_SET_IN_PARAM_ID_WIDTH 8
+/* Enum values, see field(s): */
+/* MC_CMD_KR_TUNE_TXEQ_GET_OUT/PARAM_ID */
+#define MC_CMD_KR_TUNE_TXEQ_SET_IN_PARAM_LANE_LBN 8
+#define MC_CMD_KR_TUNE_TXEQ_SET_IN_PARAM_LANE_WIDTH 3
+/* Enum values, see field(s): */
+/* MC_CMD_KR_TUNE_TXEQ_GET_OUT/PARAM_LANE */
+#define MC_CMD_KR_TUNE_TXEQ_SET_IN_RESERVED_LBN 11
+#define MC_CMD_KR_TUNE_TXEQ_SET_IN_RESERVED_WIDTH 5
+#define MC_CMD_KR_TUNE_TXEQ_SET_IN_PARAM_INITIAL_LBN 16
+#define MC_CMD_KR_TUNE_TXEQ_SET_IN_PARAM_INITIAL_WIDTH 8
+#define MC_CMD_KR_TUNE_TXEQ_SET_IN_RESERVED2_LBN 24
+#define MC_CMD_KR_TUNE_TXEQ_SET_IN_RESERVED2_WIDTH 8
+
+/* MC_CMD_KR_TUNE_TXEQ_SET_OUT msgresponse */
+#define MC_CMD_KR_TUNE_TXEQ_SET_OUT_LEN 0
+
/* MC_CMD_KR_TUNE_RECAL_IN msgrequest */
#define MC_CMD_KR_TUNE_RECAL_IN_LEN 4
/* Requested operation */
@@ -7135,6 +7512,37 @@
/* MC_CMD_KR_TUNE_RECAL_OUT msgresponse */
#define MC_CMD_KR_TUNE_RECAL_OUT_LEN 0
+/* MC_CMD_KR_TUNE_START_EYE_PLOT_IN msgrequest */
+#define MC_CMD_KR_TUNE_START_EYE_PLOT_IN_LEN 8
+/* Requested operation */
+#define MC_CMD_KR_TUNE_START_EYE_PLOT_IN_KR_TUNE_OP_OFST 0
+#define MC_CMD_KR_TUNE_START_EYE_PLOT_IN_KR_TUNE_OP_LEN 1
+/* Align the arguments to 32 bits */
+#define MC_CMD_KR_TUNE_START_EYE_PLOT_IN_KR_TUNE_RSVD_OFST 1
+#define MC_CMD_KR_TUNE_START_EYE_PLOT_IN_KR_TUNE_RSVD_LEN 3
+#define MC_CMD_KR_TUNE_START_EYE_PLOT_IN_LANE_OFST 4
+
+/* MC_CMD_KR_TUNE_START_EYE_PLOT_OUT msgresponse */
+#define MC_CMD_KR_TUNE_START_EYE_PLOT_OUT_LEN 0
+
+/* MC_CMD_KR_TUNE_POLL_EYE_PLOT_IN msgrequest */
+#define MC_CMD_KR_TUNE_POLL_EYE_PLOT_IN_LEN 4
+/* Requested operation */
+#define MC_CMD_KR_TUNE_POLL_EYE_PLOT_IN_KR_TUNE_OP_OFST 0
+#define MC_CMD_KR_TUNE_POLL_EYE_PLOT_IN_KR_TUNE_OP_LEN 1
+/* Align the arguments to 32 bits */
+#define MC_CMD_KR_TUNE_POLL_EYE_PLOT_IN_KR_TUNE_RSVD_OFST 1
+#define MC_CMD_KR_TUNE_POLL_EYE_PLOT_IN_KR_TUNE_RSVD_LEN 3
+
+/* MC_CMD_KR_TUNE_POLL_EYE_PLOT_OUT msgresponse */
+#define MC_CMD_KR_TUNE_POLL_EYE_PLOT_OUT_LENMIN 0
+#define MC_CMD_KR_TUNE_POLL_EYE_PLOT_OUT_LENMAX 252
+#define MC_CMD_KR_TUNE_POLL_EYE_PLOT_OUT_LEN(num) (0+2*(num))
+#define MC_CMD_KR_TUNE_POLL_EYE_PLOT_OUT_SAMPLES_OFST 0
+#define MC_CMD_KR_TUNE_POLL_EYE_PLOT_OUT_SAMPLES_LEN 2
+#define MC_CMD_KR_TUNE_POLL_EYE_PLOT_OUT_SAMPLES_MINNUM 0
+#define MC_CMD_KR_TUNE_POLL_EYE_PLOT_OUT_SAMPLES_MAXNUM 126
+
/***********************************/
/* MC_CMD_PCIE_TUNE
@@ -7157,6 +7565,13 @@
#define MC_CMD_PCIE_TUNE_IN_TXEQ_GET 0x2
/* enum: Override TX Driver settings */
#define MC_CMD_PCIE_TUNE_IN_TXEQ_SET 0x3
+/* enum: Start PCIe Serdes Eye diagram plot on a given lane. */
+#define MC_CMD_PCIE_TUNE_IN_START_EYE_PLOT 0x5
+/* enum: Poll PCIe Serdes Eye diagram plot. Returns one row of BER data. The
+ * caller should call this command repeatedly after starting eye plot, until no
+ * more data is returned.
+ */
+#define MC_CMD_PCIE_TUNE_IN_POLL_EYE_PLOT 0x6
/* Align the arguments to 32 bits */
#define MC_CMD_PCIE_TUNE_IN_PCIE_TUNE_RSVD_OFST 1
#define MC_CMD_PCIE_TUNE_IN_PCIE_TUNE_RSVD_LEN 3
@@ -7258,6 +7673,37 @@
#define MC_CMD_PCIE_TUNE_TXEQ_GET_OUT_PARAM_CURRENT_LBN 24
#define MC_CMD_PCIE_TUNE_TXEQ_GET_OUT_PARAM_CURRENT_WIDTH 8
+/* MC_CMD_PCIE_TUNE_START_EYE_PLOT_IN msgrequest */
+#define MC_CMD_PCIE_TUNE_START_EYE_PLOT_IN_LEN 8
+/* Requested operation */
+#define MC_CMD_PCIE_TUNE_START_EYE_PLOT_IN_PCIE_TUNE_OP_OFST 0
+#define MC_CMD_PCIE_TUNE_START_EYE_PLOT_IN_PCIE_TUNE_OP_LEN 1
+/* Align the arguments to 32 bits */
+#define MC_CMD_PCIE_TUNE_START_EYE_PLOT_IN_PCIE_TUNE_RSVD_OFST 1
+#define MC_CMD_PCIE_TUNE_START_EYE_PLOT_IN_PCIE_TUNE_RSVD_LEN 3
+#define MC_CMD_PCIE_TUNE_START_EYE_PLOT_IN_LANE_OFST 4
+
+/* MC_CMD_PCIE_TUNE_START_EYE_PLOT_OUT msgresponse */
+#define MC_CMD_PCIE_TUNE_START_EYE_PLOT_OUT_LEN 0
+
+/* MC_CMD_PCIE_TUNE_POLL_EYE_PLOT_IN msgrequest */
+#define MC_CMD_PCIE_TUNE_POLL_EYE_PLOT_IN_LEN 4
+/* Requested operation */
+#define MC_CMD_PCIE_TUNE_POLL_EYE_PLOT_IN_PCIE_TUNE_OP_OFST 0
+#define MC_CMD_PCIE_TUNE_POLL_EYE_PLOT_IN_PCIE_TUNE_OP_LEN 1
+/* Align the arguments to 32 bits */
+#define MC_CMD_PCIE_TUNE_POLL_EYE_PLOT_IN_PCIE_TUNE_RSVD_OFST 1
+#define MC_CMD_PCIE_TUNE_POLL_EYE_PLOT_IN_PCIE_TUNE_RSVD_LEN 3
+
+/* MC_CMD_PCIE_TUNE_POLL_EYE_PLOT_OUT msgresponse */
+#define MC_CMD_PCIE_TUNE_POLL_EYE_PLOT_OUT_LENMIN 0
+#define MC_CMD_PCIE_TUNE_POLL_EYE_PLOT_OUT_LENMAX 252
+#define MC_CMD_PCIE_TUNE_POLL_EYE_PLOT_OUT_LEN(num) (0+2*(num))
+#define MC_CMD_PCIE_TUNE_POLL_EYE_PLOT_OUT_SAMPLES_OFST 0
+#define MC_CMD_PCIE_TUNE_POLL_EYE_PLOT_OUT_SAMPLES_LEN 2
+#define MC_CMD_PCIE_TUNE_POLL_EYE_PLOT_OUT_SAMPLES_MINNUM 0
+#define MC_CMD_PCIE_TUNE_POLL_EYE_PLOT_OUT_SAMPLES_MAXNUM 126
+
/***********************************/
/* MC_CMD_LICENSING
@@ -7310,5 +7756,152 @@
*/
#define MC_CMD_MC2MC_PROXY 0xf4
+/* MC_CMD_MC2MC_PROXY_IN msgrequest */
+#define MC_CMD_MC2MC_PROXY_IN_LEN 0
+
+/* MC_CMD_MC2MC_PROXY_OUT msgresponse */
+#define MC_CMD_MC2MC_PROXY_OUT_LEN 0
+
+
+/***********************************/
+/* MC_CMD_GET_LICENSED_APP_STATE
+ * Query the state of an individual licensed application. (Note that the actual
+ * state may be invalidated by the MC_CMD_LICENSING OP_UPDATE_LICENSE operation
+ * or a reboot of the MC.)
+ */
+#define MC_CMD_GET_LICENSED_APP_STATE 0xf5
+
+/* MC_CMD_GET_LICENSED_APP_STATE_IN msgrequest */
+#define MC_CMD_GET_LICENSED_APP_STATE_IN_LEN 4
+/* application ID to query (LICENSED_APP_ID_xxx) */
+#define MC_CMD_GET_LICENSED_APP_STATE_IN_APP_ID_OFST 0
+
+/* MC_CMD_GET_LICENSED_APP_STATE_OUT msgresponse */
+#define MC_CMD_GET_LICENSED_APP_STATE_OUT_LEN 4
+/* state of this application */
+#define MC_CMD_GET_LICENSED_APP_STATE_OUT_STATE_OFST 0
+/* enum: no (or invalid) license is present for the application */
+#define MC_CMD_GET_LICENSED_APP_STATE_OUT_NOT_LICENSED 0x0
+/* enum: a valid license is present for the application */
+#define MC_CMD_GET_LICENSED_APP_STATE_OUT_LICENSED 0x1
+
+
+/***********************************/
+/* MC_CMD_LICENSED_APP_OP
+ * Perform an action for an individual licensed application.
+ */
+#define MC_CMD_LICENSED_APP_OP 0xf6
+
+/* MC_CMD_LICENSED_APP_OP_IN msgrequest */
+#define MC_CMD_LICENSED_APP_OP_IN_LENMIN 8
+#define MC_CMD_LICENSED_APP_OP_IN_LENMAX 252
+#define MC_CMD_LICENSED_APP_OP_IN_LEN(num) (8+4*(num))
+/* application ID */
+#define MC_CMD_LICENSED_APP_OP_IN_APP_ID_OFST 0
+/* the type of operation requested */
+#define MC_CMD_LICENSED_APP_OP_IN_OP_OFST 4
+/* enum: validate application */
+#define MC_CMD_LICENSED_APP_OP_IN_OP_VALIDATE 0x0
+/* arguments specific to this particular operation */
+#define MC_CMD_LICENSED_APP_OP_IN_ARGS_OFST 8
+#define MC_CMD_LICENSED_APP_OP_IN_ARGS_LEN 4
+#define MC_CMD_LICENSED_APP_OP_IN_ARGS_MINNUM 0
+#define MC_CMD_LICENSED_APP_OP_IN_ARGS_MAXNUM 61
+
+/* MC_CMD_LICENSED_APP_OP_OUT msgresponse */
+#define MC_CMD_LICENSED_APP_OP_OUT_LENMIN 0
+#define MC_CMD_LICENSED_APP_OP_OUT_LENMAX 252
+#define MC_CMD_LICENSED_APP_OP_OUT_LEN(num) (0+4*(num))
+/* result specific to this particular operation */
+#define MC_CMD_LICENSED_APP_OP_OUT_RESULT_OFST 0
+#define MC_CMD_LICENSED_APP_OP_OUT_RESULT_LEN 4
+#define MC_CMD_LICENSED_APP_OP_OUT_RESULT_MINNUM 0
+#define MC_CMD_LICENSED_APP_OP_OUT_RESULT_MAXNUM 63
+
+/* MC_CMD_LICENSED_APP_OP_VALIDATE_IN msgrequest */
+#define MC_CMD_LICENSED_APP_OP_VALIDATE_IN_LEN 72
+/* application ID */
+#define MC_CMD_LICENSED_APP_OP_VALIDATE_IN_APP_ID_OFST 0
+/* the type of operation requested */
+#define MC_CMD_LICENSED_APP_OP_VALIDATE_IN_OP_OFST 4
+/* validation challenge */
+#define MC_CMD_LICENSED_APP_OP_VALIDATE_IN_CHALLENGE_OFST 8
+#define MC_CMD_LICENSED_APP_OP_VALIDATE_IN_CHALLENGE_LEN 64
+
+/* MC_CMD_LICENSED_APP_OP_VALIDATE_OUT msgresponse */
+#define MC_CMD_LICENSED_APP_OP_VALIDATE_OUT_LEN 68
+/* feature expiry (time_t) */
+#define MC_CMD_LICENSED_APP_OP_VALIDATE_OUT_EXPIRY_OFST 0
+/* validation response */
+#define MC_CMD_LICENSED_APP_OP_VALIDATE_OUT_RESPONSE_OFST 4
+#define MC_CMD_LICENSED_APP_OP_VALIDATE_OUT_RESPONSE_LEN 64
+
+
+/***********************************/
+/* MC_CMD_SET_PORT_SNIFF_CONFIG
+ * Configure port sniffing for the physical port associated with the calling
+ * function. Only a privileged function may change the port sniffing
+ * configuration. A copy of all traffic delivered to the host (non-promiscuous
+ * mode) or all traffic arriving at the port (promiscuous mode) may be
+ * delivered to a specific queue, or a set of queues with RSS.
+ */
+#define MC_CMD_SET_PORT_SNIFF_CONFIG 0xf7
+
+/* MC_CMD_SET_PORT_SNIFF_CONFIG_IN msgrequest */
+#define MC_CMD_SET_PORT_SNIFF_CONFIG_IN_LEN 16
+/* configuration flags */
+#define MC_CMD_SET_PORT_SNIFF_CONFIG_IN_FLAGS_OFST 0
+#define MC_CMD_SET_PORT_SNIFF_CONFIG_IN_ENABLE_LBN 0
+#define MC_CMD_SET_PORT_SNIFF_CONFIG_IN_ENABLE_WIDTH 1
+#define MC_CMD_SET_PORT_SNIFF_CONFIG_IN_PROMISCUOUS_LBN 1
+#define MC_CMD_SET_PORT_SNIFF_CONFIG_IN_PROMISCUOUS_WIDTH 1
+/* receive queue handle (for RSS mode, this is the base queue) */
+#define MC_CMD_SET_PORT_SNIFF_CONFIG_IN_RX_QUEUE_OFST 4
+/* receive mode */
+#define MC_CMD_SET_PORT_SNIFF_CONFIG_IN_RX_MODE_OFST 8
+/* enum: receive to just the specified queue */
+#define MC_CMD_SET_PORT_SNIFF_CONFIG_IN_RX_MODE_SIMPLE 0x0
+/* enum: receive to multiple queues using RSS context */
+#define MC_CMD_SET_PORT_SNIFF_CONFIG_IN_RX_MODE_RSS 0x1
+/* RSS context (for RX_MODE_RSS) as returned by MC_CMD_RSS_CONTEXT_ALLOC. Note
+ * that these handles should be considered opaque to the host, although a value
+ * of 0xFFFFFFFF is guaranteed never to be a valid handle.
+ */
+#define MC_CMD_SET_PORT_SNIFF_CONFIG_IN_RX_CONTEXT_OFST 12
+
+/* MC_CMD_SET_PORT_SNIFF_CONFIG_OUT msgresponse */
+#define MC_CMD_SET_PORT_SNIFF_CONFIG_OUT_LEN 0
+
+
+/***********************************/
+/* MC_CMD_GET_PORT_SNIFF_CONFIG
+ * Obtain the current port sniffing configuration for the physical port
+ * associated with the calling function. Only a privileged function may read
+ * the configuration.
+ */
+#define MC_CMD_GET_PORT_SNIFF_CONFIG 0xf8
+
+/* MC_CMD_GET_PORT_SNIFF_CONFIG_IN msgrequest */
+#define MC_CMD_GET_PORT_SNIFF_CONFIG_IN_LEN 0
+
+/* MC_CMD_GET_PORT_SNIFF_CONFIG_OUT msgresponse */
+#define MC_CMD_GET_PORT_SNIFF_CONFIG_OUT_LEN 16
+/* configuration flags */
+#define MC_CMD_GET_PORT_SNIFF_CONFIG_OUT_FLAGS_OFST 0
+#define MC_CMD_GET_PORT_SNIFF_CONFIG_OUT_ENABLE_LBN 0
+#define MC_CMD_GET_PORT_SNIFF_CONFIG_OUT_ENABLE_WIDTH 1
+#define MC_CMD_GET_PORT_SNIFF_CONFIG_OUT_PROMISCUOUS_LBN 1
+#define MC_CMD_GET_PORT_SNIFF_CONFIG_OUT_PROMISCUOUS_WIDTH 1
+/* receiving queue handle (for RSS mode, this is the base queue) */
+#define MC_CMD_GET_PORT_SNIFF_CONFIG_OUT_RX_QUEUE_OFST 4
+/* receive mode */
+#define MC_CMD_GET_PORT_SNIFF_CONFIG_OUT_RX_MODE_OFST 8
+/* enum: receiving to just the specified queue */
+#define MC_CMD_GET_PORT_SNIFF_CONFIG_OUT_RX_MODE_SIMPLE 0x0
+/* enum: receiving to multiple queues using RSS context */
+#define MC_CMD_GET_PORT_SNIFF_CONFIG_OUT_RX_MODE_RSS 0x1
+/* RSS context (for RX_MODE_RSS) */
+#define MC_CMD_GET_PORT_SNIFF_CONFIG_OUT_RX_CONTEXT_OFST 12
+
#endif /* MCDI_PCOL_H */
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH net-next 03/16] sfc: Add MC BISTs to ethtool offline self test on EF10
2013-12-12 22:56 Pull request: sfc-next 2013-12-12 Ben Hutchings
2013-12-12 22:59 ` [PATCH net-next 01/16] sfc: Demote "MC Scheduler error" messages Ben Hutchings
2013-12-12 22:59 ` [PATCH net-next 02/16] sfc: Update MCDI protocol definitions Ben Hutchings
@ 2013-12-12 22:59 ` Ben Hutchings
2013-12-12 23:00 ` [PATCH net-next 04/16] sfc: Tighten the check for RX merged completion events Ben Hutchings
` (13 subsequent siblings)
16 siblings, 0 replies; 18+ messages in thread
From: Ben Hutchings @ 2013-12-12 22:59 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
From: Jon Cooper <jcooper@solarflare.com>
To run BISTs the MC goes down in to a special mode where it will only
respond to MCDI from the testing PF, and TX, RX and event queues are
torn down. Other PFs get a message as it goes down to tell them it's
going down.
When the other PFs get this message, they check the soft status
register to tell when the MC has rebooted after BIST mode and they can
start recovery.
[bwh: Convert the test result to 1 or -1 as for earlier NICs]
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/ef10.c | 84 ++++++++++++++++++++++++++++++++++-
drivers/net/ethernet/sfc/efx.c | 29 ++++++++++++
drivers/net/ethernet/sfc/enum.h | 1 +
drivers/net/ethernet/sfc/ethtool.c | 2 +
drivers/net/ethernet/sfc/mcdi.c | 34 ++++++++++++++
drivers/net/ethernet/sfc/net_driver.h | 1 +
drivers/net/ethernet/sfc/selftest.c | 2 +-
drivers/net/ethernet/sfc/selftest.h | 1 +
8 files changed, 152 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/sfc/ef10.c b/drivers/net/ethernet/sfc/ef10.c
index 676c3c057bfb..5d46d155b642 100644
--- a/drivers/net/ethernet/sfc/ef10.c
+++ b/drivers/net/ethernet/sfc/ef10.c
@@ -14,6 +14,7 @@
#include "mcdi_pcol.h"
#include "nic.h"
#include "workarounds.h"
+#include "selftest.h"
#include <linux/in.h>
#include <linux/jhash.h>
#include <linux/wait.h>
@@ -3195,6 +3196,87 @@ static int efx_ef10_mac_reconfigure(struct efx_nic *efx)
return efx_mcdi_set_mac(efx);
}
+static int efx_ef10_start_bist(struct efx_nic *efx, u32 bist_type)
+{
+ MCDI_DECLARE_BUF(inbuf, MC_CMD_START_BIST_IN_LEN);
+
+ MCDI_SET_DWORD(inbuf, START_BIST_IN_TYPE, bist_type);
+ return efx_mcdi_rpc(efx, MC_CMD_START_BIST, inbuf, sizeof(inbuf),
+ NULL, 0, NULL);
+}
+
+/* MC BISTs follow a different poll mechanism to phy BISTs.
+ * The BIST is done in the poll handler on the MC, and the MCDI command
+ * will block until the BIST is done.
+ */
+static int efx_ef10_poll_bist(struct efx_nic *efx)
+{
+ int rc;
+ MCDI_DECLARE_BUF(outbuf, MC_CMD_POLL_BIST_OUT_LEN);
+ size_t outlen;
+ u32 result;
+
+ rc = efx_mcdi_rpc(efx, MC_CMD_POLL_BIST, NULL, 0,
+ outbuf, sizeof(outbuf), &outlen);
+ if (rc != 0)
+ return rc;
+
+ if (outlen < MC_CMD_POLL_BIST_OUT_LEN)
+ return -EIO;
+
+ result = MCDI_DWORD(outbuf, POLL_BIST_OUT_RESULT);
+ switch (result) {
+ case MC_CMD_POLL_BIST_PASSED:
+ netif_dbg(efx, hw, efx->net_dev, "BIST passed.\n");
+ return 0;
+ case MC_CMD_POLL_BIST_TIMEOUT:
+ netif_err(efx, hw, efx->net_dev, "BIST timed out\n");
+ return -EIO;
+ case MC_CMD_POLL_BIST_FAILED:
+ netif_err(efx, hw, efx->net_dev, "BIST failed.\n");
+ return -EIO;
+ default:
+ netif_err(efx, hw, efx->net_dev,
+ "BIST returned unknown result %u", result);
+ return -EIO;
+ }
+}
+
+static int efx_ef10_run_bist(struct efx_nic *efx, u32 bist_type)
+{
+ int rc;
+
+ netif_dbg(efx, drv, efx->net_dev, "starting BIST type %u\n", bist_type);
+
+ rc = efx_ef10_start_bist(efx, bist_type);
+ if (rc != 0)
+ return rc;
+
+ return efx_ef10_poll_bist(efx);
+}
+
+static int
+efx_ef10_test_chip(struct efx_nic *efx, struct efx_self_tests *tests)
+{
+ int rc, rc2;
+
+ efx_reset_down(efx, RESET_TYPE_WORLD);
+
+ rc = efx_mcdi_rpc(efx, MC_CMD_ENABLE_OFFLINE_BIST,
+ NULL, 0, NULL, 0, NULL);
+ if (rc != 0)
+ goto out;
+
+ tests->memory = efx_ef10_run_bist(efx, MC_CMD_MC_MEM_BIST) ? -1 : 1;
+ tests->registers = efx_ef10_run_bist(efx, MC_CMD_REG_BIST) ? -1 : 1;
+
+ rc = efx_mcdi_reset(efx, RESET_TYPE_WORLD);
+
+out:
+ rc2 = efx_reset_up(efx, RESET_TYPE_WORLD, rc == 0);
+ return rc ? rc : rc2;
+}
+
#ifdef CONFIG_SFC_MTD
struct efx_ef10_nvram_type_info {
@@ -3345,7 +3427,7 @@ const struct efx_nic_type efx_hunt_a0_nic_type = {
.get_wol = efx_ef10_get_wol,
.set_wol = efx_ef10_set_wol,
.resume_wol = efx_port_dummy_op_void,
- /* TODO: test_chip */
+ .test_chip = efx_ef10_test_chip,
.test_nvram = efx_mcdi_nvram_test_all,
.mcdi_request = efx_ef10_mcdi_request,
.mcdi_poll_response = efx_ef10_mcdi_poll_response,
diff --git a/drivers/net/ethernet/sfc/efx.c b/drivers/net/ethernet/sfc/efx.c
index a35c63d43ae5..869418164364 100644
--- a/drivers/net/ethernet/sfc/efx.c
+++ b/drivers/net/ethernet/sfc/efx.c
@@ -83,6 +83,7 @@ const char *const efx_reset_type_names[] = {
[RESET_TYPE_DMA_ERROR] = "DMA_ERROR",
[RESET_TYPE_TX_SKIP] = "TX_SKIP",
[RESET_TYPE_MC_FAILURE] = "MC_FAILURE",
+ [RESET_TYPE_MC_BIST] = "MC_BIST",
};
/* Reset workqueue. If any NIC has a hardware failure then a reset will be
@@ -91,6 +92,12 @@ const char *const efx_reset_type_names[] = {
*/
static struct workqueue_struct *reset_workqueue;
+/* How often and how many times to poll for a reset while waiting for a
+ * BIST that another function started to complete.
+ */
+#define BIST_WAIT_DELAY_MS 100
+#define BIST_WAIT_DELAY_COUNT 100
+
/**************************************************************************
*
* Configurable values
@@ -2389,6 +2396,24 @@ int efx_try_recovery(struct efx_nic *efx)
return 0;
}
+static void efx_wait_for_bist_end(struct efx_nic *efx)
+{
+ int i;
+
+ for (i = 0; i < BIST_WAIT_DELAY_COUNT; ++i) {
+ if (efx_mcdi_poll_reboot(efx))
+ goto out;
+ msleep(BIST_WAIT_DELAY_MS);
+ }
+
+ netif_err(efx, drv, efx->net_dev, "Warning: No MC reboot after BIST mode\n");
+out:
+ /* Either way unset the BIST flag. If we found no reboot we probably
+ * won't recover, but we should try.
+ */
+ efx->mc_bist_for_other_fn = false;
+}
+
/* The worker thread exists so that code that cannot sleep can
* schedule a reset for later.
*/
@@ -2401,6 +2426,9 @@ static void efx_reset_work(struct work_struct *data)
pending = ACCESS_ONCE(efx->reset_pending);
method = fls(pending) - 1;
+ if (method == RESET_TYPE_MC_BIST)
+ efx_wait_for_bist_end(efx);
+
if ((method == RESET_TYPE_RECOVER_OR_DISABLE ||
method == RESET_TYPE_RECOVER_OR_ALL) &&
efx_try_recovery(efx))
@@ -2439,6 +2467,7 @@ void efx_schedule_reset(struct efx_nic *efx, enum reset_type type)
case RESET_TYPE_WORLD:
case RESET_TYPE_DISABLE:
case RESET_TYPE_RECOVER_OR_DISABLE:
+ case RESET_TYPE_MC_BIST:
method = type;
netif_dbg(efx, drv, efx->net_dev, "scheduling %s reset\n",
RESET_TYPE(method));
diff --git a/drivers/net/ethernet/sfc/enum.h b/drivers/net/ethernet/sfc/enum.h
index 7fdfee019092..75ef7ef6450b 100644
--- a/drivers/net/ethernet/sfc/enum.h
+++ b/drivers/net/ethernet/sfc/enum.h
@@ -165,6 +165,7 @@ enum reset_type {
RESET_TYPE_DMA_ERROR,
RESET_TYPE_TX_SKIP,
RESET_TYPE_MC_FAILURE,
+ RESET_TYPE_MC_BIST,
RESET_TYPE_MAX,
};
diff --git a/drivers/net/ethernet/sfc/ethtool.c b/drivers/net/ethernet/sfc/ethtool.c
index 1f529fa2edb1..fb8993806167 100644
--- a/drivers/net/ethernet/sfc/ethtool.c
+++ b/drivers/net/ethernet/sfc/ethtool.c
@@ -318,6 +318,8 @@ static int efx_ethtool_fill_self_tests(struct efx_nic *efx,
"eventq.int", NULL);
}
+ efx_fill_test(n++, strings, data, &tests->memory,
+ "core", 0, "memory", NULL);
efx_fill_test(n++, strings, data, &tests->registers,
"core", 0, "registers", NULL);
diff --git a/drivers/net/ethernet/sfc/mcdi.c b/drivers/net/ethernet/sfc/mcdi.c
index 25f91c0ca6a6..da14e2428944 100644
--- a/drivers/net/ethernet/sfc/mcdi.c
+++ b/drivers/net/ethernet/sfc/mcdi.c
@@ -543,6 +543,9 @@ int efx_mcdi_rpc_start(struct efx_nic *efx, unsigned cmd,
if (rc)
return rc;
+ if (efx->mc_bist_for_other_fn)
+ return -ENETDOWN;
+
efx_mcdi_acquire_sync(mcdi);
efx_mcdi_send_request(efx, cmd, inbuf, inlen);
return 0;
@@ -581,6 +584,9 @@ efx_mcdi_rpc_async(struct efx_nic *efx, unsigned int cmd,
if (rc)
return rc;
+ if (efx->mc_bist_for_other_fn)
+ return -ENETDOWN;
+
async = kmalloc(sizeof(*async) + ALIGN(max(inlen, outlen), 4),
GFP_ATOMIC);
if (!async)
@@ -834,6 +840,30 @@ static void efx_mcdi_ev_death(struct efx_nic *efx, int rc)
spin_unlock(&mcdi->iface_lock);
}
+/* The MC is going down in to BIST mode. set the BIST flag to block
+ * new MCDI, cancel any outstanding MCDI and and schedule a BIST-type reset
+ * (which doesn't actually execute a reset, it waits for the controlling
+ * function to reset it).
+ */
+static void efx_mcdi_ev_bist(struct efx_nic *efx)
+{
+ struct efx_mcdi_iface *mcdi = efx_mcdi(efx);
+
+ spin_lock(&mcdi->iface_lock);
+ efx->mc_bist_for_other_fn = true;
+ if (efx_mcdi_complete_sync(mcdi)) {
+ if (mcdi->mode == MCDI_MODE_EVENTS) {
+ mcdi->resprc = -EIO;
+ mcdi->resp_hdr_len = 0;
+ mcdi->resp_data_len = 0;
+ ++mcdi->credits;
+ }
+ }
+ mcdi->new_epoch = true;
+ efx_schedule_reset(efx, RESET_TYPE_MC_BIST);
+ spin_unlock(&mcdi->iface_lock);
+}
+
/* Called from falcon_process_eventq for MCDI events */
void efx_mcdi_process_event(struct efx_channel *channel,
efx_qword_t *event)
@@ -875,6 +905,10 @@ void efx_mcdi_process_event(struct efx_channel *channel,
netif_info(efx, hw, efx->net_dev, "MC Reboot\n");
efx_mcdi_ev_death(efx, -EIO);
break;
+ case MCDI_EVENT_CODE_MC_BIST:
+ netif_info(efx, hw, efx->net_dev, "MC entered BIST mode\n");
+ efx_mcdi_ev_bist(efx);
+ break;
case MCDI_EVENT_CODE_MAC_STATS_DMA:
/* MAC stats are gather lazily. We can ignore this. */
break;
diff --git a/drivers/net/ethernet/sfc/net_driver.h b/drivers/net/ethernet/sfc/net_driver.h
index 542a0d252ae0..d98b3f031ab5 100644
--- a/drivers/net/ethernet/sfc/net_driver.h
+++ b/drivers/net/ethernet/sfc/net_driver.h
@@ -852,6 +852,7 @@ struct efx_nic {
struct work_struct mac_work;
bool port_enabled;
+ bool mc_bist_for_other_fn;
bool port_initialized;
struct net_device *net_dev;
diff --git a/drivers/net/ethernet/sfc/selftest.c b/drivers/net/ethernet/sfc/selftest.c
index 144bbff5a4ae..26641817a9c7 100644
--- a/drivers/net/ethernet/sfc/selftest.c
+++ b/drivers/net/ethernet/sfc/selftest.c
@@ -722,7 +722,7 @@ int efx_selftest(struct efx_nic *efx, struct efx_self_tests *tests,
return rc_reset;
}
- if ((tests->registers < 0) && !rc_test)
+ if ((tests->memory < 0 || tests->registers < 0) && !rc_test)
rc_test = -EIO;
}
diff --git a/drivers/net/ethernet/sfc/selftest.h b/drivers/net/ethernet/sfc/selftest.h
index a2f4a06ffa4e..009dbe88f3be 100644
--- a/drivers/net/ethernet/sfc/selftest.h
+++ b/drivers/net/ethernet/sfc/selftest.h
@@ -38,6 +38,7 @@ struct efx_self_tests {
int eventq_dma[EFX_MAX_CHANNELS];
int eventq_int[EFX_MAX_CHANNELS];
/* offline tests */
+ int memory;
int registers;
int phy_ext[EFX_MAX_PHY_TESTS];
struct efx_loopback_self_tests loopback[LOOPBACK_TEST_MAX + 1];
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH net-next 04/16] sfc: Tighten the check for RX merged completion events
2013-12-12 22:56 Pull request: sfc-next 2013-12-12 Ben Hutchings
` (2 preceding siblings ...)
2013-12-12 22:59 ` [PATCH net-next 03/16] sfc: Add MC BISTs to ethtool offline self test on EF10 Ben Hutchings
@ 2013-12-12 23:00 ` Ben Hutchings
2013-12-12 23:00 ` [PATCH net-next 05/16] sfc: Make initial fill of RX descriptors synchronous Ben Hutchings
` (12 subsequent siblings)
16 siblings, 0 replies; 18+ messages in thread
From: Ben Hutchings @ 2013-12-12 23:00 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
The addition of RX event merging support means we don't reliably
detect dropped RX events now. Currently we will only detect them if
the previous event for the RX queue had the CONT bit set.
Only accept RX completion events as merged if the
GET_CAPABILITIES_OUT_RX_BATCHING bit is set in datapath_caps (which it
won't be for the low-latency datapath) and the CONT bit is not set on
the event.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/ef10.c | 13 +++++++++----
1 file changed, 9 insertions(+), 4 deletions(-)
diff --git a/drivers/net/ethernet/sfc/ef10.c b/drivers/net/ethernet/sfc/ef10.c
index 5d46d155b642..2f77359607d2 100644
--- a/drivers/net/ethernet/sfc/ef10.c
+++ b/drivers/net/ethernet/sfc/ef10.c
@@ -1766,6 +1766,8 @@ static int efx_ef10_handle_rx_event(struct efx_channel *channel,
((1 << ESF_DZ_RX_DSC_PTR_LBITS_WIDTH) - 1));
if (n_descs != rx_queue->scatter_n + 1) {
+ struct efx_ef10_nic_data *nic_data = efx->nic_data;
+
/* detect rx abort */
if (unlikely(n_descs == rx_queue->scatter_n)) {
WARN_ON(rx_bytes != 0);
@@ -1773,10 +1775,13 @@ static int efx_ef10_handle_rx_event(struct efx_channel *channel,
return 0;
}
- if (unlikely(rx_queue->scatter_n != 0)) {
- /* Scattered packet completions cannot be
- * merged, so something has gone wrong.
- */
+ /* Check that RX completion merging is valid, i.e.
+ * the current firmware supports it and this is a
+ * non-scattered packet.
+ */
+ if (!(nic_data->datapath_caps &
+ (1 << MC_CMD_GET_CAPABILITIES_OUT_RX_BATCHING_LBN)) ||
+ rx_queue->scatter_n != 0 || rx_cont) {
efx_ef10_handle_rx_bad_lbits(
rx_queue, next_ptr_lbits,
(rx_queue->removed_count +
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH net-next 05/16] sfc: Make initial fill of RX descriptors synchronous
2013-12-12 22:56 Pull request: sfc-next 2013-12-12 Ben Hutchings
` (3 preceding siblings ...)
2013-12-12 23:00 ` [PATCH net-next 04/16] sfc: Tighten the check for RX merged completion events Ben Hutchings
@ 2013-12-12 23:00 ` Ben Hutchings
2013-12-12 23:00 ` [PATCH net-next 06/16] sfc: Correct RX dropped count for drops while interface is down Ben Hutchings
` (11 subsequent siblings)
16 siblings, 0 replies; 18+ messages in thread
From: Ben Hutchings @ 2013-12-12 23:00 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
From: Jon Cooper <jcooper@solarflare.com>
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/ef10.c | 2 +-
drivers/net/ethernet/sfc/efx.c | 6 ++++--
drivers/net/ethernet/sfc/efx.h | 2 +-
drivers/net/ethernet/sfc/farch.c | 2 +-
drivers/net/ethernet/sfc/rx.c | 9 +++++----
5 files changed, 12 insertions(+), 9 deletions(-)
diff --git a/drivers/net/ethernet/sfc/ef10.c b/drivers/net/ethernet/sfc/ef10.c
index 2f77359607d2..e9c546bdbdfe 100644
--- a/drivers/net/ethernet/sfc/ef10.c
+++ b/drivers/net/ethernet/sfc/ef10.c
@@ -1907,7 +1907,7 @@ static void efx_ef10_handle_driver_generated_event(struct efx_channel *channel,
* events, so efx_process_channel() won't refill the
* queue. Refill it here
*/
- efx_fast_push_rx_descriptors(&channel->rx_queue);
+ efx_fast_push_rx_descriptors(&channel->rx_queue, true);
break;
default:
netif_err(efx, hw, efx->net_dev,
diff --git a/drivers/net/ethernet/sfc/efx.c b/drivers/net/ethernet/sfc/efx.c
index 869418164364..5e2454d07137 100644
--- a/drivers/net/ethernet/sfc/efx.c
+++ b/drivers/net/ethernet/sfc/efx.c
@@ -253,7 +253,7 @@ static int efx_process_channel(struct efx_channel *channel, int budget)
efx_channel_get_rx_queue(channel);
efx_rx_flush_packet(channel);
- efx_fast_push_rx_descriptors(rx_queue);
+ efx_fast_push_rx_descriptors(rx_queue, true);
}
return spent;
@@ -646,7 +646,9 @@ static void efx_start_datapath(struct efx_nic *efx)
efx_for_each_channel_rx_queue(rx_queue, channel) {
efx_init_rx_queue(rx_queue);
atomic_inc(&efx->active_queues);
- efx_nic_generate_fill_event(rx_queue);
+ efx_stop_eventq(channel);
+ efx_fast_push_rx_descriptors(rx_queue, false);
+ efx_start_eventq(channel);
}
WARN_ON(channel->rx_pkt_n_frags);
diff --git a/drivers/net/ethernet/sfc/efx.h b/drivers/net/ethernet/sfc/efx.h
index b8235ee5d7d7..a653786fbbe7 100644
--- a/drivers/net/ethernet/sfc/efx.h
+++ b/drivers/net/ethernet/sfc/efx.h
@@ -37,7 +37,7 @@ int efx_probe_rx_queue(struct efx_rx_queue *rx_queue);
void efx_remove_rx_queue(struct efx_rx_queue *rx_queue);
void efx_init_rx_queue(struct efx_rx_queue *rx_queue);
void efx_fini_rx_queue(struct efx_rx_queue *rx_queue);
-void efx_fast_push_rx_descriptors(struct efx_rx_queue *rx_queue);
+void efx_fast_push_rx_descriptors(struct efx_rx_queue *rx_queue, bool atomic);
void efx_rx_slow_fill(unsigned long context);
void __efx_rx_packet(struct efx_channel *channel);
void efx_rx_packet(struct efx_rx_queue *rx_queue, unsigned int index,
diff --git a/drivers/net/ethernet/sfc/farch.c b/drivers/net/ethernet/sfc/farch.c
index c0907d884d75..984e85ee76f6 100644
--- a/drivers/net/ethernet/sfc/farch.c
+++ b/drivers/net/ethernet/sfc/farch.c
@@ -1147,7 +1147,7 @@ static void efx_farch_handle_generated_event(struct efx_channel *channel,
/* The queue must be empty, so we won't receive any rx
* events, so efx_process_channel() won't refill the
* queue. Refill it here */
- efx_fast_push_rx_descriptors(rx_queue);
+ efx_fast_push_rx_descriptors(rx_queue, true);
} else if (rx_queue && magic == EFX_CHANNEL_MAGIC_RX_DRAIN(rx_queue)) {
efx_farch_handle_drain_event(channel);
} else if (code == _EFX_CHANNEL_MAGIC_TX_DRAIN) {
diff --git a/drivers/net/ethernet/sfc/rx.c b/drivers/net/ethernet/sfc/rx.c
index 42488df1f4ec..8671bc199a9d 100644
--- a/drivers/net/ethernet/sfc/rx.c
+++ b/drivers/net/ethernet/sfc/rx.c
@@ -149,7 +149,7 @@ static struct page *efx_reuse_page(struct efx_rx_queue *rx_queue)
* 0 on success. If a single page can be used for multiple buffers,
* then the page will either be inserted fully, or not at all.
*/
-static int efx_init_rx_buffers(struct efx_rx_queue *rx_queue)
+static int efx_init_rx_buffers(struct efx_rx_queue *rx_queue, bool atomic)
{
struct efx_nic *efx = rx_queue->efx;
struct efx_rx_buffer *rx_buf;
@@ -163,7 +163,8 @@ static int efx_init_rx_buffers(struct efx_rx_queue *rx_queue)
do {
page = efx_reuse_page(rx_queue);
if (page == NULL) {
- page = alloc_pages(__GFP_COLD | __GFP_COMP | GFP_ATOMIC,
+ page = alloc_pages(__GFP_COLD | __GFP_COMP |
+ (atomic ? GFP_ATOMIC : GFP_KERNEL),
efx->rx_buffer_order);
if (unlikely(page == NULL))
return -ENOMEM;
@@ -321,7 +322,7 @@ static void efx_discard_rx_packet(struct efx_channel *channel,
* this means this function must run from the NAPI handler, or be called
* when NAPI is disabled.
*/
-void efx_fast_push_rx_descriptors(struct efx_rx_queue *rx_queue)
+void efx_fast_push_rx_descriptors(struct efx_rx_queue *rx_queue, bool atomic)
{
struct efx_nic *efx = rx_queue->efx;
unsigned int fill_level, batch_size;
@@ -354,7 +355,7 @@ void efx_fast_push_rx_descriptors(struct efx_rx_queue *rx_queue)
do {
- rc = efx_init_rx_buffers(rx_queue);
+ rc = efx_init_rx_buffers(rx_queue, atomic);
if (unlikely(rc)) {
/* Ensure that we don't leave the rx queue empty */
if (rx_queue->added_count == rx_queue->removed_count)
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH net-next 06/16] sfc: Correct RX dropped count for drops while interface is down
2013-12-12 22:56 Pull request: sfc-next 2013-12-12 Ben Hutchings
` (4 preceding siblings ...)
2013-12-12 23:00 ` [PATCH net-next 05/16] sfc: Make initial fill of RX descriptors synchronous Ben Hutchings
@ 2013-12-12 23:00 ` Ben Hutchings
2013-12-12 23:01 ` [PATCH net-next 07/16] sfc: Report units in sensor warnings Ben Hutchings
` (10 subsequent siblings)
16 siblings, 0 replies; 18+ messages in thread
From: Ben Hutchings @ 2013-12-12 23:00 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
From: Jon Cooper <jcooper@solarflare.com>
We don't directly control RX ingress on Siena or any later
controllers, and so we cannot prevent packets from entering the RX
datapath while the RX queues are not set up. This results in
the hardware incrementing RX_NODESC_DROP_CNT, but it's not an
error and we should not include it in error stats.
When bringing an interface up or down, pull (or wait for) stats and
count the number of packets that were dropped while the interface was
down. Subtract this from the reported RX dropped count.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/ef10.c | 2 ++
drivers/net/ethernet/sfc/efx.c | 11 ++++++++
drivers/net/ethernet/sfc/falcon.c | 10 +++++++
drivers/net/ethernet/sfc/mcdi.h | 1 +
drivers/net/ethernet/sfc/mcdi_port.c | 49 +++++++++++++++++++++++++++--------
drivers/net/ethernet/sfc/net_driver.h | 5 ++++
drivers/net/ethernet/sfc/nic.c | 12 +++++++++
drivers/net/ethernet/sfc/nic.h | 1 +
drivers/net/ethernet/sfc/siena.c | 3 +++
9 files changed, 83 insertions(+), 11 deletions(-)
diff --git a/drivers/net/ethernet/sfc/ef10.c b/drivers/net/ethernet/sfc/ef10.c
index e9c546bdbdfe..c1b85edcb204 100644
--- a/drivers/net/ethernet/sfc/ef10.c
+++ b/drivers/net/ethernet/sfc/ef10.c
@@ -902,6 +902,7 @@ static int efx_ef10_try_update_nic_stats(struct efx_nic *efx)
return -EAGAIN;
/* Update derived statistics */
+ efx_nic_fix_nodesc_drop_stat(efx, &stats[EF10_STAT_rx_nodesc_drops]);
stats[EF10_STAT_rx_good_bytes] =
stats[EF10_STAT_rx_bytes] -
stats[EF10_STAT_rx_bytes_minus_good_bytes];
@@ -3423,6 +3424,7 @@ const struct efx_nic_type efx_hunt_a0_nic_type = {
.describe_stats = efx_ef10_describe_stats,
.update_stats = efx_ef10_update_stats,
.start_stats = efx_mcdi_mac_start_stats,
+ .pull_stats = efx_mcdi_mac_pull_stats,
.stop_stats = efx_mcdi_mac_stop_stats,
.set_id_led = efx_mcdi_set_id_led,
.push_irq_moderation = efx_ef10_push_irq_moderation,
diff --git a/drivers/net/ethernet/sfc/efx.c b/drivers/net/ethernet/sfc/efx.c
index 5e2454d07137..c734fba8c99c 100644
--- a/drivers/net/ethernet/sfc/efx.c
+++ b/drivers/net/ethernet/sfc/efx.c
@@ -1684,6 +1684,10 @@ static void efx_start_all(struct efx_nic *efx)
}
efx->type->start_stats(efx);
+ efx->type->pull_stats(efx);
+ spin_lock_bh(&efx->stats_lock);
+ efx->type->update_stats(efx, NULL, NULL);
+ spin_unlock_bh(&efx->stats_lock);
}
/* Flush all delayed work. Should only be called when no more delayed work
@@ -1711,6 +1715,13 @@ static void efx_stop_all(struct efx_nic *efx)
if (!efx->port_enabled)
return;
+ /* update stats before we go down so we can accurately count
+ * rx_nodesc_drops
+ */
+ efx->type->pull_stats(efx);
+ spin_lock_bh(&efx->stats_lock);
+ efx->type->update_stats(efx, NULL, NULL);
+ spin_unlock_bh(&efx->stats_lock);
efx->type->stop_stats(efx);
efx_stop_port(efx);
diff --git a/drivers/net/ethernet/sfc/falcon.c b/drivers/net/ethernet/sfc/falcon.c
index ff5d322b9b49..4a9e05c82e2a 100644
--- a/drivers/net/ethernet/sfc/falcon.c
+++ b/drivers/net/ethernet/sfc/falcon.c
@@ -2593,6 +2593,14 @@ void falcon_start_nic_stats(struct efx_nic *efx)
spin_unlock_bh(&efx->stats_lock);
}
+/* We don't acutally pull stats on falcon. Wait 10ms so that
+ * they arrive when we call this just after start_stats
+ */
+void falcon_pull_nic_stats(struct efx_nic *efx)
+{
+ msleep(10);
+}
+
void falcon_stop_nic_stats(struct efx_nic *efx)
{
struct falcon_nic_data *nic_data = efx->nic_data;
@@ -2672,6 +2680,7 @@ const struct efx_nic_type falcon_a1_nic_type = {
.describe_stats = falcon_describe_nic_stats,
.update_stats = falcon_update_nic_stats,
.start_stats = falcon_start_nic_stats,
+ .pull_stats = falcon_pull_nic_stats,
.stop_stats = falcon_stop_nic_stats,
.set_id_led = falcon_set_id_led,
.push_irq_moderation = falcon_push_irq_moderation,
@@ -2765,6 +2774,7 @@ const struct efx_nic_type falcon_b0_nic_type = {
.describe_stats = falcon_describe_nic_stats,
.update_stats = falcon_update_nic_stats,
.start_stats = falcon_start_nic_stats,
+ .pull_stats = falcon_pull_nic_stats,
.stop_stats = falcon_stop_nic_stats,
.set_id_led = falcon_set_id_led,
.push_irq_moderation = falcon_push_irq_moderation,
diff --git a/drivers/net/ethernet/sfc/mcdi.h b/drivers/net/ethernet/sfc/mcdi.h
index 15816cacb548..d861628b7ee6 100644
--- a/drivers/net/ethernet/sfc/mcdi.h
+++ b/drivers/net/ethernet/sfc/mcdi.h
@@ -301,6 +301,7 @@ int efx_mcdi_set_mac(struct efx_nic *efx);
#define EFX_MC_STATS_GENERATION_INVALID ((__force __le64)(-1))
void efx_mcdi_mac_start_stats(struct efx_nic *efx);
void efx_mcdi_mac_stop_stats(struct efx_nic *efx);
+void efx_mcdi_mac_pull_stats(struct efx_nic *efx);
bool efx_mcdi_mac_check_fault(struct efx_nic *efx);
enum reset_type efx_mcdi_map_reset_reason(enum reset_type reason);
int efx_mcdi_reset(struct efx_nic *efx, enum reset_type method);
diff --git a/drivers/net/ethernet/sfc/mcdi_port.c b/drivers/net/ethernet/sfc/mcdi_port.c
index 7b6be61d549f..7288aefc2877 100644
--- a/drivers/net/ethernet/sfc/mcdi_port.c
+++ b/drivers/net/ethernet/sfc/mcdi_port.c
@@ -927,12 +927,23 @@ bool efx_mcdi_mac_check_fault(struct efx_nic *efx)
return MCDI_DWORD(outbuf, GET_LINK_OUT_MAC_FAULT) != 0;
}
-static int efx_mcdi_mac_stats(struct efx_nic *efx, dma_addr_t dma_addr,
- u32 dma_len, int enable, int clear)
+enum efx_stats_action {
+ EFX_STATS_ENABLE,
+ EFX_STATS_DISABLE,
+ EFX_STATS_PULL,
+};
+
+static int efx_mcdi_mac_stats(struct efx_nic *efx,
+ enum efx_stats_action action, int clear)
{
MCDI_DECLARE_BUF(inbuf, MC_CMD_MAC_STATS_IN_LEN);
int rc;
- int period = enable ? 1000 : 0;
+ int change = action == EFX_STATS_PULL ? 0 : 1;
+ int enable = action == EFX_STATS_ENABLE ? 1 : 0;
+ int period = action == EFX_STATS_ENABLE ? 1000 : 0;
+ dma_addr_t dma_addr = efx->stats_buffer.dma_addr;
+ u32 dma_len = action != EFX_STATS_DISABLE ?
+ MC_CMD_MAC_NSTATS * sizeof(u64) : 0;
BUILD_BUG_ON(MC_CMD_MAC_STATS_OUT_DMA_LEN != 0);
@@ -940,8 +951,8 @@ static int efx_mcdi_mac_stats(struct efx_nic *efx, dma_addr_t dma_addr,
MCDI_POPULATE_DWORD_7(inbuf, MAC_STATS_IN_CMD,
MAC_STATS_IN_DMA, !!enable,
MAC_STATS_IN_CLEAR, clear,
- MAC_STATS_IN_PERIODIC_CHANGE, 1,
- MAC_STATS_IN_PERIODIC_ENABLE, !!enable,
+ MAC_STATS_IN_PERIODIC_CHANGE, change,
+ MAC_STATS_IN_PERIODIC_ENABLE, enable,
MAC_STATS_IN_PERIODIC_CLEAR, 0,
MAC_STATS_IN_PERIODIC_NOEVENT, 1,
MAC_STATS_IN_PERIOD_MS, period);
@@ -955,8 +966,8 @@ static int efx_mcdi_mac_stats(struct efx_nic *efx, dma_addr_t dma_addr,
return 0;
fail:
- netif_err(efx, hw, efx->net_dev, "%s: %s failed rc=%d\n",
- __func__, enable ? "enable" : "disable", rc);
+ netif_err(efx, hw, efx->net_dev, "%s: action %d failed rc=%d\n",
+ __func__, action, rc);
return rc;
}
@@ -966,13 +977,29 @@ void efx_mcdi_mac_start_stats(struct efx_nic *efx)
dma_stats[MC_CMD_MAC_GENERATION_END] = EFX_MC_STATS_GENERATION_INVALID;
- efx_mcdi_mac_stats(efx, efx->stats_buffer.dma_addr,
- MC_CMD_MAC_NSTATS * sizeof(u64), 1, 0);
+ efx_mcdi_mac_stats(efx, EFX_STATS_ENABLE, 0);
}
void efx_mcdi_mac_stop_stats(struct efx_nic *efx)
{
- efx_mcdi_mac_stats(efx, efx->stats_buffer.dma_addr, 0, 0, 0);
+ efx_mcdi_mac_stats(efx, EFX_STATS_DISABLE, 0);
+}
+
+#define EFX_MAC_STATS_WAIT_US 100
+#define EFX_MAC_STATS_WAIT_ATTEMPTS 10
+
+void efx_mcdi_mac_pull_stats(struct efx_nic *efx)
+{
+ __le64 *dma_stats = efx->stats_buffer.addr;
+ int attempts = EFX_MAC_STATS_WAIT_ATTEMPTS;
+
+ dma_stats[MC_CMD_MAC_GENERATION_END] = EFX_MC_STATS_GENERATION_INVALID;
+ efx_mcdi_mac_stats(efx, EFX_STATS_PULL, 0);
+
+ while (dma_stats[MC_CMD_MAC_GENERATION_END] ==
+ EFX_MC_STATS_GENERATION_INVALID &&
+ attempts-- != 0)
+ udelay(EFX_MAC_STATS_WAIT_US);
}
int efx_mcdi_port_probe(struct efx_nic *efx)
@@ -1003,7 +1030,7 @@ int efx_mcdi_port_probe(struct efx_nic *efx)
efx->stats_buffer.addr,
(u64)virt_to_phys(efx->stats_buffer.addr));
- efx_mcdi_mac_stats(efx, efx->stats_buffer.dma_addr, 0, 0, 1);
+ efx_mcdi_mac_stats(efx, EFX_STATS_DISABLE, 1);
return 0;
}
diff --git a/drivers/net/ethernet/sfc/net_driver.h b/drivers/net/ethernet/sfc/net_driver.h
index d98b3f031ab5..f47bac78b92c 100644
--- a/drivers/net/ethernet/sfc/net_driver.h
+++ b/drivers/net/ethernet/sfc/net_driver.h
@@ -857,6 +857,9 @@ struct efx_nic {
struct net_device *net_dev;
struct efx_buffer stats_buffer;
+ u64 rx_nodesc_drops_total;
+ u64 rx_nodesc_drops_while_down;
+ bool rx_nodesc_drops_prev_state;
unsigned int phy_type;
const struct efx_phy_operations *phy_op;
@@ -960,6 +963,7 @@ struct efx_mtd_partition {
* @update_stats: Update statistics not provided by event handling.
* Either argument may be %NULL.
* @start_stats: Start the regular fetching of statistics
+ * @pull_stats: Pull stats from the NIC and wait until they arrive.
* @stop_stats: Stop the regular fetching of statistics
* @set_id_led: Set state of identifying LED or revert to automatic function
* @push_irq_moderation: Apply interrupt moderation value
@@ -1078,6 +1082,7 @@ struct efx_nic_type {
size_t (*update_stats)(struct efx_nic *efx, u64 *full_stats,
struct rtnl_link_stats64 *core_stats);
void (*start_stats)(struct efx_nic *efx);
+ void (*pull_stats)(struct efx_nic *efx);
void (*stop_stats)(struct efx_nic *efx);
void (*set_id_led)(struct efx_nic *efx, enum efx_led_mode mode);
void (*push_irq_moderation)(struct efx_channel *channel);
diff --git a/drivers/net/ethernet/sfc/nic.c b/drivers/net/ethernet/sfc/nic.c
index 9c90bf56090f..79226b19e3c4 100644
--- a/drivers/net/ethernet/sfc/nic.c
+++ b/drivers/net/ethernet/sfc/nic.c
@@ -519,3 +519,15 @@ void efx_nic_update_stats(const struct efx_hw_stat_desc *desc, size_t count,
}
}
}
+
+void efx_nic_fix_nodesc_drop_stat(struct efx_nic *efx, u64 *rx_nodesc_drops)
+{
+ /* if down, or this is the first update after coming up */
+ if (!(efx->net_dev->flags & IFF_UP) || !efx->rx_nodesc_drops_prev_state)
+ efx->rx_nodesc_drops_while_down +=
+ *rx_nodesc_drops - efx->rx_nodesc_drops_total;
+ efx->rx_nodesc_drops_total = *rx_nodesc_drops;
+ efx->rx_nodesc_drops_prev_state = !!(efx->net_dev->flags & IFF_UP);
+ *rx_nodesc_drops -= efx->rx_nodesc_drops_while_down;
+}
+
diff --git a/drivers/net/ethernet/sfc/nic.h b/drivers/net/ethernet/sfc/nic.h
index fe71e81ee8bf..7ac9c000696f 100644
--- a/drivers/net/ethernet/sfc/nic.h
+++ b/drivers/net/ethernet/sfc/nic.h
@@ -775,6 +775,7 @@ size_t efx_nic_describe_stats(const struct efx_hw_stat_desc *desc, size_t count,
void efx_nic_update_stats(const struct efx_hw_stat_desc *desc, size_t count,
const unsigned long *mask, u64 *stats,
const void *dma_buf, bool accumulate);
+void efx_nic_fix_nodesc_drop_stat(struct efx_nic *efx, u64 *stat);
#define EFX_MAX_FLUSH_TIME 5000
diff --git a/drivers/net/ethernet/sfc/siena.c b/drivers/net/ethernet/sfc/siena.c
index d034bcd124ef..f2a7ad4c76aa 100644
--- a/drivers/net/ethernet/sfc/siena.c
+++ b/drivers/net/ethernet/sfc/siena.c
@@ -458,6 +458,8 @@ static int siena_try_update_nic_stats(struct efx_nic *efx)
return -EAGAIN;
/* Update derived statistics */
+ efx_nic_fix_nodesc_drop_stat(efx,
+ &stats[SIENA_STAT_rx_nodesc_drop_cnt]);
efx_update_diff_stat(&stats[SIENA_STAT_tx_good_bytes],
stats[SIENA_STAT_tx_bytes] -
stats[SIENA_STAT_tx_bad_bytes]);
@@ -878,6 +880,7 @@ const struct efx_nic_type siena_a0_nic_type = {
.describe_stats = siena_describe_nic_stats,
.update_stats = siena_update_nic_stats,
.start_stats = efx_mcdi_mac_start_stats,
+ .pull_stats = efx_mcdi_mac_pull_stats,
.stop_stats = efx_mcdi_mac_stop_stats,
.set_id_led = efx_mcdi_set_id_led,
.push_irq_moderation = siena_push_irq_moderation,
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH net-next 07/16] sfc: Report units in sensor warnings
2013-12-12 22:56 Pull request: sfc-next 2013-12-12 Ben Hutchings
` (5 preceding siblings ...)
2013-12-12 23:00 ` [PATCH net-next 06/16] sfc: Correct RX dropped count for drops while interface is down Ben Hutchings
@ 2013-12-12 23:01 ` Ben Hutchings
2013-12-12 23:01 ` [PATCH net-next 08/16] sfc: Revise sensor names to be more understandable and consistent Ben Hutchings
` (9 subsequent siblings)
16 siblings, 0 replies; 18+ messages in thread
From: Ben Hutchings @ 2013-12-12 23:01 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
From: Edward Cree <ecree@solarflare.com>
Add units to the "Sensor reports condition X for raw value Y" messages.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/mcdi_mon.c | 24 ++++++++++++++++++++----
1 file changed, 20 insertions(+), 4 deletions(-)
diff --git a/drivers/net/ethernet/sfc/mcdi_mon.c b/drivers/net/ethernet/sfc/mcdi_mon.c
index d72ad4fc3617..99c80f8c0a0e 100644
--- a/drivers/net/ethernet/sfc/mcdi_mon.c
+++ b/drivers/net/ethernet/sfc/mcdi_mon.c
@@ -24,6 +24,15 @@ enum efx_hwmon_type {
EFX_HWMON_IN, /* voltage */
EFX_HWMON_CURR, /* current */
EFX_HWMON_POWER, /* power */
+ EFX_HWMON_TYPES_COUNT
+};
+
+static const char *const efx_hwmon_unit[EFX_HWMON_TYPES_COUNT] = {
+ [EFX_HWMON_TEMP] = " degC",
+ [EFX_HWMON_COOL] = " rpm", /* though nonsense for a heatsink */
+ [EFX_HWMON_IN] = " mV",
+ [EFX_HWMON_CURR] = " mA",
+ [EFX_HWMON_POWER] = " W",
};
static const struct {
@@ -91,7 +100,8 @@ static const char *const sensor_status_names[] = {
void efx_mcdi_sensor_event(struct efx_nic *efx, efx_qword_t *ev)
{
unsigned int type, state, value;
- const char *name = NULL, *state_txt;
+ enum efx_hwmon_type hwmon_type = EFX_HWMON_UNKNOWN;
+ const char *name = NULL, *state_txt, *unit;
type = EFX_QWORD_FIELD(*ev, MCDI_EVENT_SENSOREVT_MONITOR);
state = EFX_QWORD_FIELD(*ev, MCDI_EVENT_SENSOREVT_STATE);
@@ -99,16 +109,22 @@ void efx_mcdi_sensor_event(struct efx_nic *efx, efx_qword_t *ev)
/* Deal gracefully with the board having more drivers than we
* know about, but do not expect new sensor states. */
- if (type < ARRAY_SIZE(efx_mcdi_sensor_type))
+ if (type < ARRAY_SIZE(efx_mcdi_sensor_type)) {
name = efx_mcdi_sensor_type[type].label;
+ hwmon_type = efx_mcdi_sensor_type[type].hwmon_type;
+ }
if (!name)
name = "No sensor name available";
EFX_BUG_ON_PARANOID(state >= ARRAY_SIZE(sensor_status_names));
state_txt = sensor_status_names[state];
+ EFX_BUG_ON_PARANOID(hwmon_type >= EFX_HWMON_TYPES_COUNT);
+ unit = efx_hwmon_unit[hwmon_type];
+ if (!unit)
+ unit = "";
netif_err(efx, hw, efx->net_dev,
- "Sensor %d (%s) reports condition '%s' for raw value %d\n",
- type, name, state_txt, value);
+ "Sensor %d (%s) reports condition '%s' for value %d%s\n",
+ type, name, state_txt, value, unit);
}
#ifdef CONFIG_SFC_MCDI_MON
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH net-next 08/16] sfc: Revise sensor names to be more understandable and consistent
2013-12-12 22:56 Pull request: sfc-next 2013-12-12 Ben Hutchings
` (6 preceding siblings ...)
2013-12-12 23:01 ` [PATCH net-next 07/16] sfc: Report units in sensor warnings Ben Hutchings
@ 2013-12-12 23:01 ` Ben Hutchings
2013-12-12 23:01 ` [PATCH net-next 09/16] sfc: Add new sensor names Ben Hutchings
` (8 subsequent siblings)
16 siblings, 0 replies; 18+ messages in thread
From: Ben Hutchings @ 2013-12-12 23:01 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
From: Edward Cree <ecree@solarflare.com>
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/mcdi_mon.c | 49 ++++++++++++++++++++-----------------
1 file changed, 26 insertions(+), 23 deletions(-)
diff --git a/drivers/net/ethernet/sfc/mcdi_mon.c b/drivers/net/ethernet/sfc/mcdi_mon.c
index 99c80f8c0a0e..2891b409d263 100644
--- a/drivers/net/ethernet/sfc/mcdi_mon.c
+++ b/drivers/net/ethernet/sfc/mcdi_mon.c
@@ -42,13 +42,13 @@ static const struct {
} efx_mcdi_sensor_type[] = {
#define SENSOR(name, label, hwmon_type, port) \
[MC_CMD_SENSOR_##name] = { label, EFX_HWMON_ ## hwmon_type, port }
- SENSOR(CONTROLLER_TEMP, "Controller ext. temp.", TEMP, -1),
+ SENSOR(CONTROLLER_TEMP, "Controller board temp.", TEMP, -1),
SENSOR(PHY_COMMON_TEMP, "PHY temp.", TEMP, -1),
- SENSOR(CONTROLLER_COOLING, "Controller cooling", COOL, -1),
+ SENSOR(CONTROLLER_COOLING, "Controller heat sink", COOL, -1),
SENSOR(PHY0_TEMP, "PHY temp.", TEMP, 0),
- SENSOR(PHY0_COOLING, "PHY cooling", COOL, 0),
+ SENSOR(PHY0_COOLING, "PHY heat sink", COOL, 0),
SENSOR(PHY1_TEMP, "PHY temp.", TEMP, 1),
- SENSOR(PHY1_COOLING, "PHY cooling", COOL, 1),
+ SENSOR(PHY1_COOLING, "PHY heat sink", COOL, 1),
SENSOR(IN_1V0, "1.0V supply", IN, -1),
SENSOR(IN_1V2, "1.2V supply", IN, -1),
SENSOR(IN_1V8, "1.8V supply", IN, -1),
@@ -56,34 +56,37 @@ static const struct {
SENSOR(IN_3V3, "3.3V supply", IN, -1),
SENSOR(IN_12V0, "12.0V supply", IN, -1),
SENSOR(IN_1V2A, "1.2V analogue supply", IN, -1),
- SENSOR(IN_VREF, "ref. voltage", IN, -1),
- SENSOR(OUT_VAOE, "AOE power supply", IN, -1),
- SENSOR(AOE_TEMP, "AOE temp.", TEMP, -1),
- SENSOR(PSU_AOE_TEMP, "AOE PSU temp.", TEMP, -1),
- SENSOR(PSU_TEMP, "Controller PSU temp.", TEMP, -1),
- SENSOR(FAN_0, NULL, COOL, -1),
- SENSOR(FAN_1, NULL, COOL, -1),
- SENSOR(FAN_2, NULL, COOL, -1),
- SENSOR(FAN_3, NULL, COOL, -1),
- SENSOR(FAN_4, NULL, COOL, -1),
+ SENSOR(IN_VREF, "Ref. voltage", IN, -1),
+ SENSOR(OUT_VAOE, "AOE FPGA supply", IN, -1),
+ SENSOR(AOE_TEMP, "AOE FPGA temp.", TEMP, -1),
+ SENSOR(PSU_AOE_TEMP, "AOE regulator temp.", TEMP, -1),
+ SENSOR(PSU_TEMP, "Controller regulator temp.",
+ TEMP, -1),
+ SENSOR(FAN_0, "Fan 0", COOL, -1),
+ SENSOR(FAN_1, "Fan 1", COOL, -1),
+ SENSOR(FAN_2, "Fan 2", COOL, -1),
+ SENSOR(FAN_3, "Fan 3", COOL, -1),
+ SENSOR(FAN_4, "Fan 4", COOL, -1),
SENSOR(IN_VAOE, "AOE input supply", IN, -1),
SENSOR(OUT_IAOE, "AOE output current", CURR, -1),
SENSOR(IN_IAOE, "AOE input current", CURR, -1),
SENSOR(NIC_POWER, "Board power use", POWER, -1),
SENSOR(IN_0V9, "0.9V supply", IN, -1),
- SENSOR(IN_I0V9, "0.9V input current", CURR, -1),
- SENSOR(IN_I1V2, "1.2V input current", CURR, -1),
- SENSOR(IN_0V9_ADC, "0.9V supply (at ADC)", IN, -1),
- SENSOR(CONTROLLER_2_TEMP, "Controller ext. temp. 2", TEMP, -1),
- SENSOR(VREG_INTERNAL_TEMP, "Voltage regulator temp.", TEMP, -1),
+ SENSOR(IN_I0V9, "0.9V supply current", CURR, -1),
+ SENSOR(IN_I1V2, "1.2V supply current", CURR, -1),
+ SENSOR(IN_0V9_ADC, "0.9V supply (ext. ADC)", IN, -1),
+ SENSOR(CONTROLLER_2_TEMP, "Controller board temp. 2", TEMP, -1),
+ SENSOR(VREG_INTERNAL_TEMP, "Regulator die temp.", TEMP, -1),
SENSOR(VREG_0V9_TEMP, "0.9V regulator temp.", TEMP, -1),
SENSOR(VREG_1V2_TEMP, "1.2V regulator temp.", TEMP, -1),
- SENSOR(CONTROLLER_VPTAT, "Controller int. temp. raw", IN, -1),
- SENSOR(CONTROLLER_INTERNAL_TEMP, "Controller int. temp.", TEMP, -1),
+ SENSOR(CONTROLLER_VPTAT,
+ "Controller PTAT voltage (int. ADC)", IN, -1),
+ SENSOR(CONTROLLER_INTERNAL_TEMP,
+ "Controller die temp. (int. ADC)", TEMP, -1),
SENSOR(CONTROLLER_VPTAT_EXTADC,
- "Controller int. temp. raw (at ADC)", IN, -1),
+ "Controller PTAT voltage (ext. ADC)", IN, -1),
SENSOR(CONTROLLER_INTERNAL_TEMP_EXTADC,
- "Controller int. temp. (via ADC)", TEMP, -1),
+ "Controller die temp. (ext. ADC)", TEMP, -1),
SENSOR(AMBIENT_TEMP, "Ambient temp.", TEMP, -1),
SENSOR(AIRFLOW, "Air flow raw", IN, -1),
#undef SENSOR
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH net-next 09/16] sfc: Add new sensor names
2013-12-12 22:56 Pull request: sfc-next 2013-12-12 Ben Hutchings
` (7 preceding siblings ...)
2013-12-12 23:01 ` [PATCH net-next 08/16] sfc: Revise sensor names to be more understandable and consistent Ben Hutchings
@ 2013-12-12 23:01 ` Ben Hutchings
2013-12-12 23:02 ` [PATCH net-next 10/16] sfc: Log all unexpected MCDI errors Ben Hutchings
` (7 subsequent siblings)
16 siblings, 0 replies; 18+ messages in thread
From: Ben Hutchings @ 2013-12-12 23:01 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/mcdi_mon.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/net/ethernet/sfc/mcdi_mon.c b/drivers/net/ethernet/sfc/mcdi_mon.c
index 2891b409d263..bc27d5b580f5 100644
--- a/drivers/net/ethernet/sfc/mcdi_mon.c
+++ b/drivers/net/ethernet/sfc/mcdi_mon.c
@@ -89,6 +89,9 @@ static const struct {
"Controller die temp. (ext. ADC)", TEMP, -1),
SENSOR(AMBIENT_TEMP, "Ambient temp.", TEMP, -1),
SENSOR(AIRFLOW, "Air flow raw", IN, -1),
+ SENSOR(VDD08D_VSS08D_CSR, "0.9V die (int. ADC)", IN, -1),
+ SENSOR(VDD08D_VSS08D_CSR_EXTADC, "0.9V die (ext. ADC)", IN, -1),
+ SENSOR(HOTPOINT_TEMP, "Controller board temp. (hotpoint)", TEMP, -1),
#undef SENSOR
};
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH net-next 10/16] sfc: Log all unexpected MCDI errors
2013-12-12 22:56 Pull request: sfc-next 2013-12-12 Ben Hutchings
` (8 preceding siblings ...)
2013-12-12 23:01 ` [PATCH net-next 09/16] sfc: Add new sensor names Ben Hutchings
@ 2013-12-12 23:02 ` Ben Hutchings
2013-12-12 23:02 ` [PATCH net-next 11/16] sfc: Map MCDI error MC_CMD_ERR_ENOTSUP to Linux EOPNOTSUPP Ben Hutchings
` (6 subsequent siblings)
16 siblings, 0 replies; 18+ messages in thread
From: Ben Hutchings @ 2013-12-12 23:02 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
From: Edward Cree <ecree@solarflare.com>
Split each of efx_mcdi_rpc, efx_mcdi_rpc_finish, and efx_mcdi_rpc_async into
a normal and a _quiet version; made the former log MCDI errors with
netif_err (and include the raw MCDI error code), and the latter never log
them at all. Changed various callers; any where some errors are expected
(but others are not) call the _quiet version and then if necessary log the
MCDI error themselves. Said logging is done by new efx_mcdi_display_error.
Callers of efx_mcdi_rpc*_quiet functions which may want to log the error
need to ensure that their outbuf is big enough to hold an MCDI error; to
this end, they now use MCDI_DECLARE_BUF_OUT_OR_ERR, which always allocates
at least 8 bytes.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/ef10.c | 39 ++--
drivers/net/ethernet/sfc/mcdi.c | 369 +++++++++++++++++++----------------
drivers/net/ethernet/sfc/mcdi.h | 18 ++
drivers/net/ethernet/sfc/mcdi_port.c | 44 +----
drivers/net/ethernet/sfc/ptp.c | 24 ++-
5 files changed, 265 insertions(+), 229 deletions(-)
diff --git a/drivers/net/ethernet/sfc/ef10.c b/drivers/net/ethernet/sfc/ef10.c
index c1b85edcb204..69a0cf7f4f47 100644
--- a/drivers/net/ethernet/sfc/ef10.c
+++ b/drivers/net/ethernet/sfc/ef10.c
@@ -278,11 +278,17 @@ fail1:
static int efx_ef10_free_vis(struct efx_nic *efx)
{
- int rc = efx_mcdi_rpc(efx, MC_CMD_FREE_VIS, NULL, 0, NULL, 0, NULL);
+ MCDI_DECLARE_BUF_OUT_OR_ERR(outbuf, 0);
+ size_t outlen;
+ int rc = efx_mcdi_rpc_quiet(efx, MC_CMD_FREE_VIS, NULL, 0,
+ outbuf, sizeof(outbuf), &outlen);
/* -EALREADY means nothing to free, so ignore */
if (rc == -EALREADY)
rc = 0;
+ if (rc)
+ efx_mcdi_display_error(efx, MC_CMD_FREE_VIS, 0, outbuf, outlen,
+ rc);
return rc;
}
@@ -1244,7 +1250,6 @@ static void efx_ef10_tx_init(struct efx_tx_queue *tx_queue)
fail:
WARN_ON(true);
- netif_err(efx, hw, efx->net_dev, "%s: failed rc=%d\n", __func__, rc);
}
static void efx_ef10_tx_fini(struct efx_tx_queue *tx_queue)
@@ -1258,7 +1263,7 @@ static void efx_ef10_tx_fini(struct efx_tx_queue *tx_queue)
MCDI_SET_DWORD(inbuf, FINI_TXQ_IN_INSTANCE,
tx_queue->queue);
- rc = efx_mcdi_rpc(efx, MC_CMD_FINI_TXQ, inbuf, sizeof(inbuf),
+ rc = efx_mcdi_rpc_quiet(efx, MC_CMD_FINI_TXQ, inbuf, sizeof(inbuf),
outbuf, sizeof(outbuf), &outlen);
if (rc && rc != -EALREADY)
@@ -1267,7 +1272,8 @@ static void efx_ef10_tx_fini(struct efx_tx_queue *tx_queue)
return;
fail:
- netif_err(efx, hw, efx->net_dev, "%s: failed rc=%d\n", __func__, rc);
+ efx_mcdi_display_error(efx, MC_CMD_FINI_TXQ, MC_CMD_FINI_TXQ_IN_LEN,
+ outbuf, outlen, rc);
}
static void efx_ef10_tx_remove(struct efx_tx_queue *tx_queue)
@@ -1482,14 +1488,9 @@ static void efx_ef10_rx_init(struct efx_rx_queue *rx_queue)
rc = efx_mcdi_rpc(efx, MC_CMD_INIT_RXQ, inbuf, inlen,
outbuf, sizeof(outbuf), &outlen);
- if (rc)
- goto fail;
+ WARN_ON(rc);
return;
-
-fail:
- WARN_ON(true);
- netif_err(efx, hw, efx->net_dev, "%s: failed rc=%d\n", __func__, rc);
}
static void efx_ef10_rx_fini(struct efx_rx_queue *rx_queue)
@@ -1503,7 +1504,7 @@ static void efx_ef10_rx_fini(struct efx_rx_queue *rx_queue)
MCDI_SET_DWORD(inbuf, FINI_RXQ_IN_INSTANCE,
efx_rx_queue_index(rx_queue));
- rc = efx_mcdi_rpc(efx, MC_CMD_FINI_RXQ, inbuf, sizeof(inbuf),
+ rc = efx_mcdi_rpc_quiet(efx, MC_CMD_FINI_RXQ, inbuf, sizeof(inbuf),
outbuf, sizeof(outbuf), &outlen);
if (rc && rc != -EALREADY)
@@ -1512,7 +1513,8 @@ static void efx_ef10_rx_fini(struct efx_rx_queue *rx_queue)
return;
fail:
- netif_err(efx, hw, efx->net_dev, "%s: failed rc=%d\n", __func__, rc);
+ efx_mcdi_display_error(efx, MC_CMD_FINI_RXQ, MC_CMD_FINI_RXQ_IN_LEN,
+ outbuf, outlen, rc);
}
static void efx_ef10_rx_remove(struct efx_rx_queue *rx_queue)
@@ -1649,15 +1651,7 @@ static int efx_ef10_ev_init(struct efx_channel *channel)
rc = efx_mcdi_rpc(efx, MC_CMD_INIT_EVQ, inbuf, inlen,
outbuf, sizeof(outbuf), &outlen);
- if (rc)
- goto fail;
-
/* IRQ return is ignored */
-
- return 0;
-
-fail:
- netif_err(efx, hw, efx->net_dev, "%s: failed rc=%d\n", __func__, rc);
return rc;
}
@@ -1671,7 +1665,7 @@ static void efx_ef10_ev_fini(struct efx_channel *channel)
MCDI_SET_DWORD(inbuf, FINI_EVQ_IN_INSTANCE, channel->channel);
- rc = efx_mcdi_rpc(efx, MC_CMD_FINI_EVQ, inbuf, sizeof(inbuf),
+ rc = efx_mcdi_rpc_quiet(efx, MC_CMD_FINI_EVQ, inbuf, sizeof(inbuf),
outbuf, sizeof(outbuf), &outlen);
if (rc && rc != -EALREADY)
@@ -1680,7 +1674,8 @@ static void efx_ef10_ev_fini(struct efx_channel *channel)
return;
fail:
- netif_err(efx, hw, efx->net_dev, "%s: failed rc=%d\n", __func__, rc);
+ efx_mcdi_display_error(efx, MC_CMD_FINI_EVQ, MC_CMD_FINI_EVQ_IN_LEN,
+ outbuf, outlen, rc);
}
static void efx_ef10_ev_remove(struct efx_channel *channel)
diff --git a/drivers/net/ethernet/sfc/mcdi.c b/drivers/net/ethernet/sfc/mcdi.c
index da14e2428944..aee67be3d8c5 100644
--- a/drivers/net/ethernet/sfc/mcdi.c
+++ b/drivers/net/ethernet/sfc/mcdi.c
@@ -42,6 +42,7 @@ struct efx_mcdi_async_param {
unsigned int cmd;
size_t inlen;
size_t outlen;
+ bool quiet;
efx_mcdi_async_completer *complete;
unsigned long cookie;
/* followed by request/response buffer */
@@ -402,8 +403,9 @@ static bool efx_mcdi_complete_async(struct efx_mcdi_iface *mcdi, bool timeout)
{
struct efx_nic *efx = mcdi->efx;
struct efx_mcdi_async_param *async;
- size_t hdr_len, data_len;
+ size_t hdr_len, data_len, err_len;
efx_dword_t *outbuf;
+ MCDI_DECLARE_BUF_OUT_OR_ERR(errbuf, 0);
int rc;
if (cmpxchg(&mcdi->state,
@@ -444,6 +446,13 @@ static bool efx_mcdi_complete_async(struct efx_mcdi_iface *mcdi, bool timeout)
outbuf = (efx_dword_t *)(async + 1);
efx->type->mcdi_read_response(efx, outbuf, hdr_len,
min(async->outlen, data_len));
+ if (!timeout && rc && !async->quiet) {
+ err_len = min(sizeof(errbuf), data_len);
+ efx->type->mcdi_read_response(efx, errbuf, hdr_len,
+ sizeof(errbuf));
+ efx_mcdi_display_error(efx, async->cmd, async->inlen, errbuf,
+ err_len, rc);
+ }
async->complete(efx, async->cookie, rc, outbuf, data_len);
kfree(async);
@@ -519,18 +528,129 @@ efx_mcdi_check_supported(struct efx_nic *efx, unsigned int cmd, size_t inlen)
return 0;
}
+static int _efx_mcdi_rpc_finish(struct efx_nic *efx, unsigned cmd, size_t inlen,
+ efx_dword_t *outbuf, size_t outlen,
+ size_t *outlen_actual, bool quiet)
+{
+ struct efx_mcdi_iface *mcdi = efx_mcdi(efx);
+ MCDI_DECLARE_BUF_OUT_OR_ERR(errbuf, 0);
+ int rc;
+
+ if (mcdi->mode == MCDI_MODE_POLL)
+ rc = efx_mcdi_poll(efx);
+ else
+ rc = efx_mcdi_await_completion(efx);
+
+ if (rc != 0) {
+ netif_err(efx, hw, efx->net_dev,
+ "MC command 0x%x inlen %d mode %d timed out\n",
+ cmd, (int)inlen, mcdi->mode);
+
+ if (mcdi->mode == MCDI_MODE_EVENTS && efx_mcdi_poll_once(efx)) {
+ netif_err(efx, hw, efx->net_dev,
+ "MCDI request was completed without an event\n");
+ rc = 0;
+ }
+
+ /* Close the race with efx_mcdi_ev_cpl() executing just too late
+ * and completing a request we've just cancelled, by ensuring
+ * that the seqno check therein fails.
+ */
+ spin_lock_bh(&mcdi->iface_lock);
+ ++mcdi->seqno;
+ ++mcdi->credits;
+ spin_unlock_bh(&mcdi->iface_lock);
+ }
+
+ if (rc != 0) {
+ if (outlen_actual)
+ *outlen_actual = 0;
+ } else {
+ size_t hdr_len, data_len, err_len;
+
+ /* At the very least we need a memory barrier here to ensure
+ * we pick up changes from efx_mcdi_ev_cpl(). Protect against
+ * a spurious efx_mcdi_ev_cpl() running concurrently by
+ * acquiring the iface_lock. */
+ spin_lock_bh(&mcdi->iface_lock);
+ rc = mcdi->resprc;
+ hdr_len = mcdi->resp_hdr_len;
+ data_len = mcdi->resp_data_len;
+ err_len = min(sizeof(errbuf), data_len);
+ spin_unlock_bh(&mcdi->iface_lock);
+
+ BUG_ON(rc > 0);
+
+ efx->type->mcdi_read_response(efx, outbuf, hdr_len,
+ min(outlen, data_len));
+ if (outlen_actual)
+ *outlen_actual = data_len;
+
+ efx->type->mcdi_read_response(efx, errbuf, hdr_len, err_len);
+
+ if (cmd == MC_CMD_REBOOT && rc == -EIO) {
+ /* Don't reset if MC_CMD_REBOOT returns EIO */
+ } else if (rc == -EIO || rc == -EINTR) {
+ netif_err(efx, hw, efx->net_dev, "MC fatal error %d\n",
+ -rc);
+ efx_schedule_reset(efx, RESET_TYPE_MC_FAILURE);
+ } else if (rc && !quiet) {
+ efx_mcdi_display_error(efx, cmd, inlen, errbuf, err_len,
+ rc);
+ }
+
+ if (rc == -EIO || rc == -EINTR) {
+ msleep(MCDI_STATUS_SLEEP_MS);
+ efx_mcdi_poll_reboot(efx);
+ mcdi->new_epoch = true;
+ }
+ }
+
+ efx_mcdi_release(mcdi);
+ return rc;
+}
+
+static int _efx_mcdi_rpc(struct efx_nic *efx, unsigned cmd,
+ const efx_dword_t *inbuf, size_t inlen,
+ efx_dword_t *outbuf, size_t outlen,
+ size_t *outlen_actual, bool quiet)
+{
+ int rc;
+
+ rc = efx_mcdi_rpc_start(efx, cmd, inbuf, inlen);
+ if (rc) {
+ if (outlen_actual)
+ *outlen_actual = 0;
+ return rc;
+ }
+ return _efx_mcdi_rpc_finish(efx, cmd, inlen, outbuf, outlen,
+ outlen_actual, quiet);
+}
+
int efx_mcdi_rpc(struct efx_nic *efx, unsigned cmd,
const efx_dword_t *inbuf, size_t inlen,
efx_dword_t *outbuf, size_t outlen,
size_t *outlen_actual)
{
- int rc;
+ return _efx_mcdi_rpc(efx, cmd, inbuf, inlen, outbuf, outlen,
+ outlen_actual, false);
+}
- rc = efx_mcdi_rpc_start(efx, cmd, inbuf, inlen);
- if (rc)
- return rc;
- return efx_mcdi_rpc_finish(efx, cmd, inlen,
- outbuf, outlen, outlen_actual);
+/* Normally, on receiving an error code in the MCDI response,
+ * efx_mcdi_rpc will log an error message containing (among other
+ * things) the raw error code, by means of efx_mcdi_display_error.
+ * This _quiet version suppresses that; if the caller wishes to log
+ * the error conditionally on the return code, it should call this
+ * function and is then responsible for calling efx_mcdi_display_error
+ * as needed.
+ */
+int efx_mcdi_rpc_quiet(struct efx_nic *efx, unsigned cmd,
+ const efx_dword_t *inbuf, size_t inlen,
+ efx_dword_t *outbuf, size_t outlen,
+ size_t *outlen_actual)
+{
+ return _efx_mcdi_rpc(efx, cmd, inbuf, inlen, outbuf, outlen,
+ outlen_actual, true);
}
int efx_mcdi_rpc_start(struct efx_nic *efx, unsigned cmd,
@@ -551,30 +671,11 @@ int efx_mcdi_rpc_start(struct efx_nic *efx, unsigned cmd,
return 0;
}
-/**
- * efx_mcdi_rpc_async - Schedule an MCDI command to run asynchronously
- * @efx: NIC through which to issue the command
- * @cmd: Command type number
- * @inbuf: Command parameters
- * @inlen: Length of command parameters, in bytes
- * @outlen: Length to allocate for response buffer, in bytes
- * @complete: Function to be called on completion or cancellation.
- * @cookie: Arbitrary value to be passed to @complete.
- *
- * This function does not sleep and therefore may be called in atomic
- * context. It will fail if event queues are disabled or if MCDI
- * event completions have been disabled due to an error.
- *
- * If it succeeds, the @complete function will be called exactly once
- * in atomic context, when one of the following occurs:
- * (a) the completion event is received (in NAPI context)
- * (b) event queues are disabled (in the process that disables them)
- * (c) the request times-out (in timer context)
- */
-int
-efx_mcdi_rpc_async(struct efx_nic *efx, unsigned int cmd,
- const efx_dword_t *inbuf, size_t inlen, size_t outlen,
- efx_mcdi_async_completer *complete, unsigned long cookie)
+static int _efx_mcdi_rpc_async(struct efx_nic *efx, unsigned int cmd,
+ const efx_dword_t *inbuf, size_t inlen,
+ size_t outlen,
+ efx_mcdi_async_completer *complete,
+ unsigned long cookie, bool quiet)
{
struct efx_mcdi_iface *mcdi = efx_mcdi(efx);
struct efx_mcdi_async_param *async;
@@ -595,6 +696,7 @@ efx_mcdi_rpc_async(struct efx_nic *efx, unsigned int cmd,
async->cmd = cmd;
async->inlen = inlen;
async->outlen = outlen;
+ async->quiet = quiet;
async->complete = complete;
async->cookie = cookie;
memcpy(async + 1, inbuf, inlen);
@@ -623,79 +725,73 @@ efx_mcdi_rpc_async(struct efx_nic *efx, unsigned int cmd,
return rc;
}
+/**
+ * efx_mcdi_rpc_async - Schedule an MCDI command to run asynchronously
+ * @efx: NIC through which to issue the command
+ * @cmd: Command type number
+ * @inbuf: Command parameters
+ * @inlen: Length of command parameters, in bytes
+ * @outlen: Length to allocate for response buffer, in bytes
+ * @complete: Function to be called on completion or cancellation.
+ * @cookie: Arbitrary value to be passed to @complete.
+ *
+ * This function does not sleep and therefore may be called in atomic
+ * context. It will fail if event queues are disabled or if MCDI
+ * event completions have been disabled due to an error.
+ *
+ * If it succeeds, the @complete function will be called exactly once
+ * in atomic context, when one of the following occurs:
+ * (a) the completion event is received (in NAPI context)
+ * (b) event queues are disabled (in the process that disables them)
+ * (c) the request times-out (in timer context)
+ */
+int
+efx_mcdi_rpc_async(struct efx_nic *efx, unsigned int cmd,
+ const efx_dword_t *inbuf, size_t inlen, size_t outlen,
+ efx_mcdi_async_completer *complete, unsigned long cookie)
+{
+ return _efx_mcdi_rpc_async(efx, cmd, inbuf, inlen, outlen, complete,
+ cookie, false);
+}
+
+int efx_mcdi_rpc_async_quiet(struct efx_nic *efx, unsigned int cmd,
+ const efx_dword_t *inbuf, size_t inlen,
+ size_t outlen, efx_mcdi_async_completer *complete,
+ unsigned long cookie)
+{
+ return _efx_mcdi_rpc_async(efx, cmd, inbuf, inlen, outlen, complete,
+ cookie, true);
+}
+
int efx_mcdi_rpc_finish(struct efx_nic *efx, unsigned cmd, size_t inlen,
efx_dword_t *outbuf, size_t outlen,
size_t *outlen_actual)
{
- struct efx_mcdi_iface *mcdi = efx_mcdi(efx);
- int rc;
-
- if (mcdi->mode == MCDI_MODE_POLL)
- rc = efx_mcdi_poll(efx);
- else
- rc = efx_mcdi_await_completion(efx);
-
- if (rc != 0) {
- netif_err(efx, hw, efx->net_dev,
- "MC command 0x%x inlen %d mode %d timed out\n",
- cmd, (int)inlen, mcdi->mode);
-
- if (mcdi->mode == MCDI_MODE_EVENTS && efx_mcdi_poll_once(efx)) {
- netif_err(efx, hw, efx->net_dev,
- "MCDI request was completed without an event\n");
- rc = 0;
- }
-
- /* Close the race with efx_mcdi_ev_cpl() executing just too late
- * and completing a request we've just cancelled, by ensuring
- * that the seqno check therein fails.
- */
- spin_lock_bh(&mcdi->iface_lock);
- ++mcdi->seqno;
- ++mcdi->credits;
- spin_unlock_bh(&mcdi->iface_lock);
- }
-
- if (rc == 0) {
- size_t hdr_len, data_len;
-
- /* At the very least we need a memory barrier here to ensure
- * we pick up changes from efx_mcdi_ev_cpl(). Protect against
- * a spurious efx_mcdi_ev_cpl() running concurrently by
- * acquiring the iface_lock. */
- spin_lock_bh(&mcdi->iface_lock);
- rc = mcdi->resprc;
- hdr_len = mcdi->resp_hdr_len;
- data_len = mcdi->resp_data_len;
- spin_unlock_bh(&mcdi->iface_lock);
-
- BUG_ON(rc > 0);
+ return _efx_mcdi_rpc_finish(efx, cmd, inlen, outbuf, outlen,
+ outlen_actual, false);
+}
- if (rc == 0) {
- efx->type->mcdi_read_response(efx, outbuf, hdr_len,
- min(outlen, data_len));
- if (outlen_actual != NULL)
- *outlen_actual = data_len;
- } else if (cmd == MC_CMD_REBOOT && rc == -EIO)
- ; /* Don't reset if MC_CMD_REBOOT returns EIO */
- else if (rc == -EIO || rc == -EINTR) {
- netif_err(efx, hw, efx->net_dev, "MC fatal error %d\n",
- -rc);
- efx_schedule_reset(efx, RESET_TYPE_MC_FAILURE);
- } else
- netif_dbg(efx, hw, efx->net_dev,
- "MC command 0x%x inlen %d failed rc=%d\n",
- cmd, (int)inlen, -rc);
+int efx_mcdi_rpc_finish_quiet(struct efx_nic *efx, unsigned cmd, size_t inlen,
+ efx_dword_t *outbuf, size_t outlen,
+ size_t *outlen_actual)
+{
+ return _efx_mcdi_rpc_finish(efx, cmd, inlen, outbuf, outlen,
+ outlen_actual, true);
+}
- if (rc == -EIO || rc == -EINTR) {
- msleep(MCDI_STATUS_SLEEP_MS);
- efx_mcdi_poll_reboot(efx);
- mcdi->new_epoch = true;
- }
- }
+void efx_mcdi_display_error(struct efx_nic *efx, unsigned cmd,
+ size_t inlen, efx_dword_t *outbuf,
+ size_t outlen, int rc)
+{
+ int code = 0, err_arg = 0;
- efx_mcdi_release(mcdi);
- return rc;
+ if (outlen >= MC_CMD_ERR_CODE_OFST + 4)
+ code = MCDI_DWORD(outbuf, ERR_CODE);
+ if (outlen >= MC_CMD_ERR_ARG_OFST + 4)
+ err_arg = MCDI_DWORD(outbuf, ERR_ARG);
+ netif_err(efx, hw, efx->net_dev,
+ "MC command 0x%x inlen %d failed rc=%d (raw=%d) arg=%d\n",
+ cmd, (int)inlen, rc, code, err_arg);
}
/* Switch to polled MCDI completions. This can be called in various
@@ -1131,13 +1227,6 @@ int efx_mcdi_log_ctrl(struct efx_nic *efx, bool evq, bool uart, u32 dest_evq)
rc = efx_mcdi_rpc(efx, MC_CMD_LOG_CTRL, inbuf, sizeof(inbuf),
NULL, 0, NULL);
- if (rc)
- goto fail;
-
- return 0;
-
-fail:
- netif_err(efx, hw, efx->net_dev, "%s: failed rc=%d\n", __func__, rc);
return rc;
}
@@ -1254,7 +1343,7 @@ fail1:
static int efx_mcdi_read_assertion(struct efx_nic *efx)
{
MCDI_DECLARE_BUF(inbuf, MC_CMD_GET_ASSERTS_IN_LEN);
- MCDI_DECLARE_BUF(outbuf, MC_CMD_GET_ASSERTS_OUT_LEN);
+ MCDI_DECLARE_BUF_OUT_OR_ERR(outbuf, MC_CMD_GET_ASSERTS_OUT_LEN);
unsigned int flags, index;
const char *reason;
size_t outlen;
@@ -1269,13 +1358,17 @@ static int efx_mcdi_read_assertion(struct efx_nic *efx)
retry = 2;
do {
MCDI_SET_DWORD(inbuf, GET_ASSERTS_IN_CLEAR, 1);
- rc = efx_mcdi_rpc(efx, MC_CMD_GET_ASSERTS,
- inbuf, MC_CMD_GET_ASSERTS_IN_LEN,
- outbuf, sizeof(outbuf), &outlen);
+ rc = efx_mcdi_rpc_quiet(efx, MC_CMD_GET_ASSERTS,
+ inbuf, MC_CMD_GET_ASSERTS_IN_LEN,
+ outbuf, sizeof(outbuf), &outlen);
} while ((rc == -EINTR || rc == -EIO) && retry-- > 0);
- if (rc)
+ if (rc) {
+ efx_mcdi_display_error(efx, MC_CMD_GET_ASSERTS,
+ MC_CMD_GET_ASSERTS_IN_LEN, outbuf,
+ outlen, rc);
return rc;
+ }
if (outlen < MC_CMD_GET_ASSERTS_OUT_LEN)
return -EIO;
@@ -1353,18 +1446,11 @@ void efx_mcdi_set_id_led(struct efx_nic *efx, enum efx_led_mode mode)
rc = efx_mcdi_rpc(efx, MC_CMD_SET_ID_LED, inbuf, sizeof(inbuf),
NULL, 0, NULL);
- if (rc)
- netif_err(efx, hw, efx->net_dev, "%s: failed rc=%d\n",
- __func__, rc);
}
static int efx_mcdi_reset_port(struct efx_nic *efx)
{
- int rc = efx_mcdi_rpc(efx, MC_CMD_ENTITY_RESET, NULL, 0, NULL, 0, NULL);
- if (rc)
- netif_err(efx, hw, efx->net_dev, "%s: failed rc=%d\n",
- __func__, rc);
- return rc;
+ return efx_mcdi_rpc(efx, MC_CMD_ENTITY_RESET, NULL, 0, NULL, 0, NULL);
}
static int efx_mcdi_reset_mc(struct efx_nic *efx)
@@ -1381,7 +1467,6 @@ static int efx_mcdi_reset_mc(struct efx_nic *efx)
return 0;
if (rc == 0)
rc = -EIO;
- netif_err(efx, hw, efx->net_dev, "%s: failed rc=%d\n", __func__, rc);
return rc;
}
@@ -1483,13 +1568,6 @@ int efx_mcdi_wol_filter_remove(struct efx_nic *efx, int id)
rc = efx_mcdi_rpc(efx, MC_CMD_WOL_FILTER_REMOVE, inbuf, sizeof(inbuf),
NULL, 0, NULL);
- if (rc)
- goto fail;
-
- return 0;
-
-fail:
- netif_err(efx, hw, efx->net_dev, "%s: failed rc=%d\n", __func__, rc);
return rc;
}
@@ -1530,13 +1608,6 @@ int efx_mcdi_wol_filter_reset(struct efx_nic *efx)
int rc;
rc = efx_mcdi_rpc(efx, MC_CMD_WOL_FILTER_RESET, NULL, 0, NULL, 0, NULL);
- if (rc)
- goto fail;
-
- return 0;
-
-fail:
- netif_err(efx, hw, efx->net_dev, "%s: failed rc=%d\n", __func__, rc);
return rc;
}
@@ -1566,13 +1637,6 @@ static int efx_mcdi_nvram_update_start(struct efx_nic *efx, unsigned int type)
rc = efx_mcdi_rpc(efx, MC_CMD_NVRAM_UPDATE_START, inbuf, sizeof(inbuf),
NULL, 0, NULL);
- if (rc)
- goto fail;
-
- return 0;
-
-fail:
- netif_err(efx, hw, efx->net_dev, "%s: failed rc=%d\n", __func__, rc);
return rc;
}
@@ -1592,14 +1656,10 @@ static int efx_mcdi_nvram_read(struct efx_nic *efx, unsigned int type,
rc = efx_mcdi_rpc(efx, MC_CMD_NVRAM_READ, inbuf, sizeof(inbuf),
outbuf, sizeof(outbuf), &outlen);
if (rc)
- goto fail;
+ return rc;
memcpy(buffer, MCDI_PTR(outbuf, NVRAM_READ_OUT_READ_BUFFER), length);
return 0;
-
-fail:
- netif_err(efx, hw, efx->net_dev, "%s: failed rc=%d\n", __func__, rc);
- return rc;
}
static int efx_mcdi_nvram_write(struct efx_nic *efx, unsigned int type,
@@ -1619,13 +1679,6 @@ static int efx_mcdi_nvram_write(struct efx_nic *efx, unsigned int type,
rc = efx_mcdi_rpc(efx, MC_CMD_NVRAM_WRITE, inbuf,
ALIGN(MC_CMD_NVRAM_WRITE_IN_LEN(length), 4),
NULL, 0, NULL);
- if (rc)
- goto fail;
-
- return 0;
-
-fail:
- netif_err(efx, hw, efx->net_dev, "%s: failed rc=%d\n", __func__, rc);
return rc;
}
@@ -1643,13 +1696,6 @@ static int efx_mcdi_nvram_erase(struct efx_nic *efx, unsigned int type,
rc = efx_mcdi_rpc(efx, MC_CMD_NVRAM_ERASE, inbuf, sizeof(inbuf),
NULL, 0, NULL);
- if (rc)
- goto fail;
-
- return 0;
-
-fail:
- netif_err(efx, hw, efx->net_dev, "%s: failed rc=%d\n", __func__, rc);
return rc;
}
@@ -1664,13 +1710,6 @@ static int efx_mcdi_nvram_update_finish(struct efx_nic *efx, unsigned int type)
rc = efx_mcdi_rpc(efx, MC_CMD_NVRAM_UPDATE_FINISH, inbuf, sizeof(inbuf),
NULL, 0, NULL);
- if (rc)
- goto fail;
-
- return 0;
-
-fail:
- netif_err(efx, hw, efx->net_dev, "%s: failed rc=%d\n", __func__, rc);
return rc;
}
diff --git a/drivers/net/ethernet/sfc/mcdi.h b/drivers/net/ethernet/sfc/mcdi.h
index d861628b7ee6..171f5f58f84a 100644
--- a/drivers/net/ethernet/sfc/mcdi.h
+++ b/drivers/net/ethernet/sfc/mcdi.h
@@ -116,12 +116,19 @@ void efx_mcdi_fini(struct efx_nic *efx);
int efx_mcdi_rpc(struct efx_nic *efx, unsigned cmd, const efx_dword_t *inbuf,
size_t inlen, efx_dword_t *outbuf, size_t outlen,
size_t *outlen_actual);
+int efx_mcdi_rpc_quiet(struct efx_nic *efx, unsigned cmd,
+ const efx_dword_t *inbuf, size_t inlen,
+ efx_dword_t *outbuf, size_t outlen,
+ size_t *outlen_actual);
int efx_mcdi_rpc_start(struct efx_nic *efx, unsigned cmd,
const efx_dword_t *inbuf, size_t inlen);
int efx_mcdi_rpc_finish(struct efx_nic *efx, unsigned cmd, size_t inlen,
efx_dword_t *outbuf, size_t outlen,
size_t *outlen_actual);
+int efx_mcdi_rpc_finish_quiet(struct efx_nic *efx, unsigned cmd,
+ size_t inlen, efx_dword_t *outbuf,
+ size_t outlen, size_t *outlen_actual);
typedef void efx_mcdi_async_completer(struct efx_nic *efx,
unsigned long cookie, int rc,
@@ -131,6 +138,15 @@ int efx_mcdi_rpc_async(struct efx_nic *efx, unsigned int cmd,
const efx_dword_t *inbuf, size_t inlen, size_t outlen,
efx_mcdi_async_completer *complete,
unsigned long cookie);
+int efx_mcdi_rpc_async_quiet(struct efx_nic *efx, unsigned int cmd,
+ const efx_dword_t *inbuf, size_t inlen,
+ size_t outlen,
+ efx_mcdi_async_completer *complete,
+ unsigned long cookie);
+
+void efx_mcdi_display_error(struct efx_nic *efx, unsigned cmd,
+ size_t inlen, efx_dword_t *outbuf,
+ size_t outlen, int rc);
int efx_mcdi_poll_reboot(struct efx_nic *efx);
void efx_mcdi_mode_poll(struct efx_nic *efx);
@@ -147,6 +163,8 @@ void efx_mcdi_sensor_event(struct efx_nic *efx, efx_qword_t *ev);
*/
#define MCDI_DECLARE_BUF(_name, _len) \
efx_dword_t _name[DIV_ROUND_UP(_len, 4)]
+#define MCDI_DECLARE_BUF_OUT_OR_ERR(_name, _len) \
+ MCDI_DECLARE_BUF(_name, max_t(size_t, _len, 8))
#define _MCDI_PTR(_buf, _offset) \
((u8 *)(_buf) + (_offset))
#define MCDI_PTR(_buf, _field) \
diff --git a/drivers/net/ethernet/sfc/mcdi_port.c b/drivers/net/ethernet/sfc/mcdi_port.c
index 7288aefc2877..91d23252f8fa 100644
--- a/drivers/net/ethernet/sfc/mcdi_port.c
+++ b/drivers/net/ethernet/sfc/mcdi_port.c
@@ -90,13 +90,6 @@ static int efx_mcdi_set_link(struct efx_nic *efx, u32 capabilities,
rc = efx_mcdi_rpc(efx, MC_CMD_SET_LINK, inbuf, sizeof(inbuf),
NULL, 0, NULL);
- if (rc)
- goto fail;
-
- return 0;
-
-fail:
- netif_err(efx, hw, efx->net_dev, "%s: failed rc=%d\n", __func__, rc);
return rc;
}
@@ -143,17 +136,13 @@ static int efx_mcdi_mdio_read(struct net_device *net_dev,
rc = efx_mcdi_rpc(efx, MC_CMD_MDIO_READ, inbuf, sizeof(inbuf),
outbuf, sizeof(outbuf), &outlen);
if (rc)
- goto fail;
+ return rc;
if (MCDI_DWORD(outbuf, MDIO_READ_OUT_STATUS) !=
MC_CMD_MDIO_STATUS_GOOD)
return -EIO;
return (u16)MCDI_DWORD(outbuf, MDIO_READ_OUT_VALUE);
-
-fail:
- netif_err(efx, hw, efx->net_dev, "%s: failed rc=%d\n", __func__, rc);
- return rc;
}
static int efx_mcdi_mdio_write(struct net_device *net_dev,
@@ -174,17 +163,13 @@ static int efx_mcdi_mdio_write(struct net_device *net_dev,
rc = efx_mcdi_rpc(efx, MC_CMD_MDIO_WRITE, inbuf, sizeof(inbuf),
outbuf, sizeof(outbuf), &outlen);
if (rc)
- goto fail;
+ return rc;
if (MCDI_DWORD(outbuf, MDIO_WRITE_OUT_STATUS) !=
MC_CMD_MDIO_STATUS_GOOD)
return -EIO;
return 0;
-
-fail:
- netif_err(efx, hw, efx->net_dev, "%s: failed rc=%d\n", __func__, rc);
- return rc;
}
static u32 mcdi_to_ethtool_cap(u32 media, u32 cap)
@@ -487,17 +472,14 @@ static bool efx_mcdi_phy_poll(struct efx_nic *efx)
rc = efx_mcdi_rpc(efx, MC_CMD_GET_LINK, NULL, 0,
outbuf, sizeof(outbuf), NULL);
- if (rc) {
- netif_err(efx, hw, efx->net_dev, "%s: failed rc=%d\n",
- __func__, rc);
+ if (rc)
efx->link_state.up = false;
- } else {
+ else
efx_mcdi_phy_decode_link(
efx, &efx->link_state,
MCDI_DWORD(outbuf, GET_LINK_OUT_LINK_SPEED),
MCDI_DWORD(outbuf, GET_LINK_OUT_FLAGS),
MCDI_DWORD(outbuf, GET_LINK_OUT_FCNTL));
- }
return !efx_link_state_equal(&efx->link_state, &old_state);
}
@@ -531,11 +513,8 @@ static void efx_mcdi_phy_get_settings(struct efx_nic *efx, struct ethtool_cmd *e
BUILD_BUG_ON(MC_CMD_GET_LINK_IN_LEN != 0);
rc = efx_mcdi_rpc(efx, MC_CMD_GET_LINK, NULL, 0,
outbuf, sizeof(outbuf), NULL);
- if (rc) {
- netif_err(efx, hw, efx->net_dev, "%s: failed rc=%d\n",
- __func__, rc);
+ if (rc)
return;
- }
ecmd->lp_advertising =
mcdi_to_ethtool_cap(phy_cfg->media,
MCDI_DWORD(outbuf, GET_LINK_OUT_LP_CAP));
@@ -918,11 +897,8 @@ bool efx_mcdi_mac_check_fault(struct efx_nic *efx)
rc = efx_mcdi_rpc(efx, MC_CMD_GET_LINK, NULL, 0,
outbuf, sizeof(outbuf), &outlength);
- if (rc) {
- netif_err(efx, hw, efx->net_dev, "%s: failed rc=%d\n",
- __func__, rc);
+ if (rc)
return true;
- }
return MCDI_DWORD(outbuf, GET_LINK_OUT_MAC_FAULT) != 0;
}
@@ -960,14 +936,6 @@ static int efx_mcdi_mac_stats(struct efx_nic *efx,
rc = efx_mcdi_rpc(efx, MC_CMD_MAC_STATS, inbuf, sizeof(inbuf),
NULL, 0, NULL);
- if (rc)
- goto fail;
-
- return 0;
-
-fail:
- netif_err(efx, hw, efx->net_dev, "%s: action %d failed rc=%d\n",
- __func__, action, rc);
return rc;
}
diff --git a/drivers/net/ethernet/sfc/ptp.c b/drivers/net/ethernet/sfc/ptp.c
index 34b25864b121..24bd2e3040f8 100644
--- a/drivers/net/ethernet/sfc/ptp.c
+++ b/drivers/net/ethernet/sfc/ptp.c
@@ -313,6 +313,8 @@ static int efx_phc_enable(struct ptp_clock_info *ptp,
static int efx_ptp_enable(struct efx_nic *efx)
{
MCDI_DECLARE_BUF(inbuf, MC_CMD_PTP_IN_ENABLE_LEN);
+ MCDI_DECLARE_BUF_OUT_OR_ERR(outbuf, 0);
+ int rc;
MCDI_SET_DWORD(inbuf, PTP_IN_OP, MC_CMD_PTP_OP_ENABLE);
MCDI_SET_DWORD(inbuf, PTP_IN_PERIPH_ID, 0);
@@ -320,8 +322,14 @@ static int efx_ptp_enable(struct efx_nic *efx)
efx->ptp_data->channel->channel);
MCDI_SET_DWORD(inbuf, PTP_IN_ENABLE_MODE, efx->ptp_data->mode);
- return efx_mcdi_rpc(efx, MC_CMD_PTP, inbuf, sizeof(inbuf),
- NULL, 0, NULL);
+ rc = efx_mcdi_rpc_quiet(efx, MC_CMD_PTP, inbuf, sizeof(inbuf),
+ outbuf, sizeof(outbuf), NULL);
+ rc = (rc == -EALREADY) ? 0 : rc;
+ if (rc)
+ efx_mcdi_display_error(efx, MC_CMD_PTP,
+ MC_CMD_PTP_IN_ENABLE_LEN,
+ outbuf, sizeof(outbuf), rc);
+ return rc;
}
/* Disable MCDI PTP support.
@@ -332,11 +340,19 @@ static int efx_ptp_enable(struct efx_nic *efx)
static int efx_ptp_disable(struct efx_nic *efx)
{
MCDI_DECLARE_BUF(inbuf, MC_CMD_PTP_IN_DISABLE_LEN);
+ MCDI_DECLARE_BUF_OUT_OR_ERR(outbuf, 0);
+ int rc;
MCDI_SET_DWORD(inbuf, PTP_IN_OP, MC_CMD_PTP_OP_DISABLE);
MCDI_SET_DWORD(inbuf, PTP_IN_PERIPH_ID, 0);
- return efx_mcdi_rpc(efx, MC_CMD_PTP, inbuf, sizeof(inbuf),
- NULL, 0, NULL);
+ rc = efx_mcdi_rpc_quiet(efx, MC_CMD_PTP, inbuf, sizeof(inbuf),
+ outbuf, sizeof(outbuf), NULL);
+ rc = (rc == -EALREADY) ? 0 : rc;
+ if (rc)
+ efx_mcdi_display_error(efx, MC_CMD_PTP,
+ MC_CMD_PTP_IN_DISABLE_LEN,
+ outbuf, sizeof(outbuf), rc);
+ return rc;
}
static void efx_ptp_deliver_rx_queue(struct sk_buff_head *q)
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH net-next 11/16] sfc: Map MCDI error MC_CMD_ERR_ENOTSUP to Linux EOPNOTSUPP
2013-12-12 22:56 Pull request: sfc-next 2013-12-12 Ben Hutchings
` (9 preceding siblings ...)
2013-12-12 23:02 ` [PATCH net-next 10/16] sfc: Log all unexpected MCDI errors Ben Hutchings
@ 2013-12-12 23:02 ` Ben Hutchings
2013-12-12 23:02 ` [PATCH net-next 12/16] sfc: Fold efx_flush_all() into efx_stop_port() and update comments Ben Hutchings
` (5 subsequent siblings)
16 siblings, 0 replies; 18+ messages in thread
From: Ben Hutchings @ 2013-12-12 23:02 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/mcdi.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/net/ethernet/sfc/mcdi.c b/drivers/net/ethernet/sfc/mcdi.c
index aee67be3d8c5..d317dfdee845 100644
--- a/drivers/net/ethernet/sfc/mcdi.c
+++ b/drivers/net/ethernet/sfc/mcdi.c
@@ -192,6 +192,8 @@ static int efx_mcdi_errno(unsigned int mcdi_err)
TRANSLATE_ERROR(EALREADY);
TRANSLATE_ERROR(ENOSPC);
#undef TRANSLATE_ERROR
+ case MC_CMD_ERR_ENOTSUP:
+ return -EOPNOTSUPP;
case MC_CMD_ERR_ALLOC_FAIL:
return -ENOBUFS;
case MC_CMD_ERR_MAC_EXIST:
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH net-next 12/16] sfc: Fold efx_flush_all() into efx_stop_port() and update comments
2013-12-12 22:56 Pull request: sfc-next 2013-12-12 Ben Hutchings
` (10 preceding siblings ...)
2013-12-12 23:02 ` [PATCH net-next 11/16] sfc: Map MCDI error MC_CMD_ERR_ENOTSUP to Linux EOPNOTSUPP Ben Hutchings
@ 2013-12-12 23:02 ` Ben Hutchings
2013-12-12 23:02 ` [PATCH net-next 13/16] sfc: Expose NVRAM_PARTITION_TYPE_LICENSE on EF10 Ben Hutchings
` (4 subsequent siblings)
16 siblings, 0 replies; 18+ messages in thread
From: Ben Hutchings @ 2013-12-12 23:02 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
efx_flush_all() is a really misleading name - it has nothing to do
with e.g. flushing DMA queues. Since it's called immediately after
efx_stop_port() and is highly dependent on what that does, combine
the two functions.
Update comments to explain what this is doing a little better.
Also update an related and erroneous comment in efx_start_port().
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/efx.c | 30 ++++++++++++------------------
1 file changed, 12 insertions(+), 18 deletions(-)
diff --git a/drivers/net/ethernet/sfc/efx.c b/drivers/net/ethernet/sfc/efx.c
index c734fba8c99c..893cd78c3a6c 100644
--- a/drivers/net/ethernet/sfc/efx.c
+++ b/drivers/net/ethernet/sfc/efx.c
@@ -1060,18 +1060,23 @@ static void efx_start_port(struct efx_nic *efx)
mutex_lock(&efx->mac_lock);
efx->port_enabled = true;
- /* efx_mac_work() might have been scheduled after efx_stop_port(),
- * and then cancelled by efx_flush_all() */
+ /* Ensure MAC ingress/egress is enabled */
efx->type->reconfigure_mac(efx);
mutex_unlock(&efx->mac_lock);
}
-/* Prevent efx_mac_work() and efx_monitor() from working */
+/* Cancel work for MAC reconfiguration, periodic hardware monitoring
+ * and the async self-test, wait for them to finish and prevent them
+ * being scheduled again. This doesn't cover online resets, which
+ * should only be cancelled when removing the device.
+ */
static void efx_stop_port(struct efx_nic *efx)
{
netif_dbg(efx, ifdown, efx->net_dev, "stop port\n");
+ EFX_ASSERT_RESET_SERIALISED(efx);
+
mutex_lock(&efx->mac_lock);
efx->port_enabled = false;
mutex_unlock(&efx->mac_lock);
@@ -1079,6 +1084,10 @@ static void efx_stop_port(struct efx_nic *efx)
/* Serialise against efx_set_multicast_list() */
netif_addr_lock_bh(efx->net_dev);
netif_addr_unlock_bh(efx->net_dev);
+
+ cancel_delayed_work_sync(&efx->monitor_work);
+ efx_selftest_async_cancel(efx);
+ cancel_work_sync(&efx->mac_work);
}
static void efx_fini_port(struct efx_nic *efx)
@@ -1690,18 +1699,6 @@ static void efx_start_all(struct efx_nic *efx)
spin_unlock_bh(&efx->stats_lock);
}
-/* Flush all delayed work. Should only be called when no more delayed work
- * will be scheduled. This doesn't flush pending online resets (efx_reset),
- * since we're holding the rtnl_lock at this point. */
-static void efx_flush_all(struct efx_nic *efx)
-{
- /* Make sure the hardware monitor and event self-test are stopped */
- cancel_delayed_work_sync(&efx->monitor_work);
- efx_selftest_async_cancel(efx);
- /* Stop scheduled port reconfigurations */
- cancel_work_sync(&efx->mac_work);
-}
-
/* Quiesce the hardware and software data path, and regular activity
* for the port without bringing the link down. Safe to call multiple
* times with the NIC in almost any state, but interrupts should be
@@ -1725,9 +1722,6 @@ static void efx_stop_all(struct efx_nic *efx)
efx->type->stop_stats(efx);
efx_stop_port(efx);
- /* Flush efx_mac_work(), refill_workqueue, monitor_work */
- efx_flush_all(efx);
-
/* Stop the kernel transmit interface. This is only valid if
* the device is stopped or detached; otherwise the watchdog
* may fire immediately.
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH net-next 13/16] sfc: Expose NVRAM_PARTITION_TYPE_LICENSE on EF10
2013-12-12 22:56 Pull request: sfc-next 2013-12-12 Ben Hutchings
` (11 preceding siblings ...)
2013-12-12 23:02 ` [PATCH net-next 12/16] sfc: Fold efx_flush_all() into efx_stop_port() and update comments Ben Hutchings
@ 2013-12-12 23:02 ` Ben Hutchings
2013-12-12 23:03 ` [PATCH net-next 14/16] sfc: Return EBUSY for filter insertion on EF10, matching Falcon/Siena Ben Hutchings
` (3 subsequent siblings)
16 siblings, 0 replies; 18+ messages in thread
From: Ben Hutchings @ 2013-12-12 23:02 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/ef10.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/net/ethernet/sfc/ef10.c b/drivers/net/ethernet/sfc/ef10.c
index 69a0cf7f4f47..45af374b1df9 100644
--- a/drivers/net/ethernet/sfc/ef10.c
+++ b/drivers/net/ethernet/sfc/ef10.c
@@ -3296,6 +3296,7 @@ static const struct efx_ef10_nvram_type_info efx_ef10_nvram_types[] = {
{ NVRAM_PARTITION_TYPE_EXPROM_CONFIG_PORT1, 0, 1, "sfc_exp_rom_cfg" },
{ NVRAM_PARTITION_TYPE_EXPROM_CONFIG_PORT2, 0, 2, "sfc_exp_rom_cfg" },
{ NVRAM_PARTITION_TYPE_EXPROM_CONFIG_PORT3, 0, 3, "sfc_exp_rom_cfg" },
+ { NVRAM_PARTITION_TYPE_LICENSE, 0, 0, "sfc_license" },
{ NVRAM_PARTITION_TYPE_PHY_MIN, 0xff, 0, "sfc_phy_fw" },
};
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH net-next 14/16] sfc: Return EBUSY for filter insertion on EF10, matching Falcon/Siena
2013-12-12 22:56 Pull request: sfc-next 2013-12-12 Ben Hutchings
` (12 preceding siblings ...)
2013-12-12 23:02 ` [PATCH net-next 13/16] sfc: Expose NVRAM_PARTITION_TYPE_LICENSE on EF10 Ben Hutchings
@ 2013-12-12 23:03 ` Ben Hutchings
2013-12-12 23:03 ` [PATCH net-next 15/16] sfc: Split PTP multicast filter insertion/removal out of efx_ptp_{start,stop}() Ben Hutchings
` (2 subsequent siblings)
16 siblings, 0 replies; 18+ messages in thread
From: Ben Hutchings @ 2013-12-12 23:03 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
The MC firmware will return error MC_CMD_ERR_ENOSPC if filter
insertion fails due to lack of resources. The net driver's filter
implementation for Falcon-architecture returns EBUSY. They should
behave consistently, so for EF10 change ENOSPC to EBUSY.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/ef10.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/net/ethernet/sfc/ef10.c b/drivers/net/ethernet/sfc/ef10.c
index 45af374b1df9..ca922b958a2e 100644
--- a/drivers/net/ethernet/sfc/ef10.c
+++ b/drivers/net/ethernet/sfc/ef10.c
@@ -2259,6 +2259,8 @@ static int efx_ef10_filter_push(struct efx_nic *efx,
outbuf, sizeof(outbuf), NULL);
if (rc == 0)
*handle = MCDI_QWORD(outbuf, FILTER_OP_OUT_HANDLE);
+ if (rc == -ENOSPC)
+ rc = -EBUSY; /* to match efx_farch_filter_insert() */
return rc;
}
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH net-next 15/16] sfc: Split PTP multicast filter insertion/removal out of efx_ptp_{start,stop}()
2013-12-12 22:56 Pull request: sfc-next 2013-12-12 Ben Hutchings
` (13 preceding siblings ...)
2013-12-12 23:03 ` [PATCH net-next 14/16] sfc: Return EBUSY for filter insertion on EF10, matching Falcon/Siena Ben Hutchings
@ 2013-12-12 23:03 ` Ben Hutchings
2013-12-12 23:03 ` [PATCH net-next 16/16] sfc: Remove dependency of PTP on having a dedicated channel Ben Hutchings
2013-12-14 6:11 ` Pull request: sfc-next 2013-12-12 David Miller
16 siblings, 0 replies; 18+ messages in thread
From: Ben Hutchings @ 2013-12-12 23:03 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/ptp.c | 56 +++++++++++++++++++++++++++++-------------
1 file changed, 39 insertions(+), 17 deletions(-)
diff --git a/drivers/net/ethernet/sfc/ptp.c b/drivers/net/ethernet/sfc/ptp.c
index 24bd2e3040f8..7ae66d900256 100644
--- a/drivers/net/ethernet/sfc/ptp.c
+++ b/drivers/net/ethernet/sfc/ptp.c
@@ -755,13 +755,27 @@ static inline void efx_ptp_process_rx(struct efx_nic *efx, struct sk_buff *skb)
local_bh_enable();
}
-static int efx_ptp_start(struct efx_nic *efx)
+static void efx_ptp_remove_multicast_filters(struct efx_nic *efx)
+{
+ struct efx_ptp_data *ptp = efx->ptp_data;
+
+ if (ptp->rxfilter_installed) {
+ efx_filter_remove_id_safe(efx, EFX_FILTER_PRI_REQUIRED,
+ ptp->rxfilter_general);
+ efx_filter_remove_id_safe(efx, EFX_FILTER_PRI_REQUIRED,
+ ptp->rxfilter_event);
+ ptp->rxfilter_installed = false;
+ }
+}
+
+static int efx_ptp_insert_multicast_filters(struct efx_nic *efx)
{
struct efx_ptp_data *ptp = efx->ptp_data;
struct efx_filter_spec rxfilter;
int rc;
- ptp->reset_required = false;
+ if (ptp->rxfilter_installed)
+ return 0;
/* Must filter on both event and general ports to ensure
* that there is no packet re-ordering.
@@ -794,23 +808,37 @@ static int efx_ptp_start(struct efx_nic *efx)
goto fail;
ptp->rxfilter_general = rc;
+ ptp->rxfilter_installed = true;
+ return 0;
+
+fail:
+ efx_filter_remove_id_safe(efx, EFX_FILTER_PRI_REQUIRED,
+ ptp->rxfilter_event);
+ return rc;
+}
+
+static int efx_ptp_start(struct efx_nic *efx)
+{
+ struct efx_ptp_data *ptp = efx->ptp_data;
+ int rc;
+
+ ptp->reset_required = false;
+
+ rc = efx_ptp_insert_multicast_filters(efx);
+ if (rc)
+ return rc;
+
rc = efx_ptp_enable(efx);
if (rc != 0)
- goto fail2;
+ goto fail;
ptp->evt_frag_idx = 0;
ptp->current_adjfreq = 0;
- ptp->rxfilter_installed = true;
return 0;
-fail2:
- efx_filter_remove_id_safe(efx, EFX_FILTER_PRI_REQUIRED,
- ptp->rxfilter_general);
fail:
- efx_filter_remove_id_safe(efx, EFX_FILTER_PRI_REQUIRED,
- ptp->rxfilter_event);
-
+ efx_ptp_remove_multicast_filters(efx);
return rc;
}
@@ -826,13 +854,7 @@ static int efx_ptp_stop(struct efx_nic *efx)
rc = efx_ptp_disable(efx);
- if (ptp->rxfilter_installed) {
- efx_filter_remove_id_safe(efx, EFX_FILTER_PRI_REQUIRED,
- ptp->rxfilter_general);
- efx_filter_remove_id_safe(efx, EFX_FILTER_PRI_REQUIRED,
- ptp->rxfilter_event);
- ptp->rxfilter_installed = false;
- }
+ efx_ptp_remove_multicast_filters(efx);
/* Make sure RX packets are really delivered */
efx_ptp_deliver_rx_queue(&efx->ptp_data->rxq);
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 18+ messages in thread
* [PATCH net-next 16/16] sfc: Remove dependency of PTP on having a dedicated channel
2013-12-12 22:56 Pull request: sfc-next 2013-12-12 Ben Hutchings
` (14 preceding siblings ...)
2013-12-12 23:03 ` [PATCH net-next 15/16] sfc: Split PTP multicast filter insertion/removal out of efx_ptp_{start,stop}() Ben Hutchings
@ 2013-12-12 23:03 ` Ben Hutchings
2013-12-14 6:11 ` Pull request: sfc-next 2013-12-12 David Miller
16 siblings, 0 replies; 18+ messages in thread
From: Ben Hutchings @ 2013-12-12 23:03 UTC (permalink / raw)
To: David Miller; +Cc: netdev, linux-net-drivers
We need a dedicated channel on Siena to ensure we can match up
the separate RX and timestamp events for each PTP packet. We won't
do this for EF10 as timestamps are delivered inline.
Pass a channel index of 0 to MC_CMD_PTP_OP_ENABLE when there is no
dedicated channel.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
---
drivers/net/ethernet/sfc/nic.h | 4 ++-
drivers/net/ethernet/sfc/ptp.c | 59 +++++++++++++++++++++++++---------------
drivers/net/ethernet/sfc/siena.c | 2 +-
3 files changed, 41 insertions(+), 24 deletions(-)
diff --git a/drivers/net/ethernet/sfc/nic.h b/drivers/net/ethernet/sfc/nic.h
index 7ac9c000696f..25178536f053 100644
--- a/drivers/net/ethernet/sfc/nic.h
+++ b/drivers/net/ethernet/sfc/nic.h
@@ -554,7 +554,9 @@ int efx_sriov_set_vf_spoofchk(struct net_device *net_dev, int vf,
bool spoofchk);
struct ethtool_ts_info;
-void efx_ptp_probe(struct efx_nic *efx);
+int efx_ptp_probe(struct efx_nic *efx, struct efx_channel *channel);
+void efx_ptp_defer_probe_with_channel(struct efx_nic *efx);
+void efx_ptp_remove(struct efx_nic *efx);
int efx_ptp_set_ts_config(struct efx_nic *efx, struct ifreq *ifr);
int efx_ptp_get_ts_config(struct efx_nic *efx, struct ifreq *ifr);
void efx_ptp_get_ts_info(struct efx_nic *efx, struct ethtool_ts_info *ts_info);
diff --git a/drivers/net/ethernet/sfc/ptp.c b/drivers/net/ethernet/sfc/ptp.c
index 7ae66d900256..aa4876edaac6 100644
--- a/drivers/net/ethernet/sfc/ptp.c
+++ b/drivers/net/ethernet/sfc/ptp.c
@@ -214,7 +214,8 @@ struct efx_ptp_timeset {
/**
* struct efx_ptp_data - Precision Time Protocol (PTP) state
- * @channel: The PTP channel
+ * @efx: The NIC context
+ * @channel: The PTP channel (Siena only)
* @rxq: Receive queue (awaiting timestamps)
* @txq: Transmit queue
* @evt_list: List of MC receive events awaiting packets
@@ -265,6 +266,7 @@ struct efx_ptp_timeset {
* @timeset: Last set of synchronisation statistics.
*/
struct efx_ptp_data {
+ struct efx_nic *efx;
struct efx_channel *channel;
struct sk_buff_head rxq;
struct sk_buff_head txq;
@@ -319,7 +321,8 @@ static int efx_ptp_enable(struct efx_nic *efx)
MCDI_SET_DWORD(inbuf, PTP_IN_OP, MC_CMD_PTP_OP_ENABLE);
MCDI_SET_DWORD(inbuf, PTP_IN_PERIPH_ID, 0);
MCDI_SET_DWORD(inbuf, PTP_IN_ENABLE_QUEUE,
- efx->ptp_data->channel->channel);
+ efx->ptp_data->channel ?
+ efx->ptp_data->channel->channel : 0);
MCDI_SET_DWORD(inbuf, PTP_IN_ENABLE_MODE, efx->ptp_data->mode);
rc = efx_mcdi_rpc_quiet(efx, MC_CMD_PTP, inbuf, sizeof(inbuf),
@@ -774,7 +777,7 @@ static int efx_ptp_insert_multicast_filters(struct efx_nic *efx)
struct efx_filter_spec rxfilter;
int rc;
- if (ptp->rxfilter_installed)
+ if (!ptp->channel || ptp->rxfilter_installed)
return 0;
/* Must filter on both event and general ports to ensure
@@ -882,7 +885,7 @@ static void efx_ptp_pps_worker(struct work_struct *work)
{
struct efx_ptp_data *ptp =
container_of(work, struct efx_ptp_data, pps_work);
- struct efx_nic *efx = ptp->channel->efx;
+ struct efx_nic *efx = ptp->efx;
struct ptp_clock_event ptp_evt;
if (efx_ptp_synchronize(efx, PTP_SYNC_ATTEMPTS))
@@ -899,7 +902,7 @@ static void efx_ptp_worker(struct work_struct *work)
{
struct efx_ptp_data *ptp_data =
container_of(work, struct efx_ptp_data, work);
- struct efx_nic *efx = ptp_data->channel->efx;
+ struct efx_nic *efx = ptp_data->efx;
struct sk_buff *skb;
struct sk_buff_head tempq;
@@ -923,31 +926,25 @@ static void efx_ptp_worker(struct work_struct *work)
efx_ptp_process_rx(efx, skb);
}
-/* Initialise PTP channel and state.
- *
- * Setting core_index to zero causes the queue to be initialised and doesn't
- * overlap with 'rxq0' because ptp.c doesn't use skb_record_rx_queue.
- */
-static int efx_ptp_probe_channel(struct efx_channel *channel)
+/* Initialise PTP state. */
+int efx_ptp_probe(struct efx_nic *efx, struct efx_channel *channel)
{
- struct efx_nic *efx = channel->efx;
struct efx_ptp_data *ptp;
int rc = 0;
unsigned int pos;
- channel->irq_moderation = 0;
- channel->rx_queue.core_index = 0;
-
ptp = kzalloc(sizeof(struct efx_ptp_data), GFP_KERNEL);
efx->ptp_data = ptp;
if (!efx->ptp_data)
return -ENOMEM;
+ ptp->efx = efx;
+ ptp->channel = channel;
+
rc = efx_nic_alloc_buffer(efx, &ptp->start, sizeof(int), GFP_KERNEL);
if (rc != 0)
goto fail1;
- ptp->channel = channel;
skb_queue_head_init(&ptp->rxq);
skb_queue_head_init(&ptp->txq);
ptp->workwq = create_singlethread_workqueue("sfc_ptp");
@@ -1014,14 +1011,27 @@ fail1:
return rc;
}
-static void efx_ptp_remove_channel(struct efx_channel *channel)
+/* Initialise PTP channel.
+ *
+ * Setting core_index to zero causes the queue to be initialised and doesn't
+ * overlap with 'rxq0' because ptp.c doesn't use skb_record_rx_queue.
+ */
+static int efx_ptp_probe_channel(struct efx_channel *channel)
{
struct efx_nic *efx = channel->efx;
+ channel->irq_moderation = 0;
+ channel->rx_queue.core_index = 0;
+
+ return efx_ptp_probe(efx, channel);
+}
+
+void efx_ptp_remove(struct efx_nic *efx)
+{
if (!efx->ptp_data)
return;
- (void)efx_ptp_disable(channel->efx);
+ (void)efx_ptp_disable(efx);
cancel_work_sync(&efx->ptp_data->work);
cancel_work_sync(&efx->ptp_data->pps_work);
@@ -1038,6 +1048,11 @@ static void efx_ptp_remove_channel(struct efx_channel *channel)
kfree(efx->ptp_data);
}
+static void efx_ptp_remove_channel(struct efx_channel *channel)
+{
+ efx_ptp_remove(channel->efx);
+}
+
static void efx_ptp_get_channel_name(struct efx_channel *channel,
char *buf, size_t len)
{
@@ -1449,7 +1464,7 @@ static int efx_phc_adjfreq(struct ptp_clock_info *ptp, s32 delta)
struct efx_ptp_data *ptp_data = container_of(ptp,
struct efx_ptp_data,
phc_clock_info);
- struct efx_nic *efx = ptp_data->channel->efx;
+ struct efx_nic *efx = ptp_data->efx;
MCDI_DECLARE_BUF(inadj, MC_CMD_PTP_IN_ADJUST_LEN);
s64 adjustment_ns;
int rc;
@@ -1482,7 +1497,7 @@ static int efx_phc_adjtime(struct ptp_clock_info *ptp, s64 delta)
struct efx_ptp_data *ptp_data = container_of(ptp,
struct efx_ptp_data,
phc_clock_info);
- struct efx_nic *efx = ptp_data->channel->efx;
+ struct efx_nic *efx = ptp_data->efx;
struct timespec delta_ts = ns_to_timespec(delta);
MCDI_DECLARE_BUF(inbuf, MC_CMD_PTP_IN_ADJUST_LEN);
@@ -1500,7 +1515,7 @@ static int efx_phc_gettime(struct ptp_clock_info *ptp, struct timespec *ts)
struct efx_ptp_data *ptp_data = container_of(ptp,
struct efx_ptp_data,
phc_clock_info);
- struct efx_nic *efx = ptp_data->channel->efx;
+ struct efx_nic *efx = ptp_data->efx;
MCDI_DECLARE_BUF(inbuf, MC_CMD_PTP_IN_READ_NIC_TIME_LEN);
MCDI_DECLARE_BUF(outbuf, MC_CMD_PTP_OUT_READ_NIC_TIME_LEN);
int rc;
@@ -1566,7 +1581,7 @@ static const struct efx_channel_type efx_ptp_channel_type = {
.keep_eventq = false,
};
-void efx_ptp_probe(struct efx_nic *efx)
+void efx_ptp_defer_probe_with_channel(struct efx_nic *efx)
{
/* Check whether PTP is implemented on this NIC. The DISABLE
* operation will succeed if and only if it is implemented.
diff --git a/drivers/net/ethernet/sfc/siena.c b/drivers/net/ethernet/sfc/siena.c
index f2a7ad4c76aa..53b4ce67be0e 100644
--- a/drivers/net/ethernet/sfc/siena.c
+++ b/drivers/net/ethernet/sfc/siena.c
@@ -259,7 +259,7 @@ static int siena_probe_nic(struct efx_nic *efx)
goto fail5;
efx_sriov_probe(efx);
- efx_ptp_probe(efx);
+ efx_ptp_defer_probe_with_channel(efx);
return 0;
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
^ permalink raw reply related [flat|nested] 18+ messages in thread
* Re: Pull request: sfc-next 2013-12-12
2013-12-12 22:56 Pull request: sfc-next 2013-12-12 Ben Hutchings
` (15 preceding siblings ...)
2013-12-12 23:03 ` [PATCH net-next 16/16] sfc: Remove dependency of PTP on having a dedicated channel Ben Hutchings
@ 2013-12-14 6:11 ` David Miller
16 siblings, 0 replies; 18+ messages in thread
From: David Miller @ 2013-12-14 6:11 UTC (permalink / raw)
To: bhutchings; +Cc: netdev, linux-net-drivers
From: Ben Hutchings <bhutchings@solarflare.com>
Date: Thu, 12 Dec 2013 22:56:17 +0000
> An assortment of changes for Linux 3.14:
>
> 1. Merge the sfc fixes that you have already merged into net.git.
> (The branch point for those was such that this does not bring in any
> other changes.)
> 2. Reduce log level for a generally useless warning message, from
> Robert Stonehouse.
> 3. Include BISTs in ethtool offline self-test for EF10 and recover from
> BISTs initiated through other functions, from Jon Cooper.
> 4. Improve a sanity check on RX completions.
> 5. Avoid incrementing RX dropped count while the interface is down, from
> Jon Cooper.
> 6. Improve hardware sensor naming and log messages, from Edward Cree.
> 7. Log all unexpected errors returned by firmware, from Edward Cree.
> 8. Expose another NVRAM partition to userland.
> 9. Some refactoring of the PTP code in preparation for EF10 support.
> 10. Various minor cleanups.
Pulled, thanks Ben.
^ permalink raw reply [flat|nested] 18+ messages in thread
end of thread, other threads:[~2013-12-14 6:11 UTC | newest]
Thread overview: 18+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2013-12-12 22:56 Pull request: sfc-next 2013-12-12 Ben Hutchings
2013-12-12 22:59 ` [PATCH net-next 01/16] sfc: Demote "MC Scheduler error" messages Ben Hutchings
2013-12-12 22:59 ` [PATCH net-next 02/16] sfc: Update MCDI protocol definitions Ben Hutchings
2013-12-12 22:59 ` [PATCH net-next 03/16] sfc: Add MC BISTs to ethtool offline self test on EF10 Ben Hutchings
2013-12-12 23:00 ` [PATCH net-next 04/16] sfc: Tighten the check for RX merged completion events Ben Hutchings
2013-12-12 23:00 ` [PATCH net-next 05/16] sfc: Make initial fill of RX descriptors synchronous Ben Hutchings
2013-12-12 23:00 ` [PATCH net-next 06/16] sfc: Correct RX dropped count for drops while interface is down Ben Hutchings
2013-12-12 23:01 ` [PATCH net-next 07/16] sfc: Report units in sensor warnings Ben Hutchings
2013-12-12 23:01 ` [PATCH net-next 08/16] sfc: Revise sensor names to be more understandable and consistent Ben Hutchings
2013-12-12 23:01 ` [PATCH net-next 09/16] sfc: Add new sensor names Ben Hutchings
2013-12-12 23:02 ` [PATCH net-next 10/16] sfc: Log all unexpected MCDI errors Ben Hutchings
2013-12-12 23:02 ` [PATCH net-next 11/16] sfc: Map MCDI error MC_CMD_ERR_ENOTSUP to Linux EOPNOTSUPP Ben Hutchings
2013-12-12 23:02 ` [PATCH net-next 12/16] sfc: Fold efx_flush_all() into efx_stop_port() and update comments Ben Hutchings
2013-12-12 23:02 ` [PATCH net-next 13/16] sfc: Expose NVRAM_PARTITION_TYPE_LICENSE on EF10 Ben Hutchings
2013-12-12 23:03 ` [PATCH net-next 14/16] sfc: Return EBUSY for filter insertion on EF10, matching Falcon/Siena Ben Hutchings
2013-12-12 23:03 ` [PATCH net-next 15/16] sfc: Split PTP multicast filter insertion/removal out of efx_ptp_{start,stop}() Ben Hutchings
2013-12-12 23:03 ` [PATCH net-next 16/16] sfc: Remove dependency of PTP on having a dedicated channel Ben Hutchings
2013-12-14 6:11 ` Pull request: sfc-next 2013-12-12 David Miller
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).