Skip to content

Commit a90ec84

Browse files
vcgomesanguy11
vcgomes
authored and
anguy11
committed
igc: Add support for PTP getcrosststamp()
i225 supports PCIe Precision Time Measurement (PTM), allowing us to support the PTP_SYS_OFFSET_PRECISE ioctl() in the driver via the getcrosststamp() function. The easiest way to expose the PTM registers would be to configure the PTM dialogs to run periodically, but the PTP_SYS_OFFSET_PRECISE ioctl() semantics are more aligned to using a kind of "one-shot" way of retrieving the PTM timestamps. But this causes a bit more code to be written: the trigger registers for the PTM dialogs are not cleared automatically. i225 can be configured to send "fake" packets with the PTM information, adding support for handling these types of packets is left for the future. PTM improves the accuracy of time synchronization, for example, using phc2sys, while a simple application is sending packets as fast as possible. First, without .getcrosststamp(): phc2sys[191.382]: enp4s0 sys offset -959 s2 freq -454 delay 4492 phc2sys[191.482]: enp4s0 sys offset 798 s2 freq +1015 delay 4069 phc2sys[191.583]: enp4s0 sys offset 962 s2 freq +1418 delay 3849 phc2sys[191.683]: enp4s0 sys offset 924 s2 freq +1669 delay 3753 phc2sys[191.783]: enp4s0 sys offset 664 s2 freq +1686 delay 3349 phc2sys[191.883]: enp4s0 sys offset 218 s2 freq +1439 delay 2585 phc2sys[191.983]: enp4s0 sys offset 761 s2 freq +2048 delay 3750 phc2sys[192.083]: enp4s0 sys offset 756 s2 freq +2271 delay 4061 phc2sys[192.183]: enp4s0 sys offset 809 s2 freq +2551 delay 4384 phc2sys[192.283]: enp4s0 sys offset -108 s2 freq +1877 delay 2480 phc2sys[192.383]: enp4s0 sys offset -1145 s2 freq +807 delay 4438 phc2sys[192.484]: enp4s0 sys offset 571 s2 freq +2180 delay 3849 phc2sys[192.584]: enp4s0 sys offset 241 s2 freq +2021 delay 3389 phc2sys[192.684]: enp4s0 sys offset 405 s2 freq +2257 delay 3829 phc2sys[192.784]: enp4s0 sys offset 17 s2 freq +1991 delay 3273 phc2sys[192.884]: enp4s0 sys offset 152 s2 freq +2131 delay 3948 phc2sys[192.984]: enp4s0 sys offset -187 s2 freq +1837 delay 3162 phc2sys[193.084]: enp4s0 sys offset -1595 s2 freq +373 delay 4557 phc2sys[193.184]: enp4s0 sys offset 107 s2 freq +1597 delay 3740 phc2sys[193.284]: enp4s0 sys offset 199 s2 freq +1721 delay 4010 phc2sys[193.385]: enp4s0 sys offset -169 s2 freq +1413 delay 3701 phc2sys[193.485]: enp4s0 sys offset -47 s2 freq +1484 delay 3581 phc2sys[193.585]: enp4s0 sys offset -65 s2 freq +1452 delay 3778 phc2sys[193.685]: enp4s0 sys offset 95 s2 freq +1592 delay 3888 phc2sys[193.785]: enp4s0 sys offset 206 s2 freq +1732 delay 4445 phc2sys[193.885]: enp4s0 sys offset -652 s2 freq +936 delay 2521 phc2sys[193.985]: enp4s0 sys offset -203 s2 freq +1189 delay 3391 phc2sys[194.085]: enp4s0 sys offset -376 s2 freq +955 delay 2951 phc2sys[194.185]: enp4s0 sys offset -134 s2 freq +1084 delay 3330 phc2sys[194.285]: enp4s0 sys offset -22 s2 freq +1156 delay 3479 phc2sys[194.386]: enp4s0 sys offset 32 s2 freq +1204 delay 3602 phc2sys[194.486]: enp4s0 sys offset 122 s2 freq +1303 delay 3731 Statistics for this run (total of 2179 lines), in nanoseconds: average: -1.12 stdev: 634.80 max: 1551 min: -2215 With .getcrosststamp() via PCIe PTM: phc2sys[367.859]: enp4s0 sys offset 6 s2 freq +1727 delay 0 phc2sys[367.959]: enp4s0 sys offset -2 s2 freq +1721 delay 0 phc2sys[368.059]: enp4s0 sys offset 5 s2 freq +1727 delay 0 phc2sys[368.160]: enp4s0 sys offset -1 s2 freq +1723 delay 0 phc2sys[368.260]: enp4s0 sys offset -4 s2 freq +1719 delay 0 phc2sys[368.360]: enp4s0 sys offset -5 s2 freq +1717 delay 0 phc2sys[368.460]: enp4s0 sys offset 1 s2 freq +1722 delay 0 phc2sys[368.560]: enp4s0 sys offset -3 s2 freq +1718 delay 0 phc2sys[368.660]: enp4s0 sys offset 5 s2 freq +1725 delay 0 phc2sys[368.760]: enp4s0 sys offset -1 s2 freq +1721 delay 0 phc2sys[368.860]: enp4s0 sys offset 0 s2 freq +1721 delay 0 phc2sys[368.960]: enp4s0 sys offset 0 s2 freq +1721 delay 0 phc2sys[369.061]: enp4s0 sys offset 4 s2 freq +1725 delay 0 phc2sys[369.161]: enp4s0 sys offset 1 s2 freq +1724 delay 0 phc2sys[369.261]: enp4s0 sys offset 4 s2 freq +1727 delay 0 phc2sys[369.361]: enp4s0 sys offset 8 s2 freq +1732 delay 0 phc2sys[369.461]: enp4s0 sys offset 7 s2 freq +1733 delay 0 phc2sys[369.561]: enp4s0 sys offset 4 s2 freq +1733 delay 0 phc2sys[369.661]: enp4s0 sys offset 1 s2 freq +1731 delay 0 phc2sys[369.761]: enp4s0 sys offset 1 s2 freq +1731 delay 0 phc2sys[369.861]: enp4s0 sys offset -5 s2 freq +1725 delay 0 phc2sys[369.961]: enp4s0 sys offset -4 s2 freq +1725 delay 0 phc2sys[370.062]: enp4s0 sys offset 2 s2 freq +1730 delay 0 phc2sys[370.162]: enp4s0 sys offset -7 s2 freq +1721 delay 0 phc2sys[370.262]: enp4s0 sys offset -3 s2 freq +1723 delay 0 phc2sys[370.362]: enp4s0 sys offset 1 s2 freq +1726 delay 0 phc2sys[370.462]: enp4s0 sys offset -3 s2 freq +1723 delay 0 phc2sys[370.562]: enp4s0 sys offset -1 s2 freq +1724 delay 0 phc2sys[370.662]: enp4s0 sys offset -4 s2 freq +1720 delay 0 phc2sys[370.762]: enp4s0 sys offset -7 s2 freq +1716 delay 0 phc2sys[370.862]: enp4s0 sys offset -2 s2 freq +1719 delay 0 Statistics for this run (total of 2179 lines), in nanoseconds: average: 0.14 stdev: 5.03 max: 48 min: -27 For reference, the statistics for runs without PCIe congestion show that the improvements from enabling PTM are less dramatic. For two runs of 16466 entries: without PTM: avg -0.04 stdev 10.57 max 39 min -42 with PTM: avg 0.01 stdev 4.20 max 19 min -16 One possible explanation is that when PTM is not enabled, and there's a lot of traffic in the PCIe fabric, some register reads will take more time than the others because of congestion on the PCIe fabric. When PTM is enabled, even if the PTM dialogs take more time to complete under heavy traffic, the time measurements do not depend on the time to read the registers. This was implemented following the i225 EAS version 0.993. Signed-off-by: Vinicius Costa Gomes <vinicius.gomes@intel.com> Tested-by: Dvora Fuxbrumer <dvorax.fuxbrumer@linux.intel.com> Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
1 parent 1b5d73f commit a90ec84

File tree

4 files changed

+234
-0
lines changed

4 files changed

+234
-0
lines changed

‎drivers/net/ethernet/intel/igc/igc.h‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -227,6 +227,7 @@ struct igc_adapter {
227227
struct timecounter tc;
228228
struct timespec64 prev_ptp_time; /* Pre-reset PTP clock */
229229
ktime_t ptp_reset_start; /* Reset time in clock mono */
230+
struct system_time_snapshot snapshot;
230231

231232
char fw_version[32];
232233

‎drivers/net/ethernet/intel/igc/igc_defines.h‎

Lines changed: 31 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -523,6 +523,37 @@
523523
#define IGC_RXCSUM_CRCOFL 0x00000800 /* CRC32 offload enable */
524524
#define IGC_RXCSUM_PCSD 0x00002000 /* packet checksum disabled */
525525

526+
/* PCIe PTM Control */
527+
#define IGC_PTM_CTRL_START_NOW BIT(29) /* Start PTM Now */
528+
#define IGC_PTM_CTRL_EN BIT(30) /* Enable PTM */
529+
#define IGC_PTM_CTRL_TRIG BIT(31) /* PTM Cycle trigger */
530+
#define IGC_PTM_CTRL_SHRT_CYC(usec) (((usec) & 0x2f) << 2)
531+
#define IGC_PTM_CTRL_PTM_TO(usec) (((usec) & 0xff) << 8)
532+
533+
#define IGC_PTM_SHORT_CYC_DEFAULT 10 /* Default Short/interrupted cycle interval */
534+
#define IGC_PTM_CYC_TIME_DEFAULT 5 /* Default PTM cycle time */
535+
#define IGC_PTM_TIMEOUT_DEFAULT 255 /* Default timeout for PTM errors */
536+
537+
/* PCIe Digital Delay */
538+
#define IGC_PCIE_DIG_DELAY_DEFAULT 0x01440000
539+
540+
/* PCIe PHY Delay */
541+
#define IGC_PCIE_PHY_DELAY_DEFAULT 0x40900000
542+
543+
#define IGC_TIMADJ_ADJUST_METH 0x40000000
544+
545+
/* PCIe PTM Status */
546+
#define IGC_PTM_STAT_VALID BIT(0) /* PTM Status */
547+
#define IGC_PTM_STAT_RET_ERR BIT(1) /* Root port timeout */
548+
#define IGC_PTM_STAT_BAD_PTM_RES BIT(2) /* PTM Response msg instead of PTM Response Data */
549+
#define IGC_PTM_STAT_T4M1_OVFL BIT(3) /* T4 minus T1 overflow */
550+
#define IGC_PTM_STAT_ADJUST_1ST BIT(4) /* 1588 timer adjusted during 1st PTM cycle */
551+
#define IGC_PTM_STAT_ADJUST_CYC BIT(5) /* 1588 timer adjusted during non-1st PTM cycle */
552+
553+
/* PCIe PTM Cycle Control */
554+
#define IGC_PTM_CYCLE_CTRL_CYC_TIME(msec) ((msec) & 0x3ff) /* PTM Cycle Time (msec) */
555+
#define IGC_PTM_CYCLE_CTRL_AUTO_CYC_EN BIT(31) /* PTM Cycle Control */
556+
526557
/* GPY211 - I225 defines */
527558
#define GPY_MMD_MASK 0xFFFF0000
528559
#define GPY_MMD_SHIFT 16

‎drivers/net/ethernet/intel/igc/igc_ptp.c‎

Lines changed: 179 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -9,13 +9,18 @@
99
#include <linux/ptp_classify.h>
1010
#include <linux/clocksource.h>
1111
#include <linux/ktime.h>
12+
#include <linux/delay.h>
13+
#include <linux/iopoll.h>
1214

1315
#define INCVALUE_MASK 0x7fffffff
1416
#define ISGN 0x80000000
1517

1618
#define IGC_SYSTIM_OVERFLOW_PERIOD (HZ * 60 * 9)
1719
#define IGC_PTP_TX_TIMEOUT (HZ * 15)
1820

21+
#define IGC_PTM_STAT_SLEEP 2
22+
#define IGC_PTM_STAT_TIMEOUT 100
23+
1924
/* SYSTIM read access for I225 */
2025
void igc_ptp_read(struct igc_adapter *adapter, struct timespec64 *ts)
2126
{
@@ -752,6 +757,147 @@ int igc_ptp_get_ts_config(struct net_device *netdev, struct ifreq *ifr)
752757
-EFAULT : 0;
753758
}
754759

760+
/* The two conditions below must be met for cross timestamping via
761+
* PCIe PTM:
762+
*
763+
* 1. We have an way to convert the timestamps in the PTM messages
764+
* to something related to the system clocks (right now, only
765+
* X86 systems with support for the Always Running Timer allow that);
766+
*
767+
* 2. We have PTM enabled in the path from the device to the PCIe root port.
768+
*/
769+
static bool igc_is_crosststamp_supported(struct igc_adapter *adapter)
770+
{
771+
return IS_ENABLED(CONFIG_X86_TSC) ? pcie_ptm_enabled(adapter->pdev) : false;
772+
}
773+
774+
static struct system_counterval_t igc_device_tstamp_to_system(u64 tstamp)
775+
{
776+
#if IS_ENABLED(CONFIG_X86_TSC)
777+
return convert_art_ns_to_tsc(tstamp);
778+
#else
779+
return (struct system_counterval_t) { };
780+
#endif
781+
}
782+
783+
static void igc_ptm_log_error(struct igc_adapter *adapter, u32 ptm_stat)
784+
{
785+
struct net_device *netdev = adapter->netdev;
786+
787+
switch (ptm_stat) {
788+
case IGC_PTM_STAT_RET_ERR:
789+
netdev_err(netdev, "PTM Error: Root port timeout\n");
790+
break;
791+
case IGC_PTM_STAT_BAD_PTM_RES:
792+
netdev_err(netdev, "PTM Error: Bad response, PTM Response Data expected\n");
793+
break;
794+
case IGC_PTM_STAT_T4M1_OVFL:
795+
netdev_err(netdev, "PTM Error: T4 minus T1 overflow\n");
796+
break;
797+
case IGC_PTM_STAT_ADJUST_1ST:
798+
netdev_err(netdev, "PTM Error: 1588 timer adjusted during first PTM cycle\n");
799+
break;
800+
case IGC_PTM_STAT_ADJUST_CYC:
801+
netdev_err(netdev, "PTM Error: 1588 timer adjusted during non-first PTM cycle\n");
802+
break;
803+
default:
804+
netdev_err(netdev, "PTM Error: Unknown error (%#x)\n", ptm_stat);
805+
break;
806+
}
807+
}
808+
809+
static int igc_phc_get_syncdevicetime(ktime_t *device,
810+
struct system_counterval_t *system,
811+
void *ctx)
812+
{
813+
u32 stat, t2_curr_h, t2_curr_l, ctrl;
814+
struct igc_adapter *adapter = ctx;
815+
struct igc_hw *hw = &adapter->hw;
816+
int err, count = 100;
817+
ktime_t t1, t2_curr;
818+
819+
/* Get a snapshot of system clocks to use as historic value. */
820+
ktime_get_snapshot(&adapter->snapshot);
821+
822+
do {
823+
/* Doing this in a loop because in the event of a
824+
* badly timed (ha!) system clock adjustment, we may
825+
* get PTM errors from the PCI root, but these errors
826+
* are transitory. Repeating the process returns valid
827+
* data eventually.
828+
*/
829+
830+
/* To "manually" start the PTM cycle we need to clear and
831+
* then set again the TRIG bit.
832+
*/
833+
ctrl = rd32(IGC_PTM_CTRL);
834+
ctrl &= ~IGC_PTM_CTRL_TRIG;
835+
wr32(IGC_PTM_CTRL, ctrl);
836+
ctrl |= IGC_PTM_CTRL_TRIG;
837+
wr32(IGC_PTM_CTRL, ctrl);
838+
839+
/* The cycle only starts "for real" when software notifies
840+
* that it has read the registers, this is done by setting
841+
* VALID bit.
842+
*/
843+
wr32(IGC_PTM_STAT, IGC_PTM_STAT_VALID);
844+
845+
err = readx_poll_timeout(rd32, IGC_PTM_STAT, stat,
846+
stat, IGC_PTM_STAT_SLEEP,
847+
IGC_PTM_STAT_TIMEOUT);
848+
if (err < 0) {
849+
netdev_err(adapter->netdev, "Timeout reading IGC_PTM_STAT register\n");
850+
return err;
851+
}
852+
853+
if ((stat & IGC_PTM_STAT_VALID) == IGC_PTM_STAT_VALID)
854+
break;
855+
856+
if (stat & ~IGC_PTM_STAT_VALID) {
857+
/* An error occurred, log it. */
858+
igc_ptm_log_error(adapter, stat);
859+
/* The STAT register is write-1-to-clear (W1C),
860+
* so write the previous error status to clear it.
861+
*/
862+
wr32(IGC_PTM_STAT, stat);
863+
continue;
864+
}
865+
} while (--count);
866+
867+
if (!count) {
868+
netdev_err(adapter->netdev, "Exceeded number of tries for PTM cycle\n");
869+
return -ETIMEDOUT;
870+
}
871+
872+
t1 = ktime_set(rd32(IGC_PTM_T1_TIM0_H), rd32(IGC_PTM_T1_TIM0_L));
873+
874+
t2_curr_l = rd32(IGC_PTM_CURR_T2_L);
875+
t2_curr_h = rd32(IGC_PTM_CURR_T2_H);
876+
877+
/* FIXME: When the register that tells the endianness of the
878+
* PTM registers are implemented, check them here and add the
879+
* appropriate conversion.
880+
*/
881+
t2_curr_h = swab32(t2_curr_h);
882+
883+
t2_curr = ((s64)t2_curr_h << 32 | t2_curr_l);
884+
885+
*device = t1;
886+
*system = igc_device_tstamp_to_system(t2_curr);
887+
888+
return 0;
889+
}
890+
891+
static int igc_ptp_getcrosststamp(struct ptp_clock_info *ptp,
892+
struct system_device_crosststamp *cts)
893+
{
894+
struct igc_adapter *adapter = container_of(ptp, struct igc_adapter,
895+
ptp_caps);
896+
897+
return get_device_system_crosststamp(igc_phc_get_syncdevicetime,
898+
adapter, &adapter->snapshot, cts);
899+
}
900+
755901
/**
756902
* igc_ptp_init - Initialize PTP functionality
757903
* @adapter: Board private structure
@@ -788,6 +934,11 @@ void igc_ptp_init(struct igc_adapter *adapter)
788934
adapter->ptp_caps.n_per_out = IGC_N_PEROUT;
789935
adapter->ptp_caps.n_pins = IGC_N_SDP;
790936
adapter->ptp_caps.verify = igc_ptp_verify_pin;
937+
938+
if (!igc_is_crosststamp_supported(adapter))
939+
break;
940+
941+
adapter->ptp_caps.getcrosststamp = igc_ptp_getcrosststamp;
791942
break;
792943
default:
793944
adapter->ptp_clock = NULL;
@@ -878,7 +1029,9 @@ void igc_ptp_stop(struct igc_adapter *adapter)
8781029
void igc_ptp_reset(struct igc_adapter *adapter)
8791030
{
8801031
struct igc_hw *hw = &adapter->hw;
1032+
u32 cycle_ctrl, ctrl;
8811033
unsigned long flags;
1034+
u32 timadj;
8821035

8831036
/* reset the tstamp_config */
8841037
igc_ptp_set_timestamp_mode(adapter, &adapter->tstamp_config);
@@ -887,12 +1040,38 @@ void igc_ptp_reset(struct igc_adapter *adapter)
8871040

8881041
switch (adapter->hw.mac.type) {
8891042
case igc_i225:
1043+
timadj = rd32(IGC_TIMADJ);
1044+
timadj |= IGC_TIMADJ_ADJUST_METH;
1045+
wr32(IGC_TIMADJ, timadj);
1046+
8901047
wr32(IGC_TSAUXC, 0x0);
8911048
wr32(IGC_TSSDP, 0x0);
8921049
wr32(IGC_TSIM,
8931050
IGC_TSICR_INTERRUPTS |
8941051
(adapter->pps_sys_wrap_on ? IGC_TSICR_SYS_WRAP : 0));
8951052
wr32(IGC_IMS, IGC_IMS_TS);
1053+
1054+
if (!igc_is_crosststamp_supported(adapter))
1055+
break;
1056+
1057+
wr32(IGC_PCIE_DIG_DELAY, IGC_PCIE_DIG_DELAY_DEFAULT);
1058+
wr32(IGC_PCIE_PHY_DELAY, IGC_PCIE_PHY_DELAY_DEFAULT);
1059+
1060+
cycle_ctrl = IGC_PTM_CYCLE_CTRL_CYC_TIME(IGC_PTM_CYC_TIME_DEFAULT);
1061+
1062+
wr32(IGC_PTM_CYCLE_CTRL, cycle_ctrl);
1063+
1064+
ctrl = IGC_PTM_CTRL_EN |
1065+
IGC_PTM_CTRL_START_NOW |
1066+
IGC_PTM_CTRL_SHRT_CYC(IGC_PTM_SHORT_CYC_DEFAULT) |
1067+
IGC_PTM_CTRL_PTM_TO(IGC_PTM_TIMEOUT_DEFAULT) |
1068+
IGC_PTM_CTRL_TRIG;
1069+
1070+
wr32(IGC_PTM_CTRL, ctrl);
1071+
1072+
/* Force the first cycle to run. */
1073+
wr32(IGC_PTM_STAT, IGC_PTM_STAT_VALID);
1074+
8961075
break;
8971076
default:
8981077
/* No work to do. */

‎drivers/net/ethernet/intel/igc/igc_regs.h‎

Lines changed: 23 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -245,6 +245,29 @@
245245
#define IGC_TXSTMPL 0x0B618 /* Tx timestamp value Low - RO */
246246
#define IGC_TXSTMPH 0x0B61C /* Tx timestamp value High - RO */
247247

248+
#define IGC_TIMADJ 0x0B60C /* Time Adjustment Offset Register */
249+
250+
/* PCIe Registers */
251+
#define IGC_PTM_CTRL 0x12540 /* PTM Control */
252+
#define IGC_PTM_STAT 0x12544 /* PTM Status */
253+
#define IGC_PTM_CYCLE_CTRL 0x1254C /* PTM Cycle Control */
254+
255+
/* PTM Time registers */
256+
#define IGC_PTM_T1_TIM0_L 0x12558 /* T1 on Timer 0 Low */
257+
#define IGC_PTM_T1_TIM0_H 0x1255C /* T1 on Timer 0 High */
258+
259+
#define IGC_PTM_CURR_T2_L 0x1258C /* Current T2 Low */
260+
#define IGC_PTM_CURR_T2_H 0x12590 /* Current T2 High */
261+
#define IGC_PTM_PREV_T2_L 0x12584 /* Previous T2 Low */
262+
#define IGC_PTM_PREV_T2_H 0x12588 /* Previous T2 High */
263+
#define IGC_PTM_PREV_T4M1 0x12578 /* T4 Minus T1 on previous PTM Cycle */
264+
#define IGC_PTM_CURR_T4M1 0x1257C /* T4 Minus T1 on this PTM Cycle */
265+
#define IGC_PTM_PREV_T3M2 0x12580 /* T3 Minus T2 on previous PTM Cycle */
266+
#define IGC_PTM_TDELAY 0x12594 /* PTM PCIe Link Delay */
267+
268+
#define IGC_PCIE_DIG_DELAY 0x12550 /* PCIe Digital Delay */
269+
#define IGC_PCIE_PHY_DELAY 0x12554 /* PCIe PHY Delay */
270+
248271
/* Management registers */
249272
#define IGC_MANC 0x05820 /* Management Control - RW */
250273

0 commit comments

Comments
 (0)