| Message ID | 20240706153807.3390950-1-faizal.abdul.rahim@linux.intel.com (mailing list archive) |
|---|---|
| State | Accepted |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [iwl-net,v2,1/1] igc: Fix packet still tx after gate close by reducing i226 MAC retry buffer | expand |
Faizal Rahim <faizal.abdul.rahim@linux.intel.com> writes: > Testing uncovered that even when the taprio gate is closed, some packets > still transmit. > > According to i225/6 hardware errata [1], traffic might overflow the > planned QBV window. This happens because MAC maintains an internal buffer, > primarily for supporting half duplex retries. Therefore, even when the > gate closes, residual MAC data in the buffer may still transmit. > > To mitigate this for i226, reduce the MAC's internal buffer from 192 bytes > to the recommended 88 bytes by modifying the RETX_CTL register value. > > This follows guidelines from: > [1] Ethernet Controller I225/I22 Spec Update Rev 2.1 Errata Item 9: > TSN: Packet Transmission Might Cross Qbv Window > [2] I225/6 SW User Manual Rev 1.2.4: Section 8.11.5 Retry Buffer Control > > Note that the RETX_CTL register can't be used in TSN mode because half > duplex feature cannot coexist with TSN. > > Test Steps: > 1. Send taprio cmd to board A: > tc qdisc replace dev enp1s0 parent root handle 100 taprio \ > num_tc 4 \ > map 3 2 1 0 3 3 3 3 3 3 3 3 3 3 3 3 \ > queues 1@0 1@1 1@2 1@3 \ > base-time 0 \ > sched-entry S 0x07 500000 \ > sched-entry S 0x0f 500000 \ > flags 0x2 \ > txtime-delay 0 > > Note that for TC3, gate should open for 500us and close for another > 500us. > > 3. Take tcpdump log on Board B. > > 4. Send udp packets via UDP tai app from Board A to Board B. > > 5. Analyze tcpdump log via wireshark log on Board B. Ensure that the > total time from the first to the last packet received during one cycle > for TC3 does not exceed 500us. > > Signed-off-by: Faizal Rahim <faizal.abdul.rahim@linux.intel.com> > --- > v1: https://patchwork.kernel.org/project/netdevbpf/patch/20240701100058.3301229-1-faizal.abdul.rahim@linux.intel.com/ > > Changelog: > v1 -> v2 > - Update commit description (Paul). > - Rename qbvfullth -> qbvfullthreshold (Paul). > --- Looks fine. Acked-by: Vinicius Costa Gomes <vinicius.gomes@intel.com> Cheers,
On 06/07/2024 18:38, Faizal Rahim wrote: > Testing uncovered that even when the taprio gate is closed, some packets > still transmit. > > According to i225/6 hardware errata [1], traffic might overflow the > planned QBV window. This happens because MAC maintains an internal buffer, > primarily for supporting half duplex retries. Therefore, even when the > gate closes, residual MAC data in the buffer may still transmit. > > To mitigate this for i226, reduce the MAC's internal buffer from 192 bytes > to the recommended 88 bytes by modifying the RETX_CTL register value. > > This follows guidelines from: > [1] Ethernet Controller I225/I22 Spec Update Rev 2.1 Errata Item 9: > TSN: Packet Transmission Might Cross Qbv Window > [2] I225/6 SW User Manual Rev 1.2.4: Section 8.11.5 Retry Buffer Control > > Note that the RETX_CTL register can't be used in TSN mode because half > duplex feature cannot coexist with TSN. > > Test Steps: > 1. Send taprio cmd to board A: > tc qdisc replace dev enp1s0 parent root handle 100 taprio \ > num_tc 4 \ > map 3 2 1 0 3 3 3 3 3 3 3 3 3 3 3 3 \ > queues 1@0 1@1 1@2 1@3 \ > base-time 0 \ > sched-entry S 0x07 500000 \ > sched-entry S 0x0f 500000 \ > flags 0x2 \ > txtime-delay 0 > > Note that for TC3, gate should open for 500us and close for another > 500us. > > 3. Take tcpdump log on Board B. > > 4. Send udp packets via UDP tai app from Board A to Board B. > > 5. Analyze tcpdump log via wireshark log on Board B. Ensure that the > total time from the first to the last packet received during one cycle > for TC3 does not exceed 500us. > > Signed-off-by: Faizal Rahim <faizal.abdul.rahim@linux.intel.com> > Acked-by: Vinicius Costa Gomes <vinicius.gomes@intel.com> > --- > v1: https://patchwork.kernel.org/project/netdevbpf/patch/20240701100058.3301229-1-faizal.abdul.rahim@linux.intel.com/ > > Changelog: > v1 -> v2 > - Update commit description (Paul). > - Rename qbvfullth -> qbvfullthreshold (Paul). > --- > drivers/net/ethernet/intel/igc/igc_defines.h | 6 ++++ > drivers/net/ethernet/intel/igc/igc_tsn.c | 34 ++++++++++++++++++++ > 2 files changed, 40 insertions(+) > Tested-by: Mor Bar-Gabay <morx.bar.gabay@intel.com>
diff --git a/drivers/net/ethernet/intel/igc/igc_defines.h b/drivers/net/ethernet/intel/igc/igc_defines.h index 5f92b3c7c3d4..511384f3ec5c 100644 --- a/drivers/net/ethernet/intel/igc/igc_defines.h +++ b/drivers/net/ethernet/intel/igc/igc_defines.h @@ -404,6 +404,12 @@ #define IGC_DTXMXPKTSZ_TSN 0x19 /* 1600 bytes of max TX DMA packet size */ #define IGC_DTXMXPKTSZ_DEFAULT 0x98 /* 9728-byte Jumbo frames */ +/* Retry Buffer Control */ +#define IGC_RETX_CTL 0x041C +#define IGC_RETX_CTL_WATERMARK_MASK 0xF +#define IGC_RETX_CTL_QBVFULLTH_SHIFT 8 /* QBV Retry Buffer Full Threshold */ +#define IGC_RETX_CTL_QBVFULLEN 0x1000 /* Enable QBV Retry Buffer Full Threshold */ + /* Transmit Scheduling Latency */ /* Latency between transmission scheduling (LaunchTime) and the time * the packet is transmitted to the network in nanosecond. diff --git a/drivers/net/ethernet/intel/igc/igc_tsn.c b/drivers/net/ethernet/intel/igc/igc_tsn.c index 22cefb1eeedf..9df3680b9fcb 100644 --- a/drivers/net/ethernet/intel/igc/igc_tsn.c +++ b/drivers/net/ethernet/intel/igc/igc_tsn.c @@ -78,6 +78,15 @@ void igc_tsn_adjust_txtime_offset(struct igc_adapter *adapter) wr32(IGC_GTXOFFSET, txoffset); } +static void igc_tsn_restore_retx_default(struct igc_adapter *adapter) +{ + struct igc_hw *hw = &adapter->hw; + u32 retxctl; + + retxctl = rd32(IGC_RETX_CTL) & IGC_RETX_CTL_WATERMARK_MASK; + wr32(IGC_RETX_CTL, retxctl); +} + /* Returns the TSN specific registers to their default values after * the adapter is reset. */ @@ -91,6 +100,9 @@ static int igc_tsn_disable_offload(struct igc_adapter *adapter) wr32(IGC_TXPBS, I225_TXPBSIZE_DEFAULT); wr32(IGC_DTXMXPKTSZ, IGC_DTXMXPKTSZ_DEFAULT); + if (igc_is_device_id_i226(hw)) + igc_tsn_restore_retx_default(adapter); + tqavctrl = rd32(IGC_TQAVCTRL); tqavctrl &= ~(IGC_TQAVCTRL_TRANSMIT_MODE_TSN | IGC_TQAVCTRL_ENHANCED_QAV | IGC_TQAVCTRL_FUTSCDDIS); @@ -111,6 +123,25 @@ static int igc_tsn_disable_offload(struct igc_adapter *adapter) return 0; } +/* To partially fix i226 HW errata, reduce MAC internal buffering from 192 Bytes + * to 88 Bytes by setting RETX_CTL register using the recommendation from: + * a) Ethernet Controller I225/I22 Specification Update Rev 2.1 + * Item 9: TSN: Packet Transmission Might Cross the Qbv Window + * b) I225/6 SW User Manual Rev 1.2.4: Section 8.11.5 Retry Buffer Control + */ +static void igc_tsn_set_retx_qbvfullthreshold(struct igc_adapter *adapter) +{ + struct igc_hw *hw = &adapter->hw; + u32 retxctl, watermark; + + retxctl = rd32(IGC_RETX_CTL); + watermark = retxctl & IGC_RETX_CTL_WATERMARK_MASK; + /* Set QBVFULLTH value using watermark and set QBVFULLEN */ + retxctl |= (watermark << IGC_RETX_CTL_QBVFULLTH_SHIFT) | + IGC_RETX_CTL_QBVFULLEN; + wr32(IGC_RETX_CTL, retxctl); +} + static int igc_tsn_enable_offload(struct igc_adapter *adapter) { struct igc_hw *hw = &adapter->hw; @@ -123,6 +154,9 @@ static int igc_tsn_enable_offload(struct igc_adapter *adapter) wr32(IGC_DTXMXPKTSZ, IGC_DTXMXPKTSZ_TSN); wr32(IGC_TXPBS, IGC_TXPBSIZE_TSN); + if (igc_is_device_id_i226(hw)) + igc_tsn_set_retx_qbvfullthreshold(adapter); + for (i = 0; i < adapter->num_tx_queues; i++) { struct igc_ring *ring = adapter->tx_ring[i]; u32 txqctl = 0;
Testing uncovered that even when the taprio gate is closed, some packets still transmit. According to i225/6 hardware errata [1], traffic might overflow the planned QBV window. This happens because MAC maintains an internal buffer, primarily for supporting half duplex retries. Therefore, even when the gate closes, residual MAC data in the buffer may still transmit. To mitigate this for i226, reduce the MAC's internal buffer from 192 bytes to the recommended 88 bytes by modifying the RETX_CTL register value. This follows guidelines from: [1] Ethernet Controller I225/I22 Spec Update Rev 2.1 Errata Item 9: TSN: Packet Transmission Might Cross Qbv Window [2] I225/6 SW User Manual Rev 1.2.4: Section 8.11.5 Retry Buffer Control Note that the RETX_CTL register can't be used in TSN mode because half duplex feature cannot coexist with TSN. Test Steps: 1. Send taprio cmd to board A: tc qdisc replace dev enp1s0 parent root handle 100 taprio \ num_tc 4 \ map 3 2 1 0 3 3 3 3 3 3 3 3 3 3 3 3 \ queues 1@0 1@1 1@2 1@3 \ base-time 0 \ sched-entry S 0x07 500000 \ sched-entry S 0x0f 500000 \ flags 0x2 \ txtime-delay 0 Note that for TC3, gate should open for 500us and close for another 500us. 3. Take tcpdump log on Board B. 4. Send udp packets via UDP tai app from Board A to Board B. 5. Analyze tcpdump log via wireshark log on Board B. Ensure that the total time from the first to the last packet received during one cycle for TC3 does not exceed 500us. Signed-off-by: Faizal Rahim <faizal.abdul.rahim@linux.intel.com> --- v1: https://patchwork.kernel.org/project/netdevbpf/patch/20240701100058.3301229-1-faizal.abdul.rahim@linux.intel.com/ Changelog: v1 -> v2 - Update commit description (Paul). - Rename qbvfullth -> qbvfullthreshold (Paul). --- drivers/net/ethernet/intel/igc/igc_defines.h | 6 ++++ drivers/net/ethernet/intel/igc/igc_tsn.c | 34 ++++++++++++++++++++ 2 files changed, 40 insertions(+)