* [PATCH net-next v6 00/14] Add PPE driver for Qualcomm IPQ9574 SoC
@ 2025-07-20 10:57 Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 01/14] dt-bindings: net: Add PPE " Luo Jie
` (13 more replies)
0 siblings, 14 replies; 17+ messages in thread
From: Luo Jie @ 2025-07-20 10:57 UTC (permalink / raw)
To: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Rob Herring, Krzysztof Kozlowski, Conor Dooley,
Lei Wei, Suruchi Agarwal, Pavithra R, Simon Horman,
Jonathan Corbet, Kees Cook, Gustavo A. R. Silva, Philipp Zabel
Cc: linux-arm-msm, netdev, devicetree, linux-kernel, linux-doc,
linux-hardening, quic_kkumarcs, quic_linchen, Luo Jie
The PPE (packet process engine) hardware block is available in Qualcomm
IPQ chipsets that support PPE architecture, such as IPQ9574 and IPQ5332.
The PPE in the IPQ9574 SoC includes six ethernet ports (6 GMAC and 6
XGMAC), which are used to connect with external PHY devices by PCS. The
PPE also includes packet processing offload capabilities for various
networking functions such as route and bridge flows, VLANs, different
tunnel protocols and VPN. It also includes an L2 switch function for
bridging packets among the 6 ethernet ports and the CPU port. The CPU
port enables packet transfer between the ethernet ports and the ARM
cores in the SoC, using the ethernet DMA.
This patch series is the first part of a three part series that will
together enable Ethernet function for IPQ9574 SoC. While support is
initially being added for IPQ9574 SoC, the driver will be easily
extendable to enable Ethernet support for other IPQ SoC such as IPQ5332.
The driver can also be extended later for adding support for L2/L3
network offload features that the PPE can support. The functionality
to be enabled by each of the three series (to be posted sequentially)
is as below:
Part 1: The PPE patch series (this series), which enables the platform
driver, probe and initialization/configuration of different PPE hardware
blocks.
Part 2: The PPE MAC patch series, which enables the phylink operations
for the PPE ethernet ports.
Part 3: The PPE EDMA patch series, which enables the Rx/Tx Ethernet DMA
and netdevice driver for the 6 PPE ethernet ports.
A more detailed description of the functions enabled by part 1 is below:
1. Initialize PPE device hardware functions such as buffer management,
queue management, scheduler and clocks in order to bring up PPE
device.
2. Enable platform driver and probe functions
3. Register debugfs file to provide access to various PPE packet
counters. These statistics are recorded by the various hardware
process counters, such as port RX/TX, CPU code and hardware queue
counters.
4. A detailed introduction of PPE along with the PPE hardware diagram
in the first two patches (dt-bindings and documentation).
Below is a reference to an earlier RFC discussion with the community
about enabling ethernet driver support for Qualcomm IPQ9574 SoC. This
writeup can help provide a higher level architectural view of various
other drivers that support the PPE such as clock and PCS drivers.
Topic: RFC: Advice on adding support for Qualcomm IPQ9574 SoC Ethernet.
https://lore.kernel.org/linux-arm-msm/d2929bd2-bc9e-4733-a89f-2a187e8bf917@quicinc.com/
Signed-off-by: Luo Jie <quic_luoj@quicinc.com>
---
Changes in v6:
- DT binding improvements:
* Add the missing ethernet-ports node.
* Update to use 'ethernet-port' as the port node name.
* Ensure the port node numbering is constrained appropriately.
* Enhance the commit message to add a brief overview of the switch schema
hierarchy and properties.
- Refactor the RSS hash configuration code to eliminate redundant code.
- Update copyright to remove year as per guidelines in all driver files.
- Update 'M' to 'MHz' in the documentation patch for clarity.
- Link to v5: https://lore.kernel.org/r/20250626-qcom_ipq_ppe-v5-0-95bdc6b8f6ff@quicinc.com
Changes in v5:
- DT binding improvements:
* Update example by using clock & reset ID instead of magic number.
* Fix the duplicate '>' symbol.
* Convert the interrupt number to a fixed value and add the missing but
unused interrupts to complete the hardware description.
* Constraint the port node with clock and reset.
- Add debugfs based queue drop counter.
- Cosmetic update to commit messages to wrap the text correctly.
- Link to v4: https://lore.kernel.org/r/20250513-qcom_ipq_ppe-v4-0-4fbe40cbbb71@quicinc.com
Changes in v4:
- Simplify statements regarding module load in documentation as per comments.
- Improve data structure definitions for scheduler patch for clarity.
- Replace u32p_replace_bits() with FIELD_MODIFY().
- Enhance the comment of the PPE scheduler for BM and QM configurations.
- Debugfs improvements:
*Remove print related macros and inline the code instead.
*Return error codes from register read/write wherever applicable.
*Split the hardware counter display file into separate files.
- Link to v3: https://lore.kernel.org/r/20250209-qcom_ipq_ppe-v3-0-453ea18d3271@quicinc.com
Changes in v3:
- Add the top-level ref ethernet-switch.yaml and remove node definition
ethernet-ports in the DT binding file.
- Remove unnecessary error message for devm_kzalloc().
- Reverse the mapping of BM ceiling bits.
- Fix multicast queue start/end configurations.
- Declare the SoC-specific PPE configuration variables as const.
- Fix kernel documentation errors.
- Fix the compile errors reported by gcc-14.
- Improve the commit message of L2 bridge initialization and debugfs patches.
- Link to v2: https://lore.kernel.org/r/20250108-qcom_ipq_ppe-v2-0-7394dbda7199@quicinc.com
Changes in v2:
- Represent the PPE hardware hierarchy in dtbindings, add PPE hardware diagram.
- Remove all SoC specific hardware properties from dtbindings since driver
maintains them.
- Move out the PCS (UNIPHY) handling into a separate PCS driver posted
separately at below.
https://lore.kernel.org/all/20250108-ipq_pcs_net-next-v4-0-0de14cd2902b@quicinc.com
- Move out the PPE MAC patches into a separate series to limit patch count to
15 or less. (PPE MAC patches will be posted sequentially after this series).
- Rename the hardware initialization related files from ppe_ops.c[h] to
ppe_config.c[h]
- Improve PPE driver documentation and diagram.
- Fix dtbinding check errors.
- Link to v1: https://lore.kernel.org/r/20240110114033.32575-1-quic_luoj@quicinc.com
---
Lei Wei (2):
docs: networking: Add PPE driver documentation for Qualcomm IPQ9574 SoC
net: ethernet: qualcomm: Initialize PPE L2 bridge settings
Luo Jie (12):
dt-bindings: net: Add PPE for Qualcomm IPQ9574 SoC
net: ethernet: qualcomm: Add PPE driver for IPQ9574 SoC
net: ethernet: qualcomm: Initialize PPE buffer management for IPQ9574
net: ethernet: qualcomm: Initialize PPE queue management for IPQ9574
net: ethernet: qualcomm: Initialize the PPE scheduler settings
net: ethernet: qualcomm: Initialize PPE queue settings
net: ethernet: qualcomm: Initialize PPE service code settings
net: ethernet: qualcomm: Initialize PPE port control settings
net: ethernet: qualcomm: Initialize PPE RSS hash settings
net: ethernet: qualcomm: Initialize PPE queue to Ethernet DMA ring mapping
net: ethernet: qualcomm: Add PPE debugfs support for PPE counters
MAINTAINERS: Add maintainer for Qualcomm PPE driver
.../devicetree/bindings/net/qcom,ipq9574-ppe.yaml | 529 +++++
.../networking/device_drivers/ethernet/index.rst | 1 +
.../device_drivers/ethernet/qualcomm/ppe/ppe.rst | 194 ++
MAINTAINERS | 8 +
drivers/net/ethernet/qualcomm/Kconfig | 15 +
drivers/net/ethernet/qualcomm/Makefile | 1 +
drivers/net/ethernet/qualcomm/ppe/Makefile | 7 +
drivers/net/ethernet/qualcomm/ppe/ppe.c | 234 +++
drivers/net/ethernet/qualcomm/ppe/ppe.h | 39 +
drivers/net/ethernet/qualcomm/ppe/ppe_config.c | 2033 ++++++++++++++++++++
drivers/net/ethernet/qualcomm/ppe/ppe_config.h | 317 +++
drivers/net/ethernet/qualcomm/ppe/ppe_debugfs.c | 847 ++++++++
drivers/net/ethernet/qualcomm/ppe/ppe_debugfs.h | 16 +
drivers/net/ethernet/qualcomm/ppe/ppe_regs.h | 591 ++++++
14 files changed, 4832 insertions(+)
---
base-commit: 4701ee5044fb3992f1c910630a9673c2dc600ce5
change-id: 20250717-qcom_ipq_ppe-7684dbc38fa4
Best regards,
--
Luo Jie <quic_luoj@quicinc.com>
^ permalink raw reply [flat|nested] 17+ messages in thread
* [PATCH net-next v6 01/14] dt-bindings: net: Add PPE for Qualcomm IPQ9574 SoC
2025-07-20 10:57 [PATCH net-next v6 00/14] Add PPE driver for Qualcomm IPQ9574 SoC Luo Jie
@ 2025-07-20 10:57 ` Luo Jie
2025-07-21 19:35 ` Rob Herring
2025-07-20 10:57 ` [PATCH net-next v6 02/14] docs: networking: Add PPE driver documentation " Luo Jie
` (12 subsequent siblings)
13 siblings, 1 reply; 17+ messages in thread
From: Luo Jie @ 2025-07-20 10:57 UTC (permalink / raw)
To: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Rob Herring, Krzysztof Kozlowski, Conor Dooley,
Lei Wei, Suruchi Agarwal, Pavithra R, Simon Horman,
Jonathan Corbet, Kees Cook, Gustavo A. R. Silva, Philipp Zabel
Cc: linux-arm-msm, netdev, devicetree, linux-kernel, linux-doc,
linux-hardening, quic_kkumarcs, quic_linchen, Luo Jie
The PPE (packet process engine) hardware block is available in Qualcomm
IPQ chipsets that support PPE architecture, such as IPQ9574. The PPE in
the IPQ9574 SoC includes six ethernet ports (6 GMAC and 6 XGMAC), which
are used to connect with external PHY devices by PCS. It includes an L2
switch function for bridging packets among the 6 ethernet ports and the
CPU port. The CPU port enables packet transfer between the ethernet ports
and the ARM cores in the SoC, using the ethernet DMA.
The PPE also includes packet processing offload capabilities for various
networking functions such as route and bridge flows, VLANs, different
tunnel protocols and VPN.
The PPE switch is modeled according to the ethernet switch schema, with
additional properties defined for the switch node for interrupts, clocks,
resets, interconnects and Ethernet DMA. The switch port node is extended
with additional properties for clocks and resets.
Signed-off-by: Luo Jie <quic_luoj@quicinc.com>
---
.../devicetree/bindings/net/qcom,ipq9574-ppe.yaml | 529 +++++++++++++++++++++
1 file changed, 529 insertions(+)
diff --git a/Documentation/devicetree/bindings/net/qcom,ipq9574-ppe.yaml b/Documentation/devicetree/bindings/net/qcom,ipq9574-ppe.yaml
new file mode 100644
index 000000000000..d48169a8ba7c
--- /dev/null
+++ b/Documentation/devicetree/bindings/net/qcom,ipq9574-ppe.yaml
@@ -0,0 +1,529 @@
+# SPDX-License-Identifier: GPL-2.0 OR BSD-2-Clause
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/net/qcom,ipq9574-ppe.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Qualcomm IPQ packet process engine (PPE)
+
+maintainers:
+ - Luo Jie <quic_luoj@quicinc.com>
+ - Lei Wei <quic_leiwei@quicinc.com>
+ - Suruchi Agarwal <quic_suruchia@quicinc.com>
+ - Pavithra R <quic_pavir@quicinc.com>
+
+description: |
+ The Ethernet functionality in the PPE (Packet Process Engine) is comprised
+ of three components, the switch core, port wrapper and Ethernet DMA.
+
+ The Switch core in the IPQ9574 PPE has maximum of 6 front panel ports and
+ two FIFO interfaces. One of the two FIFO interfaces is used for Ethernet
+ port to host CPU communication using Ethernet DMA. The other is used
+ communicating to the EIP engine which is used for IPsec offload. On the
+ IPQ9574, the PPE includes 6 GMAC/XGMACs that can be connected with external
+ Ethernet PHY. Switch core also includes BM (Buffer Management), QM (Queue
+ Management) and SCH (Scheduler) modules for supporting the packet processing.
+
+ The port wrapper provides connections from the 6 GMAC/XGMACS to UNIPHY (PCS)
+ supporting various modes such as SGMII/QSGMII/PSGMII/USXGMII/10G-BASER. There
+ are 3 UNIPHY (PCS) instances supported on the IPQ9574.
+
+ Ethernet DMA is used to transmit and receive packets between the six Ethernet
+ ports and ARM host CPU.
+
+ The follow diagram shows the PPE hardware block along with its connectivity
+ to the external hardware blocks such clock hardware blocks (CMNPLL, GCC,
+ NSS clock controller) and ethernet PCS/PHY blocks. For depicting the PHY
+ connectivity, one 4x1 Gbps PHY (QCA8075) and two 10 GBps PHYs are used as an
+ example.
+
+ +---------+
+ | 48 MHZ |
+ +----+----+
+ |(clock)
+ v
+ +----+----+
+ +------| CMN PLL |
+ | +----+----+
+ | |(clock)
+ | v
+ | +----+----+ +----+----+ (clock) +----+----+
+ | +---| NSSCC | | GCC |--------->| MDIO |
+ | | +----+----+ +----+----+ +----+----+
+ | | |(clock & reset) |(clock)
+ | | v v
+ | | +----+---------------------+--+----------+----------+---------+
+ | | | +-----+ |EDMA FIFO | | EIP FIFO|
+ | | | | SCH | +----------+ +---------+
+ | | | +-----+ | | |
+ | | | +------+ +------+ +-------------------+ |
+ | | | | BM | | QM | IPQ9574-PPE | L2/L3 Process | |
+ | | | +------+ +------+ +-------------------+ |
+ | | | | |
+ | | | +-------+ +-------+ +-------+ +-------+ +-------+ +-------+ |
+ | | | | MAC0 | | MAC1 | | MAC2 | | MAC3 | | XGMAC4| |XGMAC5 | |
+ | | | +---+---+ +---+---+ +---+---+ +---+---+ +---+---+ +---+---+ |
+ | | | | | | | | | |
+ | | +-----+---------+---------+---------+---------+---------+-----+
+ | | | | | | | |
+ | | +---+---------+---------+---------+---+ +---+---+ +---+---+
+ +--+---->| PCS0 | | PCS1 | | PCS2 |
+ |(clock) +---+---------+---------+---------+---+ +---+---+ +---+---+
+ | | | | | | |
+ | +---+---------+---------+---------+---+ +---+---+ +---+---+
+ +------->| QCA8075 PHY | | PHY4 | | PHY5 |
+ (clock) +-------------------------------------+ +-------+ +-------+
+
+properties:
+ compatible:
+ enum:
+ - qcom,ipq9574-ppe
+
+ reg:
+ maxItems: 1
+
+ clocks:
+ items:
+ - description: PPE core clock from NSS clock controller
+ - description: PPE APB (Advanced Peripheral Bus) clock from NSS clock controller
+ - description: PPE ingress process engine clock from NSS clock controller
+ - description: PPE BM, QM and scheduler clock from NSS clock controller
+
+ clock-names:
+ items:
+ - const: ppe
+ - const: apb
+ - const: ipe
+ - const: btq
+
+ resets:
+ maxItems: 1
+ description: PPE reset, which is necessary before configuring PPE hardware
+
+ interrupts:
+ maxItems: 1
+ description: PPE switch miscellaneous interrupt
+
+ interconnects:
+ items:
+ - description: Clock path leading to PPE switch core function
+ - description: Clock path leading to PPE register access
+ - description: Clock path leading to QoS generation
+ - description: Clock path leading to timeout reference
+ - description: Clock path leading to NSS NOC from memory NOC
+ - description: Clock path leading to memory NOC from NSS NOC
+ - description: Clock path leading to enhanced memory NOC from NSS NOC
+
+ interconnect-names:
+ items:
+ - const: ppe
+ - const: ppe_cfg
+ - const: qos_gen
+ - const: timeout_ref
+ - const: nssnoc_memnoc
+ - const: memnoc_nssnoc
+ - const: memnoc_nssnoc_1
+
+ ethernet-dma:
+ type: object
+ additionalProperties: false
+ description:
+ EDMA (Ethernet DMA) is used to transmit packets between PPE and ARM
+ host CPU. There are 32 TX descriptor rings, 32 TX completion rings,
+ 24 RX descriptor rings and 8 RX fill rings supported.
+
+ properties:
+ clocks:
+ items:
+ - description: EDMA system clock from NSS Clock Controller
+ - description: EDMA APB (Advanced Peripheral Bus) clock from
+ NSS Clock Controller
+
+ clock-names:
+ items:
+ - const: sys
+ - const: apb
+
+ resets:
+ maxItems: 1
+ description: EDMA reset from NSS clock controller
+
+ interrupts:
+ minItems: 65
+ maxItems: 65
+
+ interrupt-names:
+ minItems: 65
+ maxItems: 65
+ description:
+ Interrupts "txcmpl_[0-31]" are the Ethernet DMA TX completion ring interrupts.
+ Interrupts "rxfill_[0-7]" are the Ethernet DMA RX fill ring interrupts.
+ Interrupts "rxdesc_[0-23]" are the Ethernet DMA RX Descriptor ring interrupts.
+ Interrupt "misc" is the Ethernet DMA miscellaneous error interrupt.
+
+ required:
+ - clocks
+ - clock-names
+ - resets
+ - interrupts
+ - interrupt-names
+
+patternProperties:
+ "^(ethernet-)?ports$":
+ patternProperties:
+ "^ethernet-port@[1-6]+$":
+ type: object
+ unevaluatedProperties: false
+ $ref: ethernet-switch-port.yaml#
+
+ properties:
+ reg:
+ minimum: 1
+ maximum: 6
+ description: PPE Ethernet port ID
+
+ clocks:
+ items:
+ - description: Port MAC clock from NSS clock controller
+ - description: Port RX clock from NSS clock controller
+ - description: Port TX clock from NSS clock controller
+
+ clock-names:
+ items:
+ - const: mac
+ - const: rx
+ - const: tx
+
+ resets:
+ items:
+ - description: Port MAC reset from NSS clock controller
+ - description: Port RX reset from NSS clock controller
+ - description: Port TX reset from NSS clock controller
+
+ reset-names:
+ items:
+ - const: mac
+ - const: rx
+ - const: tx
+
+ required:
+ - reg
+ - clocks
+ - clock-names
+ - resets
+ - reset-names
+
+required:
+ - compatible
+ - reg
+ - clocks
+ - clock-names
+ - resets
+ - interconnects
+ - interconnect-names
+ - ethernet-dma
+
+allOf:
+ - $ref: ethernet-switch.yaml
+
+unevaluatedProperties: false
+
+examples:
+ - |
+ #include <dt-bindings/clock/qcom,ipq9574-gcc.h>
+ #include <dt-bindings/clock/qcom,ipq9574-nsscc.h>
+ #include <dt-bindings/interconnect/qcom,ipq9574.h>
+ #include <dt-bindings/interrupt-controller/arm-gic.h>
+ #include <dt-bindings/reset/qcom,ipq9574-nsscc.h>
+
+ ethernet-switch@3a000000 {
+ compatible = "qcom,ipq9574-ppe";
+ reg = <0x3a000000 0xbef800>;
+ clocks = <&nsscc NSS_CC_PPE_SWITCH_CLK>,
+ <&nsscc NSS_CC_PPE_SWITCH_CFG_CLK>,
+ <&nsscc NSS_CC_PPE_SWITCH_IPE_CLK>,
+ <&nsscc NSS_CC_PPE_SWITCH_BTQ_CLK>;
+ clock-names = "ppe",
+ "apb",
+ "ipe",
+ "btq";
+ resets = <&nsscc PPE_FULL_RESET>;
+ interrupts = <GIC_SPI 498 IRQ_TYPE_LEVEL_HIGH>;
+ interconnects = <&nsscc MASTER_NSSNOC_PPE &nsscc SLAVE_NSSNOC_PPE>,
+ <&nsscc MASTER_NSSNOC_PPE_CFG &nsscc SLAVE_NSSNOC_PPE_CFG>,
+ <&gcc MASTER_NSSNOC_QOSGEN_REF &gcc SLAVE_NSSNOC_QOSGEN_REF>,
+ <&gcc MASTER_NSSNOC_TIMEOUT_REF &gcc SLAVE_NSSNOC_TIMEOUT_REF>,
+ <&gcc MASTER_MEM_NOC_NSSNOC &gcc SLAVE_MEM_NOC_NSSNOC>,
+ <&gcc MASTER_NSSNOC_MEMNOC &gcc SLAVE_NSSNOC_MEMNOC>,
+ <&gcc MASTER_NSSNOC_MEM_NOC_1 &gcc SLAVE_NSSNOC_MEM_NOC_1>;
+ interconnect-names = "ppe",
+ "ppe_cfg",
+ "qos_gen",
+ "timeout_ref",
+ "nssnoc_memnoc",
+ "memnoc_nssnoc",
+ "memnoc_nssnoc_1";
+
+ ethernet-dma {
+ clocks = <&nsscc NSS_CC_PPE_EDMA_CLK>,
+ <&nsscc NSS_CC_PPE_EDMA_CFG_CLK>;
+ clock-names = "sys",
+ "apb";
+ resets = <&nsscc EDMA_HW_RESET>;
+ interrupts = <GIC_SPI 363 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 364 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 365 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 366 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 367 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 368 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 369 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 370 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 371 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 372 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 373 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 374 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 375 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 376 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 377 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 378 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 379 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 380 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 381 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 382 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 383 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 384 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 509 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 508 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 507 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 506 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 505 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 504 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 503 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 502 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 501 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 500 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 355 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 356 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 357 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 358 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 359 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 360 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 361 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 362 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 331 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 332 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 333 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 334 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 335 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 336 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 337 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 338 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 339 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 340 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 341 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 342 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 343 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 344 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 345 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 346 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 347 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 348 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 349 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 350 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 351 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 352 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 353 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 354 IRQ_TYPE_LEVEL_HIGH>,
+ <GIC_SPI 499 IRQ_TYPE_LEVEL_HIGH>;
+ interrupt-names = "txcmpl_0",
+ "txcmpl_1",
+ "txcmpl_2",
+ "txcmpl_3",
+ "txcmpl_4",
+ "txcmpl_5",
+ "txcmpl_6",
+ "txcmpl_7",
+ "txcmpl_8",
+ "txcmpl_9",
+ "txcmpl_10",
+ "txcmpl_11",
+ "txcmpl_12",
+ "txcmpl_13",
+ "txcmpl_14",
+ "txcmpl_15",
+ "txcmpl_16",
+ "txcmpl_17",
+ "txcmpl_18",
+ "txcmpl_19",
+ "txcmpl_20",
+ "txcmpl_21",
+ "txcmpl_22",
+ "txcmpl_23",
+ "txcmpl_24",
+ "txcmpl_25",
+ "txcmpl_26",
+ "txcmpl_27",
+ "txcmpl_28",
+ "txcmpl_29",
+ "txcmpl_30",
+ "txcmpl_31",
+ "rxfill_0",
+ "rxfill_1",
+ "rxfill_2",
+ "rxfill_3",
+ "rxfill_4",
+ "rxfill_5",
+ "rxfill_6",
+ "rxfill_7",
+ "rxdesc_0",
+ "rxdesc_1",
+ "rxdesc_2",
+ "rxdesc_3",
+ "rxdesc_4",
+ "rxdesc_5",
+ "rxdesc_6",
+ "rxdesc_7",
+ "rxdesc_8",
+ "rxdesc_9",
+ "rxdesc_10",
+ "rxdesc_11",
+ "rxdesc_12",
+ "rxdesc_13",
+ "rxdesc_14",
+ "rxdesc_15",
+ "rxdesc_16",
+ "rxdesc_17",
+ "rxdesc_18",
+ "rxdesc_19",
+ "rxdesc_20",
+ "rxdesc_21",
+ "rxdesc_22",
+ "rxdesc_23",
+ "misc";
+ };
+
+ ethernet-ports {
+ #address-cells = <1>;
+ #size-cells = <0>;
+
+ ethernet-port@1 {
+ reg = <1>;
+ phy-mode = "qsgmii";
+ managed = "in-band-status";
+ phy-handle = <&phy0>;
+ pcs-handle = <&pcs0_ch0>;
+ clocks = <&nsscc NSS_CC_PORT1_MAC_CLK>,
+ <&nsscc NSS_CC_PORT1_RX_CLK>,
+ <&nsscc NSS_CC_PORT1_TX_CLK>;
+ clock-names = "mac",
+ "rx",
+ "tx";
+ resets = <&nsscc PORT1_MAC_ARES>,
+ <&nsscc PORT1_RX_ARES>,
+ <&nsscc PORT1_TX_ARES>;
+ reset-names = "mac",
+ "rx",
+ "tx";
+ };
+
+ ethernet-port@2 {
+ reg = <2>;
+ phy-mode = "qsgmii";
+ managed = "in-band-status";
+ phy-handle = <&phy1>;
+ pcs-handle = <&pcs0_ch1>;
+ clocks = <&nsscc NSS_CC_PORT2_MAC_CLK>,
+ <&nsscc NSS_CC_PORT2_RX_CLK>,
+ <&nsscc NSS_CC_PORT2_TX_CLK>;
+ clock-names = "mac",
+ "rx",
+ "tx";
+ resets = <&nsscc PORT2_MAC_ARES>,
+ <&nsscc PORT2_RX_ARES>,
+ <&nsscc PORT2_TX_ARES>;
+ reset-names = "mac",
+ "rx",
+ "tx";
+ };
+
+ ethernet-port@3 {
+ reg = <3>;
+ phy-mode = "qsgmii";
+ managed = "in-band-status";
+ phy-handle = <&phy2>;
+ pcs-handle = <&pcs0_ch2>;
+ clocks = <&nsscc NSS_CC_PORT3_MAC_CLK>,
+ <&nsscc NSS_CC_PORT3_RX_CLK>,
+ <&nsscc NSS_CC_PORT3_TX_CLK>;
+ clock-names = "mac",
+ "rx",
+ "tx";
+ resets = <&nsscc PORT3_MAC_ARES>,
+ <&nsscc PORT3_RX_ARES>,
+ <&nsscc PORT3_TX_ARES>;
+ reset-names = "mac",
+ "rx",
+ "tx";
+ };
+
+ ethernet-port@4 {
+ reg = <4>;
+ phy-mode = "qsgmii";
+ managed = "in-band-status";
+ phy-handle = <&phy3>;
+ pcs-handle = <&pcs0_ch3>;
+ clocks = <&nsscc NSS_CC_PORT4_MAC_CLK>,
+ <&nsscc NSS_CC_PORT4_RX_CLK>,
+ <&nsscc NSS_CC_PORT4_TX_CLK>;
+ clock-names = "mac",
+ "rx",
+ "tx";
+ resets = <&nsscc PORT4_MAC_ARES>,
+ <&nsscc PORT4_RX_ARES>,
+ <&nsscc PORT4_TX_ARES>;
+ reset-names = "mac",
+ "rx",
+ "tx";
+ };
+
+ ethernet-port@5 {
+ reg = <5>;
+ phy-mode = "usxgmii";
+ managed = "in-band-status";
+ phy-handle = <&phy4>;
+ pcs-handle = <&pcs1_ch0>;
+ clocks = <&nsscc NSS_CC_PORT5_MAC_CLK>,
+ <&nsscc NSS_CC_PORT5_RX_CLK>,
+ <&nsscc NSS_CC_PORT5_TX_CLK>;
+ clock-names = "mac",
+ "rx",
+ "tx";
+ resets = <&nsscc PORT5_MAC_ARES>,
+ <&nsscc PORT5_RX_ARES>,
+ <&nsscc PORT5_TX_ARES>;
+ reset-names = "mac",
+ "rx",
+ "tx";
+ };
+
+ ethernet-port@6 {
+ reg = <6>;
+ phy-mode = "usxgmii";
+ managed = "in-band-status";
+ phy-handle = <&phy5>;
+ pcs-handle = <&pcs2_ch0>;
+ clocks = <&nsscc NSS_CC_PORT6_MAC_CLK>,
+ <&nsscc NSS_CC_PORT6_RX_CLK>,
+ <&nsscc NSS_CC_PORT6_TX_CLK>;
+ clock-names = "mac",
+ "rx",
+ "tx";
+ resets = <&nsscc PORT6_MAC_ARES>,
+ <&nsscc PORT6_RX_ARES>,
+ <&nsscc PORT6_TX_ARES>;
+ reset-names = "mac",
+ "rx",
+ "tx";
+ };
+ };
+ };
--
2.34.1
^ permalink raw reply related [flat|nested] 17+ messages in thread
* [PATCH net-next v6 02/14] docs: networking: Add PPE driver documentation for Qualcomm IPQ9574 SoC
2025-07-20 10:57 [PATCH net-next v6 00/14] Add PPE driver for Qualcomm IPQ9574 SoC Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 01/14] dt-bindings: net: Add PPE " Luo Jie
@ 2025-07-20 10:57 ` Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 03/14] net: ethernet: qualcomm: Add PPE driver for " Luo Jie
` (11 subsequent siblings)
13 siblings, 0 replies; 17+ messages in thread
From: Luo Jie @ 2025-07-20 10:57 UTC (permalink / raw)
To: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Rob Herring, Krzysztof Kozlowski, Conor Dooley,
Lei Wei, Suruchi Agarwal, Pavithra R, Simon Horman,
Jonathan Corbet, Kees Cook, Gustavo A. R. Silva, Philipp Zabel
Cc: linux-arm-msm, netdev, devicetree, linux-kernel, linux-doc,
linux-hardening, quic_kkumarcs, quic_linchen, Luo Jie
From: Lei Wei <quic_leiwei@quicinc.com>
Add description and high-level diagram for PPE, driver overview and
module enable/debug information.
Signed-off-by: Lei Wei <quic_leiwei@quicinc.com>
Signed-off-by: Luo Jie <quic_luoj@quicinc.com>
---
.../networking/device_drivers/ethernet/index.rst | 1 +
.../device_drivers/ethernet/qualcomm/ppe/ppe.rst | 194 +++++++++++++++++++++
2 files changed, 195 insertions(+)
diff --git a/Documentation/networking/device_drivers/ethernet/index.rst b/Documentation/networking/device_drivers/ethernet/index.rst
index 40ac552641a3..0b0a3eef6aae 100644
--- a/Documentation/networking/device_drivers/ethernet/index.rst
+++ b/Documentation/networking/device_drivers/ethernet/index.rst
@@ -50,6 +50,7 @@ Contents:
neterion/s2io
netronome/nfp
pensando/ionic
+ qualcomm/ppe/ppe
smsc/smc9
stmicro/stmmac
ti/cpsw
diff --git a/Documentation/networking/device_drivers/ethernet/qualcomm/ppe/ppe.rst b/Documentation/networking/device_drivers/ethernet/qualcomm/ppe/ppe.rst
new file mode 100644
index 000000000000..4ab299a28969
--- /dev/null
+++ b/Documentation/networking/device_drivers/ethernet/qualcomm/ppe/ppe.rst
@@ -0,0 +1,194 @@
+.. SPDX-License-Identifier: GPL-2.0
+
+===============================================
+PPE Ethernet Driver for Qualcomm IPQ SoC Family
+===============================================
+
+Copyright (c) Qualcomm Technologies, Inc. and/or its subsidiaries.
+
+Author: Lei Wei <quic_leiwei@quicinc.com>
+
+
+Contents
+========
+
+- `PPE Overview`_
+- `PPE Driver Overview`_
+- `PPE Driver Supported SoCs`_
+- `Enabling the Driver`_
+- `Debugging`_
+
+
+PPE Overview
+============
+
+IPQ (Qualcomm Internet Processor) SoC (System-on-Chip) series is Qualcomm's series of
+networking SoC for Wi-Fi access points. The PPE (Packet Process Engine) is the Ethernet
+packet process engine in the IPQ SoC.
+
+Below is a simplified hardware diagram of IPQ9574 SoC which includes the PPE engine and
+other blocks which are in the SoC but outside the PPE engine. These blocks work together
+to enable the Ethernet for the IPQ SoC::
+
+ +------+ +------+ +------+ +------+ +------+ +------+ start +-------+
+ |netdev| |netdev| |netdev| |netdev| |netdev| |netdev|<------|PHYLINK|
+ +------+ +------+ +------+ +------+ +------+ +------+ stop +-+-+-+-+
+ | | | ^
+ +-------+ +-------------------------+--------+----------------------+ | | |
+ | GCC | | | EDMA | | | | |
+ +---+---+ | PPE +---+----+ | | | |
+ | clk | | | | | |
+ +-------->| +-----------------------+------+-----+---------------+ | | | |
+ | | Switch Core |Port0 | |Port7(EIP FIFO)| | | | |
+ | | +---+--+ +------+--------+ | | | |
+ | | | | | | | | |
+ +-------+ | | +------+---------------+----+ | | | | |
+ |CMN PLL| | | +---+ +---+ +----+ | +--------+ | | | | | |
+ +---+---+ | | |BM | |QM | |SCH | | | L2/L3 | ....... | | | | | |
+ | | | | +---+ +---+ +----+ | +--------+ | | | | | |
+ | | | | +------+--------------------+ | | | | |
+ | | | | | | | | | |
+ | v | | +-----+-+-----+-+-----+-+-+---+--+-----+-+-----+ | | | | |
+ | +------+ | | |Port1| |Port2| |Port3| |Port4| |Port5| |Port6| | | | | |
+ | |NSSCC | | | +-----+ +-----+ +-----+ +-----+ +-----+ +-----+ | | mac| | |
+ | +-+-+--+ | | |MAC0 | |MAC1 | |MAC2 | |MAC3 | |MAC4 | |MAC5 | | |<---+ | |
+ | ^ | |clk | | +-----+-+-----+-+-----+-+-----+--+-----+-+-----+ | | ops | |
+ | | | +------>| +----|------|-------|-------|---------|--------|-----+ | | |
+ | | | +---------------------------------------------------------+ | |
+ | | | | | | | | | | |
+ | | | MII clk | QSGMII USXGMII USXGMII | |
+ | | +--------------->| | | | | | | |
+ | | +-------------------------+ +---------+ +---------+ | |
+ | |125/312.5MHz clk| (PCS0) | | (PCS1) | | (PCS2) | pcs ops | |
+ | +----------------+ UNIPHY0 | | UNIPHY1 | | UNIPHY2 |<--------+ |
+ +----------------->| | | | | | |
+ | 31.25MHz ref clk +-------------------------+ +---------+ +---------+ |
+ | | | | | | | |
+ | +-----------------------------------------------------+ |
+ |25/50MHz ref clk| +-------------------------+ +------+ +------+ | link |
+ +--------------->| | QUAD PHY | | PHY4 | | PHY5 | |---------+
+ | +-------------------------+ +------+ +------+ | change
+ | |
+ | MDIO bus |
+ +-----------------------------------------------------+
+
+The CMN (Common) PLL, NSSCC (Networking Sub System Clock Controller) and GCC (Global
+Clock Controller) blocks are in the SoC and act as clock providers.
+
+The UNIPHY block is in the SoC and provides the PCS (Physical Coding Sublayer) and
+XPCS (10-Gigabit Physical Coding Sublayer) functions to support different interface
+modes between the PPE MAC and the external PHY.
+
+This documentation focuses on the descriptions of PPE engine and the PPE driver.
+
+The Ethernet functionality in the PPE (Packet Process Engine) is comprised of three
+components: the switch core, port wrapper and Ethernet DMA.
+
+The Switch core in the IPQ9574 PPE has maximum of 6 front panel ports and two FIFO
+interfaces. One of the two FIFO interfaces is used for Ethernet port to host CPU
+communication using Ethernet DMA. The other one is used to communicate to the EIP
+engine which is used for IPsec offload. On the IPQ9574, the PPE includes 6 GMAC/XGMACs
+that can be connected with external Ethernet PHY. Switch core also includes BM (Buffer
+Management), QM (Queue Management) and SCH (Scheduler) modules for supporting the
+packet processing.
+
+The port wrapper provides connections from the 6 GMAC/XGMACS to UNIPHY (PCS) supporting
+various modes such as SGMII/QSGMII/PSGMII/USXGMII/10G-BASER. There are 3 UNIPHY (PCS)
+instances supported on the IPQ9574.
+
+Ethernet DMA is used to transmit and receive packets between the Ethernet subsystem
+and ARM host CPU.
+
+The following lists the main blocks in the PPE engine which will be driven by this
+PPE driver:
+
+- BM
+ BM is the hardware buffer manager for the PPE switch ports.
+- QM
+ Queue Manager for managing the egress hardware queues of the PPE switch ports.
+- SCH
+ The scheduler which manages the hardware traffic scheduling for the PPE switch ports.
+- L2
+ The L2 block performs the packet bridging in the switch core. The bridge domain is
+ represented by the VSI (Virtual Switch Instance) domain in PPE. FDB learning can be
+ enabled based on the VSI domain and bridge forwarding occurs within the VSI domain.
+- MAC
+ The PPE in the IPQ9574 supports up to six MACs (MAC0 to MAC5) which are corresponding
+ to six switch ports (port1 to port6). The MAC block is connected with external PHY
+ through the UNIPHY PCS block. Each MAC block includes the GMAC and XGMAC blocks and
+ the switch port can select to use GMAC or XMAC through a MUX selection according to
+ the external PHY's capability.
+- EDMA (Ethernet DMA)
+ The Ethernet DMA is used to transmit and receive Ethernet packets between the PPE
+ ports and the ARM cores.
+
+The received packet on a PPE MAC port can be forwarded to another PPE MAC port. It can
+be also forwarded to internal switch port0 so that the packet can be delivered to the
+ARM cores using the Ethernet DMA (EDMA) engine. The Ethernet DMA driver will deliver the
+packet to the corresponding 'netdevice' interface.
+
+The software instantiations of the PPE MAC (netdevice), PCS and external PHYs interact
+with the Linux PHYLINK framework to manage the connectivity between the PPE ports and
+the connected PHYs, and the port link states. This is also illustrated in above diagram.
+
+
+PPE Driver Overview
+===================
+PPE driver is Ethernet driver for the Qualcomm IPQ SoC. It is a single platform driver
+which includes the PPE part and Ethernet DMA part. The PPE part initializes and drives the
+various blocks in PPE switch core such as BM/QM/L2 blocks and the PPE MACs. The EDMA part
+drives the Ethernet DMA for packet transfer between PPE ports and ARM cores, and enables
+the netdevice driver for the PPE ports.
+
+The PPE driver files in drivers/net/ethernet/qualcomm/ppe/ are listed as below:
+
+- Makefile
+- ppe.c
+- ppe.h
+- ppe_config.c
+- ppe_config.h
+- ppe_debugfs.c
+- ppe_debugfs.h
+- ppe_regs.h
+
+The ppe.c file contains the main PPE platform driver and undertakes the initialization of
+PPE switch core blocks such as QM, BM and L2. The configuration APIs for these hardware
+blocks are provided in the ppe_config.c file.
+
+The ppe.h defines the PPE device data structure which will be used by PPE driver functions.
+
+The ppe_debugfs.c enables the PPE statistics counters such as PPE port Rx and Tx counters,
+CPU code counters and queue counters.
+
+
+PPE Driver Supported SoCs
+=========================
+
+The PPE driver supports the following IPQ SoC:
+
+- IPQ9574
+
+
+Enabling the Driver
+===================
+
+The driver is located in the menu structure at::
+
+ -> Device Drivers
+ -> Network device support (NETDEVICES [=y])
+ -> Ethernet driver support
+ -> Qualcomm devices
+ -> Qualcomm Technologies, Inc. PPE Ethernet support
+
+If the driver is built as a module, the module will be called qcom-ppe.
+
+The PPE driver functionally depends on the CMN PLL and NSSCC clock controller drivers.
+Please make sure the dependent modules are installed before installing the PPE driver
+module.
+
+
+Debugging
+=========
+
+The PPE hardware counters can be accessed using debugfs interface from the
+``/sys/kernel/debug/ppe/`` directory.
--
2.34.1
^ permalink raw reply related [flat|nested] 17+ messages in thread
* [PATCH net-next v6 03/14] net: ethernet: qualcomm: Add PPE driver for IPQ9574 SoC
2025-07-20 10:57 [PATCH net-next v6 00/14] Add PPE driver for Qualcomm IPQ9574 SoC Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 01/14] dt-bindings: net: Add PPE " Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 02/14] docs: networking: Add PPE driver documentation " Luo Jie
@ 2025-07-20 10:57 ` Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 04/14] net: ethernet: qualcomm: Initialize PPE buffer management for IPQ9574 Luo Jie
` (10 subsequent siblings)
13 siblings, 0 replies; 17+ messages in thread
From: Luo Jie @ 2025-07-20 10:57 UTC (permalink / raw)
To: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Rob Herring, Krzysztof Kozlowski, Conor Dooley,
Lei Wei, Suruchi Agarwal, Pavithra R, Simon Horman,
Jonathan Corbet, Kees Cook, Gustavo A. R. Silva, Philipp Zabel
Cc: linux-arm-msm, netdev, devicetree, linux-kernel, linux-doc,
linux-hardening, quic_kkumarcs, quic_linchen, Luo Jie
The PPE (Packet Process Engine) hardware block is available on Qualcomm
IPQ SoC that support PPE architecture, such as IPQ9574.
The PPE in IPQ9574 includes six integrated ethernet MAC for 6 PPE ports,
buffer management, queue management and scheduler functions. The MACs
can connect with the external PHY or switch devices using the UNIPHY PCS
block available in the SoC.
The PPE also includes various packet processing offload capabilities
such as L3 routing and L2 bridging, VLAN and tunnel processing offload.
It also includes Ethernet DMA function for transferring packets between
ARM cores and PPE ethernet ports.
This patch adds the base source files and Makefiles for the PPE driver
such as platform driver registration, clock initialization, and PPE
reset routines.
Signed-off-by: Luo Jie <quic_luoj@quicinc.com>
---
drivers/net/ethernet/qualcomm/Kconfig | 15 ++
drivers/net/ethernet/qualcomm/Makefile | 1 +
drivers/net/ethernet/qualcomm/ppe/Makefile | 7 +
drivers/net/ethernet/qualcomm/ppe/ppe.c | 218 +++++++++++++++++++++++++++++
drivers/net/ethernet/qualcomm/ppe/ppe.h | 36 +++++
5 files changed, 277 insertions(+)
diff --git a/drivers/net/ethernet/qualcomm/Kconfig b/drivers/net/ethernet/qualcomm/Kconfig
index a4434eb38950..6e56b022fc2d 100644
--- a/drivers/net/ethernet/qualcomm/Kconfig
+++ b/drivers/net/ethernet/qualcomm/Kconfig
@@ -60,6 +60,21 @@ config QCOM_EMAC
low power, Receive-Side Scaling (RSS), and IEEE 1588-2008
Precision Clock Synchronization Protocol.
+config QCOM_PPE
+ tristate "Qualcomm Technologies, Inc. PPE Ethernet support"
+ depends on HAS_IOMEM && OF
+ depends on COMMON_CLK
+ select REGMAP_MMIO
+ help
+ This driver supports the Qualcomm Technologies, Inc. packet
+ process engine (PPE) available with IPQ SoC. The PPE includes
+ the ethernet MACs, Ethernet DMA (EDMA) and switch core that
+ supports L3 flow offload, L2 switch function, RSS and tunnel
+ offload.
+
+ To compile this driver as a module, choose M here. The module
+ will be called qcom-ppe.
+
source "drivers/net/ethernet/qualcomm/rmnet/Kconfig"
endif # NET_VENDOR_QUALCOMM
diff --git a/drivers/net/ethernet/qualcomm/Makefile b/drivers/net/ethernet/qualcomm/Makefile
index 9250976dd884..166a59aea363 100644
--- a/drivers/net/ethernet/qualcomm/Makefile
+++ b/drivers/net/ethernet/qualcomm/Makefile
@@ -11,4 +11,5 @@ qcauart-objs := qca_uart.o
obj-y += emac/
+obj-$(CONFIG_QCOM_PPE) += ppe/
obj-$(CONFIG_RMNET) += rmnet/
diff --git a/drivers/net/ethernet/qualcomm/ppe/Makefile b/drivers/net/ethernet/qualcomm/ppe/Makefile
new file mode 100644
index 000000000000..63d50d3b4f2e
--- /dev/null
+++ b/drivers/net/ethernet/qualcomm/ppe/Makefile
@@ -0,0 +1,7 @@
+# SPDX-License-Identifier: GPL-2.0-only
+#
+# Makefile for the device driver of PPE (Packet Process Engine) in IPQ SoC
+#
+
+obj-$(CONFIG_QCOM_PPE) += qcom-ppe.o
+qcom-ppe-objs := ppe.o
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe.c b/drivers/net/ethernet/qualcomm/ppe/ppe.c
new file mode 100644
index 000000000000..932339045cdf
--- /dev/null
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe.c
@@ -0,0 +1,218 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (c) Qualcomm Technologies, Inc. and/or its subsidiaries.
+ */
+
+/* PPE platform device probe, DTSI parser and PPE clock initializations. */
+
+#include <linux/clk.h>
+#include <linux/interconnect.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/regmap.h>
+#include <linux/reset.h>
+
+#include "ppe.h"
+
+#define PPE_PORT_MAX 8
+#define PPE_CLK_RATE 353000000
+
+/* ICC clocks for enabling PPE device. The avg_bw and peak_bw with value 0
+ * will be updated by the clock rate of PPE.
+ */
+static const struct icc_bulk_data ppe_icc_data[] = {
+ {
+ .name = "ppe",
+ .avg_bw = 0,
+ .peak_bw = 0,
+ },
+ {
+ .name = "ppe_cfg",
+ .avg_bw = 0,
+ .peak_bw = 0,
+ },
+ {
+ .name = "qos_gen",
+ .avg_bw = 6000,
+ .peak_bw = 6000,
+ },
+ {
+ .name = "timeout_ref",
+ .avg_bw = 6000,
+ .peak_bw = 6000,
+ },
+ {
+ .name = "nssnoc_memnoc",
+ .avg_bw = 533333,
+ .peak_bw = 533333,
+ },
+ {
+ .name = "memnoc_nssnoc",
+ .avg_bw = 533333,
+ .peak_bw = 533333,
+ },
+ {
+ .name = "memnoc_nssnoc_1",
+ .avg_bw = 533333,
+ .peak_bw = 533333,
+ },
+};
+
+static const struct regmap_range ppe_readable_ranges[] = {
+ regmap_reg_range(0x0, 0x1ff), /* Global */
+ regmap_reg_range(0x400, 0x5ff), /* LPI CSR */
+ regmap_reg_range(0x1000, 0x11ff), /* GMAC0 */
+ regmap_reg_range(0x1200, 0x13ff), /* GMAC1 */
+ regmap_reg_range(0x1400, 0x15ff), /* GMAC2 */
+ regmap_reg_range(0x1600, 0x17ff), /* GMAC3 */
+ regmap_reg_range(0x1800, 0x19ff), /* GMAC4 */
+ regmap_reg_range(0x1a00, 0x1bff), /* GMAC5 */
+ regmap_reg_range(0xb000, 0xefff), /* PRX CSR */
+ regmap_reg_range(0xf000, 0x1efff), /* IPE */
+ regmap_reg_range(0x20000, 0x5ffff), /* PTX CSR */
+ regmap_reg_range(0x60000, 0x9ffff), /* IPE L2 CSR */
+ regmap_reg_range(0xb0000, 0xeffff), /* IPO CSR */
+ regmap_reg_range(0x100000, 0x17ffff), /* IPE PC */
+ regmap_reg_range(0x180000, 0x1bffff), /* PRE IPO CSR */
+ regmap_reg_range(0x1d0000, 0x1dffff), /* Tunnel parser */
+ regmap_reg_range(0x1e0000, 0x1effff), /* Ingress parse */
+ regmap_reg_range(0x200000, 0x2fffff), /* IPE L3 */
+ regmap_reg_range(0x300000, 0x3fffff), /* IPE tunnel */
+ regmap_reg_range(0x400000, 0x4fffff), /* Scheduler */
+ regmap_reg_range(0x500000, 0x503fff), /* XGMAC0 */
+ regmap_reg_range(0x504000, 0x507fff), /* XGMAC1 */
+ regmap_reg_range(0x508000, 0x50bfff), /* XGMAC2 */
+ regmap_reg_range(0x50c000, 0x50ffff), /* XGMAC3 */
+ regmap_reg_range(0x510000, 0x513fff), /* XGMAC4 */
+ regmap_reg_range(0x514000, 0x517fff), /* XGMAC5 */
+ regmap_reg_range(0x600000, 0x6fffff), /* BM */
+ regmap_reg_range(0x800000, 0x9fffff), /* QM */
+ regmap_reg_range(0xb00000, 0xbef800), /* EDMA */
+};
+
+static const struct regmap_access_table ppe_reg_table = {
+ .yes_ranges = ppe_readable_ranges,
+ .n_yes_ranges = ARRAY_SIZE(ppe_readable_ranges),
+};
+
+static const struct regmap_config regmap_config_ipq9574 = {
+ .reg_bits = 32,
+ .reg_stride = 4,
+ .val_bits = 32,
+ .rd_table = &ppe_reg_table,
+ .wr_table = &ppe_reg_table,
+ .max_register = 0xbef800,
+ .fast_io = true,
+};
+
+static int ppe_clock_init_and_reset(struct ppe_device *ppe_dev)
+{
+ unsigned long ppe_rate = ppe_dev->clk_rate;
+ struct device *dev = ppe_dev->dev;
+ struct reset_control *rstc;
+ struct clk_bulk_data *clks;
+ struct clk *clk;
+ int ret, i;
+
+ for (i = 0; i < ppe_dev->num_icc_paths; i++) {
+ ppe_dev->icc_paths[i].name = ppe_icc_data[i].name;
+ ppe_dev->icc_paths[i].avg_bw = ppe_icc_data[i].avg_bw ? :
+ Bps_to_icc(ppe_rate);
+ ppe_dev->icc_paths[i].peak_bw = ppe_icc_data[i].peak_bw ? :
+ Bps_to_icc(ppe_rate);
+ }
+
+ ret = devm_of_icc_bulk_get(dev, ppe_dev->num_icc_paths,
+ ppe_dev->icc_paths);
+ if (ret)
+ return ret;
+
+ ret = icc_bulk_set_bw(ppe_dev->num_icc_paths, ppe_dev->icc_paths);
+ if (ret)
+ return ret;
+
+ /* The PPE clocks have a common parent clock. Setting the clock
+ * rate of "ppe" ensures the clock rate of all PPE clocks is
+ * configured to the same rate.
+ */
+ clk = devm_clk_get(dev, "ppe");
+ if (IS_ERR(clk))
+ return PTR_ERR(clk);
+
+ ret = clk_set_rate(clk, ppe_rate);
+ if (ret)
+ return ret;
+
+ ret = devm_clk_bulk_get_all_enabled(dev, &clks);
+ if (ret < 0)
+ return ret;
+
+ /* Reset the PPE. */
+ rstc = devm_reset_control_get_exclusive(dev, NULL);
+ if (IS_ERR(rstc))
+ return PTR_ERR(rstc);
+
+ ret = reset_control_assert(rstc);
+ if (ret)
+ return ret;
+
+ /* The delay 10 ms of assert is necessary for resetting PPE. */
+ usleep_range(10000, 11000);
+
+ return reset_control_deassert(rstc);
+}
+
+static int qcom_ppe_probe(struct platform_device *pdev)
+{
+ struct device *dev = &pdev->dev;
+ struct ppe_device *ppe_dev;
+ void __iomem *base;
+ int ret, num_icc;
+
+ num_icc = ARRAY_SIZE(ppe_icc_data);
+ ppe_dev = devm_kzalloc(dev, struct_size(ppe_dev, icc_paths, num_icc),
+ GFP_KERNEL);
+ if (!ppe_dev)
+ return -ENOMEM;
+
+ base = devm_platform_ioremap_resource(pdev, 0);
+ if (IS_ERR(base))
+ return dev_err_probe(dev, PTR_ERR(base), "PPE ioremap failed\n");
+
+ ppe_dev->regmap = devm_regmap_init_mmio(dev, base, ®map_config_ipq9574);
+ if (IS_ERR(ppe_dev->regmap))
+ return dev_err_probe(dev, PTR_ERR(ppe_dev->regmap),
+ "PPE initialize regmap failed\n");
+ ppe_dev->dev = dev;
+ ppe_dev->clk_rate = PPE_CLK_RATE;
+ ppe_dev->num_ports = PPE_PORT_MAX;
+ ppe_dev->num_icc_paths = num_icc;
+
+ ret = ppe_clock_init_and_reset(ppe_dev);
+ if (ret)
+ return dev_err_probe(dev, ret, "PPE clock config failed\n");
+
+ platform_set_drvdata(pdev, ppe_dev);
+
+ return 0;
+}
+
+static const struct of_device_id qcom_ppe_of_match[] = {
+ { .compatible = "qcom,ipq9574-ppe" },
+ {}
+};
+MODULE_DEVICE_TABLE(of, qcom_ppe_of_match);
+
+static struct platform_driver qcom_ppe_driver = {
+ .driver = {
+ .name = "qcom_ppe",
+ .of_match_table = qcom_ppe_of_match,
+ },
+ .probe = qcom_ppe_probe,
+};
+module_platform_driver(qcom_ppe_driver);
+
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("Qualcomm Technologies, Inc. IPQ PPE driver");
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe.h b/drivers/net/ethernet/qualcomm/ppe/ppe.h
new file mode 100644
index 000000000000..779f39c9f098
--- /dev/null
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe.h
@@ -0,0 +1,36 @@
+/* SPDX-License-Identifier: GPL-2.0-only
+ *
+ * Copyright (c) Qualcomm Technologies, Inc. and/or its subsidiaries.
+ */
+
+#ifndef __PPE_H__
+#define __PPE_H__
+
+#include <linux/compiler.h>
+#include <linux/interconnect.h>
+
+struct device;
+struct regmap;
+
+/**
+ * struct ppe_device - PPE device private data.
+ * @dev: PPE device structure.
+ * @regmap: PPE register map.
+ * @clk_rate: PPE clock rate.
+ * @num_ports: Number of PPE ports.
+ * @num_icc_paths: Number of interconnect paths.
+ * @icc_paths: Interconnect path array.
+ *
+ * PPE device is the instance of PPE hardware, which is used to
+ * configure PPE packet process modules such as BM (buffer management),
+ * QM (queue management), and scheduler.
+ */
+struct ppe_device {
+ struct device *dev;
+ struct regmap *regmap;
+ unsigned long clk_rate;
+ unsigned int num_ports;
+ unsigned int num_icc_paths;
+ struct icc_bulk_data icc_paths[] __counted_by(num_icc_paths);
+};
+#endif
--
2.34.1
^ permalink raw reply related [flat|nested] 17+ messages in thread
* [PATCH net-next v6 04/14] net: ethernet: qualcomm: Initialize PPE buffer management for IPQ9574
2025-07-20 10:57 [PATCH net-next v6 00/14] Add PPE driver for Qualcomm IPQ9574 SoC Luo Jie
` (2 preceding siblings ...)
2025-07-20 10:57 ` [PATCH net-next v6 03/14] net: ethernet: qualcomm: Add PPE driver for " Luo Jie
@ 2025-07-20 10:57 ` Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 05/14] net: ethernet: qualcomm: Initialize PPE queue " Luo Jie
` (9 subsequent siblings)
13 siblings, 0 replies; 17+ messages in thread
From: Luo Jie @ 2025-07-20 10:57 UTC (permalink / raw)
To: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Rob Herring, Krzysztof Kozlowski, Conor Dooley,
Lei Wei, Suruchi Agarwal, Pavithra R, Simon Horman,
Jonathan Corbet, Kees Cook, Gustavo A. R. Silva, Philipp Zabel
Cc: linux-arm-msm, netdev, devicetree, linux-kernel, linux-doc,
linux-hardening, quic_kkumarcs, quic_linchen, Luo Jie
The BM (Buffer Management) config controls the pause frame generated
on the PPE port. There are maximum 15 BM ports and 4 groups supported,
all BM ports are assigned to group 0 by default. The number of hardware
buffers configured for the port influence the threshold of the flow
control for that port.
Signed-off-by: Luo Jie <quic_luoj@quicinc.com>
---
drivers/net/ethernet/qualcomm/ppe/Makefile | 2 +-
drivers/net/ethernet/qualcomm/ppe/ppe.c | 5 +
drivers/net/ethernet/qualcomm/ppe/ppe_config.c | 199 +++++++++++++++++++++++++
drivers/net/ethernet/qualcomm/ppe/ppe_config.h | 12 ++
drivers/net/ethernet/qualcomm/ppe/ppe_regs.h | 59 ++++++++
5 files changed, 276 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/qualcomm/ppe/Makefile b/drivers/net/ethernet/qualcomm/ppe/Makefile
index 63d50d3b4f2e..410a7bb54cfe 100644
--- a/drivers/net/ethernet/qualcomm/ppe/Makefile
+++ b/drivers/net/ethernet/qualcomm/ppe/Makefile
@@ -4,4 +4,4 @@
#
obj-$(CONFIG_QCOM_PPE) += qcom-ppe.o
-qcom-ppe-objs := ppe.o
+qcom-ppe-objs := ppe.o ppe_config.o
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe.c b/drivers/net/ethernet/qualcomm/ppe/ppe.c
index 932339045cdf..12d0bddc4c87 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe.c
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe.c
@@ -15,6 +15,7 @@
#include <linux/reset.h>
#include "ppe.h"
+#include "ppe_config.h"
#define PPE_PORT_MAX 8
#define PPE_CLK_RATE 353000000
@@ -194,6 +195,10 @@ static int qcom_ppe_probe(struct platform_device *pdev)
if (ret)
return dev_err_probe(dev, ret, "PPE clock config failed\n");
+ ret = ppe_hw_config(ppe_dev);
+ if (ret)
+ return dev_err_probe(dev, ret, "PPE HW config failed\n");
+
platform_set_drvdata(pdev, ppe_dev);
return 0;
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_config.c b/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
new file mode 100644
index 000000000000..45b031d4dd46
--- /dev/null
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
@@ -0,0 +1,199 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (c) Qualcomm Technologies, Inc. and/or its subsidiaries.
+ */
+
+/* PPE HW initialization configs such as BM(buffer management),
+ * QM(queue management) and scheduler configs.
+ */
+
+#include <linux/bitfield.h>
+#include <linux/bits.h>
+#include <linux/device.h>
+#include <linux/regmap.h>
+
+#include "ppe.h"
+#include "ppe_config.h"
+#include "ppe_regs.h"
+
+/**
+ * struct ppe_bm_port_config - PPE BM port configuration.
+ * @port_id_start: The fist BM port ID to configure.
+ * @port_id_end: The last BM port ID to configure.
+ * @pre_alloc: BM port dedicated buffer number.
+ * @in_fly_buf: Buffer number for receiving the packet after pause frame sent.
+ * @ceil: Ceil to generate the back pressure.
+ * @weight: Weight value.
+ * @resume_offset: Resume offset from the threshold value.
+ * @resume_ceil: Ceil to resume from the back pressure state.
+ * @dynamic: Dynamic threshold used or not.
+ *
+ * The is for configuring the threshold that impacts the port
+ * flow control.
+ */
+struct ppe_bm_port_config {
+ unsigned int port_id_start;
+ unsigned int port_id_end;
+ unsigned int pre_alloc;
+ unsigned int in_fly_buf;
+ unsigned int ceil;
+ unsigned int weight;
+ unsigned int resume_offset;
+ unsigned int resume_ceil;
+ bool dynamic;
+};
+
+/* There are total 2048 buffers available in PPE, out of which some
+ * buffers are reserved for some specific purposes per PPE port. The
+ * rest of the pool of 1550 buffers are assigned to the general 'group0'
+ * which is shared among all ports of the PPE.
+ */
+static const int ipq9574_ppe_bm_group_config = 1550;
+
+/* The buffer configurations per PPE port. There are 15 BM ports and
+ * 4 BM groups supported by PPE. BM port (0-7) is for EDMA port 0,
+ * BM port (8-13) is for PPE physical port 1-6 and BM port 14 is for
+ * EIP port.
+ */
+static const struct ppe_bm_port_config ipq9574_ppe_bm_port_config[] = {
+ {
+ /* Buffer configuration for the BM port ID 0 of EDMA. */
+ .port_id_start = 0,
+ .port_id_end = 0,
+ .pre_alloc = 0,
+ .in_fly_buf = 100,
+ .ceil = 1146,
+ .weight = 7,
+ .resume_offset = 8,
+ .resume_ceil = 0,
+ .dynamic = true,
+ },
+ {
+ /* Buffer configuration for the BM port ID 1-7 of EDMA. */
+ .port_id_start = 1,
+ .port_id_end = 7,
+ .pre_alloc = 0,
+ .in_fly_buf = 100,
+ .ceil = 250,
+ .weight = 4,
+ .resume_offset = 36,
+ .resume_ceil = 0,
+ .dynamic = true,
+ },
+ {
+ /* Buffer configuration for the BM port ID 8-13 of PPE ports. */
+ .port_id_start = 8,
+ .port_id_end = 13,
+ .pre_alloc = 0,
+ .in_fly_buf = 128,
+ .ceil = 250,
+ .weight = 4,
+ .resume_offset = 36,
+ .resume_ceil = 0,
+ .dynamic = true,
+ },
+ {
+ /* Buffer configuration for the BM port ID 14 of EIP. */
+ .port_id_start = 14,
+ .port_id_end = 14,
+ .pre_alloc = 0,
+ .in_fly_buf = 40,
+ .ceil = 250,
+ .weight = 4,
+ .resume_offset = 36,
+ .resume_ceil = 0,
+ .dynamic = true,
+ },
+};
+
+static int ppe_config_bm_threshold(struct ppe_device *ppe_dev, int bm_port_id,
+ const struct ppe_bm_port_config port_cfg)
+{
+ u32 reg, val, bm_fc_val[2];
+ int ret;
+
+ reg = PPE_BM_PORT_FC_CFG_TBL_ADDR + PPE_BM_PORT_FC_CFG_TBL_INC * bm_port_id;
+ ret = regmap_bulk_read(ppe_dev->regmap, reg,
+ bm_fc_val, ARRAY_SIZE(bm_fc_val));
+ if (ret)
+ return ret;
+
+ /* Configure BM flow control related threshold. */
+ PPE_BM_PORT_FC_SET_WEIGHT(bm_fc_val, port_cfg.weight);
+ PPE_BM_PORT_FC_SET_RESUME_OFFSET(bm_fc_val, port_cfg.resume_offset);
+ PPE_BM_PORT_FC_SET_RESUME_THRESHOLD(bm_fc_val, port_cfg.resume_ceil);
+ PPE_BM_PORT_FC_SET_DYNAMIC(bm_fc_val, port_cfg.dynamic);
+ PPE_BM_PORT_FC_SET_REACT_LIMIT(bm_fc_val, port_cfg.in_fly_buf);
+ PPE_BM_PORT_FC_SET_PRE_ALLOC(bm_fc_val, port_cfg.pre_alloc);
+
+ /* Configure low/high bits of the ceiling for the BM port. */
+ val = FIELD_GET(GENMASK(2, 0), port_cfg.ceil);
+ PPE_BM_PORT_FC_SET_CEILING_LOW(bm_fc_val, val);
+ val = FIELD_GET(GENMASK(10, 3), port_cfg.ceil);
+ PPE_BM_PORT_FC_SET_CEILING_HIGH(bm_fc_val, val);
+
+ ret = regmap_bulk_write(ppe_dev->regmap, reg,
+ bm_fc_val, ARRAY_SIZE(bm_fc_val));
+ if (ret)
+ return ret;
+
+ /* Assign the default group ID 0 to the BM port. */
+ val = FIELD_PREP(PPE_BM_PORT_GROUP_ID_SHARED_GROUP_ID, 0);
+ reg = PPE_BM_PORT_GROUP_ID_ADDR + PPE_BM_PORT_GROUP_ID_INC * bm_port_id;
+ ret = regmap_update_bits(ppe_dev->regmap, reg,
+ PPE_BM_PORT_GROUP_ID_SHARED_GROUP_ID,
+ val);
+ if (ret)
+ return ret;
+
+ /* Enable BM port flow control. */
+ reg = PPE_BM_PORT_FC_MODE_ADDR + PPE_BM_PORT_FC_MODE_INC * bm_port_id;
+
+ return regmap_set_bits(ppe_dev->regmap, reg, PPE_BM_PORT_FC_MODE_EN);
+}
+
+/* Configure the buffer threshold for the port flow control function. */
+static int ppe_config_bm(struct ppe_device *ppe_dev)
+{
+ const struct ppe_bm_port_config *port_cfg;
+ unsigned int i, bm_port_id, port_cfg_cnt;
+ u32 reg, val;
+ int ret;
+
+ /* Configure the allocated buffer number only for group 0.
+ * The buffer number of group 1-3 is already cleared to 0
+ * after PPE reset during the probe of PPE driver.
+ */
+ reg = PPE_BM_SHARED_GROUP_CFG_ADDR;
+ val = FIELD_PREP(PPE_BM_SHARED_GROUP_CFG_SHARED_LIMIT,
+ ipq9574_ppe_bm_group_config);
+ ret = regmap_update_bits(ppe_dev->regmap, reg,
+ PPE_BM_SHARED_GROUP_CFG_SHARED_LIMIT,
+ val);
+ if (ret)
+ goto bm_config_fail;
+
+ /* Configure buffer thresholds for the BM ports. */
+ port_cfg = ipq9574_ppe_bm_port_config;
+ port_cfg_cnt = ARRAY_SIZE(ipq9574_ppe_bm_port_config);
+ for (i = 0; i < port_cfg_cnt; i++) {
+ for (bm_port_id = port_cfg[i].port_id_start;
+ bm_port_id <= port_cfg[i].port_id_end; bm_port_id++) {
+ ret = ppe_config_bm_threshold(ppe_dev, bm_port_id,
+ port_cfg[i]);
+ if (ret)
+ goto bm_config_fail;
+ }
+ }
+
+ return 0;
+
+bm_config_fail:
+ dev_err(ppe_dev->dev, "PPE BM config error %d\n", ret);
+ return ret;
+}
+
+int ppe_hw_config(struct ppe_device *ppe_dev)
+{
+ return ppe_config_bm(ppe_dev);
+}
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_config.h b/drivers/net/ethernet/qualcomm/ppe/ppe_config.h
new file mode 100644
index 000000000000..8e032910bfe7
--- /dev/null
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_config.h
@@ -0,0 +1,12 @@
+/* SPDX-License-Identifier: GPL-2.0-only
+ *
+ * Copyright (c) Qualcomm Technologies, Inc. and/or its subsidiaries.
+ */
+
+#ifndef __PPE_CONFIG_H__
+#define __PPE_CONFIG_H__
+
+#include "ppe.h"
+
+int ppe_hw_config(struct ppe_device *ppe_dev);
+#endif
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h b/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
new file mode 100644
index 000000000000..b89d717fdae8
--- /dev/null
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
@@ -0,0 +1,59 @@
+/* SPDX-License-Identifier: GPL-2.0-only
+ *
+ * Copyright (c) Qualcomm Technologies, Inc. and/or its subsidiaries.
+ */
+
+/* PPE hardware register and table declarations. */
+#ifndef __PPE_REGS_H__
+#define __PPE_REGS_H__
+
+#include <linux/bitfield.h>
+
+/* There are 15 BM ports and 4 BM groups supported by PPE.
+ * BM port (0-7) is for EDMA port 0, BM port (8-13) is for
+ * PPE physical port 1-6 and BM port 14 is for EIP port.
+ */
+#define PPE_BM_PORT_FC_MODE_ADDR 0x600100
+#define PPE_BM_PORT_FC_MODE_ENTRIES 15
+#define PPE_BM_PORT_FC_MODE_INC 0x4
+#define PPE_BM_PORT_FC_MODE_EN BIT(0)
+
+#define PPE_BM_PORT_GROUP_ID_ADDR 0x600180
+#define PPE_BM_PORT_GROUP_ID_ENTRIES 15
+#define PPE_BM_PORT_GROUP_ID_INC 0x4
+#define PPE_BM_PORT_GROUP_ID_SHARED_GROUP_ID GENMASK(1, 0)
+
+#define PPE_BM_SHARED_GROUP_CFG_ADDR 0x600290
+#define PPE_BM_SHARED_GROUP_CFG_ENTRIES 4
+#define PPE_BM_SHARED_GROUP_CFG_INC 0x4
+#define PPE_BM_SHARED_GROUP_CFG_SHARED_LIMIT GENMASK(10, 0)
+
+#define PPE_BM_PORT_FC_CFG_TBL_ADDR 0x601000
+#define PPE_BM_PORT_FC_CFG_TBL_ENTRIES 15
+#define PPE_BM_PORT_FC_CFG_TBL_INC 0x10
+#define PPE_BM_PORT_FC_W0_REACT_LIMIT GENMASK(8, 0)
+#define PPE_BM_PORT_FC_W0_RESUME_THRESHOLD GENMASK(17, 9)
+#define PPE_BM_PORT_FC_W0_RESUME_OFFSET GENMASK(28, 18)
+#define PPE_BM_PORT_FC_W0_CEILING_LOW GENMASK(31, 29)
+#define PPE_BM_PORT_FC_W1_CEILING_HIGH GENMASK(7, 0)
+#define PPE_BM_PORT_FC_W1_WEIGHT GENMASK(10, 8)
+#define PPE_BM_PORT_FC_W1_DYNAMIC BIT(11)
+#define PPE_BM_PORT_FC_W1_PRE_ALLOC GENMASK(22, 12)
+
+#define PPE_BM_PORT_FC_SET_REACT_LIMIT(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_BM_PORT_FC_W0_REACT_LIMIT, tbl_cfg, value)
+#define PPE_BM_PORT_FC_SET_RESUME_THRESHOLD(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_BM_PORT_FC_W0_RESUME_THRESHOLD, tbl_cfg, value)
+#define PPE_BM_PORT_FC_SET_RESUME_OFFSET(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_BM_PORT_FC_W0_RESUME_OFFSET, tbl_cfg, value)
+#define PPE_BM_PORT_FC_SET_CEILING_LOW(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_BM_PORT_FC_W0_CEILING_LOW, tbl_cfg, value)
+#define PPE_BM_PORT_FC_SET_CEILING_HIGH(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_BM_PORT_FC_W1_CEILING_HIGH, (tbl_cfg) + 0x1, value)
+#define PPE_BM_PORT_FC_SET_WEIGHT(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_BM_PORT_FC_W1_WEIGHT, (tbl_cfg) + 0x1, value)
+#define PPE_BM_PORT_FC_SET_DYNAMIC(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_BM_PORT_FC_W1_DYNAMIC, (tbl_cfg) + 0x1, value)
+#define PPE_BM_PORT_FC_SET_PRE_ALLOC(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_BM_PORT_FC_W1_PRE_ALLOC, (tbl_cfg) + 0x1, value)
+#endif
--
2.34.1
^ permalink raw reply related [flat|nested] 17+ messages in thread
* [PATCH net-next v6 05/14] net: ethernet: qualcomm: Initialize PPE queue management for IPQ9574
2025-07-20 10:57 [PATCH net-next v6 00/14] Add PPE driver for Qualcomm IPQ9574 SoC Luo Jie
` (3 preceding siblings ...)
2025-07-20 10:57 ` [PATCH net-next v6 04/14] net: ethernet: qualcomm: Initialize PPE buffer management for IPQ9574 Luo Jie
@ 2025-07-20 10:57 ` Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 06/14] net: ethernet: qualcomm: Initialize the PPE scheduler settings Luo Jie
` (8 subsequent siblings)
13 siblings, 0 replies; 17+ messages in thread
From: Luo Jie @ 2025-07-20 10:57 UTC (permalink / raw)
To: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Rob Herring, Krzysztof Kozlowski, Conor Dooley,
Lei Wei, Suruchi Agarwal, Pavithra R, Simon Horman,
Jonathan Corbet, Kees Cook, Gustavo A. R. Silva, Philipp Zabel
Cc: linux-arm-msm, netdev, devicetree, linux-kernel, linux-doc,
linux-hardening, quic_kkumarcs, quic_linchen, Luo Jie
QM (queue management) configurations decide the length of PPE queues
and the queue depth for these queues which are used to drop packets
in events of congestion.
There are two types of PPE queues - unicast queues (0-255) and multicast
queues (256-299). These queue types are used to forward different types
of traffic, and are configured with different lengths.
Signed-off-by: Luo Jie <quic_luoj@quicinc.com>
---
drivers/net/ethernet/qualcomm/ppe/ppe_config.c | 184 ++++++++++++++++++++++++-
drivers/net/ethernet/qualcomm/ppe/ppe_regs.h | 85 ++++++++++++
2 files changed, 268 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_config.c b/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
index 45b031d4dd46..53887069b432 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
@@ -43,6 +43,29 @@ struct ppe_bm_port_config {
bool dynamic;
};
+/**
+ * struct ppe_qm_queue_config - PPE queue config.
+ * @queue_start: PPE start of queue ID.
+ * @queue_end: PPE end of queue ID.
+ * @prealloc_buf: Queue dedicated buffer number.
+ * @ceil: Ceil to start drop packet from queue.
+ * @weight: Weight value.
+ * @resume_offset: Resume offset from the threshold.
+ * @dynamic: Threshold value is decided dynamically or statically.
+ *
+ * Queue configuration decides the threshold to drop packet from PPE
+ * hardware queue.
+ */
+struct ppe_qm_queue_config {
+ unsigned int queue_start;
+ unsigned int queue_end;
+ unsigned int prealloc_buf;
+ unsigned int ceil;
+ unsigned int weight;
+ unsigned int resume_offset;
+ bool dynamic;
+};
+
/* There are total 2048 buffers available in PPE, out of which some
* buffers are reserved for some specific purposes per PPE port. The
* rest of the pool of 1550 buffers are assigned to the general 'group0'
@@ -106,6 +129,40 @@ static const struct ppe_bm_port_config ipq9574_ppe_bm_port_config[] = {
},
};
+/* QM fetches the packet from PPE buffer management for transmitting the
+ * packet out. The QM group configuration limits the total number of buffers
+ * enqueued by all PPE hardware queues.
+ * There are total 2048 buffers available, out of which some buffers are
+ * dedicated to hardware exception handlers. The remaining buffers are
+ * assigned to the general 'group0', which is the group assigned to all
+ * queues by default.
+ */
+static const int ipq9574_ppe_qm_group_config = 2000;
+
+/* Default QM settings for unicast and multicast queues for IPQ9754. */
+static const struct ppe_qm_queue_config ipq9574_ppe_qm_queue_config[] = {
+ {
+ /* QM settings for unicast queues 0 to 255. */
+ .queue_start = 0,
+ .queue_end = 255,
+ .prealloc_buf = 0,
+ .ceil = 1200,
+ .weight = 7,
+ .resume_offset = 36,
+ .dynamic = true,
+ },
+ {
+ /* QM settings for multicast queues 256 to 299. */
+ .queue_start = 256,
+ .queue_end = 299,
+ .prealloc_buf = 0,
+ .ceil = 250,
+ .weight = 0,
+ .resume_offset = 36,
+ .dynamic = false,
+ },
+};
+
static int ppe_config_bm_threshold(struct ppe_device *ppe_dev, int bm_port_id,
const struct ppe_bm_port_config port_cfg)
{
@@ -193,7 +250,132 @@ static int ppe_config_bm(struct ppe_device *ppe_dev)
return ret;
}
+/* Configure PPE hardware queue depth, which is decided by the threshold
+ * of queue.
+ */
+static int ppe_config_qm(struct ppe_device *ppe_dev)
+{
+ const struct ppe_qm_queue_config *queue_cfg;
+ int ret, i, queue_id, queue_cfg_count;
+ u32 reg, multicast_queue_cfg[5];
+ u32 unicast_queue_cfg[4];
+ u32 group_cfg[3];
+
+ /* Assign the buffer number to the group 0 by default. */
+ reg = PPE_AC_GRP_CFG_TBL_ADDR;
+ ret = regmap_bulk_read(ppe_dev->regmap, reg,
+ group_cfg, ARRAY_SIZE(group_cfg));
+ if (ret)
+ goto qm_config_fail;
+
+ PPE_AC_GRP_SET_BUF_LIMIT(group_cfg, ipq9574_ppe_qm_group_config);
+
+ ret = regmap_bulk_write(ppe_dev->regmap, reg,
+ group_cfg, ARRAY_SIZE(group_cfg));
+ if (ret)
+ goto qm_config_fail;
+
+ queue_cfg = ipq9574_ppe_qm_queue_config;
+ queue_cfg_count = ARRAY_SIZE(ipq9574_ppe_qm_queue_config);
+ for (i = 0; i < queue_cfg_count; i++) {
+ queue_id = queue_cfg[i].queue_start;
+
+ /* Configure threshold for dropping packets separately for
+ * unicast and multicast PPE queues.
+ */
+ while (queue_id <= queue_cfg[i].queue_end) {
+ if (queue_id < PPE_AC_UNICAST_QUEUE_CFG_TBL_ENTRIES) {
+ reg = PPE_AC_UNICAST_QUEUE_CFG_TBL_ADDR +
+ PPE_AC_UNICAST_QUEUE_CFG_TBL_INC * queue_id;
+
+ ret = regmap_bulk_read(ppe_dev->regmap, reg,
+ unicast_queue_cfg,
+ ARRAY_SIZE(unicast_queue_cfg));
+ if (ret)
+ goto qm_config_fail;
+
+ PPE_AC_UNICAST_QUEUE_SET_EN(unicast_queue_cfg, true);
+ PPE_AC_UNICAST_QUEUE_SET_GRP_ID(unicast_queue_cfg, 0);
+ PPE_AC_UNICAST_QUEUE_SET_PRE_LIMIT(unicast_queue_cfg,
+ queue_cfg[i].prealloc_buf);
+ PPE_AC_UNICAST_QUEUE_SET_DYNAMIC(unicast_queue_cfg,
+ queue_cfg[i].dynamic);
+ PPE_AC_UNICAST_QUEUE_SET_WEIGHT(unicast_queue_cfg,
+ queue_cfg[i].weight);
+ PPE_AC_UNICAST_QUEUE_SET_THRESHOLD(unicast_queue_cfg,
+ queue_cfg[i].ceil);
+ PPE_AC_UNICAST_QUEUE_SET_GRN_RESUME(unicast_queue_cfg,
+ queue_cfg[i].resume_offset);
+
+ ret = regmap_bulk_write(ppe_dev->regmap, reg,
+ unicast_queue_cfg,
+ ARRAY_SIZE(unicast_queue_cfg));
+ if (ret)
+ goto qm_config_fail;
+ } else {
+ reg = PPE_AC_MULTICAST_QUEUE_CFG_TBL_ADDR +
+ PPE_AC_MULTICAST_QUEUE_CFG_TBL_INC * queue_id;
+
+ ret = regmap_bulk_read(ppe_dev->regmap, reg,
+ multicast_queue_cfg,
+ ARRAY_SIZE(multicast_queue_cfg));
+ if (ret)
+ goto qm_config_fail;
+
+ PPE_AC_MULTICAST_QUEUE_SET_EN(multicast_queue_cfg, true);
+ PPE_AC_MULTICAST_QUEUE_SET_GRN_GRP_ID(multicast_queue_cfg, 0);
+ PPE_AC_MULTICAST_QUEUE_SET_GRN_PRE_LIMIT(multicast_queue_cfg,
+ queue_cfg[i].prealloc_buf);
+ PPE_AC_MULTICAST_QUEUE_SET_GRN_THRESHOLD(multicast_queue_cfg,
+ queue_cfg[i].ceil);
+ PPE_AC_MULTICAST_QUEUE_SET_GRN_RESUME(multicast_queue_cfg,
+ queue_cfg[i].resume_offset);
+
+ ret = regmap_bulk_write(ppe_dev->regmap, reg,
+ multicast_queue_cfg,
+ ARRAY_SIZE(multicast_queue_cfg));
+ if (ret)
+ goto qm_config_fail;
+ }
+
+ /* Enable enqueue. */
+ reg = PPE_ENQ_OPR_TBL_ADDR + PPE_ENQ_OPR_TBL_INC * queue_id;
+ ret = regmap_clear_bits(ppe_dev->regmap, reg,
+ PPE_ENQ_OPR_TBL_ENQ_DISABLE);
+ if (ret)
+ goto qm_config_fail;
+
+ /* Enable dequeue. */
+ reg = PPE_DEQ_OPR_TBL_ADDR + PPE_DEQ_OPR_TBL_INC * queue_id;
+ ret = regmap_clear_bits(ppe_dev->regmap, reg,
+ PPE_DEQ_OPR_TBL_DEQ_DISABLE);
+ if (ret)
+ goto qm_config_fail;
+
+ queue_id++;
+ }
+ }
+
+ /* Enable queue counter for all PPE hardware queues. */
+ ret = regmap_set_bits(ppe_dev->regmap, PPE_EG_BRIDGE_CONFIG_ADDR,
+ PPE_EG_BRIDGE_CONFIG_QUEUE_CNT_EN);
+ if (ret)
+ goto qm_config_fail;
+
+ return 0;
+
+qm_config_fail:
+ dev_err(ppe_dev->dev, "PPE QM config error %d\n", ret);
+ return ret;
+}
+
int ppe_hw_config(struct ppe_device *ppe_dev)
{
- return ppe_config_bm(ppe_dev);
+ int ret;
+
+ ret = ppe_config_bm(ppe_dev);
+ if (ret)
+ return ret;
+
+ return ppe_config_qm(ppe_dev);
}
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h b/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
index b89d717fdae8..ca256fe2a321 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
@@ -9,6 +9,16 @@
#include <linux/bitfield.h>
+/* PPE queue counters enable/disable control. */
+#define PPE_EG_BRIDGE_CONFIG_ADDR 0x20044
+#define PPE_EG_BRIDGE_CONFIG_QUEUE_CNT_EN BIT(2)
+
+/* Table addresses for per-queue dequeue setting. */
+#define PPE_DEQ_OPR_TBL_ADDR 0x430000
+#define PPE_DEQ_OPR_TBL_ENTRIES 300
+#define PPE_DEQ_OPR_TBL_INC 0x10
+#define PPE_DEQ_OPR_TBL_DEQ_DISABLE BIT(0)
+
/* There are 15 BM ports and 4 BM groups supported by PPE.
* BM port (0-7) is for EDMA port 0, BM port (8-13) is for
* PPE physical port 1-6 and BM port 14 is for EIP port.
@@ -56,4 +66,79 @@
FIELD_MODIFY(PPE_BM_PORT_FC_W1_DYNAMIC, (tbl_cfg) + 0x1, value)
#define PPE_BM_PORT_FC_SET_PRE_ALLOC(tbl_cfg, value) \
FIELD_MODIFY(PPE_BM_PORT_FC_W1_PRE_ALLOC, (tbl_cfg) + 0x1, value)
+
+/* PPE unicast queue (0-255) configurations. */
+#define PPE_AC_UNICAST_QUEUE_CFG_TBL_ADDR 0x848000
+#define PPE_AC_UNICAST_QUEUE_CFG_TBL_ENTRIES 256
+#define PPE_AC_UNICAST_QUEUE_CFG_TBL_INC 0x10
+#define PPE_AC_UNICAST_QUEUE_CFG_W0_EN BIT(0)
+#define PPE_AC_UNICAST_QUEUE_CFG_W0_WRED_EN BIT(1)
+#define PPE_AC_UNICAST_QUEUE_CFG_W0_FC_EN BIT(2)
+#define PPE_AC_UNICAST_QUEUE_CFG_W0_CLR_AWARE BIT(3)
+#define PPE_AC_UNICAST_QUEUE_CFG_W0_GRP_ID GENMASK(5, 4)
+#define PPE_AC_UNICAST_QUEUE_CFG_W0_PRE_LIMIT GENMASK(16, 6)
+#define PPE_AC_UNICAST_QUEUE_CFG_W0_DYNAMIC BIT(17)
+#define PPE_AC_UNICAST_QUEUE_CFG_W0_WEIGHT GENMASK(20, 18)
+#define PPE_AC_UNICAST_QUEUE_CFG_W0_THRESHOLD GENMASK(31, 21)
+#define PPE_AC_UNICAST_QUEUE_CFG_W3_GRN_RESUME GENMASK(23, 13)
+
+#define PPE_AC_UNICAST_QUEUE_SET_EN(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_AC_UNICAST_QUEUE_CFG_W0_EN, tbl_cfg, value)
+#define PPE_AC_UNICAST_QUEUE_SET_GRP_ID(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_AC_UNICAST_QUEUE_CFG_W0_GRP_ID, tbl_cfg, value)
+#define PPE_AC_UNICAST_QUEUE_SET_PRE_LIMIT(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_AC_UNICAST_QUEUE_CFG_W0_PRE_LIMIT, tbl_cfg, value)
+#define PPE_AC_UNICAST_QUEUE_SET_DYNAMIC(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_AC_UNICAST_QUEUE_CFG_W0_DYNAMIC, tbl_cfg, value)
+#define PPE_AC_UNICAST_QUEUE_SET_WEIGHT(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_AC_UNICAST_QUEUE_CFG_W0_WEIGHT, tbl_cfg, value)
+#define PPE_AC_UNICAST_QUEUE_SET_THRESHOLD(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_AC_UNICAST_QUEUE_CFG_W0_THRESHOLD, tbl_cfg, value)
+#define PPE_AC_UNICAST_QUEUE_SET_GRN_RESUME(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_AC_UNICAST_QUEUE_CFG_W3_GRN_RESUME, (tbl_cfg) + 0x3, value)
+
+/* PPE multicast queue (256-299) configurations. */
+#define PPE_AC_MULTICAST_QUEUE_CFG_TBL_ADDR 0x84a000
+#define PPE_AC_MULTICAST_QUEUE_CFG_TBL_ENTRIES 44
+#define PPE_AC_MULTICAST_QUEUE_CFG_TBL_INC 0x10
+#define PPE_AC_MULTICAST_QUEUE_CFG_W0_EN BIT(0)
+#define PPE_AC_MULTICAST_QUEUE_CFG_W0_FC_EN BIT(1)
+#define PPE_AC_MULTICAST_QUEUE_CFG_W0_CLR_AWARE BIT(2)
+#define PPE_AC_MULTICAST_QUEUE_CFG_W0_GRP_ID GENMASK(4, 3)
+#define PPE_AC_MULTICAST_QUEUE_CFG_W0_PRE_LIMIT GENMASK(15, 5)
+#define PPE_AC_MULTICAST_QUEUE_CFG_W0_THRESHOLD GENMASK(26, 16)
+#define PPE_AC_MULTICAST_QUEUE_CFG_W2_RESUME GENMASK(17, 7)
+
+#define PPE_AC_MULTICAST_QUEUE_SET_EN(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_AC_MULTICAST_QUEUE_CFG_W0_EN, tbl_cfg, value)
+#define PPE_AC_MULTICAST_QUEUE_SET_GRN_GRP_ID(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_AC_MULTICAST_QUEUE_CFG_W0_GRP_ID, tbl_cfg, value)
+#define PPE_AC_MULTICAST_QUEUE_SET_GRN_PRE_LIMIT(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_AC_MULTICAST_QUEUE_CFG_W0_PRE_LIMIT, tbl_cfg, value)
+#define PPE_AC_MULTICAST_QUEUE_SET_GRN_THRESHOLD(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_AC_MULTICAST_QUEUE_CFG_W0_THRESHOLD, tbl_cfg, value)
+#define PPE_AC_MULTICAST_QUEUE_SET_GRN_RESUME(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_AC_MULTICAST_QUEUE_CFG_W2_RESUME, (tbl_cfg) + 0x2, value)
+
+/* PPE admission control group (0-3) configurations */
+#define PPE_AC_GRP_CFG_TBL_ADDR 0x84c000
+#define PPE_AC_GRP_CFG_TBL_ENTRIES 0x4
+#define PPE_AC_GRP_CFG_TBL_INC 0x10
+#define PPE_AC_GRP_W0_AC_EN BIT(0)
+#define PPE_AC_GRP_W0_AC_FC_EN BIT(1)
+#define PPE_AC_GRP_W0_CLR_AWARE BIT(2)
+#define PPE_AC_GRP_W0_THRESHOLD_LOW GENMASK(31, 25)
+#define PPE_AC_GRP_W1_THRESHOLD_HIGH GENMASK(3, 0)
+#define PPE_AC_GRP_W1_BUF_LIMIT GENMASK(14, 4)
+#define PPE_AC_GRP_W2_RESUME_GRN GENMASK(15, 5)
+#define PPE_AC_GRP_W2_PRE_ALLOC GENMASK(26, 16)
+
+#define PPE_AC_GRP_SET_BUF_LIMIT(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_AC_GRP_W1_BUF_LIMIT, (tbl_cfg) + 0x1, value)
+
+/* Table addresses for per-queue enqueue setting. */
+#define PPE_ENQ_OPR_TBL_ADDR 0x85c000
+#define PPE_ENQ_OPR_TBL_ENTRIES 300
+#define PPE_ENQ_OPR_TBL_INC 0x10
+#define PPE_ENQ_OPR_TBL_ENQ_DISABLE BIT(0)
#endif
--
2.34.1
^ permalink raw reply related [flat|nested] 17+ messages in thread
* [PATCH net-next v6 06/14] net: ethernet: qualcomm: Initialize the PPE scheduler settings
2025-07-20 10:57 [PATCH net-next v6 00/14] Add PPE driver for Qualcomm IPQ9574 SoC Luo Jie
` (4 preceding siblings ...)
2025-07-20 10:57 ` [PATCH net-next v6 05/14] net: ethernet: qualcomm: Initialize PPE queue " Luo Jie
@ 2025-07-20 10:57 ` Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 07/14] net: ethernet: qualcomm: Initialize PPE queue settings Luo Jie
` (7 subsequent siblings)
13 siblings, 0 replies; 17+ messages in thread
From: Luo Jie @ 2025-07-20 10:57 UTC (permalink / raw)
To: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Rob Herring, Krzysztof Kozlowski, Conor Dooley,
Lei Wei, Suruchi Agarwal, Pavithra R, Simon Horman,
Jonathan Corbet, Kees Cook, Gustavo A. R. Silva, Philipp Zabel
Cc: linux-arm-msm, netdev, devicetree, linux-kernel, linux-doc,
linux-hardening, quic_kkumarcs, quic_linchen, Luo Jie
The PPE scheduler settings determine the priority of scheduling the
packet across the different hardware queues per PPE port.
Signed-off-by: Luo Jie <quic_luoj@quicinc.com>
---
drivers/net/ethernet/qualcomm/ppe/ppe_config.c | 804 ++++++++++++++++++++++++-
drivers/net/ethernet/qualcomm/ppe/ppe_config.h | 37 ++
drivers/net/ethernet/qualcomm/ppe/ppe_regs.h | 97 +++
3 files changed, 937 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_config.c b/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
index 53887069b432..3c45d0d1c560 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
@@ -16,6 +16,8 @@
#include "ppe_config.h"
#include "ppe_regs.h"
+#define PPE_QUEUE_SCH_PRI_NUM 8
+
/**
* struct ppe_bm_port_config - PPE BM port configuration.
* @port_id_start: The fist BM port ID to configure.
@@ -66,6 +68,76 @@ struct ppe_qm_queue_config {
bool dynamic;
};
+/**
+ * enum ppe_scheduler_direction - PPE scheduler direction for packet.
+ * @PPE_SCH_INGRESS: Scheduler for the packet on ingress,
+ * @PPE_SCH_EGRESS: Scheduler for the packet on egress,
+ */
+enum ppe_scheduler_direction {
+ PPE_SCH_INGRESS = 0,
+ PPE_SCH_EGRESS = 1,
+};
+
+/**
+ * struct ppe_scheduler_bm_config - PPE arbitration for buffer config.
+ * @valid: Arbitration entry valid or not.
+ * @dir: Arbitration entry for egress or ingress.
+ * @port: Port ID to use arbitration entry.
+ * @backup_port_valid: Backup port valid or not.
+ * @backup_port: Backup port ID to use.
+ *
+ * Configure the scheduler settings for accessing and releasing the PPE buffers.
+ */
+struct ppe_scheduler_bm_config {
+ bool valid;
+ enum ppe_scheduler_direction dir;
+ unsigned int port;
+ bool backup_port_valid;
+ unsigned int backup_port;
+};
+
+/**
+ * struct ppe_scheduler_qm_config - PPE arbitration for scheduler config.
+ * @ensch_port_bmp: Port bit map for enqueue scheduler.
+ * @ensch_port: Port ID to enqueue scheduler.
+ * @desch_port: Port ID to dequeue scheduler.
+ * @desch_backup_port_valid: Dequeue for the backup port valid or not.
+ * @desch_backup_port: Backup port ID to dequeue scheduler.
+ *
+ * Configure the scheduler settings for enqueuing and dequeuing packets on
+ * the PPE port.
+ */
+struct ppe_scheduler_qm_config {
+ unsigned int ensch_port_bmp;
+ unsigned int ensch_port;
+ unsigned int desch_port;
+ bool desch_backup_port_valid;
+ unsigned int desch_backup_port;
+};
+
+/**
+ * struct ppe_scheduler_port_config - PPE port scheduler config.
+ * @port: Port ID to be scheduled.
+ * @flow_level: Scheduler flow level or not.
+ * @node_id: Node ID, for level 0, queue ID is used.
+ * @loop_num: Loop number of scheduler config.
+ * @pri_max: Max priority configured.
+ * @flow_id: Strict priority ID.
+ * @drr_node_id: Node ID for scheduler.
+ *
+ * PPE port scheduler configuration which decides the priority in the
+ * packet scheduler for the egress port.
+ */
+struct ppe_scheduler_port_config {
+ unsigned int port;
+ bool flow_level;
+ unsigned int node_id;
+ unsigned int loop_num;
+ unsigned int pri_max;
+ unsigned int flow_id;
+ unsigned int drr_node_id;
+};
+
/* There are total 2048 buffers available in PPE, out of which some
* buffers are reserved for some specific purposes per PPE port. The
* rest of the pool of 1550 buffers are assigned to the general 'group0'
@@ -163,6 +235,603 @@ static const struct ppe_qm_queue_config ipq9574_ppe_qm_queue_config[] = {
},
};
+/* PPE scheduler configuration for BM includes multiple entries. Each entry
+ * indicates the primary port to be assigned the buffers for the ingress or
+ * to release the buffers for the egress. Backup port ID will be used when
+ * the primary port ID is down.
+ */
+static const struct ppe_scheduler_bm_config ipq9574_ppe_sch_bm_config[] = {
+ {true, PPE_SCH_INGRESS, 0, false, 0},
+ {true, PPE_SCH_EGRESS, 0, false, 0},
+ {true, PPE_SCH_INGRESS, 5, false, 0},
+ {true, PPE_SCH_EGRESS, 5, false, 0},
+ {true, PPE_SCH_INGRESS, 6, false, 0},
+ {true, PPE_SCH_EGRESS, 6, false, 0},
+ {true, PPE_SCH_INGRESS, 1, false, 0},
+ {true, PPE_SCH_EGRESS, 1, false, 0},
+ {true, PPE_SCH_INGRESS, 0, false, 0},
+ {true, PPE_SCH_EGRESS, 0, false, 0},
+ {true, PPE_SCH_INGRESS, 5, false, 0},
+ {true, PPE_SCH_EGRESS, 5, false, 0},
+ {true, PPE_SCH_INGRESS, 6, false, 0},
+ {true, PPE_SCH_EGRESS, 6, false, 0},
+ {true, PPE_SCH_INGRESS, 7, false, 0},
+ {true, PPE_SCH_EGRESS, 7, false, 0},
+ {true, PPE_SCH_INGRESS, 0, false, 0},
+ {true, PPE_SCH_EGRESS, 0, false, 0},
+ {true, PPE_SCH_INGRESS, 1, false, 0},
+ {true, PPE_SCH_EGRESS, 1, false, 0},
+ {true, PPE_SCH_INGRESS, 5, false, 0},
+ {true, PPE_SCH_EGRESS, 5, false, 0},
+ {true, PPE_SCH_INGRESS, 6, false, 0},
+ {true, PPE_SCH_EGRESS, 6, false, 0},
+ {true, PPE_SCH_INGRESS, 2, false, 0},
+ {true, PPE_SCH_EGRESS, 2, false, 0},
+ {true, PPE_SCH_INGRESS, 0, false, 0},
+ {true, PPE_SCH_EGRESS, 0, false, 0},
+ {true, PPE_SCH_INGRESS, 5, false, 0},
+ {true, PPE_SCH_EGRESS, 5, false, 0},
+ {true, PPE_SCH_INGRESS, 6, false, 0},
+ {true, PPE_SCH_EGRESS, 6, false, 0},
+ {true, PPE_SCH_INGRESS, 1, false, 0},
+ {true, PPE_SCH_EGRESS, 1, false, 0},
+ {true, PPE_SCH_INGRESS, 3, false, 0},
+ {true, PPE_SCH_EGRESS, 3, false, 0},
+ {true, PPE_SCH_INGRESS, 0, false, 0},
+ {true, PPE_SCH_EGRESS, 0, false, 0},
+ {true, PPE_SCH_INGRESS, 5, false, 0},
+ {true, PPE_SCH_EGRESS, 5, false, 0},
+ {true, PPE_SCH_INGRESS, 6, false, 0},
+ {true, PPE_SCH_EGRESS, 6, false, 0},
+ {true, PPE_SCH_INGRESS, 7, false, 0},
+ {true, PPE_SCH_EGRESS, 7, false, 0},
+ {true, PPE_SCH_INGRESS, 0, false, 0},
+ {true, PPE_SCH_EGRESS, 0, false, 0},
+ {true, PPE_SCH_INGRESS, 1, false, 0},
+ {true, PPE_SCH_EGRESS, 1, false, 0},
+ {true, PPE_SCH_INGRESS, 5, false, 0},
+ {true, PPE_SCH_EGRESS, 5, false, 0},
+ {true, PPE_SCH_INGRESS, 6, false, 0},
+ {true, PPE_SCH_EGRESS, 6, false, 0},
+ {true, PPE_SCH_INGRESS, 4, false, 0},
+ {true, PPE_SCH_EGRESS, 4, false, 0},
+ {true, PPE_SCH_INGRESS, 0, false, 0},
+ {true, PPE_SCH_EGRESS, 0, false, 0},
+ {true, PPE_SCH_INGRESS, 5, false, 0},
+ {true, PPE_SCH_EGRESS, 5, false, 0},
+ {true, PPE_SCH_INGRESS, 6, false, 0},
+ {true, PPE_SCH_EGRESS, 6, false, 0},
+ {true, PPE_SCH_INGRESS, 1, false, 0},
+ {true, PPE_SCH_EGRESS, 1, false, 0},
+ {true, PPE_SCH_INGRESS, 0, false, 0},
+ {true, PPE_SCH_EGRESS, 0, false, 0},
+ {true, PPE_SCH_INGRESS, 5, false, 0},
+ {true, PPE_SCH_EGRESS, 5, false, 0},
+ {true, PPE_SCH_INGRESS, 6, false, 0},
+ {true, PPE_SCH_EGRESS, 6, false, 0},
+ {true, PPE_SCH_INGRESS, 2, false, 0},
+ {true, PPE_SCH_EGRESS, 2, false, 0},
+ {true, PPE_SCH_INGRESS, 0, false, 0},
+ {true, PPE_SCH_EGRESS, 0, false, 0},
+ {true, PPE_SCH_INGRESS, 7, false, 0},
+ {true, PPE_SCH_EGRESS, 7, false, 0},
+ {true, PPE_SCH_INGRESS, 5, false, 0},
+ {true, PPE_SCH_EGRESS, 5, false, 0},
+ {true, PPE_SCH_INGRESS, 6, false, 0},
+ {true, PPE_SCH_EGRESS, 6, false, 0},
+ {true, PPE_SCH_INGRESS, 1, false, 0},
+ {true, PPE_SCH_EGRESS, 1, false, 0},
+ {true, PPE_SCH_INGRESS, 0, false, 0},
+ {true, PPE_SCH_EGRESS, 0, false, 0},
+ {true, PPE_SCH_INGRESS, 5, false, 0},
+ {true, PPE_SCH_EGRESS, 5, false, 0},
+ {true, PPE_SCH_INGRESS, 6, false, 0},
+ {true, PPE_SCH_EGRESS, 6, false, 0},
+ {true, PPE_SCH_INGRESS, 3, false, 0},
+ {true, PPE_SCH_EGRESS, 3, false, 0},
+ {true, PPE_SCH_INGRESS, 1, false, 0},
+ {true, PPE_SCH_EGRESS, 1, false, 0},
+ {true, PPE_SCH_INGRESS, 0, false, 0},
+ {true, PPE_SCH_EGRESS, 0, false, 0},
+ {true, PPE_SCH_INGRESS, 5, false, 0},
+ {true, PPE_SCH_EGRESS, 5, false, 0},
+ {true, PPE_SCH_INGRESS, 6, false, 0},
+ {true, PPE_SCH_EGRESS, 6, false, 0},
+ {true, PPE_SCH_INGRESS, 4, false, 0},
+ {true, PPE_SCH_EGRESS, 4, false, 0},
+ {true, PPE_SCH_INGRESS, 7, false, 0},
+ {true, PPE_SCH_EGRESS, 7, false, 0},
+};
+
+/* PPE scheduler configuration for QM includes multiple entries. Each entry
+ * contains ports to be dispatched for enqueueing and dequeueing. The backup
+ * port for dequeueing is supported to be used when the primary port for
+ * dequeueing is down.
+ */
+static const struct ppe_scheduler_qm_config ipq9574_ppe_sch_qm_config[] = {
+ {0x98, 6, 0, true, 1},
+ {0x94, 5, 6, true, 3},
+ {0x86, 0, 5, true, 4},
+ {0x8C, 1, 6, true, 0},
+ {0x1C, 7, 5, true, 1},
+ {0x98, 2, 6, true, 0},
+ {0x1C, 5, 7, true, 1},
+ {0x34, 3, 6, true, 0},
+ {0x8C, 4, 5, true, 1},
+ {0x98, 2, 6, true, 0},
+ {0x8C, 5, 4, true, 1},
+ {0xA8, 0, 6, true, 2},
+ {0x98, 5, 1, true, 0},
+ {0x98, 6, 5, true, 2},
+ {0x89, 1, 6, true, 4},
+ {0xA4, 3, 0, true, 1},
+ {0x8C, 5, 6, true, 4},
+ {0xA8, 0, 2, true, 1},
+ {0x98, 6, 5, true, 0},
+ {0xC4, 4, 3, true, 1},
+ {0x94, 6, 5, true, 0},
+ {0x1C, 7, 6, true, 1},
+ {0x98, 2, 5, true, 0},
+ {0x1C, 6, 7, true, 1},
+ {0x1C, 5, 6, true, 0},
+ {0x94, 3, 5, true, 1},
+ {0x8C, 4, 6, true, 0},
+ {0x94, 1, 5, true, 3},
+ {0x94, 6, 1, true, 0},
+ {0xD0, 3, 5, true, 2},
+ {0x98, 6, 0, true, 1},
+ {0x94, 5, 6, true, 3},
+ {0x94, 1, 5, true, 0},
+ {0x98, 2, 6, true, 1},
+ {0x8C, 4, 5, true, 0},
+ {0x1C, 7, 6, true, 1},
+ {0x8C, 0, 5, true, 4},
+ {0x89, 1, 6, true, 2},
+ {0x98, 5, 0, true, 1},
+ {0x94, 6, 5, true, 3},
+ {0x92, 0, 6, true, 2},
+ {0x98, 1, 5, true, 0},
+ {0x98, 6, 2, true, 1},
+ {0xD0, 0, 5, true, 3},
+ {0x94, 6, 0, true, 1},
+ {0x8C, 5, 6, true, 4},
+ {0x8C, 1, 5, true, 0},
+ {0x1C, 6, 7, true, 1},
+ {0x1C, 5, 6, true, 0},
+ {0xB0, 2, 3, true, 1},
+ {0xC4, 4, 5, true, 0},
+ {0x8C, 6, 4, true, 1},
+ {0xA4, 3, 6, true, 0},
+ {0x1C, 5, 7, true, 1},
+ {0x4C, 0, 5, true, 4},
+ {0x8C, 6, 0, true, 1},
+ {0x34, 7, 6, true, 3},
+ {0x94, 5, 0, true, 1},
+ {0x98, 6, 5, true, 2},
+};
+
+static const struct ppe_scheduler_port_config ppe_port_sch_config[] = {
+ {
+ .port = 0,
+ .flow_level = true,
+ .node_id = 0,
+ .loop_num = 1,
+ .pri_max = 1,
+ .flow_id = 0,
+ .drr_node_id = 0,
+ },
+ {
+ .port = 0,
+ .flow_level = false,
+ .node_id = 0,
+ .loop_num = 8,
+ .pri_max = 8,
+ .flow_id = 0,
+ .drr_node_id = 0,
+ },
+ {
+ .port = 0,
+ .flow_level = false,
+ .node_id = 8,
+ .loop_num = 8,
+ .pri_max = 8,
+ .flow_id = 0,
+ .drr_node_id = 0,
+ },
+ {
+ .port = 0,
+ .flow_level = false,
+ .node_id = 16,
+ .loop_num = 8,
+ .pri_max = 8,
+ .flow_id = 0,
+ .drr_node_id = 0,
+ },
+ {
+ .port = 0,
+ .flow_level = false,
+ .node_id = 24,
+ .loop_num = 8,
+ .pri_max = 8,
+ .flow_id = 0,
+ .drr_node_id = 0,
+ },
+ {
+ .port = 0,
+ .flow_level = false,
+ .node_id = 32,
+ .loop_num = 8,
+ .pri_max = 8,
+ .flow_id = 0,
+ .drr_node_id = 0,
+ },
+ {
+ .port = 0,
+ .flow_level = false,
+ .node_id = 40,
+ .loop_num = 8,
+ .pri_max = 8,
+ .flow_id = 0,
+ .drr_node_id = 0,
+ },
+ {
+ .port = 0,
+ .flow_level = false,
+ .node_id = 48,
+ .loop_num = 8,
+ .pri_max = 8,
+ .flow_id = 0,
+ .drr_node_id = 0,
+ },
+ {
+ .port = 0,
+ .flow_level = false,
+ .node_id = 56,
+ .loop_num = 8,
+ .pri_max = 8,
+ .flow_id = 0,
+ .drr_node_id = 0,
+ },
+ {
+ .port = 0,
+ .flow_level = false,
+ .node_id = 256,
+ .loop_num = 8,
+ .pri_max = 8,
+ .flow_id = 0,
+ .drr_node_id = 0,
+ },
+ {
+ .port = 0,
+ .flow_level = false,
+ .node_id = 264,
+ .loop_num = 8,
+ .pri_max = 8,
+ .flow_id = 0,
+ .drr_node_id = 0,
+ },
+ {
+ .port = 1,
+ .flow_level = true,
+ .node_id = 36,
+ .loop_num = 2,
+ .pri_max = 0,
+ .flow_id = 1,
+ .drr_node_id = 8,
+ },
+ {
+ .port = 1,
+ .flow_level = false,
+ .node_id = 144,
+ .loop_num = 16,
+ .pri_max = 8,
+ .flow_id = 36,
+ .drr_node_id = 48,
+ },
+ {
+ .port = 1,
+ .flow_level = false,
+ .node_id = 272,
+ .loop_num = 4,
+ .pri_max = 4,
+ .flow_id = 36,
+ .drr_node_id = 48,
+ },
+ {
+ .port = 2,
+ .flow_level = true,
+ .node_id = 40,
+ .loop_num = 2,
+ .pri_max = 0,
+ .flow_id = 2,
+ .drr_node_id = 12,
+ },
+ {
+ .port = 2,
+ .flow_level = false,
+ .node_id = 160,
+ .loop_num = 16,
+ .pri_max = 8,
+ .flow_id = 40,
+ .drr_node_id = 64,
+ },
+ {
+ .port = 2,
+ .flow_level = false,
+ .node_id = 276,
+ .loop_num = 4,
+ .pri_max = 4,
+ .flow_id = 40,
+ .drr_node_id = 64,
+ },
+ {
+ .port = 3,
+ .flow_level = true,
+ .node_id = 44,
+ .loop_num = 2,
+ .pri_max = 0,
+ .flow_id = 3,
+ .drr_node_id = 16,
+ },
+ {
+ .port = 3,
+ .flow_level = false,
+ .node_id = 176,
+ .loop_num = 16,
+ .pri_max = 8,
+ .flow_id = 44,
+ .drr_node_id = 80,
+ },
+ {
+ .port = 3,
+ .flow_level = false,
+ .node_id = 280,
+ .loop_num = 4,
+ .pri_max = 4,
+ .flow_id = 44,
+ .drr_node_id = 80,
+ },
+ {
+ .port = 4,
+ .flow_level = true,
+ .node_id = 48,
+ .loop_num = 2,
+ .pri_max = 0,
+ .flow_id = 4,
+ .drr_node_id = 20,
+ },
+ {
+ .port = 4,
+ .flow_level = false,
+ .node_id = 192,
+ .loop_num = 16,
+ .pri_max = 8,
+ .flow_id = 48,
+ .drr_node_id = 96,
+ },
+ {
+ .port = 4,
+ .flow_level = false,
+ .node_id = 284,
+ .loop_num = 4,
+ .pri_max = 4,
+ .flow_id = 48,
+ .drr_node_id = 96,
+ },
+ {
+ .port = 5,
+ .flow_level = true,
+ .node_id = 52,
+ .loop_num = 2,
+ .pri_max = 0,
+ .flow_id = 5,
+ .drr_node_id = 24,
+ },
+ {
+ .port = 5,
+ .flow_level = false,
+ .node_id = 208,
+ .loop_num = 16,
+ .pri_max = 8,
+ .flow_id = 52,
+ .drr_node_id = 112,
+ },
+ {
+ .port = 5,
+ .flow_level = false,
+ .node_id = 288,
+ .loop_num = 4,
+ .pri_max = 4,
+ .flow_id = 52,
+ .drr_node_id = 112,
+ },
+ {
+ .port = 6,
+ .flow_level = true,
+ .node_id = 56,
+ .loop_num = 2,
+ .pri_max = 0,
+ .flow_id = 6,
+ .drr_node_id = 28,
+ },
+ {
+ .port = 6,
+ .flow_level = false,
+ .node_id = 224,
+ .loop_num = 16,
+ .pri_max = 8,
+ .flow_id = 56,
+ .drr_node_id = 128,
+ },
+ {
+ .port = 6,
+ .flow_level = false,
+ .node_id = 292,
+ .loop_num = 4,
+ .pri_max = 4,
+ .flow_id = 56,
+ .drr_node_id = 128,
+ },
+ {
+ .port = 7,
+ .flow_level = true,
+ .node_id = 60,
+ .loop_num = 2,
+ .pri_max = 0,
+ .flow_id = 7,
+ .drr_node_id = 32,
+ },
+ {
+ .port = 7,
+ .flow_level = false,
+ .node_id = 240,
+ .loop_num = 16,
+ .pri_max = 8,
+ .flow_id = 60,
+ .drr_node_id = 144,
+ },
+ {
+ .port = 7,
+ .flow_level = false,
+ .node_id = 296,
+ .loop_num = 4,
+ .pri_max = 4,
+ .flow_id = 60,
+ .drr_node_id = 144,
+ },
+};
+
+/* Set the PPE queue level scheduler configuration. */
+static int ppe_scheduler_l0_queue_map_set(struct ppe_device *ppe_dev,
+ int node_id, int port,
+ struct ppe_scheduler_cfg scheduler_cfg)
+{
+ u32 val, reg;
+ int ret;
+
+ reg = PPE_L0_FLOW_MAP_TBL_ADDR + node_id * PPE_L0_FLOW_MAP_TBL_INC;
+ val = FIELD_PREP(PPE_L0_FLOW_MAP_TBL_FLOW_ID, scheduler_cfg.flow_id);
+ val |= FIELD_PREP(PPE_L0_FLOW_MAP_TBL_C_PRI, scheduler_cfg.pri);
+ val |= FIELD_PREP(PPE_L0_FLOW_MAP_TBL_E_PRI, scheduler_cfg.pri);
+ val |= FIELD_PREP(PPE_L0_FLOW_MAP_TBL_C_NODE_WT, scheduler_cfg.drr_node_wt);
+ val |= FIELD_PREP(PPE_L0_FLOW_MAP_TBL_E_NODE_WT, scheduler_cfg.drr_node_wt);
+
+ ret = regmap_write(ppe_dev->regmap, reg, val);
+ if (ret)
+ return ret;
+
+ reg = PPE_L0_C_FLOW_CFG_TBL_ADDR +
+ (scheduler_cfg.flow_id * PPE_QUEUE_SCH_PRI_NUM + scheduler_cfg.pri) *
+ PPE_L0_C_FLOW_CFG_TBL_INC;
+ val = FIELD_PREP(PPE_L0_C_FLOW_CFG_TBL_NODE_ID, scheduler_cfg.drr_node_id);
+ val |= FIELD_PREP(PPE_L0_C_FLOW_CFG_TBL_NODE_CREDIT_UNIT, scheduler_cfg.unit_is_packet);
+
+ ret = regmap_write(ppe_dev->regmap, reg, val);
+ if (ret)
+ return ret;
+
+ reg = PPE_L0_E_FLOW_CFG_TBL_ADDR +
+ (scheduler_cfg.flow_id * PPE_QUEUE_SCH_PRI_NUM + scheduler_cfg.pri) *
+ PPE_L0_E_FLOW_CFG_TBL_INC;
+ val = FIELD_PREP(PPE_L0_E_FLOW_CFG_TBL_NODE_ID, scheduler_cfg.drr_node_id);
+ val |= FIELD_PREP(PPE_L0_E_FLOW_CFG_TBL_NODE_CREDIT_UNIT, scheduler_cfg.unit_is_packet);
+
+ ret = regmap_write(ppe_dev->regmap, reg, val);
+ if (ret)
+ return ret;
+
+ reg = PPE_L0_FLOW_PORT_MAP_TBL_ADDR + node_id * PPE_L0_FLOW_PORT_MAP_TBL_INC;
+ val = FIELD_PREP(PPE_L0_FLOW_PORT_MAP_TBL_PORT_NUM, port);
+
+ ret = regmap_write(ppe_dev->regmap, reg, val);
+ if (ret)
+ return ret;
+
+ reg = PPE_L0_COMP_CFG_TBL_ADDR + node_id * PPE_L0_COMP_CFG_TBL_INC;
+ val = FIELD_PREP(PPE_L0_COMP_CFG_TBL_NODE_METER_LEN, scheduler_cfg.frame_mode);
+
+ return regmap_update_bits(ppe_dev->regmap, reg,
+ PPE_L0_COMP_CFG_TBL_NODE_METER_LEN,
+ val);
+}
+
+/* Set the PPE flow level scheduler configuration. */
+static int ppe_scheduler_l1_queue_map_set(struct ppe_device *ppe_dev,
+ int node_id, int port,
+ struct ppe_scheduler_cfg scheduler_cfg)
+{
+ u32 val, reg;
+ int ret;
+
+ val = FIELD_PREP(PPE_L1_FLOW_MAP_TBL_FLOW_ID, scheduler_cfg.flow_id);
+ val |= FIELD_PREP(PPE_L1_FLOW_MAP_TBL_C_PRI, scheduler_cfg.pri);
+ val |= FIELD_PREP(PPE_L1_FLOW_MAP_TBL_E_PRI, scheduler_cfg.pri);
+ val |= FIELD_PREP(PPE_L1_FLOW_MAP_TBL_C_NODE_WT, scheduler_cfg.drr_node_wt);
+ val |= FIELD_PREP(PPE_L1_FLOW_MAP_TBL_E_NODE_WT, scheduler_cfg.drr_node_wt);
+ reg = PPE_L1_FLOW_MAP_TBL_ADDR + node_id * PPE_L1_FLOW_MAP_TBL_INC;
+
+ ret = regmap_write(ppe_dev->regmap, reg, val);
+ if (ret)
+ return ret;
+
+ val = FIELD_PREP(PPE_L1_C_FLOW_CFG_TBL_NODE_ID, scheduler_cfg.drr_node_id);
+ val |= FIELD_PREP(PPE_L1_C_FLOW_CFG_TBL_NODE_CREDIT_UNIT, scheduler_cfg.unit_is_packet);
+ reg = PPE_L1_C_FLOW_CFG_TBL_ADDR +
+ (scheduler_cfg.flow_id * PPE_QUEUE_SCH_PRI_NUM + scheduler_cfg.pri) *
+ PPE_L1_C_FLOW_CFG_TBL_INC;
+
+ ret = regmap_write(ppe_dev->regmap, reg, val);
+ if (ret)
+ return ret;
+
+ val = FIELD_PREP(PPE_L1_E_FLOW_CFG_TBL_NODE_ID, scheduler_cfg.drr_node_id);
+ val |= FIELD_PREP(PPE_L1_E_FLOW_CFG_TBL_NODE_CREDIT_UNIT, scheduler_cfg.unit_is_packet);
+ reg = PPE_L1_E_FLOW_CFG_TBL_ADDR +
+ (scheduler_cfg.flow_id * PPE_QUEUE_SCH_PRI_NUM + scheduler_cfg.pri) *
+ PPE_L1_E_FLOW_CFG_TBL_INC;
+
+ ret = regmap_write(ppe_dev->regmap, reg, val);
+ if (ret)
+ return ret;
+
+ val = FIELD_PREP(PPE_L1_FLOW_PORT_MAP_TBL_PORT_NUM, port);
+ reg = PPE_L1_FLOW_PORT_MAP_TBL_ADDR + node_id * PPE_L1_FLOW_PORT_MAP_TBL_INC;
+
+ ret = regmap_write(ppe_dev->regmap, reg, val);
+ if (ret)
+ return ret;
+
+ reg = PPE_L1_COMP_CFG_TBL_ADDR + node_id * PPE_L1_COMP_CFG_TBL_INC;
+ val = FIELD_PREP(PPE_L1_COMP_CFG_TBL_NODE_METER_LEN, scheduler_cfg.frame_mode);
+
+ return regmap_update_bits(ppe_dev->regmap, reg, PPE_L1_COMP_CFG_TBL_NODE_METER_LEN, val);
+}
+
+/**
+ * ppe_queue_scheduler_set - Configure scheduler for PPE hardware queue
+ * @ppe_dev: PPE device
+ * @node_id: PPE queue ID or flow ID
+ * @flow_level: Flow level scheduler or queue level scheduler
+ * @port: PPE port ID set scheduler configuration
+ * @scheduler_cfg: PPE scheduler configuration
+ *
+ * PPE scheduler configuration supports queue level and flow level on
+ * the PPE egress port.
+ *
+ * Return: 0 on success, negative error code on failure.
+ */
+int ppe_queue_scheduler_set(struct ppe_device *ppe_dev,
+ int node_id, bool flow_level, int port,
+ struct ppe_scheduler_cfg scheduler_cfg)
+{
+ if (flow_level)
+ return ppe_scheduler_l1_queue_map_set(ppe_dev, node_id,
+ port, scheduler_cfg);
+
+ return ppe_scheduler_l0_queue_map_set(ppe_dev, node_id,
+ port, scheduler_cfg);
+}
+
static int ppe_config_bm_threshold(struct ppe_device *ppe_dev, int bm_port_id,
const struct ppe_bm_port_config port_cfg)
{
@@ -369,6 +1038,135 @@ static int ppe_config_qm(struct ppe_device *ppe_dev)
return ret;
}
+static int ppe_node_scheduler_config(struct ppe_device *ppe_dev,
+ const struct ppe_scheduler_port_config config)
+{
+ struct ppe_scheduler_cfg sch_cfg;
+ int ret, i;
+
+ for (i = 0; i < config.loop_num; i++) {
+ if (!config.pri_max) {
+ /* Round robin scheduler without priority. */
+ sch_cfg.flow_id = config.flow_id;
+ sch_cfg.pri = 0;
+ sch_cfg.drr_node_id = config.drr_node_id;
+ } else {
+ sch_cfg.flow_id = config.flow_id + (i / config.pri_max);
+ sch_cfg.pri = i % config.pri_max;
+ sch_cfg.drr_node_id = config.drr_node_id + i;
+ }
+
+ /* Scheduler weight, must be more than 0. */
+ sch_cfg.drr_node_wt = 1;
+ /* Byte based to be scheduled. */
+ sch_cfg.unit_is_packet = false;
+ /* Frame + CRC calculated. */
+ sch_cfg.frame_mode = PPE_SCH_WITH_FRAME_CRC;
+
+ ret = ppe_queue_scheduler_set(ppe_dev, config.node_id + i,
+ config.flow_level,
+ config.port,
+ sch_cfg);
+ if (ret)
+ return ret;
+ }
+
+ return 0;
+}
+
+/* Initialize scheduler settings for PPE buffer utilization and dispatching
+ * packet on PPE queue.
+ */
+static int ppe_config_scheduler(struct ppe_device *ppe_dev)
+{
+ const struct ppe_scheduler_port_config *port_cfg;
+ const struct ppe_scheduler_qm_config *qm_cfg;
+ const struct ppe_scheduler_bm_config *bm_cfg;
+ int ret, i, count;
+ u32 val, reg;
+
+ count = ARRAY_SIZE(ipq9574_ppe_sch_bm_config);
+ bm_cfg = ipq9574_ppe_sch_bm_config;
+
+ /* Configure the depth of BM scheduler entries. */
+ val = FIELD_PREP(PPE_BM_SCH_CTRL_SCH_DEPTH, count);
+ val |= FIELD_PREP(PPE_BM_SCH_CTRL_SCH_OFFSET, 0);
+ val |= FIELD_PREP(PPE_BM_SCH_CTRL_SCH_EN, 1);
+
+ ret = regmap_write(ppe_dev->regmap, PPE_BM_SCH_CTRL_ADDR, val);
+ if (ret)
+ goto sch_config_fail;
+
+ /* Configure each BM scheduler entry with the valid ingress port and
+ * egress port, the second port takes effect when the specified port
+ * is in the inactive state.
+ */
+ for (i = 0; i < count; i++) {
+ val = FIELD_PREP(PPE_BM_SCH_CFG_TBL_VALID, bm_cfg[i].valid);
+ val |= FIELD_PREP(PPE_BM_SCH_CFG_TBL_DIR, bm_cfg[i].dir);
+ val |= FIELD_PREP(PPE_BM_SCH_CFG_TBL_PORT_NUM, bm_cfg[i].port);
+ val |= FIELD_PREP(PPE_BM_SCH_CFG_TBL_SECOND_PORT_VALID,
+ bm_cfg[i].backup_port_valid);
+ val |= FIELD_PREP(PPE_BM_SCH_CFG_TBL_SECOND_PORT,
+ bm_cfg[i].backup_port);
+
+ reg = PPE_BM_SCH_CFG_TBL_ADDR + i * PPE_BM_SCH_CFG_TBL_INC;
+ ret = regmap_write(ppe_dev->regmap, reg, val);
+ if (ret)
+ goto sch_config_fail;
+ }
+
+ count = ARRAY_SIZE(ipq9574_ppe_sch_qm_config);
+ qm_cfg = ipq9574_ppe_sch_qm_config;
+
+ /* Configure the depth of QM scheduler entries. */
+ val = FIELD_PREP(PPE_PSCH_SCH_DEPTH_CFG_SCH_DEPTH, count);
+ ret = regmap_write(ppe_dev->regmap, PPE_PSCH_SCH_DEPTH_CFG_ADDR, val);
+ if (ret)
+ goto sch_config_fail;
+
+ /* Configure each QM scheduler entry with enqueue port and dequeue
+ * port, the second port takes effect when the specified dequeue
+ * port is in the inactive port.
+ */
+ for (i = 0; i < count; i++) {
+ val = FIELD_PREP(PPE_PSCH_SCH_CFG_TBL_ENS_PORT_BITMAP,
+ qm_cfg[i].ensch_port_bmp);
+ val |= FIELD_PREP(PPE_PSCH_SCH_CFG_TBL_ENS_PORT,
+ qm_cfg[i].ensch_port);
+ val |= FIELD_PREP(PPE_PSCH_SCH_CFG_TBL_DES_PORT,
+ qm_cfg[i].desch_port);
+ val |= FIELD_PREP(PPE_PSCH_SCH_CFG_TBL_DES_SECOND_PORT_EN,
+ qm_cfg[i].desch_backup_port_valid);
+ val |= FIELD_PREP(PPE_PSCH_SCH_CFG_TBL_DES_SECOND_PORT,
+ qm_cfg[i].desch_backup_port);
+
+ reg = PPE_PSCH_SCH_CFG_TBL_ADDR + i * PPE_PSCH_SCH_CFG_TBL_INC;
+ ret = regmap_write(ppe_dev->regmap, reg, val);
+ if (ret)
+ goto sch_config_fail;
+ }
+
+ count = ARRAY_SIZE(ppe_port_sch_config);
+ port_cfg = ppe_port_sch_config;
+
+ /* Configure scheduler per PPE queue or flow. */
+ for (i = 0; i < count; i++) {
+ if (port_cfg[i].port >= ppe_dev->num_ports)
+ break;
+
+ ret = ppe_node_scheduler_config(ppe_dev, port_cfg[i]);
+ if (ret)
+ goto sch_config_fail;
+ }
+
+ return 0;
+
+sch_config_fail:
+ dev_err(ppe_dev->dev, "PPE scheduler arbitration config error %d\n", ret);
+ return ret;
+};
+
int ppe_hw_config(struct ppe_device *ppe_dev)
{
int ret;
@@ -377,5 +1175,9 @@ int ppe_hw_config(struct ppe_device *ppe_dev)
if (ret)
return ret;
- return ppe_config_qm(ppe_dev);
+ ret = ppe_config_qm(ppe_dev);
+ if (ret)
+ return ret;
+
+ return ppe_config_scheduler(ppe_dev);
}
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_config.h b/drivers/net/ethernet/qualcomm/ppe/ppe_config.h
index 8e032910bfe7..e2c703c0b99d 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe_config.h
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_config.h
@@ -8,5 +8,42 @@
#include "ppe.h"
+/**
+ * enum ppe_scheduler_frame_mode - PPE scheduler frame mode.
+ * @PPE_SCH_WITH_IPG_PREAMBLE_FRAME_CRC: The scheduled frame includes IPG,
+ * preamble, Ethernet packet and CRC.
+ * @PPE_SCH_WITH_FRAME_CRC: The scheduled frame includes Ethernet frame and CRC
+ * excluding IPG and preamble.
+ * @PPE_SCH_WITH_L3_PAYLOAD: The scheduled frame includes layer 3 packet data.
+ */
+enum ppe_scheduler_frame_mode {
+ PPE_SCH_WITH_IPG_PREAMBLE_FRAME_CRC = 0,
+ PPE_SCH_WITH_FRAME_CRC = 1,
+ PPE_SCH_WITH_L3_PAYLOAD = 2,
+};
+
+/**
+ * struct ppe_scheduler_cfg - PPE scheduler configuration.
+ * @flow_id: PPE flow ID.
+ * @pri: Scheduler priority.
+ * @drr_node_id: Node ID for scheduled traffic.
+ * @drr_node_wt: Weight for scheduled traffic.
+ * @unit_is_packet: Packet based or byte based unit for scheduled traffic.
+ * @frame_mode: Packet mode to be scheduled.
+ *
+ * PPE scheduler supports commit rate and exceed rate configurations.
+ */
+struct ppe_scheduler_cfg {
+ int flow_id;
+ int pri;
+ int drr_node_id;
+ int drr_node_wt;
+ bool unit_is_packet;
+ enum ppe_scheduler_frame_mode frame_mode;
+};
+
int ppe_hw_config(struct ppe_device *ppe_dev);
+int ppe_queue_scheduler_set(struct ppe_device *ppe_dev,
+ int node_id, bool flow_level, int port,
+ struct ppe_scheduler_cfg scheduler_cfg);
#endif
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h b/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
index ca256fe2a321..2e43fbc56845 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
@@ -9,16 +9,113 @@
#include <linux/bitfield.h>
+/* PPE scheduler configurations for buffer manager block. */
+#define PPE_BM_SCH_CTRL_ADDR 0xb000
+#define PPE_BM_SCH_CTRL_INC 4
+#define PPE_BM_SCH_CTRL_SCH_DEPTH GENMASK(7, 0)
+#define PPE_BM_SCH_CTRL_SCH_OFFSET GENMASK(14, 8)
+#define PPE_BM_SCH_CTRL_SCH_EN BIT(31)
+
+#define PPE_BM_SCH_CFG_TBL_ADDR 0xc000
+#define PPE_BM_SCH_CFG_TBL_ENTRIES 128
+#define PPE_BM_SCH_CFG_TBL_INC 0x10
+#define PPE_BM_SCH_CFG_TBL_PORT_NUM GENMASK(3, 0)
+#define PPE_BM_SCH_CFG_TBL_DIR BIT(4)
+#define PPE_BM_SCH_CFG_TBL_VALID BIT(5)
+#define PPE_BM_SCH_CFG_TBL_SECOND_PORT_VALID BIT(6)
+#define PPE_BM_SCH_CFG_TBL_SECOND_PORT GENMASK(11, 8)
+
/* PPE queue counters enable/disable control. */
#define PPE_EG_BRIDGE_CONFIG_ADDR 0x20044
#define PPE_EG_BRIDGE_CONFIG_QUEUE_CNT_EN BIT(2)
+/* Port scheduler global config. */
+#define PPE_PSCH_SCH_DEPTH_CFG_ADDR 0x400000
+#define PPE_PSCH_SCH_DEPTH_CFG_INC 4
+#define PPE_PSCH_SCH_DEPTH_CFG_SCH_DEPTH GENMASK(7, 0)
+
+/* PPE queue level scheduler configurations. */
+#define PPE_L0_FLOW_MAP_TBL_ADDR 0x402000
+#define PPE_L0_FLOW_MAP_TBL_ENTRIES 300
+#define PPE_L0_FLOW_MAP_TBL_INC 0x10
+#define PPE_L0_FLOW_MAP_TBL_FLOW_ID GENMASK(5, 0)
+#define PPE_L0_FLOW_MAP_TBL_C_PRI GENMASK(8, 6)
+#define PPE_L0_FLOW_MAP_TBL_E_PRI GENMASK(11, 9)
+#define PPE_L0_FLOW_MAP_TBL_C_NODE_WT GENMASK(21, 12)
+#define PPE_L0_FLOW_MAP_TBL_E_NODE_WT GENMASK(31, 22)
+
+#define PPE_L0_C_FLOW_CFG_TBL_ADDR 0x404000
+#define PPE_L0_C_FLOW_CFG_TBL_ENTRIES 512
+#define PPE_L0_C_FLOW_CFG_TBL_INC 0x10
+#define PPE_L0_C_FLOW_CFG_TBL_NODE_ID GENMASK(7, 0)
+#define PPE_L0_C_FLOW_CFG_TBL_NODE_CREDIT_UNIT BIT(8)
+
+#define PPE_L0_E_FLOW_CFG_TBL_ADDR 0x406000
+#define PPE_L0_E_FLOW_CFG_TBL_ENTRIES 512
+#define PPE_L0_E_FLOW_CFG_TBL_INC 0x10
+#define PPE_L0_E_FLOW_CFG_TBL_NODE_ID GENMASK(7, 0)
+#define PPE_L0_E_FLOW_CFG_TBL_NODE_CREDIT_UNIT BIT(8)
+
+#define PPE_L0_FLOW_PORT_MAP_TBL_ADDR 0x408000
+#define PPE_L0_FLOW_PORT_MAP_TBL_ENTRIES 300
+#define PPE_L0_FLOW_PORT_MAP_TBL_INC 0x10
+#define PPE_L0_FLOW_PORT_MAP_TBL_PORT_NUM GENMASK(3, 0)
+
+#define PPE_L0_COMP_CFG_TBL_ADDR 0x428000
+#define PPE_L0_COMP_CFG_TBL_ENTRIES 300
+#define PPE_L0_COMP_CFG_TBL_INC 0x10
+#define PPE_L0_COMP_CFG_TBL_SHAPER_METER_LEN GENMASK(1, 0)
+#define PPE_L0_COMP_CFG_TBL_NODE_METER_LEN GENMASK(3, 2)
+
/* Table addresses for per-queue dequeue setting. */
#define PPE_DEQ_OPR_TBL_ADDR 0x430000
#define PPE_DEQ_OPR_TBL_ENTRIES 300
#define PPE_DEQ_OPR_TBL_INC 0x10
#define PPE_DEQ_OPR_TBL_DEQ_DISABLE BIT(0)
+/* PPE flow level scheduler configurations. */
+#define PPE_L1_FLOW_MAP_TBL_ADDR 0x440000
+#define PPE_L1_FLOW_MAP_TBL_ENTRIES 64
+#define PPE_L1_FLOW_MAP_TBL_INC 0x10
+#define PPE_L1_FLOW_MAP_TBL_FLOW_ID GENMASK(3, 0)
+#define PPE_L1_FLOW_MAP_TBL_C_PRI GENMASK(6, 4)
+#define PPE_L1_FLOW_MAP_TBL_E_PRI GENMASK(9, 7)
+#define PPE_L1_FLOW_MAP_TBL_C_NODE_WT GENMASK(19, 10)
+#define PPE_L1_FLOW_MAP_TBL_E_NODE_WT GENMASK(29, 20)
+
+#define PPE_L1_C_FLOW_CFG_TBL_ADDR 0x442000
+#define PPE_L1_C_FLOW_CFG_TBL_ENTRIES 64
+#define PPE_L1_C_FLOW_CFG_TBL_INC 0x10
+#define PPE_L1_C_FLOW_CFG_TBL_NODE_ID GENMASK(5, 0)
+#define PPE_L1_C_FLOW_CFG_TBL_NODE_CREDIT_UNIT BIT(6)
+
+#define PPE_L1_E_FLOW_CFG_TBL_ADDR 0x444000
+#define PPE_L1_E_FLOW_CFG_TBL_ENTRIES 64
+#define PPE_L1_E_FLOW_CFG_TBL_INC 0x10
+#define PPE_L1_E_FLOW_CFG_TBL_NODE_ID GENMASK(5, 0)
+#define PPE_L1_E_FLOW_CFG_TBL_NODE_CREDIT_UNIT BIT(6)
+
+#define PPE_L1_FLOW_PORT_MAP_TBL_ADDR 0x446000
+#define PPE_L1_FLOW_PORT_MAP_TBL_ENTRIES 64
+#define PPE_L1_FLOW_PORT_MAP_TBL_INC 0x10
+#define PPE_L1_FLOW_PORT_MAP_TBL_PORT_NUM GENMASK(3, 0)
+
+#define PPE_L1_COMP_CFG_TBL_ADDR 0x46a000
+#define PPE_L1_COMP_CFG_TBL_ENTRIES 64
+#define PPE_L1_COMP_CFG_TBL_INC 0x10
+#define PPE_L1_COMP_CFG_TBL_SHAPER_METER_LEN GENMASK(1, 0)
+#define PPE_L1_COMP_CFG_TBL_NODE_METER_LEN GENMASK(3, 2)
+
+/* PPE port scheduler configurations for egress. */
+#define PPE_PSCH_SCH_CFG_TBL_ADDR 0x47a000
+#define PPE_PSCH_SCH_CFG_TBL_ENTRIES 128
+#define PPE_PSCH_SCH_CFG_TBL_INC 0x10
+#define PPE_PSCH_SCH_CFG_TBL_DES_PORT GENMASK(3, 0)
+#define PPE_PSCH_SCH_CFG_TBL_ENS_PORT GENMASK(7, 4)
+#define PPE_PSCH_SCH_CFG_TBL_ENS_PORT_BITMAP GENMASK(15, 8)
+#define PPE_PSCH_SCH_CFG_TBL_DES_SECOND_PORT_EN BIT(16)
+#define PPE_PSCH_SCH_CFG_TBL_DES_SECOND_PORT GENMASK(20, 17)
+
/* There are 15 BM ports and 4 BM groups supported by PPE.
* BM port (0-7) is for EDMA port 0, BM port (8-13) is for
* PPE physical port 1-6 and BM port 14 is for EIP port.
--
2.34.1
^ permalink raw reply related [flat|nested] 17+ messages in thread
* [PATCH net-next v6 07/14] net: ethernet: qualcomm: Initialize PPE queue settings
2025-07-20 10:57 [PATCH net-next v6 00/14] Add PPE driver for Qualcomm IPQ9574 SoC Luo Jie
` (5 preceding siblings ...)
2025-07-20 10:57 ` [PATCH net-next v6 06/14] net: ethernet: qualcomm: Initialize the PPE scheduler settings Luo Jie
@ 2025-07-20 10:57 ` Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 08/14] net: ethernet: qualcomm: Initialize PPE service code settings Luo Jie
` (6 subsequent siblings)
13 siblings, 0 replies; 17+ messages in thread
From: Luo Jie @ 2025-07-20 10:57 UTC (permalink / raw)
To: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Rob Herring, Krzysztof Kozlowski, Conor Dooley,
Lei Wei, Suruchi Agarwal, Pavithra R, Simon Horman,
Jonathan Corbet, Kees Cook, Gustavo A. R. Silva, Philipp Zabel
Cc: linux-arm-msm, netdev, devicetree, linux-kernel, linux-doc,
linux-hardening, quic_kkumarcs, quic_linchen, Luo Jie
Configure unicast and multicast hardware queues for the PPE ports to
enable packet forwarding between the ports.
Each PPE port is assigned with a range of queues. The queue ID selection
for the packet is decided by the queue base and queue offset that is
configured based on the internal priority and the RSS hash value of the
packet.
Signed-off-by: Luo Jie <quic_luoj@quicinc.com>
---
drivers/net/ethernet/qualcomm/ppe/ppe_config.c | 356 ++++++++++++++++++++++++-
drivers/net/ethernet/qualcomm/ppe/ppe_config.h | 63 +++++
drivers/net/ethernet/qualcomm/ppe/ppe_regs.h | 21 ++
3 files changed, 439 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_config.c b/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
index 3c45d0d1c560..9037702460b5 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
@@ -138,6 +138,34 @@ struct ppe_scheduler_port_config {
unsigned int drr_node_id;
};
+/**
+ * struct ppe_port_schedule_resource - PPE port scheduler resource.
+ * @ucastq_start: Unicast queue start ID.
+ * @ucastq_end: Unicast queue end ID.
+ * @mcastq_start: Multicast queue start ID.
+ * @mcastq_end: Multicast queue end ID.
+ * @flow_id_start: Flow start ID.
+ * @flow_id_end: Flow end ID.
+ * @l0node_start: Scheduler node start ID for queue level.
+ * @l0node_end: Scheduler node end ID for queue level.
+ * @l1node_start: Scheduler node start ID for flow level.
+ * @l1node_end: Scheduler node end ID for flow level.
+ *
+ * PPE scheduler resource allocated among the PPE ports.
+ */
+struct ppe_port_schedule_resource {
+ unsigned int ucastq_start;
+ unsigned int ucastq_end;
+ unsigned int mcastq_start;
+ unsigned int mcastq_end;
+ unsigned int flow_id_start;
+ unsigned int flow_id_end;
+ unsigned int l0node_start;
+ unsigned int l0node_end;
+ unsigned int l1node_start;
+ unsigned int l1node_end;
+};
+
/* There are total 2048 buffers available in PPE, out of which some
* buffers are reserved for some specific purposes per PPE port. The
* rest of the pool of 1550 buffers are assigned to the general 'group0'
@@ -701,6 +729,111 @@ static const struct ppe_scheduler_port_config ppe_port_sch_config[] = {
},
};
+/* The scheduler resource is applied to each PPE port, The resource
+ * includes the unicast & multicast queues, flow nodes and DRR nodes.
+ */
+static const struct ppe_port_schedule_resource ppe_scheduler_res[] = {
+ { .ucastq_start = 0,
+ .ucastq_end = 63,
+ .mcastq_start = 256,
+ .mcastq_end = 271,
+ .flow_id_start = 0,
+ .flow_id_end = 0,
+ .l0node_start = 0,
+ .l0node_end = 7,
+ .l1node_start = 0,
+ .l1node_end = 0,
+ },
+ { .ucastq_start = 144,
+ .ucastq_end = 159,
+ .mcastq_start = 272,
+ .mcastq_end = 275,
+ .flow_id_start = 36,
+ .flow_id_end = 39,
+ .l0node_start = 48,
+ .l0node_end = 63,
+ .l1node_start = 8,
+ .l1node_end = 11,
+ },
+ { .ucastq_start = 160,
+ .ucastq_end = 175,
+ .mcastq_start = 276,
+ .mcastq_end = 279,
+ .flow_id_start = 40,
+ .flow_id_end = 43,
+ .l0node_start = 64,
+ .l0node_end = 79,
+ .l1node_start = 12,
+ .l1node_end = 15,
+ },
+ { .ucastq_start = 176,
+ .ucastq_end = 191,
+ .mcastq_start = 280,
+ .mcastq_end = 283,
+ .flow_id_start = 44,
+ .flow_id_end = 47,
+ .l0node_start = 80,
+ .l0node_end = 95,
+ .l1node_start = 16,
+ .l1node_end = 19,
+ },
+ { .ucastq_start = 192,
+ .ucastq_end = 207,
+ .mcastq_start = 284,
+ .mcastq_end = 287,
+ .flow_id_start = 48,
+ .flow_id_end = 51,
+ .l0node_start = 96,
+ .l0node_end = 111,
+ .l1node_start = 20,
+ .l1node_end = 23,
+ },
+ { .ucastq_start = 208,
+ .ucastq_end = 223,
+ .mcastq_start = 288,
+ .mcastq_end = 291,
+ .flow_id_start = 52,
+ .flow_id_end = 55,
+ .l0node_start = 112,
+ .l0node_end = 127,
+ .l1node_start = 24,
+ .l1node_end = 27,
+ },
+ { .ucastq_start = 224,
+ .ucastq_end = 239,
+ .mcastq_start = 292,
+ .mcastq_end = 295,
+ .flow_id_start = 56,
+ .flow_id_end = 59,
+ .l0node_start = 128,
+ .l0node_end = 143,
+ .l1node_start = 28,
+ .l1node_end = 31,
+ },
+ { .ucastq_start = 240,
+ .ucastq_end = 255,
+ .mcastq_start = 296,
+ .mcastq_end = 299,
+ .flow_id_start = 60,
+ .flow_id_end = 63,
+ .l0node_start = 144,
+ .l0node_end = 159,
+ .l1node_start = 32,
+ .l1node_end = 35,
+ },
+ { .ucastq_start = 64,
+ .ucastq_end = 143,
+ .mcastq_start = 0,
+ .mcastq_end = 0,
+ .flow_id_start = 1,
+ .flow_id_end = 35,
+ .l0node_start = 8,
+ .l0node_end = 47,
+ .l1node_start = 1,
+ .l1node_end = 7,
+ },
+};
+
/* Set the PPE queue level scheduler configuration. */
static int ppe_scheduler_l0_queue_map_set(struct ppe_device *ppe_dev,
int node_id, int port,
@@ -832,6 +965,149 @@ int ppe_queue_scheduler_set(struct ppe_device *ppe_dev,
port, scheduler_cfg);
}
+/**
+ * ppe_queue_ucast_base_set - Set PPE unicast queue base ID and profile ID
+ * @ppe_dev: PPE device
+ * @queue_dst: PPE queue destination configuration
+ * @queue_base: PPE queue base ID
+ * @profile_id: Profile ID
+ *
+ * The PPE unicast queue base ID and profile ID are configured based on the
+ * destination port information that can be service code or CPU code or the
+ * destination port.
+ *
+ * Return: 0 on success, negative error code on failure.
+ */
+int ppe_queue_ucast_base_set(struct ppe_device *ppe_dev,
+ struct ppe_queue_ucast_dest queue_dst,
+ int queue_base, int profile_id)
+{
+ int index, profile_size;
+ u32 val, reg;
+
+ profile_size = queue_dst.src_profile << 8;
+ if (queue_dst.service_code_en)
+ index = PPE_QUEUE_BASE_SERVICE_CODE + profile_size +
+ queue_dst.service_code;
+ else if (queue_dst.cpu_code_en)
+ index = PPE_QUEUE_BASE_CPU_CODE + profile_size +
+ queue_dst.cpu_code;
+ else
+ index = profile_size + queue_dst.dest_port;
+
+ val = FIELD_PREP(PPE_UCAST_QUEUE_MAP_TBL_PROFILE_ID, profile_id);
+ val |= FIELD_PREP(PPE_UCAST_QUEUE_MAP_TBL_QUEUE_ID, queue_base);
+ reg = PPE_UCAST_QUEUE_MAP_TBL_ADDR + index * PPE_UCAST_QUEUE_MAP_TBL_INC;
+
+ return regmap_write(ppe_dev->regmap, reg, val);
+}
+
+/**
+ * ppe_queue_ucast_offset_pri_set - Set PPE unicast queue offset based on priority
+ * @ppe_dev: PPE device
+ * @profile_id: Profile ID
+ * @priority: PPE internal priority to be used to set queue offset
+ * @queue_offset: Queue offset used for calculating the destination queue ID
+ *
+ * The PPE unicast queue offset is configured based on the PPE
+ * internal priority.
+ *
+ * Return: 0 on success, negative error code on failure.
+ */
+int ppe_queue_ucast_offset_pri_set(struct ppe_device *ppe_dev,
+ int profile_id,
+ int priority,
+ int queue_offset)
+{
+ u32 val, reg;
+ int index;
+
+ index = (profile_id << 4) + priority;
+ val = FIELD_PREP(PPE_UCAST_PRIORITY_MAP_TBL_CLASS, queue_offset);
+ reg = PPE_UCAST_PRIORITY_MAP_TBL_ADDR + index * PPE_UCAST_PRIORITY_MAP_TBL_INC;
+
+ return regmap_write(ppe_dev->regmap, reg, val);
+}
+
+/**
+ * ppe_queue_ucast_offset_hash_set - Set PPE unicast queue offset based on hash
+ * @ppe_dev: PPE device
+ * @profile_id: Profile ID
+ * @rss_hash: Packet hash value to be used to set queue offset
+ * @queue_offset: Queue offset used for calculating the destination queue ID
+ *
+ * The PPE unicast queue offset is configured based on the RSS hash value.
+ *
+ * Return: 0 on success, negative error code on failure.
+ */
+int ppe_queue_ucast_offset_hash_set(struct ppe_device *ppe_dev,
+ int profile_id,
+ int rss_hash,
+ int queue_offset)
+{
+ u32 val, reg;
+ int index;
+
+ index = (profile_id << 8) + rss_hash;
+ val = FIELD_PREP(PPE_UCAST_HASH_MAP_TBL_HASH, queue_offset);
+ reg = PPE_UCAST_HASH_MAP_TBL_ADDR + index * PPE_UCAST_HASH_MAP_TBL_INC;
+
+ return regmap_write(ppe_dev->regmap, reg, val);
+}
+
+/**
+ * ppe_port_resource_get - Get PPE resource per port
+ * @ppe_dev: PPE device
+ * @port: PPE port
+ * @type: Resource type
+ * @res_start: Resource start ID returned
+ * @res_end: Resource end ID returned
+ *
+ * PPE resource is assigned per PPE port, which is acquired for QoS scheduler.
+ *
+ * Return: 0 on success, negative error code on failure.
+ */
+int ppe_port_resource_get(struct ppe_device *ppe_dev, int port,
+ enum ppe_resource_type type,
+ int *res_start, int *res_end)
+{
+ struct ppe_port_schedule_resource res;
+
+ /* The reserved resource with the maximum port ID of PPE is
+ * also allowed to be acquired.
+ */
+ if (port > ppe_dev->num_ports)
+ return -EINVAL;
+
+ res = ppe_scheduler_res[port];
+ switch (type) {
+ case PPE_RES_UCAST:
+ *res_start = res.ucastq_start;
+ *res_end = res.ucastq_end;
+ break;
+ case PPE_RES_MCAST:
+ *res_start = res.mcastq_start;
+ *res_end = res.mcastq_end;
+ break;
+ case PPE_RES_FLOW_ID:
+ *res_start = res.flow_id_start;
+ *res_end = res.flow_id_end;
+ break;
+ case PPE_RES_L0_NODE:
+ *res_start = res.l0node_start;
+ *res_end = res.l0node_end;
+ break;
+ case PPE_RES_L1_NODE:
+ *res_start = res.l1node_start;
+ *res_end = res.l1node_end;
+ break;
+ default:
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
static int ppe_config_bm_threshold(struct ppe_device *ppe_dev, int bm_port_id,
const struct ppe_bm_port_config port_cfg)
{
@@ -1167,6 +1443,80 @@ static int ppe_config_scheduler(struct ppe_device *ppe_dev)
return ret;
};
+/* Configure PPE queue destination of each PPE port. */
+static int ppe_queue_dest_init(struct ppe_device *ppe_dev)
+{
+ int ret, port_id, index, q_base, q_offset, res_start, res_end, pri_max;
+ struct ppe_queue_ucast_dest queue_dst;
+
+ for (port_id = 0; port_id < ppe_dev->num_ports; port_id++) {
+ memset(&queue_dst, 0, sizeof(queue_dst));
+
+ ret = ppe_port_resource_get(ppe_dev, port_id, PPE_RES_UCAST,
+ &res_start, &res_end);
+ if (ret)
+ return ret;
+
+ q_base = res_start;
+ queue_dst.dest_port = port_id;
+
+ /* Configure queue base ID and profile ID that is same as
+ * physical port ID.
+ */
+ ret = ppe_queue_ucast_base_set(ppe_dev, queue_dst,
+ q_base, port_id);
+ if (ret)
+ return ret;
+
+ /* Queue priority range supported by each PPE port */
+ ret = ppe_port_resource_get(ppe_dev, port_id, PPE_RES_L0_NODE,
+ &res_start, &res_end);
+ if (ret)
+ return ret;
+
+ pri_max = res_end - res_start;
+
+ /* Redirect ARP reply packet with the max priority on CPU port,
+ * which keeps the ARP reply directed to CPU (CPU code is 101)
+ * with highest priority queue of EDMA.
+ */
+ if (port_id == 0) {
+ memset(&queue_dst, 0, sizeof(queue_dst));
+
+ queue_dst.cpu_code_en = true;
+ queue_dst.cpu_code = 101;
+ ret = ppe_queue_ucast_base_set(ppe_dev, queue_dst,
+ q_base + pri_max,
+ 0);
+ if (ret)
+ return ret;
+ }
+
+ /* Initialize the queue offset of internal priority. */
+ for (index = 0; index < PPE_QUEUE_INTER_PRI_NUM; index++) {
+ q_offset = index > pri_max ? pri_max : index;
+
+ ret = ppe_queue_ucast_offset_pri_set(ppe_dev, port_id,
+ index, q_offset);
+ if (ret)
+ return ret;
+ }
+
+ /* Initialize the queue offset of RSS hash as 0 to avoid the
+ * random hardware value that will lead to the unexpected
+ * destination queue generated.
+ */
+ for (index = 0; index < PPE_QUEUE_HASH_NUM; index++) {
+ ret = ppe_queue_ucast_offset_hash_set(ppe_dev, port_id,
+ index, 0);
+ if (ret)
+ return ret;
+ }
+ }
+
+ return 0;
+}
+
int ppe_hw_config(struct ppe_device *ppe_dev)
{
int ret;
@@ -1179,5 +1529,9 @@ int ppe_hw_config(struct ppe_device *ppe_dev)
if (ret)
return ret;
- return ppe_config_scheduler(ppe_dev);
+ ret = ppe_config_scheduler(ppe_dev);
+ if (ret)
+ return ret;
+
+ return ppe_queue_dest_init(ppe_dev);
}
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_config.h b/drivers/net/ethernet/qualcomm/ppe/ppe_config.h
index e2c703c0b99d..b4fd5f833bac 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe_config.h
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_config.h
@@ -8,6 +8,16 @@
#include "ppe.h"
+/* There are different table index ranges for configuring queue base ID of
+ * the destination port, CPU code and service code.
+ */
+#define PPE_QUEUE_BASE_DEST_PORT 0
+#define PPE_QUEUE_BASE_CPU_CODE 1024
+#define PPE_QUEUE_BASE_SERVICE_CODE 2048
+
+#define PPE_QUEUE_INTER_PRI_NUM 16
+#define PPE_QUEUE_HASH_NUM 256
+
/**
* enum ppe_scheduler_frame_mode - PPE scheduler frame mode.
* @PPE_SCH_WITH_IPG_PREAMBLE_FRAME_CRC: The scheduled frame includes IPG,
@@ -42,8 +52,61 @@ struct ppe_scheduler_cfg {
enum ppe_scheduler_frame_mode frame_mode;
};
+/**
+ * enum ppe_resource_type - PPE resource type.
+ * @PPE_RES_UCAST: Unicast queue resource.
+ * @PPE_RES_MCAST: Multicast queue resource.
+ * @PPE_RES_L0_NODE: Level 0 for queue based node resource.
+ * @PPE_RES_L1_NODE: Level 1 for flow based node resource.
+ * @PPE_RES_FLOW_ID: Flow based node resource.
+ */
+enum ppe_resource_type {
+ PPE_RES_UCAST,
+ PPE_RES_MCAST,
+ PPE_RES_L0_NODE,
+ PPE_RES_L1_NODE,
+ PPE_RES_FLOW_ID,
+};
+
+/**
+ * struct ppe_queue_ucast_dest - PPE unicast queue destination.
+ * @src_profile: Source profile.
+ * @service_code_en: Enable service code to map the queue base ID.
+ * @service_code: Service code.
+ * @cpu_code_en: Enable CPU code to map the queue base ID.
+ * @cpu_code: CPU code.
+ * @dest_port: destination port.
+ *
+ * PPE egress queue ID is decided by the service code if enabled, otherwise
+ * by the CPU code if enabled, or by destination port if both service code
+ * and CPU code are disabled.
+ */
+struct ppe_queue_ucast_dest {
+ int src_profile;
+ bool service_code_en;
+ int service_code;
+ bool cpu_code_en;
+ int cpu_code;
+ int dest_port;
+};
+
int ppe_hw_config(struct ppe_device *ppe_dev);
int ppe_queue_scheduler_set(struct ppe_device *ppe_dev,
int node_id, bool flow_level, int port,
struct ppe_scheduler_cfg scheduler_cfg);
+int ppe_queue_ucast_base_set(struct ppe_device *ppe_dev,
+ struct ppe_queue_ucast_dest queue_dst,
+ int queue_base,
+ int profile_id);
+int ppe_queue_ucast_offset_pri_set(struct ppe_device *ppe_dev,
+ int profile_id,
+ int priority,
+ int queue_offset);
+int ppe_queue_ucast_offset_hash_set(struct ppe_device *ppe_dev,
+ int profile_id,
+ int rss_hash,
+ int queue_offset);
+int ppe_port_resource_get(struct ppe_device *ppe_dev, int port,
+ enum ppe_resource_type type,
+ int *res_start, int *res_end);
#endif
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h b/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
index 2e43fbc56845..3776e619e70f 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
@@ -164,6 +164,27 @@
#define PPE_BM_PORT_FC_SET_PRE_ALLOC(tbl_cfg, value) \
FIELD_MODIFY(PPE_BM_PORT_FC_W1_PRE_ALLOC, (tbl_cfg) + 0x1, value)
+/* The queue base configurations based on destination port,
+ * service code or CPU code.
+ */
+#define PPE_UCAST_QUEUE_MAP_TBL_ADDR 0x810000
+#define PPE_UCAST_QUEUE_MAP_TBL_ENTRIES 3072
+#define PPE_UCAST_QUEUE_MAP_TBL_INC 0x10
+#define PPE_UCAST_QUEUE_MAP_TBL_PROFILE_ID GENMASK(3, 0)
+#define PPE_UCAST_QUEUE_MAP_TBL_QUEUE_ID GENMASK(11, 4)
+
+/* The queue offset configurations based on RSS hash value. */
+#define PPE_UCAST_HASH_MAP_TBL_ADDR 0x830000
+#define PPE_UCAST_HASH_MAP_TBL_ENTRIES 4096
+#define PPE_UCAST_HASH_MAP_TBL_INC 0x10
+#define PPE_UCAST_HASH_MAP_TBL_HASH GENMASK(7, 0)
+
+/* The queue offset configurations based on PPE internal priority. */
+#define PPE_UCAST_PRIORITY_MAP_TBL_ADDR 0x842000
+#define PPE_UCAST_PRIORITY_MAP_TBL_ENTRIES 256
+#define PPE_UCAST_PRIORITY_MAP_TBL_INC 0x10
+#define PPE_UCAST_PRIORITY_MAP_TBL_CLASS GENMASK(3, 0)
+
/* PPE unicast queue (0-255) configurations. */
#define PPE_AC_UNICAST_QUEUE_CFG_TBL_ADDR 0x848000
#define PPE_AC_UNICAST_QUEUE_CFG_TBL_ENTRIES 256
--
2.34.1
^ permalink raw reply related [flat|nested] 17+ messages in thread
* [PATCH net-next v6 08/14] net: ethernet: qualcomm: Initialize PPE service code settings
2025-07-20 10:57 [PATCH net-next v6 00/14] Add PPE driver for Qualcomm IPQ9574 SoC Luo Jie
` (6 preceding siblings ...)
2025-07-20 10:57 ` [PATCH net-next v6 07/14] net: ethernet: qualcomm: Initialize PPE queue settings Luo Jie
@ 2025-07-20 10:57 ` Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 09/14] net: ethernet: qualcomm: Initialize PPE port control settings Luo Jie
` (5 subsequent siblings)
13 siblings, 0 replies; 17+ messages in thread
From: Luo Jie @ 2025-07-20 10:57 UTC (permalink / raw)
To: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Rob Herring, Krzysztof Kozlowski, Conor Dooley,
Lei Wei, Suruchi Agarwal, Pavithra R, Simon Horman,
Jonathan Corbet, Kees Cook, Gustavo A. R. Silva, Philipp Zabel
Cc: linux-arm-msm, netdev, devicetree, linux-kernel, linux-doc,
linux-hardening, quic_kkumarcs, quic_linchen, Luo Jie
PPE service code is a special code (0-255) that is defined by PPE for
PPE's packet processing stages, as per the network functions required
for the packet.
For packet being sent out by ARM cores on Ethernet ports, The service
code 1 is used as the default service code. This service code is used
to bypass most of packet processing stages of the PPE before the packet
transmitted out PPE port, since the software network stack has already
processed the packet.
Signed-off-by: Luo Jie <quic_luoj@quicinc.com>
---
drivers/net/ethernet/qualcomm/ppe/ppe_config.c | 95 +++++++++++++++-
drivers/net/ethernet/qualcomm/ppe/ppe_config.h | 145 +++++++++++++++++++++++++
drivers/net/ethernet/qualcomm/ppe/ppe_regs.h | 53 +++++++++
3 files changed, 292 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_config.c b/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
index 9037702460b5..39a01f25f5ef 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
@@ -8,6 +8,7 @@
*/
#include <linux/bitfield.h>
+#include <linux/bitmap.h>
#include <linux/bits.h>
#include <linux/device.h>
#include <linux/regmap.h>
@@ -1108,6 +1109,75 @@ int ppe_port_resource_get(struct ppe_device *ppe_dev, int port,
return 0;
}
+/**
+ * ppe_sc_config_set - Set PPE service code configuration
+ * @ppe_dev: PPE device
+ * @sc: Service ID, 0-255 supported by PPE
+ * @cfg: Service code configuration
+ *
+ * PPE service code is used by the PPE during its packet processing stages,
+ * to perform or bypass certain selected packet operations on the packet.
+ *
+ * Return: 0 on success, negative error code on failure.
+ */
+int ppe_sc_config_set(struct ppe_device *ppe_dev, int sc, struct ppe_sc_cfg cfg)
+{
+ u32 val, reg, servcode_val[2] = {};
+ unsigned long bitmap_value;
+ int ret;
+
+ val = FIELD_PREP(PPE_IN_L2_SERVICE_TBL_DST_PORT_ID_VALID, cfg.dest_port_valid);
+ val |= FIELD_PREP(PPE_IN_L2_SERVICE_TBL_DST_PORT_ID, cfg.dest_port);
+ val |= FIELD_PREP(PPE_IN_L2_SERVICE_TBL_DST_DIRECTION, cfg.is_src);
+
+ bitmap_value = bitmap_read(cfg.bitmaps.egress, 0, PPE_SC_BYPASS_EGRESS_SIZE);
+ val |= FIELD_PREP(PPE_IN_L2_SERVICE_TBL_DST_BYPASS_BITMAP, bitmap_value);
+ val |= FIELD_PREP(PPE_IN_L2_SERVICE_TBL_RX_CNT_EN,
+ test_bit(PPE_SC_BYPASS_COUNTER_RX, cfg.bitmaps.counter));
+ val |= FIELD_PREP(PPE_IN_L2_SERVICE_TBL_TX_CNT_EN,
+ test_bit(PPE_SC_BYPASS_COUNTER_TX, cfg.bitmaps.counter));
+ reg = PPE_IN_L2_SERVICE_TBL_ADDR + PPE_IN_L2_SERVICE_TBL_INC * sc;
+
+ ret = regmap_write(ppe_dev->regmap, reg, val);
+ if (ret)
+ return ret;
+
+ bitmap_value = bitmap_read(cfg.bitmaps.ingress, 0, PPE_SC_BYPASS_INGRESS_SIZE);
+ PPE_SERVICE_SET_BYPASS_BITMAP(servcode_val, bitmap_value);
+ PPE_SERVICE_SET_RX_CNT_EN(servcode_val,
+ test_bit(PPE_SC_BYPASS_COUNTER_RX_VLAN, cfg.bitmaps.counter));
+ reg = PPE_SERVICE_TBL_ADDR + PPE_SERVICE_TBL_INC * sc;
+
+ ret = regmap_bulk_write(ppe_dev->regmap, reg,
+ servcode_val, ARRAY_SIZE(servcode_val));
+ if (ret)
+ return ret;
+
+ reg = PPE_EG_SERVICE_TBL_ADDR + PPE_EG_SERVICE_TBL_INC * sc;
+ ret = regmap_bulk_read(ppe_dev->regmap, reg,
+ servcode_val, ARRAY_SIZE(servcode_val));
+ if (ret)
+ return ret;
+
+ PPE_EG_SERVICE_SET_NEXT_SERVCODE(servcode_val, cfg.next_service_code);
+ PPE_EG_SERVICE_SET_UPDATE_ACTION(servcode_val, cfg.eip_field_update_bitmap);
+ PPE_EG_SERVICE_SET_HW_SERVICE(servcode_val, cfg.eip_hw_service);
+ PPE_EG_SERVICE_SET_OFFSET_SEL(servcode_val, cfg.eip_offset_sel);
+ PPE_EG_SERVICE_SET_TX_CNT_EN(servcode_val,
+ test_bit(PPE_SC_BYPASS_COUNTER_TX_VLAN, cfg.bitmaps.counter));
+
+ ret = regmap_bulk_write(ppe_dev->regmap, reg,
+ servcode_val, ARRAY_SIZE(servcode_val));
+ if (ret)
+ return ret;
+
+ bitmap_value = bitmap_read(cfg.bitmaps.tunnel, 0, PPE_SC_BYPASS_TUNNEL_SIZE);
+ val = FIELD_PREP(PPE_TL_SERVICE_TBL_BYPASS_BITMAP, bitmap_value);
+ reg = PPE_TL_SERVICE_TBL_ADDR + PPE_TL_SERVICE_TBL_INC * sc;
+
+ return regmap_write(ppe_dev->regmap, reg, val);
+}
+
static int ppe_config_bm_threshold(struct ppe_device *ppe_dev, int bm_port_id,
const struct ppe_bm_port_config port_cfg)
{
@@ -1517,6 +1587,25 @@ static int ppe_queue_dest_init(struct ppe_device *ppe_dev)
return 0;
}
+/* Initialize the service code 1 used by CPU port. */
+static int ppe_servcode_init(struct ppe_device *ppe_dev)
+{
+ struct ppe_sc_cfg sc_cfg = {};
+
+ bitmap_zero(sc_cfg.bitmaps.counter, PPE_SC_BYPASS_COUNTER_SIZE);
+ bitmap_zero(sc_cfg.bitmaps.tunnel, PPE_SC_BYPASS_TUNNEL_SIZE);
+
+ bitmap_fill(sc_cfg.bitmaps.ingress, PPE_SC_BYPASS_INGRESS_SIZE);
+ clear_bit(PPE_SC_BYPASS_INGRESS_FAKE_MAC_HEADER, sc_cfg.bitmaps.ingress);
+ clear_bit(PPE_SC_BYPASS_INGRESS_SERVICE_CODE, sc_cfg.bitmaps.ingress);
+ clear_bit(PPE_SC_BYPASS_INGRESS_FAKE_L2_PROTO, sc_cfg.bitmaps.ingress);
+
+ bitmap_fill(sc_cfg.bitmaps.egress, PPE_SC_BYPASS_EGRESS_SIZE);
+ clear_bit(PPE_SC_BYPASS_EGRESS_ACL_POST_ROUTING_CHECK, sc_cfg.bitmaps.egress);
+
+ return ppe_sc_config_set(ppe_dev, PPE_EDMA_SC_BYPASS_ID, sc_cfg);
+}
+
int ppe_hw_config(struct ppe_device *ppe_dev)
{
int ret;
@@ -1533,5 +1622,9 @@ int ppe_hw_config(struct ppe_device *ppe_dev)
if (ret)
return ret;
- return ppe_queue_dest_init(ppe_dev);
+ ret = ppe_queue_dest_init(ppe_dev);
+ if (ret)
+ return ret;
+
+ return ppe_servcode_init(ppe_dev);
}
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_config.h b/drivers/net/ethernet/qualcomm/ppe/ppe_config.h
index b4fd5f833bac..2b3f7e39cc7e 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe_config.h
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_config.h
@@ -6,6 +6,8 @@
#ifndef __PPE_CONFIG_H__
#define __PPE_CONFIG_H__
+#include <linux/types.h>
+
#include "ppe.h"
/* There are different table index ranges for configuring queue base ID of
@@ -18,6 +20,9 @@
#define PPE_QUEUE_INTER_PRI_NUM 16
#define PPE_QUEUE_HASH_NUM 256
+/* The service code is used by EDMA port to transmit packet to PPE. */
+#define PPE_EDMA_SC_BYPASS_ID 1
+
/**
* enum ppe_scheduler_frame_mode - PPE scheduler frame mode.
* @PPE_SCH_WITH_IPG_PREAMBLE_FRAME_CRC: The scheduled frame includes IPG,
@@ -90,6 +95,144 @@ struct ppe_queue_ucast_dest {
int dest_port;
};
+/* Hardware bitmaps for bypassing features of the ingress packet. */
+enum ppe_sc_ingress_type {
+ PPE_SC_BYPASS_INGRESS_VLAN_TAG_FMT_CHECK = 0,
+ PPE_SC_BYPASS_INGRESS_VLAN_MEMBER_CHECK = 1,
+ PPE_SC_BYPASS_INGRESS_VLAN_TRANSLATE = 2,
+ PPE_SC_BYPASS_INGRESS_MY_MAC_CHECK = 3,
+ PPE_SC_BYPASS_INGRESS_DIP_LOOKUP = 4,
+ PPE_SC_BYPASS_INGRESS_FLOW_LOOKUP = 5,
+ PPE_SC_BYPASS_INGRESS_FLOW_ACTION = 6,
+ PPE_SC_BYPASS_INGRESS_ACL = 7,
+ PPE_SC_BYPASS_INGRESS_FAKE_MAC_HEADER = 8,
+ PPE_SC_BYPASS_INGRESS_SERVICE_CODE = 9,
+ PPE_SC_BYPASS_INGRESS_WRONG_PKT_FMT_L2 = 10,
+ PPE_SC_BYPASS_INGRESS_WRONG_PKT_FMT_L3_IPV4 = 11,
+ PPE_SC_BYPASS_INGRESS_WRONG_PKT_FMT_L3_IPV6 = 12,
+ PPE_SC_BYPASS_INGRESS_WRONG_PKT_FMT_L4 = 13,
+ PPE_SC_BYPASS_INGRESS_FLOW_SERVICE_CODE = 14,
+ PPE_SC_BYPASS_INGRESS_ACL_SERVICE_CODE = 15,
+ PPE_SC_BYPASS_INGRESS_FAKE_L2_PROTO = 16,
+ PPE_SC_BYPASS_INGRESS_PPPOE_TERMINATION = 17,
+ PPE_SC_BYPASS_INGRESS_DEFAULT_VLAN = 18,
+ PPE_SC_BYPASS_INGRESS_DEFAULT_PCP = 19,
+ PPE_SC_BYPASS_INGRESS_VSI_ASSIGN = 20,
+ /* Values 21-23 are not specified by hardware. */
+ PPE_SC_BYPASS_INGRESS_VLAN_ASSIGN_FAIL = 24,
+ PPE_SC_BYPASS_INGRESS_SOURCE_GUARD = 25,
+ PPE_SC_BYPASS_INGRESS_MRU_MTU_CHECK = 26,
+ PPE_SC_BYPASS_INGRESS_FLOW_SRC_CHECK = 27,
+ PPE_SC_BYPASS_INGRESS_FLOW_QOS = 28,
+ /* This must be last as it determines the size of the BITMAP. */
+ PPE_SC_BYPASS_INGRESS_SIZE,
+};
+
+/* Hardware bitmaps for bypassing features of the egress packet. */
+enum ppe_sc_egress_type {
+ PPE_SC_BYPASS_EGRESS_VLAN_MEMBER_CHECK = 0,
+ PPE_SC_BYPASS_EGRESS_VLAN_TRANSLATE = 1,
+ PPE_SC_BYPASS_EGRESS_VLAN_TAG_FMT_CTRL = 2,
+ PPE_SC_BYPASS_EGRESS_FDB_LEARN = 3,
+ PPE_SC_BYPASS_EGRESS_FDB_REFRESH = 4,
+ PPE_SC_BYPASS_EGRESS_L2_SOURCE_SECURITY = 5,
+ PPE_SC_BYPASS_EGRESS_MANAGEMENT_FWD = 6,
+ PPE_SC_BYPASS_EGRESS_BRIDGING_FWD = 7,
+ PPE_SC_BYPASS_EGRESS_IN_STP_FLTR = 8,
+ PPE_SC_BYPASS_EGRESS_EG_STP_FLTR = 9,
+ PPE_SC_BYPASS_EGRESS_SOURCE_FLTR = 10,
+ PPE_SC_BYPASS_EGRESS_POLICER = 11,
+ PPE_SC_BYPASS_EGRESS_L2_PKT_EDIT = 12,
+ PPE_SC_BYPASS_EGRESS_L3_PKT_EDIT = 13,
+ PPE_SC_BYPASS_EGRESS_ACL_POST_ROUTING_CHECK = 14,
+ PPE_SC_BYPASS_EGRESS_PORT_ISOLATION = 15,
+ PPE_SC_BYPASS_EGRESS_PRE_ACL_QOS = 16,
+ PPE_SC_BYPASS_EGRESS_POST_ACL_QOS = 17,
+ PPE_SC_BYPASS_EGRESS_DSCP_QOS = 18,
+ PPE_SC_BYPASS_EGRESS_PCP_QOS = 19,
+ PPE_SC_BYPASS_EGRESS_PREHEADER_QOS = 20,
+ PPE_SC_BYPASS_EGRESS_FAKE_MAC_DROP = 21,
+ PPE_SC_BYPASS_EGRESS_TUNL_CONTEXT = 22,
+ PPE_SC_BYPASS_EGRESS_FLOW_POLICER = 23,
+ /* This must be last as it determines the size of the BITMAP. */
+ PPE_SC_BYPASS_EGRESS_SIZE,
+};
+
+/* Hardware bitmaps for bypassing counter of packet. */
+enum ppe_sc_counter_type {
+ PPE_SC_BYPASS_COUNTER_RX_VLAN = 0,
+ PPE_SC_BYPASS_COUNTER_RX = 1,
+ PPE_SC_BYPASS_COUNTER_TX_VLAN = 2,
+ PPE_SC_BYPASS_COUNTER_TX = 3,
+ /* This must be last as it determines the size of the BITMAP. */
+ PPE_SC_BYPASS_COUNTER_SIZE,
+};
+
+/* Hardware bitmaps for bypassing features of tunnel packet. */
+enum ppe_sc_tunnel_type {
+ PPE_SC_BYPASS_TUNNEL_SERVICE_CODE = 0,
+ PPE_SC_BYPASS_TUNNEL_TUNNEL_HANDLE = 1,
+ PPE_SC_BYPASS_TUNNEL_L3_IF_CHECK = 2,
+ PPE_SC_BYPASS_TUNNEL_VLAN_CHECK = 3,
+ PPE_SC_BYPASS_TUNNEL_DMAC_CHECK = 4,
+ PPE_SC_BYPASS_TUNNEL_UDP_CSUM_0_CHECK = 5,
+ PPE_SC_BYPASS_TUNNEL_TBL_DE_ACCE_CHECK = 6,
+ PPE_SC_BYPASS_TUNNEL_PPPOE_MC_TERM_CHECK = 7,
+ PPE_SC_BYPASS_TUNNEL_TTL_EXCEED_CHECK = 8,
+ PPE_SC_BYPASS_TUNNEL_MAP_SRC_CHECK = 9,
+ PPE_SC_BYPASS_TUNNEL_MAP_DST_CHECK = 10,
+ PPE_SC_BYPASS_TUNNEL_LPM_DST_LOOKUP = 11,
+ PPE_SC_BYPASS_TUNNEL_LPM_LOOKUP = 12,
+ PPE_SC_BYPASS_TUNNEL_WRONG_PKT_FMT_L2 = 13,
+ PPE_SC_BYPASS_TUNNEL_WRONG_PKT_FMT_L3_IPV4 = 14,
+ PPE_SC_BYPASS_TUNNEL_WRONG_PKT_FMT_L3_IPV6 = 15,
+ PPE_SC_BYPASS_TUNNEL_WRONG_PKT_FMT_L4 = 16,
+ PPE_SC_BYPASS_TUNNEL_WRONG_PKT_FMT_TUNNEL = 17,
+ /* Values 18-19 are not specified by hardware. */
+ PPE_SC_BYPASS_TUNNEL_PRE_IPO = 20,
+ /* This must be last as it determines the size of the BITMAP. */
+ PPE_SC_BYPASS_TUNNEL_SIZE,
+};
+
+/**
+ * struct ppe_sc_bypass - PPE service bypass bitmaps
+ * @ingress: Bitmap of features that can be bypassed on the ingress packet.
+ * @egress: Bitmap of features that can be bypassed on the egress packet.
+ * @counter: Bitmap of features that can be bypassed on the counter type.
+ * @tunnel: Bitmap of features that can be bypassed on the tunnel packet.
+ */
+struct ppe_sc_bypass {
+ DECLARE_BITMAP(ingress, PPE_SC_BYPASS_INGRESS_SIZE);
+ DECLARE_BITMAP(egress, PPE_SC_BYPASS_EGRESS_SIZE);
+ DECLARE_BITMAP(counter, PPE_SC_BYPASS_COUNTER_SIZE);
+ DECLARE_BITMAP(tunnel, PPE_SC_BYPASS_TUNNEL_SIZE);
+};
+
+/**
+ * struct ppe_sc_cfg - PPE service code configuration.
+ * @dest_port_valid: Generate destination port or not.
+ * @dest_port: Destination port ID.
+ * @bitmaps: Bitmap of bypass features.
+ * @is_src: Destination port acts as source port, packet sent to CPU.
+ * @next_service_code: New service code generated.
+ * @eip_field_update_bitmap: Fields updated as actions taken for EIP.
+ * @eip_hw_service: Selected hardware functions for EIP.
+ * @eip_offset_sel: Packet offset selection, using packet's layer 4 offset
+ * or using packet's layer 3 offset for EIP.
+ *
+ * Service code is generated during the packet passing through PPE.
+ */
+struct ppe_sc_cfg {
+ bool dest_port_valid;
+ int dest_port;
+ struct ppe_sc_bypass bitmaps;
+ bool is_src;
+ int next_service_code;
+ int eip_field_update_bitmap;
+ int eip_hw_service;
+ int eip_offset_sel;
+};
+
int ppe_hw_config(struct ppe_device *ppe_dev);
int ppe_queue_scheduler_set(struct ppe_device *ppe_dev,
int node_id, bool flow_level, int port,
@@ -109,4 +252,6 @@ int ppe_queue_ucast_offset_hash_set(struct ppe_device *ppe_dev,
int ppe_port_resource_get(struct ppe_device *ppe_dev, int port,
enum ppe_resource_type type,
int *res_start, int *res_end);
+int ppe_sc_config_set(struct ppe_device *ppe_dev, int sc,
+ struct ppe_sc_cfg cfg);
#endif
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h b/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
index 3776e619e70f..4cb76313db87 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
@@ -25,10 +25,63 @@
#define PPE_BM_SCH_CFG_TBL_SECOND_PORT_VALID BIT(6)
#define PPE_BM_SCH_CFG_TBL_SECOND_PORT GENMASK(11, 8)
+/* PPE service code configuration for the ingress direction functions,
+ * including bypass configuration for relevant PPE switch core functions
+ * such as flow entry lookup bypass.
+ */
+#define PPE_SERVICE_TBL_ADDR 0x15000
+#define PPE_SERVICE_TBL_ENTRIES 256
+#define PPE_SERVICE_TBL_INC 0x10
+#define PPE_SERVICE_W0_BYPASS_BITMAP GENMASK(31, 0)
+#define PPE_SERVICE_W1_RX_CNT_EN BIT(0)
+
+#define PPE_SERVICE_SET_BYPASS_BITMAP(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_SERVICE_W0_BYPASS_BITMAP, tbl_cfg, value)
+#define PPE_SERVICE_SET_RX_CNT_EN(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_SERVICE_W1_RX_CNT_EN, (tbl_cfg) + 0x1, value)
+
/* PPE queue counters enable/disable control. */
#define PPE_EG_BRIDGE_CONFIG_ADDR 0x20044
#define PPE_EG_BRIDGE_CONFIG_QUEUE_CNT_EN BIT(2)
+/* PPE service code configuration on the egress direction. */
+#define PPE_EG_SERVICE_TBL_ADDR 0x43000
+#define PPE_EG_SERVICE_TBL_ENTRIES 256
+#define PPE_EG_SERVICE_TBL_INC 0x10
+#define PPE_EG_SERVICE_W0_UPDATE_ACTION GENMASK(31, 0)
+#define PPE_EG_SERVICE_W1_NEXT_SERVCODE GENMASK(7, 0)
+#define PPE_EG_SERVICE_W1_HW_SERVICE GENMASK(13, 8)
+#define PPE_EG_SERVICE_W1_OFFSET_SEL BIT(14)
+#define PPE_EG_SERVICE_W1_TX_CNT_EN BIT(15)
+
+#define PPE_EG_SERVICE_SET_UPDATE_ACTION(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_EG_SERVICE_W0_UPDATE_ACTION, tbl_cfg, value)
+#define PPE_EG_SERVICE_SET_NEXT_SERVCODE(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_EG_SERVICE_W1_NEXT_SERVCODE, (tbl_cfg) + 0x1, value)
+#define PPE_EG_SERVICE_SET_HW_SERVICE(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_EG_SERVICE_W1_HW_SERVICE, (tbl_cfg) + 0x1, value)
+#define PPE_EG_SERVICE_SET_OFFSET_SEL(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_EG_SERVICE_W1_OFFSET_SEL, (tbl_cfg) + 0x1, value)
+#define PPE_EG_SERVICE_SET_TX_CNT_EN(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_EG_SERVICE_W1_TX_CNT_EN, (tbl_cfg) + 0x1, value)
+
+/* PPE service code configuration for destination port and counter. */
+#define PPE_IN_L2_SERVICE_TBL_ADDR 0x66000
+#define PPE_IN_L2_SERVICE_TBL_ENTRIES 256
+#define PPE_IN_L2_SERVICE_TBL_INC 0x10
+#define PPE_IN_L2_SERVICE_TBL_DST_PORT_ID_VALID BIT(0)
+#define PPE_IN_L2_SERVICE_TBL_DST_PORT_ID GENMASK(4, 1)
+#define PPE_IN_L2_SERVICE_TBL_DST_DIRECTION BIT(5)
+#define PPE_IN_L2_SERVICE_TBL_DST_BYPASS_BITMAP GENMASK(29, 6)
+#define PPE_IN_L2_SERVICE_TBL_RX_CNT_EN BIT(30)
+#define PPE_IN_L2_SERVICE_TBL_TX_CNT_EN BIT(31)
+
+/* PPE service code configuration for the tunnel packet. */
+#define PPE_TL_SERVICE_TBL_ADDR 0x306000
+#define PPE_TL_SERVICE_TBL_ENTRIES 256
+#define PPE_TL_SERVICE_TBL_INC 4
+#define PPE_TL_SERVICE_TBL_BYPASS_BITMAP GENMASK(31, 0)
+
/* Port scheduler global config. */
#define PPE_PSCH_SCH_DEPTH_CFG_ADDR 0x400000
#define PPE_PSCH_SCH_DEPTH_CFG_INC 4
--
2.34.1
^ permalink raw reply related [flat|nested] 17+ messages in thread
* [PATCH net-next v6 09/14] net: ethernet: qualcomm: Initialize PPE port control settings
2025-07-20 10:57 [PATCH net-next v6 00/14] Add PPE driver for Qualcomm IPQ9574 SoC Luo Jie
` (7 preceding siblings ...)
2025-07-20 10:57 ` [PATCH net-next v6 08/14] net: ethernet: qualcomm: Initialize PPE service code settings Luo Jie
@ 2025-07-20 10:57 ` Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 10/14] net: ethernet: qualcomm: Initialize PPE RSS hash settings Luo Jie
` (4 subsequent siblings)
13 siblings, 0 replies; 17+ messages in thread
From: Luo Jie @ 2025-07-20 10:57 UTC (permalink / raw)
To: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Rob Herring, Krzysztof Kozlowski, Conor Dooley,
Lei Wei, Suruchi Agarwal, Pavithra R, Simon Horman,
Jonathan Corbet, Kees Cook, Gustavo A. R. Silva, Philipp Zabel
Cc: linux-arm-msm, netdev, devicetree, linux-kernel, linux-doc,
linux-hardening, quic_kkumarcs, quic_linchen, Luo Jie
Configure the default action as drop when the packet size is more than
the configured MTU of physical port. Also enable port specific counters
in PPE.
Signed-off-by: Luo Jie <quic_luoj@quicinc.com>
---
drivers/net/ethernet/qualcomm/ppe/ppe_config.c | 86 +++++++++++++++++++++++++-
drivers/net/ethernet/qualcomm/ppe/ppe_config.h | 15 +++++
drivers/net/ethernet/qualcomm/ppe/ppe_regs.h | 47 ++++++++++++++
3 files changed, 147 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_config.c b/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
index 39a01f25f5ef..8b2b9986b895 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
@@ -1178,6 +1178,44 @@ int ppe_sc_config_set(struct ppe_device *ppe_dev, int sc, struct ppe_sc_cfg cfg)
return regmap_write(ppe_dev->regmap, reg, val);
}
+/**
+ * ppe_counter_enable_set - Set PPE port counter enabled
+ * @ppe_dev: PPE device
+ * @port: PPE port ID
+ *
+ * Enable PPE counters on the given port for the unicast packet, multicast
+ * packet and VLAN packet received and transmitted by PPE.
+ *
+ * Return: 0 on success, negative error code on failure.
+ */
+int ppe_counter_enable_set(struct ppe_device *ppe_dev, int port)
+{
+ u32 reg, mru_mtu_val[3];
+ int ret;
+
+ reg = PPE_MRU_MTU_CTRL_TBL_ADDR + PPE_MRU_MTU_CTRL_TBL_INC * port;
+ ret = regmap_bulk_read(ppe_dev->regmap, reg,
+ mru_mtu_val, ARRAY_SIZE(mru_mtu_val));
+ if (ret)
+ return ret;
+
+ PPE_MRU_MTU_CTRL_SET_RX_CNT_EN(mru_mtu_val, true);
+ PPE_MRU_MTU_CTRL_SET_TX_CNT_EN(mru_mtu_val, true);
+ ret = regmap_bulk_write(ppe_dev->regmap, reg,
+ mru_mtu_val, ARRAY_SIZE(mru_mtu_val));
+ if (ret)
+ return ret;
+
+ reg = PPE_MC_MTU_CTRL_TBL_ADDR + PPE_MC_MTU_CTRL_TBL_INC * port;
+ ret = regmap_set_bits(ppe_dev->regmap, reg, PPE_MC_MTU_CTRL_TBL_TX_CNT_EN);
+ if (ret)
+ return ret;
+
+ reg = PPE_PORT_EG_VLAN_TBL_ADDR + PPE_PORT_EG_VLAN_TBL_INC * port;
+
+ return regmap_set_bits(ppe_dev->regmap, reg, PPE_PORT_EG_VLAN_TBL_TX_COUNTING_EN);
+}
+
static int ppe_config_bm_threshold(struct ppe_device *ppe_dev, int bm_port_id,
const struct ppe_bm_port_config port_cfg)
{
@@ -1606,6 +1644,48 @@ static int ppe_servcode_init(struct ppe_device *ppe_dev)
return ppe_sc_config_set(ppe_dev, PPE_EDMA_SC_BYPASS_ID, sc_cfg);
}
+/* Initialize PPE port configurations. */
+static int ppe_port_config_init(struct ppe_device *ppe_dev)
+{
+ u32 reg, val, mru_mtu_val[3];
+ int i, ret;
+
+ /* MTU and MRU settings are not required for CPU port 0. */
+ for (i = 1; i < ppe_dev->num_ports; i++) {
+ /* Enable Ethernet port counter */
+ ret = ppe_counter_enable_set(ppe_dev, i);
+ if (ret)
+ return ret;
+
+ reg = PPE_MRU_MTU_CTRL_TBL_ADDR + PPE_MRU_MTU_CTRL_TBL_INC * i;
+ ret = regmap_bulk_read(ppe_dev->regmap, reg,
+ mru_mtu_val, ARRAY_SIZE(mru_mtu_val));
+ if (ret)
+ return ret;
+
+ /* Drop the packet when the packet size is more than
+ * the MTU or MRU of the physical interface.
+ */
+ PPE_MRU_MTU_CTRL_SET_MRU_CMD(mru_mtu_val, PPE_ACTION_DROP);
+ PPE_MRU_MTU_CTRL_SET_MTU_CMD(mru_mtu_val, PPE_ACTION_DROP);
+ ret = regmap_bulk_write(ppe_dev->regmap, reg,
+ mru_mtu_val, ARRAY_SIZE(mru_mtu_val));
+ if (ret)
+ return ret;
+
+ reg = PPE_MC_MTU_CTRL_TBL_ADDR + PPE_MC_MTU_CTRL_TBL_INC * i;
+ val = FIELD_PREP(PPE_MC_MTU_CTRL_TBL_MTU_CMD, PPE_ACTION_DROP);
+ ret = regmap_update_bits(ppe_dev->regmap, reg,
+ PPE_MC_MTU_CTRL_TBL_MTU_CMD,
+ val);
+ if (ret)
+ return ret;
+ }
+
+ /* Enable CPU port counters. */
+ return ppe_counter_enable_set(ppe_dev, 0);
+}
+
int ppe_hw_config(struct ppe_device *ppe_dev)
{
int ret;
@@ -1626,5 +1706,9 @@ int ppe_hw_config(struct ppe_device *ppe_dev)
if (ret)
return ret;
- return ppe_servcode_init(ppe_dev);
+ ret = ppe_servcode_init(ppe_dev);
+ if (ret)
+ return ret;
+
+ return ppe_port_config_init(ppe_dev);
}
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_config.h b/drivers/net/ethernet/qualcomm/ppe/ppe_config.h
index 2b3f7e39cc7e..84fa447742e3 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe_config.h
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_config.h
@@ -233,6 +233,20 @@ struct ppe_sc_cfg {
int eip_offset_sel;
};
+/**
+ * enum ppe_action_type - PPE action of the received packet.
+ * @PPE_ACTION_FORWARD: Packet forwarded per L2/L3 process.
+ * @PPE_ACTION_DROP: Packet dropped by PPE.
+ * @PPE_ACTION_COPY_TO_CPU: Packet copied to CPU port per multicast queue.
+ * @PPE_ACTION_REDIRECT_TO_CPU: Packet redirected to CPU port per unicast queue.
+ */
+enum ppe_action_type {
+ PPE_ACTION_FORWARD = 0,
+ PPE_ACTION_DROP = 1,
+ PPE_ACTION_COPY_TO_CPU = 2,
+ PPE_ACTION_REDIRECT_TO_CPU = 3,
+};
+
int ppe_hw_config(struct ppe_device *ppe_dev);
int ppe_queue_scheduler_set(struct ppe_device *ppe_dev,
int node_id, bool flow_level, int port,
@@ -254,4 +268,5 @@ int ppe_port_resource_get(struct ppe_device *ppe_dev, int port,
int *res_start, int *res_end);
int ppe_sc_config_set(struct ppe_device *ppe_dev, int sc,
struct ppe_sc_cfg cfg);
+int ppe_counter_enable_set(struct ppe_device *ppe_dev, int port);
#endif
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h b/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
index 4cb76313db87..c26bee83252f 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
@@ -40,6 +40,18 @@
#define PPE_SERVICE_SET_RX_CNT_EN(tbl_cfg, value) \
FIELD_MODIFY(PPE_SERVICE_W1_RX_CNT_EN, (tbl_cfg) + 0x1, value)
+/* PPE port egress VLAN configurations. */
+#define PPE_PORT_EG_VLAN_TBL_ADDR 0x20020
+#define PPE_PORT_EG_VLAN_TBL_ENTRIES 8
+#define PPE_PORT_EG_VLAN_TBL_INC 4
+#define PPE_PORT_EG_VLAN_TBL_VLAN_TYPE BIT(0)
+#define PPE_PORT_EG_VLAN_TBL_CTAG_MODE GENMASK(2, 1)
+#define PPE_PORT_EG_VLAN_TBL_STAG_MODE GENMASK(4, 3)
+#define PPE_PORT_EG_VLAN_TBL_VSI_TAG_MODE_EN BIT(5)
+#define PPE_PORT_EG_VLAN_TBL_PCP_PROP_CMD BIT(6)
+#define PPE_PORT_EG_VLAN_TBL_DEI_PROP_CMD BIT(7)
+#define PPE_PORT_EG_VLAN_TBL_TX_COUNTING_EN BIT(8)
+
/* PPE queue counters enable/disable control. */
#define PPE_EG_BRIDGE_CONFIG_ADDR 0x20044
#define PPE_EG_BRIDGE_CONFIG_QUEUE_CNT_EN BIT(2)
@@ -65,6 +77,41 @@
#define PPE_EG_SERVICE_SET_TX_CNT_EN(tbl_cfg, value) \
FIELD_MODIFY(PPE_EG_SERVICE_W1_TX_CNT_EN, (tbl_cfg) + 0x1, value)
+/* PPE port control configurations for the traffic to the multicast queues. */
+#define PPE_MC_MTU_CTRL_TBL_ADDR 0x60a00
+#define PPE_MC_MTU_CTRL_TBL_ENTRIES 8
+#define PPE_MC_MTU_CTRL_TBL_INC 4
+#define PPE_MC_MTU_CTRL_TBL_MTU GENMASK(13, 0)
+#define PPE_MC_MTU_CTRL_TBL_MTU_CMD GENMASK(15, 14)
+#define PPE_MC_MTU_CTRL_TBL_TX_CNT_EN BIT(16)
+
+/* PPE port control configurations for the traffic to the unicast queues. */
+#define PPE_MRU_MTU_CTRL_TBL_ADDR 0x65000
+#define PPE_MRU_MTU_CTRL_TBL_ENTRIES 256
+#define PPE_MRU_MTU_CTRL_TBL_INC 0x10
+#define PPE_MRU_MTU_CTRL_W0_MRU GENMASK(13, 0)
+#define PPE_MRU_MTU_CTRL_W0_MRU_CMD GENMASK(15, 14)
+#define PPE_MRU_MTU_CTRL_W0_MTU GENMASK(29, 16)
+#define PPE_MRU_MTU_CTRL_W0_MTU_CMD GENMASK(31, 30)
+#define PPE_MRU_MTU_CTRL_W1_RX_CNT_EN BIT(0)
+#define PPE_MRU_MTU_CTRL_W1_TX_CNT_EN BIT(1)
+#define PPE_MRU_MTU_CTRL_W1_SRC_PROFILE GENMASK(3, 2)
+#define PPE_MRU_MTU_CTRL_W1_INNER_PREC_LOW BIT(31)
+#define PPE_MRU_MTU_CTRL_W2_INNER_PREC_HIGH GENMASK(1, 0)
+
+#define PPE_MRU_MTU_CTRL_SET_MRU(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_MRU_MTU_CTRL_W0_MRU, tbl_cfg, value)
+#define PPE_MRU_MTU_CTRL_SET_MRU_CMD(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_MRU_MTU_CTRL_W0_MRU_CMD, tbl_cfg, value)
+#define PPE_MRU_MTU_CTRL_SET_MTU(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_MRU_MTU_CTRL_W0_MTU, tbl_cfg, value)
+#define PPE_MRU_MTU_CTRL_SET_MTU_CMD(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_MRU_MTU_CTRL_W0_MTU_CMD, tbl_cfg, value)
+#define PPE_MRU_MTU_CTRL_SET_RX_CNT_EN(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_MRU_MTU_CTRL_W1_RX_CNT_EN, (tbl_cfg) + 0x1, value)
+#define PPE_MRU_MTU_CTRL_SET_TX_CNT_EN(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_MRU_MTU_CTRL_W1_TX_CNT_EN, (tbl_cfg) + 0x1, value)
+
/* PPE service code configuration for destination port and counter. */
#define PPE_IN_L2_SERVICE_TBL_ADDR 0x66000
#define PPE_IN_L2_SERVICE_TBL_ENTRIES 256
--
2.34.1
^ permalink raw reply related [flat|nested] 17+ messages in thread
* [PATCH net-next v6 10/14] net: ethernet: qualcomm: Initialize PPE RSS hash settings
2025-07-20 10:57 [PATCH net-next v6 00/14] Add PPE driver for Qualcomm IPQ9574 SoC Luo Jie
` (8 preceding siblings ...)
2025-07-20 10:57 ` [PATCH net-next v6 09/14] net: ethernet: qualcomm: Initialize PPE port control settings Luo Jie
@ 2025-07-20 10:57 ` Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 11/14] net: ethernet: qualcomm: Initialize PPE queue to Ethernet DMA ring mapping Luo Jie
` (3 subsequent siblings)
13 siblings, 0 replies; 17+ messages in thread
From: Luo Jie @ 2025-07-20 10:57 UTC (permalink / raw)
To: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Rob Herring, Krzysztof Kozlowski, Conor Dooley,
Lei Wei, Suruchi Agarwal, Pavithra R, Simon Horman,
Jonathan Corbet, Kees Cook, Gustavo A. R. Silva, Philipp Zabel
Cc: linux-arm-msm, netdev, devicetree, linux-kernel, linux-doc,
linux-hardening, quic_kkumarcs, quic_linchen, Luo Jie
The PPE RSS hash is generated during PPE receive, based on the packet
content (3 tuples or 5 tuples) and as per the configured RSS seed. The
hash is then used to select the queue to transmit the packet to the
ARM CPU.
This patch initializes the RSS hash settings that are used to generate
the hash for the packet during PPE packet receive.
Signed-off-by: Luo Jie <quic_luoj@quicinc.com>
---
drivers/net/ethernet/qualcomm/ppe/ppe_config.c | 198 ++++++++++++++++++++++++-
drivers/net/ethernet/qualcomm/ppe/ppe_config.h | 39 +++++
drivers/net/ethernet/qualcomm/ppe/ppe_regs.h | 40 +++++
3 files changed, 276 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_config.c b/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
index 8b2b9986b895..778cb07f4ba4 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
@@ -1216,6 +1216,147 @@ int ppe_counter_enable_set(struct ppe_device *ppe_dev, int port)
return regmap_set_bits(ppe_dev->regmap, reg, PPE_PORT_EG_VLAN_TBL_TX_COUNTING_EN);
}
+static int ppe_rss_hash_ipv4_config(struct ppe_device *ppe_dev, int index,
+ struct ppe_rss_hash_cfg cfg)
+{
+ u32 reg, val;
+
+ switch (index) {
+ case 0:
+ val = cfg.hash_sip_mix[0];
+ break;
+ case 1:
+ val = cfg.hash_dip_mix[0];
+ break;
+ case 2:
+ val = cfg.hash_protocol_mix;
+ break;
+ case 3:
+ val = cfg.hash_dport_mix;
+ break;
+ case 4:
+ val = cfg.hash_sport_mix;
+ break;
+ default:
+ return -EINVAL;
+ }
+
+ reg = PPE_RSS_HASH_MIX_IPV4_ADDR + index * PPE_RSS_HASH_MIX_IPV4_INC;
+
+ return regmap_update_bits(ppe_dev->regmap, reg,
+ PPE_RSS_HASH_MIX_IPV4_VAL,
+ FIELD_PREP(PPE_RSS_HASH_MIX_IPV4_VAL, val));
+}
+
+static int ppe_rss_hash_ipv6_config(struct ppe_device *ppe_dev, int index,
+ struct ppe_rss_hash_cfg cfg)
+{
+ u32 reg, val;
+
+ switch (index) {
+ case 0 ... 3:
+ val = cfg.hash_sip_mix[index];
+ break;
+ case 4 ... 7:
+ val = cfg.hash_dip_mix[index - 4];
+ break;
+ case 8:
+ val = cfg.hash_protocol_mix;
+ break;
+ case 9:
+ val = cfg.hash_dport_mix;
+ break;
+ case 10:
+ val = cfg.hash_sport_mix;
+ break;
+ default:
+ return -EINVAL;
+ }
+
+ reg = PPE_RSS_HASH_MIX_ADDR + index * PPE_RSS_HASH_MIX_INC;
+
+ return regmap_update_bits(ppe_dev->regmap, reg,
+ PPE_RSS_HASH_MIX_VAL,
+ FIELD_PREP(PPE_RSS_HASH_MIX_VAL, val));
+}
+
+/**
+ * ppe_rss_hash_config_set - Configure the PPE hash settings for the packet received.
+ * @ppe_dev: PPE device.
+ * @mode: Configure RSS hash for the packet type IPv4 and IPv6.
+ * @cfg: RSS hash configuration.
+ *
+ * PPE RSS hash settings are configured for the packet type IPv4 and IPv6.
+ *
+ * Return: 0 on success, negative error code on failure.
+ */
+int ppe_rss_hash_config_set(struct ppe_device *ppe_dev, int mode,
+ struct ppe_rss_hash_cfg cfg)
+{
+ u32 val, reg;
+ int i, ret;
+
+ if (mode & PPE_RSS_HASH_MODE_IPV4) {
+ val = FIELD_PREP(PPE_RSS_HASH_MASK_IPV4_HASH_MASK, cfg.hash_mask);
+ val |= FIELD_PREP(PPE_RSS_HASH_MASK_IPV4_FRAGMENT, cfg.hash_fragment_mode);
+ ret = regmap_write(ppe_dev->regmap, PPE_RSS_HASH_MASK_IPV4_ADDR, val);
+ if (ret)
+ return ret;
+
+ val = FIELD_PREP(PPE_RSS_HASH_SEED_IPV4_VAL, cfg.hash_seed);
+ ret = regmap_write(ppe_dev->regmap, PPE_RSS_HASH_SEED_IPV4_ADDR, val);
+ if (ret)
+ return ret;
+
+ for (i = 0; i < PPE_RSS_HASH_MIX_IPV4_ENTRIES; i++) {
+ ret = ppe_rss_hash_ipv4_config(ppe_dev, i, cfg);
+ if (ret)
+ return ret;
+ }
+
+ for (i = 0; i < PPE_RSS_HASH_FIN_IPV4_ENTRIES; i++) {
+ val = FIELD_PREP(PPE_RSS_HASH_FIN_IPV4_INNER, cfg.hash_fin_inner[i]);
+ val |= FIELD_PREP(PPE_RSS_HASH_FIN_IPV4_OUTER, cfg.hash_fin_outer[i]);
+ reg = PPE_RSS_HASH_FIN_IPV4_ADDR + i * PPE_RSS_HASH_FIN_IPV4_INC;
+
+ ret = regmap_write(ppe_dev->regmap, reg, val);
+ if (ret)
+ return ret;
+ }
+ }
+
+ if (mode & PPE_RSS_HASH_MODE_IPV6) {
+ val = FIELD_PREP(PPE_RSS_HASH_MASK_HASH_MASK, cfg.hash_mask);
+ val |= FIELD_PREP(PPE_RSS_HASH_MASK_FRAGMENT, cfg.hash_fragment_mode);
+ ret = regmap_write(ppe_dev->regmap, PPE_RSS_HASH_MASK_ADDR, val);
+ if (ret)
+ return ret;
+
+ val = FIELD_PREP(PPE_RSS_HASH_SEED_VAL, cfg.hash_seed);
+ ret = regmap_write(ppe_dev->regmap, PPE_RSS_HASH_SEED_ADDR, val);
+ if (ret)
+ return ret;
+
+ for (i = 0; i < PPE_RSS_HASH_MIX_ENTRIES; i++) {
+ ret = ppe_rss_hash_ipv6_config(ppe_dev, i, cfg);
+ if (ret)
+ return ret;
+ }
+
+ for (i = 0; i < PPE_RSS_HASH_FIN_ENTRIES; i++) {
+ val = FIELD_PREP(PPE_RSS_HASH_FIN_INNER, cfg.hash_fin_inner[i]);
+ val |= FIELD_PREP(PPE_RSS_HASH_FIN_OUTER, cfg.hash_fin_outer[i]);
+ reg = PPE_RSS_HASH_FIN_ADDR + i * PPE_RSS_HASH_FIN_INC;
+
+ ret = regmap_write(ppe_dev->regmap, reg, val);
+ if (ret)
+ return ret;
+ }
+ }
+
+ return 0;
+}
+
static int ppe_config_bm_threshold(struct ppe_device *ppe_dev, int bm_port_id,
const struct ppe_bm_port_config port_cfg)
{
@@ -1686,6 +1827,57 @@ static int ppe_port_config_init(struct ppe_device *ppe_dev)
return ppe_counter_enable_set(ppe_dev, 0);
}
+/* Initialize the PPE RSS configuration for IPv4 and IPv6 packet receive.
+ * RSS settings are to calculate the random RSS hash value generated during
+ * packet receive. This hash is then used to generate the queue offset used
+ * to determine the queue used to transmit the packet.
+ */
+static int ppe_rss_hash_init(struct ppe_device *ppe_dev)
+{
+ u16 fins[PPE_RSS_HASH_TUPLES] = { 0x205, 0x264, 0x227, 0x245, 0x201 };
+ u8 ips[PPE_RSS_HASH_IP_LENGTH] = { 0x13, 0xb, 0x13, 0xb };
+ struct ppe_rss_hash_cfg hash_cfg;
+ int i, ret;
+
+ hash_cfg.hash_seed = get_random_u32();
+ hash_cfg.hash_mask = 0xfff;
+
+ /* Use 5 tuple as RSS hash key for the first fragment of TCP, UDP
+ * and UDP-Lite packets.
+ */
+ hash_cfg.hash_fragment_mode = false;
+
+ /* The final common seed configs used to calculate the RSS has value,
+ * which is available for both IPv4 and IPv6 packet.
+ */
+ for (i = 0; i < ARRAY_SIZE(fins); i++) {
+ hash_cfg.hash_fin_inner[i] = fins[i] & 0x1f;
+ hash_cfg.hash_fin_outer[i] = fins[i] >> 5;
+ }
+
+ /* RSS seeds for IP protocol, L4 destination & source port and
+ * destination & source IP used to calculate the RSS hash value.
+ */
+ hash_cfg.hash_protocol_mix = 0x13;
+ hash_cfg.hash_dport_mix = 0xb;
+ hash_cfg.hash_sport_mix = 0x13;
+ hash_cfg.hash_dip_mix[0] = 0xb;
+ hash_cfg.hash_sip_mix[0] = 0x13;
+
+ /* Configure RSS seed configs for IPv4 packet. */
+ ret = ppe_rss_hash_config_set(ppe_dev, PPE_RSS_HASH_MODE_IPV4, hash_cfg);
+ if (ret)
+ return ret;
+
+ for (i = 0; i < ARRAY_SIZE(ips); i++) {
+ hash_cfg.hash_sip_mix[i] = ips[i];
+ hash_cfg.hash_dip_mix[i] = ips[i];
+ }
+
+ /* Configure RSS seed configs for IPv6 packet. */
+ return ppe_rss_hash_config_set(ppe_dev, PPE_RSS_HASH_MODE_IPV6, hash_cfg);
+}
+
int ppe_hw_config(struct ppe_device *ppe_dev)
{
int ret;
@@ -1710,5 +1902,9 @@ int ppe_hw_config(struct ppe_device *ppe_dev)
if (ret)
return ret;
- return ppe_port_config_init(ppe_dev);
+ ret = ppe_port_config_init(ppe_dev);
+ if (ret)
+ return ret;
+
+ return ppe_rss_hash_init(ppe_dev);
}
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_config.h b/drivers/net/ethernet/qualcomm/ppe/ppe_config.h
index 84fa447742e3..eb4a82375bb2 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe_config.h
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_config.h
@@ -23,6 +23,12 @@
/* The service code is used by EDMA port to transmit packet to PPE. */
#define PPE_EDMA_SC_BYPASS_ID 1
+/* The PPE RSS hash configured for IPv4 and IPv6 packet separately. */
+#define PPE_RSS_HASH_MODE_IPV4 BIT(0)
+#define PPE_RSS_HASH_MODE_IPV6 BIT(1)
+#define PPE_RSS_HASH_IP_LENGTH 4
+#define PPE_RSS_HASH_TUPLES 5
+
/**
* enum ppe_scheduler_frame_mode - PPE scheduler frame mode.
* @PPE_SCH_WITH_IPG_PREAMBLE_FRAME_CRC: The scheduled frame includes IPG,
@@ -247,6 +253,37 @@ enum ppe_action_type {
PPE_ACTION_REDIRECT_TO_CPU = 3,
};
+/**
+ * struct ppe_rss_hash_cfg - PPE RSS hash configuration.
+ * @hash_mask: Mask of the generated hash value.
+ * @hash_fragment_mode: Hash generation mode for the first fragment of TCP,
+ * UDP and UDP-Lite packets, to use either 3 tuple or 5 tuple for RSS hash
+ * key computation.
+ * @hash_seed: Seed to generate RSS hash.
+ * @hash_sip_mix: Source IP selection.
+ * @hash_dip_mix: Destination IP selection.
+ * @hash_protocol_mix: Protocol selection.
+ * @hash_sport_mix: Source L4 port selection.
+ * @hash_dport_mix: Destination L4 port selection.
+ * @hash_fin_inner: RSS hash value first selection.
+ * @hash_fin_outer: RSS hash value second selection.
+ *
+ * PPE RSS hash value is generated for the packet based on the RSS hash
+ * configured.
+ */
+struct ppe_rss_hash_cfg {
+ u32 hash_mask;
+ bool hash_fragment_mode;
+ u32 hash_seed;
+ u8 hash_sip_mix[PPE_RSS_HASH_IP_LENGTH];
+ u8 hash_dip_mix[PPE_RSS_HASH_IP_LENGTH];
+ u8 hash_protocol_mix;
+ u8 hash_sport_mix;
+ u8 hash_dport_mix;
+ u8 hash_fin_inner[PPE_RSS_HASH_TUPLES];
+ u8 hash_fin_outer[PPE_RSS_HASH_TUPLES];
+};
+
int ppe_hw_config(struct ppe_device *ppe_dev);
int ppe_queue_scheduler_set(struct ppe_device *ppe_dev,
int node_id, bool flow_level, int port,
@@ -269,4 +306,6 @@ int ppe_port_resource_get(struct ppe_device *ppe_dev, int port,
int ppe_sc_config_set(struct ppe_device *ppe_dev, int sc,
struct ppe_sc_cfg cfg);
int ppe_counter_enable_set(struct ppe_device *ppe_dev, int port);
+int ppe_rss_hash_config_set(struct ppe_device *ppe_dev, int mode,
+ struct ppe_rss_hash_cfg hash_cfg);
#endif
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h b/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
index c26bee83252f..3e9cccedc6be 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
@@ -16,6 +16,46 @@
#define PPE_BM_SCH_CTRL_SCH_OFFSET GENMASK(14, 8)
#define PPE_BM_SCH_CTRL_SCH_EN BIT(31)
+/* RSS settings are to calculate the random RSS hash value generated during
+ * packet receive to ARM cores. This hash is then used to generate the queue
+ * offset used to determine the queue used to transmit the packet to ARM cores.
+ */
+#define PPE_RSS_HASH_MASK_ADDR 0xb4318
+#define PPE_RSS_HASH_MASK_HASH_MASK GENMASK(20, 0)
+#define PPE_RSS_HASH_MASK_FRAGMENT BIT(28)
+
+#define PPE_RSS_HASH_SEED_ADDR 0xb431c
+#define PPE_RSS_HASH_SEED_VAL GENMASK(31, 0)
+
+#define PPE_RSS_HASH_MIX_ADDR 0xb4320
+#define PPE_RSS_HASH_MIX_ENTRIES 11
+#define PPE_RSS_HASH_MIX_INC 4
+#define PPE_RSS_HASH_MIX_VAL GENMASK(4, 0)
+
+#define PPE_RSS_HASH_FIN_ADDR 0xb4350
+#define PPE_RSS_HASH_FIN_ENTRIES 5
+#define PPE_RSS_HASH_FIN_INC 4
+#define PPE_RSS_HASH_FIN_INNER GENMASK(4, 0)
+#define PPE_RSS_HASH_FIN_OUTER GENMASK(9, 5)
+
+#define PPE_RSS_HASH_MASK_IPV4_ADDR 0xb4380
+#define PPE_RSS_HASH_MASK_IPV4_HASH_MASK GENMASK(20, 0)
+#define PPE_RSS_HASH_MASK_IPV4_FRAGMENT BIT(28)
+
+#define PPE_RSS_HASH_SEED_IPV4_ADDR 0xb4384
+#define PPE_RSS_HASH_SEED_IPV4_VAL GENMASK(31, 0)
+
+#define PPE_RSS_HASH_MIX_IPV4_ADDR 0xb4390
+#define PPE_RSS_HASH_MIX_IPV4_ENTRIES 5
+#define PPE_RSS_HASH_MIX_IPV4_INC 4
+#define PPE_RSS_HASH_MIX_IPV4_VAL GENMASK(4, 0)
+
+#define PPE_RSS_HASH_FIN_IPV4_ADDR 0xb43b0
+#define PPE_RSS_HASH_FIN_IPV4_ENTRIES 5
+#define PPE_RSS_HASH_FIN_IPV4_INC 4
+#define PPE_RSS_HASH_FIN_IPV4_INNER GENMASK(4, 0)
+#define PPE_RSS_HASH_FIN_IPV4_OUTER GENMASK(9, 5)
+
#define PPE_BM_SCH_CFG_TBL_ADDR 0xc000
#define PPE_BM_SCH_CFG_TBL_ENTRIES 128
#define PPE_BM_SCH_CFG_TBL_INC 0x10
--
2.34.1
^ permalink raw reply related [flat|nested] 17+ messages in thread
* [PATCH net-next v6 11/14] net: ethernet: qualcomm: Initialize PPE queue to Ethernet DMA ring mapping
2025-07-20 10:57 [PATCH net-next v6 00/14] Add PPE driver for Qualcomm IPQ9574 SoC Luo Jie
` (9 preceding siblings ...)
2025-07-20 10:57 ` [PATCH net-next v6 10/14] net: ethernet: qualcomm: Initialize PPE RSS hash settings Luo Jie
@ 2025-07-20 10:57 ` Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 12/14] net: ethernet: qualcomm: Initialize PPE L2 bridge settings Luo Jie
` (2 subsequent siblings)
13 siblings, 0 replies; 17+ messages in thread
From: Luo Jie @ 2025-07-20 10:57 UTC (permalink / raw)
To: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Rob Herring, Krzysztof Kozlowski, Conor Dooley,
Lei Wei, Suruchi Agarwal, Pavithra R, Simon Horman,
Jonathan Corbet, Kees Cook, Gustavo A. R. Silva, Philipp Zabel
Cc: linux-arm-msm, netdev, devicetree, linux-kernel, linux-doc,
linux-hardening, quic_kkumarcs, quic_linchen, Luo Jie
Configure the selected queues to map with an Ethernet DMA ring for the
packet to receive on ARM cores.
As default initialization, all queues assigned to CPU port 0 are mapped
to the EDMA ring 0. This configuration is later updated during Ethernet
DMA initialization.
Signed-off-by: Luo Jie <quic_luoj@quicinc.com>
---
drivers/net/ethernet/qualcomm/ppe/ppe_config.c | 47 +++++++++++++++++++++++++-
drivers/net/ethernet/qualcomm/ppe/ppe_config.h | 6 ++++
drivers/net/ethernet/qualcomm/ppe/ppe_regs.h | 5 +++
3 files changed, 57 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_config.c b/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
index 778cb07f4ba4..b9e99fd28831 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
@@ -1357,6 +1357,28 @@ int ppe_rss_hash_config_set(struct ppe_device *ppe_dev, int mode,
return 0;
}
+/**
+ * ppe_ring_queue_map_set - Set the PPE queue to Ethernet DMA ring mapping
+ * @ppe_dev: PPE device
+ * @ring_id: Ethernet DMA ring ID
+ * @queue_map: Bit map of queue IDs to given Ethernet DMA ring
+ *
+ * Configure the mapping from a set of PPE queues to a given Ethernet DMA ring.
+ *
+ * Return: 0 on success, negative error code on failure.
+ */
+int ppe_ring_queue_map_set(struct ppe_device *ppe_dev, int ring_id, u32 *queue_map)
+{
+ u32 reg, queue_bitmap_val[PPE_RING_TO_QUEUE_BITMAP_WORD_CNT];
+
+ memcpy(queue_bitmap_val, queue_map, sizeof(queue_bitmap_val));
+ reg = PPE_RING_Q_MAP_TBL_ADDR + PPE_RING_Q_MAP_TBL_INC * ring_id;
+
+ return regmap_bulk_write(ppe_dev->regmap, reg,
+ queue_bitmap_val,
+ ARRAY_SIZE(queue_bitmap_val));
+}
+
static int ppe_config_bm_threshold(struct ppe_device *ppe_dev, int bm_port_id,
const struct ppe_bm_port_config port_cfg)
{
@@ -1878,6 +1900,25 @@ static int ppe_rss_hash_init(struct ppe_device *ppe_dev)
return ppe_rss_hash_config_set(ppe_dev, PPE_RSS_HASH_MODE_IPV6, hash_cfg);
}
+/* Initialize mapping between PPE queues assigned to CPU port 0
+ * to Ethernet DMA ring 0.
+ */
+static int ppe_queues_to_ring_init(struct ppe_device *ppe_dev)
+{
+ u32 queue_bmap[PPE_RING_TO_QUEUE_BITMAP_WORD_CNT] = {};
+ int ret, queue_id, queue_max;
+
+ ret = ppe_port_resource_get(ppe_dev, 0, PPE_RES_UCAST,
+ &queue_id, &queue_max);
+ if (ret)
+ return ret;
+
+ for (; queue_id <= queue_max; queue_id++)
+ queue_bmap[queue_id / 32] |= BIT_MASK(queue_id % 32);
+
+ return ppe_ring_queue_map_set(ppe_dev, 0, queue_bmap);
+}
+
int ppe_hw_config(struct ppe_device *ppe_dev)
{
int ret;
@@ -1906,5 +1947,9 @@ int ppe_hw_config(struct ppe_device *ppe_dev)
if (ret)
return ret;
- return ppe_rss_hash_init(ppe_dev);
+ ret = ppe_rss_hash_init(ppe_dev);
+ if (ret)
+ return ret;
+
+ return ppe_queues_to_ring_init(ppe_dev);
}
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_config.h b/drivers/net/ethernet/qualcomm/ppe/ppe_config.h
index eb4a82375bb2..4bb45ca40144 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe_config.h
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_config.h
@@ -29,6 +29,9 @@
#define PPE_RSS_HASH_IP_LENGTH 4
#define PPE_RSS_HASH_TUPLES 5
+/* PPE supports 300 queues, each bit presents as one queue. */
+#define PPE_RING_TO_QUEUE_BITMAP_WORD_CNT 10
+
/**
* enum ppe_scheduler_frame_mode - PPE scheduler frame mode.
* @PPE_SCH_WITH_IPG_PREAMBLE_FRAME_CRC: The scheduled frame includes IPG,
@@ -308,4 +311,7 @@ int ppe_sc_config_set(struct ppe_device *ppe_dev, int sc,
int ppe_counter_enable_set(struct ppe_device *ppe_dev, int port);
int ppe_rss_hash_config_set(struct ppe_device *ppe_dev, int mode,
struct ppe_rss_hash_cfg hash_cfg);
+int ppe_ring_queue_map_set(struct ppe_device *ppe_dev,
+ int ring_id,
+ u32 *queue_map);
#endif
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h b/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
index 3e9cccedc6be..224beda046d4 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
@@ -207,6 +207,11 @@
#define PPE_L0_COMP_CFG_TBL_SHAPER_METER_LEN GENMASK(1, 0)
#define PPE_L0_COMP_CFG_TBL_NODE_METER_LEN GENMASK(3, 2)
+/* PPE queue to Ethernet DMA ring mapping table. */
+#define PPE_RING_Q_MAP_TBL_ADDR 0x42a000
+#define PPE_RING_Q_MAP_TBL_ENTRIES 24
+#define PPE_RING_Q_MAP_TBL_INC 0x40
+
/* Table addresses for per-queue dequeue setting. */
#define PPE_DEQ_OPR_TBL_ADDR 0x430000
#define PPE_DEQ_OPR_TBL_ENTRIES 300
--
2.34.1
^ permalink raw reply related [flat|nested] 17+ messages in thread
* [PATCH net-next v6 12/14] net: ethernet: qualcomm: Initialize PPE L2 bridge settings
2025-07-20 10:57 [PATCH net-next v6 00/14] Add PPE driver for Qualcomm IPQ9574 SoC Luo Jie
` (10 preceding siblings ...)
2025-07-20 10:57 ` [PATCH net-next v6 11/14] net: ethernet: qualcomm: Initialize PPE queue to Ethernet DMA ring mapping Luo Jie
@ 2025-07-20 10:57 ` Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 13/14] net: ethernet: qualcomm: Add PPE debugfs support for PPE counters Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 14/14] MAINTAINERS: Add maintainer for Qualcomm PPE driver Luo Jie
13 siblings, 0 replies; 17+ messages in thread
From: Luo Jie @ 2025-07-20 10:57 UTC (permalink / raw)
To: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Rob Herring, Krzysztof Kozlowski, Conor Dooley,
Lei Wei, Suruchi Agarwal, Pavithra R, Simon Horman,
Jonathan Corbet, Kees Cook, Gustavo A. R. Silva, Philipp Zabel
Cc: linux-arm-msm, netdev, devicetree, linux-kernel, linux-doc,
linux-hardening, quic_kkumarcs, quic_linchen, Luo Jie
From: Lei Wei <quic_leiwei@quicinc.com>
Initialize the L2 bridge settings for the PPE ports to only enable
L2 frame forwarding between CPU port and PPE Ethernet ports.
The per-port L2 bridge settings are initialized as follows:
For PPE CPU port, the PPE bridge TX is enabled and FDB learning is
disabled. For PPE physical ports, the default L2 forwarding action
is initialized to forward to CPU port only.
L2/FDB learning and forwarding will not be enabled for PPE physical
ports yet, since the port's VSI (Virtual Switch Instance) and VSI
membership are not yet configured, which are required for FDB
forwarding. The VSI and FDB forwarding will later be enabled when
switchdev is enabled.
Signed-off-by: Lei Wei <quic_leiwei@quicinc.com>
Signed-off-by: Luo Jie <quic_luoj@quicinc.com>
---
drivers/net/ethernet/qualcomm/ppe/ppe_config.c | 80 +++++++++++++++++++++++++-
drivers/net/ethernet/qualcomm/ppe/ppe_regs.h | 50 ++++++++++++++++
2 files changed, 129 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_config.c b/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
index b9e99fd28831..89109a699ac3 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_config.c
@@ -1919,6 +1919,80 @@ static int ppe_queues_to_ring_init(struct ppe_device *ppe_dev)
return ppe_ring_queue_map_set(ppe_dev, 0, queue_bmap);
}
+/* Initialize PPE bridge settings to only enable L2 frame receive and
+ * transmit between CPU port and PPE Ethernet ports.
+ */
+static int ppe_bridge_init(struct ppe_device *ppe_dev)
+{
+ u32 reg, mask, port_cfg[4], vsi_cfg[2];
+ int ret, i;
+
+ /* Configure the following settings for CPU port0:
+ * a.) Enable Bridge TX
+ * b.) Disable FDB new address learning
+ * c.) Disable station move address learning
+ */
+ mask = PPE_PORT_BRIDGE_TXMAC_EN;
+ mask |= PPE_PORT_BRIDGE_NEW_LRN_EN;
+ mask |= PPE_PORT_BRIDGE_STA_MOVE_LRN_EN;
+ ret = regmap_update_bits(ppe_dev->regmap,
+ PPE_PORT_BRIDGE_CTRL_ADDR,
+ mask,
+ PPE_PORT_BRIDGE_TXMAC_EN);
+ if (ret)
+ return ret;
+
+ for (i = 1; i < ppe_dev->num_ports; i++) {
+ /* Enable invalid VSI forwarding for all the physical ports
+ * to CPU port0, in case no VSI is assigned to the physical
+ * port.
+ */
+ reg = PPE_L2_VP_PORT_TBL_ADDR + PPE_L2_VP_PORT_TBL_INC * i;
+ ret = regmap_bulk_read(ppe_dev->regmap, reg,
+ port_cfg, ARRAY_SIZE(port_cfg));
+
+ if (ret)
+ return ret;
+
+ PPE_L2_PORT_SET_INVALID_VSI_FWD_EN(port_cfg, true);
+ PPE_L2_PORT_SET_DST_INFO(port_cfg, 0);
+
+ ret = regmap_bulk_write(ppe_dev->regmap, reg,
+ port_cfg, ARRAY_SIZE(port_cfg));
+ if (ret)
+ return ret;
+ }
+
+ for (i = 0; i < PPE_VSI_TBL_ENTRIES; i++) {
+ /* Set the VSI forward membership to include only CPU port0.
+ * FDB learning and forwarding take place only after switchdev
+ * is supported later to create the VSI and join the physical
+ * ports to the VSI port member.
+ */
+ reg = PPE_VSI_TBL_ADDR + PPE_VSI_TBL_INC * i;
+ ret = regmap_bulk_read(ppe_dev->regmap, reg,
+ vsi_cfg, ARRAY_SIZE(vsi_cfg));
+ if (ret)
+ return ret;
+
+ PPE_VSI_SET_MEMBER_PORT_BITMAP(vsi_cfg, BIT(0));
+ PPE_VSI_SET_UUC_BITMAP(vsi_cfg, BIT(0));
+ PPE_VSI_SET_UMC_BITMAP(vsi_cfg, BIT(0));
+ PPE_VSI_SET_BC_BITMAP(vsi_cfg, BIT(0));
+ PPE_VSI_SET_NEW_ADDR_LRN_EN(vsi_cfg, true);
+ PPE_VSI_SET_NEW_ADDR_FWD_CMD(vsi_cfg, PPE_ACTION_FORWARD);
+ PPE_VSI_SET_STATION_MOVE_LRN_EN(vsi_cfg, true);
+ PPE_VSI_SET_STATION_MOVE_FWD_CMD(vsi_cfg, PPE_ACTION_FORWARD);
+
+ ret = regmap_bulk_write(ppe_dev->regmap, reg,
+ vsi_cfg, ARRAY_SIZE(vsi_cfg));
+ if (ret)
+ return ret;
+ }
+
+ return 0;
+}
+
int ppe_hw_config(struct ppe_device *ppe_dev)
{
int ret;
@@ -1951,5 +2025,9 @@ int ppe_hw_config(struct ppe_device *ppe_dev)
if (ret)
return ret;
- return ppe_queues_to_ring_init(ppe_dev);
+ ret = ppe_queues_to_ring_init(ppe_dev);
+ if (ret)
+ return ret;
+
+ return ppe_bridge_init(ppe_dev);
}
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h b/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
index 224beda046d4..6fc63f82ee80 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
@@ -117,6 +117,14 @@
#define PPE_EG_SERVICE_SET_TX_CNT_EN(tbl_cfg, value) \
FIELD_MODIFY(PPE_EG_SERVICE_W1_TX_CNT_EN, (tbl_cfg) + 0x1, value)
+/* PPE port bridge configuration */
+#define PPE_PORT_BRIDGE_CTRL_ADDR 0x60300
+#define PPE_PORT_BRIDGE_CTRL_ENTRIES 8
+#define PPE_PORT_BRIDGE_CTRL_INC 4
+#define PPE_PORT_BRIDGE_NEW_LRN_EN BIT(0)
+#define PPE_PORT_BRIDGE_STA_MOVE_LRN_EN BIT(3)
+#define PPE_PORT_BRIDGE_TXMAC_EN BIT(16)
+
/* PPE port control configurations for the traffic to the multicast queues. */
#define PPE_MC_MTU_CTRL_TBL_ADDR 0x60a00
#define PPE_MC_MTU_CTRL_TBL_ENTRIES 8
@@ -125,6 +133,36 @@
#define PPE_MC_MTU_CTRL_TBL_MTU_CMD GENMASK(15, 14)
#define PPE_MC_MTU_CTRL_TBL_TX_CNT_EN BIT(16)
+/* PPE VSI configurations */
+#define PPE_VSI_TBL_ADDR 0x63800
+#define PPE_VSI_TBL_ENTRIES 64
+#define PPE_VSI_TBL_INC 0x10
+#define PPE_VSI_W0_MEMBER_PORT_BITMAP GENMASK(7, 0)
+#define PPE_VSI_W0_UUC_BITMAP GENMASK(15, 8)
+#define PPE_VSI_W0_UMC_BITMAP GENMASK(23, 16)
+#define PPE_VSI_W0_BC_BITMAP GENMASK(31, 24)
+#define PPE_VSI_W1_NEW_ADDR_LRN_EN BIT(0)
+#define PPE_VSI_W1_NEW_ADDR_FWD_CMD GENMASK(2, 1)
+#define PPE_VSI_W1_STATION_MOVE_LRN_EN BIT(3)
+#define PPE_VSI_W1_STATION_MOVE_FWD_CMD GENMASK(5, 4)
+
+#define PPE_VSI_SET_MEMBER_PORT_BITMAP(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_VSI_W0_MEMBER_PORT_BITMAP, tbl_cfg, value)
+#define PPE_VSI_SET_UUC_BITMAP(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_VSI_W0_UUC_BITMAP, tbl_cfg, value)
+#define PPE_VSI_SET_UMC_BITMAP(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_VSI_W0_UMC_BITMAP, tbl_cfg, value)
+#define PPE_VSI_SET_BC_BITMAP(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_VSI_W0_BC_BITMAP, tbl_cfg, value)
+#define PPE_VSI_SET_NEW_ADDR_LRN_EN(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_VSI_W1_NEW_ADDR_LRN_EN, (tbl_cfg) + 0x1, value)
+#define PPE_VSI_SET_NEW_ADDR_FWD_CMD(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_VSI_W1_NEW_ADDR_FWD_CMD, (tbl_cfg) + 0x1, value)
+#define PPE_VSI_SET_STATION_MOVE_LRN_EN(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_VSI_W1_STATION_MOVE_LRN_EN, (tbl_cfg) + 0x1, value)
+#define PPE_VSI_SET_STATION_MOVE_FWD_CMD(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_VSI_W1_STATION_MOVE_FWD_CMD, (tbl_cfg) + 0x1, value)
+
/* PPE port control configurations for the traffic to the unicast queues. */
#define PPE_MRU_MTU_CTRL_TBL_ADDR 0x65000
#define PPE_MRU_MTU_CTRL_TBL_ENTRIES 256
@@ -163,6 +201,18 @@
#define PPE_IN_L2_SERVICE_TBL_RX_CNT_EN BIT(30)
#define PPE_IN_L2_SERVICE_TBL_TX_CNT_EN BIT(31)
+/* L2 Port configurations */
+#define PPE_L2_VP_PORT_TBL_ADDR 0x98000
+#define PPE_L2_VP_PORT_TBL_ENTRIES 256
+#define PPE_L2_VP_PORT_TBL_INC 0x10
+#define PPE_L2_VP_PORT_W0_INVALID_VSI_FWD_EN BIT(0)
+#define PPE_L2_VP_PORT_W0_DST_INFO GENMASK(9, 2)
+
+#define PPE_L2_PORT_SET_INVALID_VSI_FWD_EN(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_L2_VP_PORT_W0_INVALID_VSI_FWD_EN, tbl_cfg, value)
+#define PPE_L2_PORT_SET_DST_INFO(tbl_cfg, value) \
+ FIELD_MODIFY(PPE_L2_VP_PORT_W0_DST_INFO, tbl_cfg, value)
+
/* PPE service code configuration for the tunnel packet. */
#define PPE_TL_SERVICE_TBL_ADDR 0x306000
#define PPE_TL_SERVICE_TBL_ENTRIES 256
--
2.34.1
^ permalink raw reply related [flat|nested] 17+ messages in thread
* [PATCH net-next v6 13/14] net: ethernet: qualcomm: Add PPE debugfs support for PPE counters
2025-07-20 10:57 [PATCH net-next v6 00/14] Add PPE driver for Qualcomm IPQ9574 SoC Luo Jie
` (11 preceding siblings ...)
2025-07-20 10:57 ` [PATCH net-next v6 12/14] net: ethernet: qualcomm: Initialize PPE L2 bridge settings Luo Jie
@ 2025-07-20 10:57 ` Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 14/14] MAINTAINERS: Add maintainer for Qualcomm PPE driver Luo Jie
13 siblings, 0 replies; 17+ messages in thread
From: Luo Jie @ 2025-07-20 10:57 UTC (permalink / raw)
To: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Rob Herring, Krzysztof Kozlowski, Conor Dooley,
Lei Wei, Suruchi Agarwal, Pavithra R, Simon Horman,
Jonathan Corbet, Kees Cook, Gustavo A. R. Silva, Philipp Zabel
Cc: linux-arm-msm, netdev, devicetree, linux-kernel, linux-doc,
linux-hardening, quic_kkumarcs, quic_linchen, Luo Jie
The PPE hardware counters maintain counters for packets handled by the
various functional blocks of PPE. They help in tracing the packets
passed through PPE and debugging any packet drops.
The counters displayed by this debugfs file are ones that are common
for all Ethernet ports, and they do not include the counters that are
specific for a MAC port. Hence they cannot be displayed using ethtool.
The per-MAC counters will be supported using "ethtool -S" along with
the netdevice driver.
The PPE hardware various type counters are made available through the
debugfs files under directory "/sys/kernel/debug/ppe/".
Signed-off-by: Luo Jie <quic_luoj@quicinc.com>
---
drivers/net/ethernet/qualcomm/ppe/Makefile | 2 +-
drivers/net/ethernet/qualcomm/ppe/ppe.c | 11 +
drivers/net/ethernet/qualcomm/ppe/ppe.h | 3 +
drivers/net/ethernet/qualcomm/ppe/ppe_debugfs.c | 847 ++++++++++++++++++++++++
drivers/net/ethernet/qualcomm/ppe/ppe_debugfs.h | 16 +
drivers/net/ethernet/qualcomm/ppe/ppe_regs.h | 134 ++++
6 files changed, 1012 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/qualcomm/ppe/Makefile b/drivers/net/ethernet/qualcomm/ppe/Makefile
index 410a7bb54cfe..9e60b2400c16 100644
--- a/drivers/net/ethernet/qualcomm/ppe/Makefile
+++ b/drivers/net/ethernet/qualcomm/ppe/Makefile
@@ -4,4 +4,4 @@
#
obj-$(CONFIG_QCOM_PPE) += qcom-ppe.o
-qcom-ppe-objs := ppe.o ppe_config.o
+qcom-ppe-objs := ppe.o ppe_config.o ppe_debugfs.o
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe.c b/drivers/net/ethernet/qualcomm/ppe/ppe.c
index 12d0bddc4c87..5f757ddc4406 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe.c
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe.c
@@ -16,6 +16,7 @@
#include "ppe.h"
#include "ppe_config.h"
+#include "ppe_debugfs.h"
#define PPE_PORT_MAX 8
#define PPE_CLK_RATE 353000000
@@ -199,11 +200,20 @@ static int qcom_ppe_probe(struct platform_device *pdev)
if (ret)
return dev_err_probe(dev, ret, "PPE HW config failed\n");
+ ppe_debugfs_setup(ppe_dev);
platform_set_drvdata(pdev, ppe_dev);
return 0;
}
+static void qcom_ppe_remove(struct platform_device *pdev)
+{
+ struct ppe_device *ppe_dev;
+
+ ppe_dev = platform_get_drvdata(pdev);
+ ppe_debugfs_teardown(ppe_dev);
+}
+
static const struct of_device_id qcom_ppe_of_match[] = {
{ .compatible = "qcom,ipq9574-ppe" },
{}
@@ -216,6 +226,7 @@ static struct platform_driver qcom_ppe_driver = {
.of_match_table = qcom_ppe_of_match,
},
.probe = qcom_ppe_probe,
+ .remove = qcom_ppe_remove,
};
module_platform_driver(qcom_ppe_driver);
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe.h b/drivers/net/ethernet/qualcomm/ppe/ppe.h
index 779f39c9f098..27458f0bc206 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe.h
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe.h
@@ -11,6 +11,7 @@
struct device;
struct regmap;
+struct dentry;
/**
* struct ppe_device - PPE device private data.
@@ -18,6 +19,7 @@ struct regmap;
* @regmap: PPE register map.
* @clk_rate: PPE clock rate.
* @num_ports: Number of PPE ports.
+ * @debugfs_root: Debugfs root entry.
* @num_icc_paths: Number of interconnect paths.
* @icc_paths: Interconnect path array.
*
@@ -30,6 +32,7 @@ struct ppe_device {
struct regmap *regmap;
unsigned long clk_rate;
unsigned int num_ports;
+ struct dentry *debugfs_root;
unsigned int num_icc_paths;
struct icc_bulk_data icc_paths[] __counted_by(num_icc_paths);
};
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_debugfs.c b/drivers/net/ethernet/qualcomm/ppe/ppe_debugfs.c
new file mode 100644
index 000000000000..fd959a76ff43
--- /dev/null
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_debugfs.c
@@ -0,0 +1,847 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (c) Qualcomm Technologies, Inc. and/or its subsidiaries.
+ */
+
+/* PPE debugfs routines for display of PPE counters useful for debug. */
+
+#include <linux/bitfield.h>
+#include <linux/debugfs.h>
+#include <linux/dev_printk.h>
+#include <linux/device.h>
+#include <linux/regmap.h>
+#include <linux/seq_file.h>
+
+#include "ppe.h"
+#include "ppe_config.h"
+#include "ppe_debugfs.h"
+#include "ppe_regs.h"
+
+#define PPE_PKT_CNT_TBL_SIZE 3
+#define PPE_DROP_PKT_CNT_TBL_SIZE 5
+
+#define PPE_W0_PKT_CNT GENMASK(31, 0)
+#define PPE_W2_DROP_PKT_CNT_LOW GENMASK(31, 8)
+#define PPE_W3_DROP_PKT_CNT_HIGH GENMASK(7, 0)
+
+#define PPE_GET_PKT_CNT(tbl_cnt) \
+ FIELD_GET(PPE_W0_PKT_CNT, *(tbl_cnt))
+#define PPE_GET_DROP_PKT_CNT_LOW(tbl_cnt) \
+ FIELD_GET(PPE_W2_DROP_PKT_CNT_LOW, *((tbl_cnt) + 0x2))
+#define PPE_GET_DROP_PKT_CNT_HIGH(tbl_cnt) \
+ FIELD_GET(PPE_W3_DROP_PKT_CNT_HIGH, *((tbl_cnt) + 0x3))
+
+/**
+ * enum ppe_cnt_size_type - PPE counter size type
+ * @PPE_PKT_CNT_SIZE_1WORD: Counter size with single register
+ * @PPE_PKT_CNT_SIZE_3WORD: Counter size with table of 3 words
+ * @PPE_PKT_CNT_SIZE_5WORD: Counter size with table of 5 words
+ *
+ * PPE takes the different register size to record the packet counters.
+ * It uses single register, or register table with 3 words or 5 words.
+ * The counter with table size 5 words also records the drop counter.
+ * There are also some other counter types occupying sizes less than 32
+ * bits, which is not covered by this enumeration type.
+ */
+enum ppe_cnt_size_type {
+ PPE_PKT_CNT_SIZE_1WORD,
+ PPE_PKT_CNT_SIZE_3WORD,
+ PPE_PKT_CNT_SIZE_5WORD,
+};
+
+/**
+ * enum ppe_cnt_type - PPE counter type.
+ * @PPE_CNT_BM: Packet counter processed by BM.
+ * @PPE_CNT_PARSE: Packet counter parsed on ingress.
+ * @PPE_CNT_PORT_RX: Packet counter on the ingress port.
+ * @PPE_CNT_VLAN_RX: VLAN packet counter received.
+ * @PPE_CNT_L2_FWD: Packet counter processed by L2 forwarding.
+ * @PPE_CNT_CPU_CODE: Packet counter marked with various CPU codes.
+ * @PPE_CNT_VLAN_TX: VLAN packet counter transmitted.
+ * @PPE_CNT_PORT_TX: Packet counter on the egress port.
+ * @PPE_CNT_QM: Packet counter processed by QM.
+ */
+enum ppe_cnt_type {
+ PPE_CNT_BM,
+ PPE_CNT_PARSE,
+ PPE_CNT_PORT_RX,
+ PPE_CNT_VLAN_RX,
+ PPE_CNT_L2_FWD,
+ PPE_CNT_CPU_CODE,
+ PPE_CNT_VLAN_TX,
+ PPE_CNT_PORT_TX,
+ PPE_CNT_QM,
+};
+
+/**
+ * struct ppe_debugfs_entry - PPE debugfs entry.
+ * @name: Debugfs file name.
+ * @counter_type: PPE packet counter type.
+ * @ppe: PPE device.
+ *
+ * The PPE debugfs entry is used to create the debugfs file and passed
+ * to debugfs_create_file() as private data.
+ */
+struct ppe_debugfs_entry {
+ const char *name;
+ enum ppe_cnt_type counter_type;
+ struct ppe_device *ppe;
+};
+
+static const struct ppe_debugfs_entry debugfs_files[] = {
+ {
+ .name = "bm",
+ .counter_type = PPE_CNT_BM,
+ },
+ {
+ .name = "parse",
+ .counter_type = PPE_CNT_PARSE,
+ },
+ {
+ .name = "port_rx",
+ .counter_type = PPE_CNT_PORT_RX,
+ },
+ {
+ .name = "vlan_rx",
+ .counter_type = PPE_CNT_VLAN_RX,
+ },
+ {
+ .name = "l2_forward",
+ .counter_type = PPE_CNT_L2_FWD,
+ },
+ {
+ .name = "cpu_code",
+ .counter_type = PPE_CNT_CPU_CODE,
+ },
+ {
+ .name = "vlan_tx",
+ .counter_type = PPE_CNT_VLAN_TX,
+ },
+ {
+ .name = "port_tx",
+ .counter_type = PPE_CNT_PORT_TX,
+ },
+ {
+ .name = "qm",
+ .counter_type = PPE_CNT_QM,
+ },
+};
+
+static int ppe_pkt_cnt_get(struct ppe_device *ppe_dev, u32 reg,
+ enum ppe_cnt_size_type cnt_type,
+ u32 *cnt, u32 *drop_cnt)
+{
+ u32 drop_pkt_cnt[PPE_DROP_PKT_CNT_TBL_SIZE];
+ u32 pkt_cnt[PPE_PKT_CNT_TBL_SIZE];
+ u32 value;
+ int ret;
+
+ switch (cnt_type) {
+ case PPE_PKT_CNT_SIZE_1WORD:
+ ret = regmap_read(ppe_dev->regmap, reg, &value);
+ if (ret)
+ return ret;
+
+ *cnt = value;
+ break;
+ case PPE_PKT_CNT_SIZE_3WORD:
+ ret = regmap_bulk_read(ppe_dev->regmap, reg,
+ pkt_cnt, ARRAY_SIZE(pkt_cnt));
+ if (ret)
+ return ret;
+
+ *cnt = PPE_GET_PKT_CNT(pkt_cnt);
+ break;
+ case PPE_PKT_CNT_SIZE_5WORD:
+ ret = regmap_bulk_read(ppe_dev->regmap, reg,
+ drop_pkt_cnt, ARRAY_SIZE(drop_pkt_cnt));
+ if (ret)
+ return ret;
+
+ *cnt = PPE_GET_PKT_CNT(drop_pkt_cnt);
+
+ /* Drop counter with low 24 bits. */
+ value = PPE_GET_DROP_PKT_CNT_LOW(drop_pkt_cnt);
+ *drop_cnt = FIELD_PREP(GENMASK(23, 0), value);
+
+ /* Drop counter with high 8 bits. */
+ value = PPE_GET_DROP_PKT_CNT_HIGH(drop_pkt_cnt);
+ *drop_cnt |= FIELD_PREP(GENMASK(31, 24), value);
+ break;
+ }
+
+ return 0;
+}
+
+static void ppe_tbl_pkt_cnt_clear(struct ppe_device *ppe_dev, u32 reg,
+ enum ppe_cnt_size_type cnt_type)
+{
+ u32 drop_pkt_cnt[PPE_DROP_PKT_CNT_TBL_SIZE] = {};
+ u32 pkt_cnt[PPE_PKT_CNT_TBL_SIZE] = {};
+
+ switch (cnt_type) {
+ case PPE_PKT_CNT_SIZE_1WORD:
+ regmap_write(ppe_dev->regmap, reg, 0);
+ break;
+ case PPE_PKT_CNT_SIZE_3WORD:
+ regmap_bulk_write(ppe_dev->regmap, reg,
+ pkt_cnt, ARRAY_SIZE(pkt_cnt));
+ break;
+ case PPE_PKT_CNT_SIZE_5WORD:
+ regmap_bulk_write(ppe_dev->regmap, reg,
+ drop_pkt_cnt, ARRAY_SIZE(drop_pkt_cnt));
+ break;
+ }
+}
+
+static int ppe_bm_counter_get(struct ppe_device *ppe_dev, struct seq_file *seq)
+{
+ u32 reg, val, pkt_cnt, pkt_cnt1;
+ int ret, i, tag;
+
+ seq_printf(seq, "%-24s", "BM SILENT_DROP:");
+ tag = 0;
+ for (i = 0; i < PPE_DROP_CNT_TBL_ENTRIES; i++) {
+ reg = PPE_DROP_CNT_TBL_ADDR + i * PPE_DROP_CNT_TBL_INC;
+ ret = ppe_pkt_cnt_get(ppe_dev, reg, PPE_PKT_CNT_SIZE_1WORD,
+ &pkt_cnt, NULL);
+ if (ret) {
+ dev_err(ppe_dev->dev, "CNT ERROR %d\n", ret);
+ return ret;
+ }
+
+ if (pkt_cnt > 0) {
+ if (!((++tag) % 4))
+ seq_printf(seq, "\n%-24s", "");
+
+ seq_printf(seq, "%10u(%s=%04d)", pkt_cnt, "port", i);
+ }
+ }
+
+ seq_putc(seq, '\n');
+
+ /* The number of packets dropped because hardware buffers were
+ * available only partially for the packet.
+ */
+ seq_printf(seq, "%-24s", "BM OVERFLOW_DROP:");
+ tag = 0;
+ for (i = 0; i < PPE_DROP_STAT_TBL_ENTRIES; i++) {
+ reg = PPE_DROP_STAT_TBL_ADDR + PPE_DROP_STAT_TBL_INC * i;
+
+ ret = ppe_pkt_cnt_get(ppe_dev, reg, PPE_PKT_CNT_SIZE_3WORD,
+ &pkt_cnt, NULL);
+ if (ret) {
+ dev_err(ppe_dev->dev, "CNT ERROR %d\n", ret);
+ return ret;
+ }
+
+ if (pkt_cnt > 0) {
+ if (!((++tag) % 4))
+ seq_printf(seq, "\n%-24s", "");
+
+ seq_printf(seq, "%10u(%s=%04d)", pkt_cnt, "port", i);
+ }
+ }
+
+ seq_putc(seq, '\n');
+
+ /* The number of currently occupied buffers, that can't be flushed. */
+ seq_printf(seq, "%-24s", "BM USED/REACT:");
+ tag = 0;
+ for (i = 0; i < PPE_BM_USED_CNT_TBL_ENTRIES; i++) {
+ reg = PPE_BM_USED_CNT_TBL_ADDR + i * PPE_BM_USED_CNT_TBL_INC;
+ ret = regmap_read(ppe_dev->regmap, reg, &val);
+ if (ret) {
+ dev_err(ppe_dev->dev, "CNT ERROR %d\n", ret);
+ return ret;
+ }
+
+ /* The number of PPE buffers used for caching the received
+ * packets before the pause frame sent.
+ */
+ pkt_cnt = FIELD_GET(PPE_BM_USED_CNT_VAL, val);
+
+ reg = PPE_BM_REACT_CNT_TBL_ADDR + i * PPE_BM_REACT_CNT_TBL_INC;
+ ret = regmap_read(ppe_dev->regmap, reg, &val);
+ if (ret) {
+ dev_err(ppe_dev->dev, "CNT ERROR %d\n", ret);
+ return ret;
+ }
+
+ /* The number of PPE buffers used for caching the received
+ * packets after pause frame sent out.
+ */
+ pkt_cnt1 = FIELD_GET(PPE_BM_REACT_CNT_VAL, val);
+
+ if (pkt_cnt > 0 || pkt_cnt1 > 0) {
+ if (!((++tag) % 4))
+ seq_printf(seq, "\n%-24s", "");
+
+ seq_printf(seq, "%10u/%u(%s=%04d)", pkt_cnt, pkt_cnt1,
+ "port", i);
+ }
+ }
+
+ seq_putc(seq, '\n');
+
+ return 0;
+}
+
+/* The number of packets processed by the ingress parser module of PPE. */
+static int ppe_parse_pkt_counter_get(struct ppe_device *ppe_dev,
+ struct seq_file *seq)
+{
+ u32 reg, cnt = 0, tunnel_cnt = 0;
+ int i, ret, tag = 0;
+
+ seq_printf(seq, "%-24s", "PARSE TPRX/IPRX:");
+ for (i = 0; i < PPE_IPR_PKT_CNT_TBL_ENTRIES; i++) {
+ reg = PPE_TPR_PKT_CNT_TBL_ADDR + i * PPE_TPR_PKT_CNT_TBL_INC;
+ ret = ppe_pkt_cnt_get(ppe_dev, reg, PPE_PKT_CNT_SIZE_1WORD,
+ &tunnel_cnt, NULL);
+ if (ret) {
+ dev_err(ppe_dev->dev, "CNT ERROR %d\n", ret);
+ return ret;
+ }
+
+ reg = PPE_IPR_PKT_CNT_TBL_ADDR + i * PPE_IPR_PKT_CNT_TBL_INC;
+ ret = ppe_pkt_cnt_get(ppe_dev, reg, PPE_PKT_CNT_SIZE_1WORD,
+ &cnt, NULL);
+ if (ret) {
+ dev_err(ppe_dev->dev, "CNT ERROR %d\n", ret);
+ return ret;
+ }
+
+ if (tunnel_cnt > 0 || cnt > 0) {
+ if (!((++tag) % 4))
+ seq_printf(seq, "\n%-24s", "");
+
+ seq_printf(seq, "%10u/%u(%s=%04d)", tunnel_cnt, cnt,
+ "port", i);
+ }
+ }
+
+ seq_putc(seq, '\n');
+
+ return 0;
+}
+
+/* The number of packets received or dropped on the ingress port. */
+static int ppe_port_rx_counter_get(struct ppe_device *ppe_dev,
+ struct seq_file *seq)
+{
+ u32 reg, pkt_cnt = 0, drop_cnt = 0;
+ int ret, i, tag;
+
+ seq_printf(seq, "%-24s", "PORT RX/RX_DROP:");
+ tag = 0;
+ for (i = 0; i < PPE_PHY_PORT_RX_CNT_TBL_ENTRIES; i++) {
+ reg = PPE_PHY_PORT_RX_CNT_TBL_ADDR + PPE_PHY_PORT_RX_CNT_TBL_INC * i;
+ ret = ppe_pkt_cnt_get(ppe_dev, reg, PPE_PKT_CNT_SIZE_5WORD,
+ &pkt_cnt, &drop_cnt);
+ if (ret) {
+ dev_err(ppe_dev->dev, "CNT ERROR %d\n", ret);
+ return ret;
+ }
+
+ if (pkt_cnt > 0) {
+ if (!((++tag) % 4))
+ seq_printf(seq, "\n%-24s", "");
+
+ seq_printf(seq, "%10u/%u(%s=%04d)", pkt_cnt, drop_cnt,
+ "port", i);
+ }
+ }
+
+ seq_putc(seq, '\n');
+
+ seq_printf(seq, "%-24s", "VPORT RX/RX_DROP:");
+ tag = 0;
+ for (i = 0; i < PPE_PORT_RX_CNT_TBL_ENTRIES; i++) {
+ reg = PPE_PORT_RX_CNT_TBL_ADDR + PPE_PORT_RX_CNT_TBL_INC * i;
+ ret = ppe_pkt_cnt_get(ppe_dev, reg, PPE_PKT_CNT_SIZE_5WORD,
+ &pkt_cnt, &drop_cnt);
+ if (ret) {
+ dev_err(ppe_dev->dev, "CNT ERROR %d\n", ret);
+ return ret;
+ }
+
+ if (pkt_cnt > 0) {
+ if (!((++tag) % 4))
+ seq_printf(seq, "\n%-24s", "");
+
+ seq_printf(seq, "%10u/%u(%s=%04d)", pkt_cnt, drop_cnt,
+ "port", i);
+ }
+ }
+
+ seq_putc(seq, '\n');
+
+ return 0;
+}
+
+/* The number of packets received or dropped by layer 2 processing. */
+static int ppe_l2_counter_get(struct ppe_device *ppe_dev,
+ struct seq_file *seq)
+{
+ u32 reg, pkt_cnt = 0, drop_cnt = 0;
+ int ret, i, tag = 0;
+
+ seq_printf(seq, "%-24s", "L2 RX/RX_DROP:");
+ for (i = 0; i < PPE_PRE_L2_CNT_TBL_ENTRIES; i++) {
+ reg = PPE_PRE_L2_CNT_TBL_ADDR + PPE_PRE_L2_CNT_TBL_INC * i;
+ ret = ppe_pkt_cnt_get(ppe_dev, reg, PPE_PKT_CNT_SIZE_5WORD,
+ &pkt_cnt, &drop_cnt);
+ if (ret) {
+ dev_err(ppe_dev->dev, "CNT ERROR %d\n", ret);
+ return ret;
+ }
+
+ if (pkt_cnt > 0) {
+ if (!((++tag) % 4))
+ seq_printf(seq, "\n%-24s", "");
+
+ seq_printf(seq, "%10u/%u(%s=%04d)", pkt_cnt, drop_cnt,
+ "vsi", i);
+ }
+ }
+
+ seq_putc(seq, '\n');
+
+ return 0;
+}
+
+/* The number of VLAN packets received by PPE. */
+static int ppe_vlan_rx_counter_get(struct ppe_device *ppe_dev,
+ struct seq_file *seq)
+{
+ u32 reg, pkt_cnt = 0;
+ int ret, i, tag = 0;
+
+ seq_printf(seq, "%-24s", "VLAN RX:");
+ for (i = 0; i < PPE_VLAN_CNT_TBL_ENTRIES; i++) {
+ reg = PPE_VLAN_CNT_TBL_ADDR + PPE_VLAN_CNT_TBL_INC * i;
+
+ ret = ppe_pkt_cnt_get(ppe_dev, reg, PPE_PKT_CNT_SIZE_3WORD,
+ &pkt_cnt, NULL);
+ if (ret) {
+ dev_err(ppe_dev->dev, "CNT ERROR %d\n", ret);
+ return ret;
+ }
+
+ if (pkt_cnt > 0) {
+ if (!((++tag) % 4))
+ seq_printf(seq, "\n%-24s", "");
+
+ seq_printf(seq, "%10u(%s=%04d)", pkt_cnt, "vsi", i);
+ }
+ }
+
+ seq_putc(seq, '\n');
+
+ return 0;
+}
+
+/* The number of packets handed to CPU by PPE. */
+static int ppe_cpu_code_counter_get(struct ppe_device *ppe_dev,
+ struct seq_file *seq)
+{
+ u32 reg, pkt_cnt = 0;
+ int ret, i;
+
+ seq_printf(seq, "%-24s", "CPU CODE:");
+ for (i = 0; i < PPE_DROP_CPU_CNT_TBL_ENTRIES; i++) {
+ reg = PPE_DROP_CPU_CNT_TBL_ADDR + PPE_DROP_CPU_CNT_TBL_INC * i;
+
+ ret = ppe_pkt_cnt_get(ppe_dev, reg, PPE_PKT_CNT_SIZE_3WORD,
+ &pkt_cnt, NULL);
+ if (ret) {
+ dev_err(ppe_dev->dev, "CNT ERROR %d\n", ret);
+ return ret;
+ }
+
+ if (!pkt_cnt)
+ continue;
+
+ /* There are 256 CPU codes saved in the first 256 entries
+ * of register table, and 128 drop codes for each PPE port
+ * (0-7), the total entries is 256 + 8 * 128.
+ */
+ if (i < 256)
+ seq_printf(seq, "%10u(cpucode:%d)", pkt_cnt, i);
+ else
+ seq_printf(seq, "%10u(port=%04d),dropcode:%d", pkt_cnt,
+ (i - 256) % 8, (i - 256) / 8);
+ seq_putc(seq, '\n');
+ seq_printf(seq, "%-24s", "");
+ }
+
+ seq_putc(seq, '\n');
+
+ return 0;
+}
+
+/* The number of packets forwarded by VLAN on the egress direction. */
+static int ppe_vlan_tx_counter_get(struct ppe_device *ppe_dev,
+ struct seq_file *seq)
+{
+ u32 reg, pkt_cnt = 0;
+ int ret, i, tag = 0;
+
+ seq_printf(seq, "%-24s", "VLAN TX:");
+ for (i = 0; i < PPE_EG_VSI_COUNTER_TBL_ENTRIES; i++) {
+ reg = PPE_EG_VSI_COUNTER_TBL_ADDR + PPE_EG_VSI_COUNTER_TBL_INC * i;
+
+ ret = ppe_pkt_cnt_get(ppe_dev, reg, PPE_PKT_CNT_SIZE_3WORD,
+ &pkt_cnt, NULL);
+ if (ret) {
+ dev_err(ppe_dev->dev, "CNT ERROR %d\n", ret);
+ return ret;
+ }
+
+ if (pkt_cnt > 0) {
+ if (!((++tag) % 4))
+ seq_printf(seq, "\n%-24s", "");
+
+ seq_printf(seq, "%10u(%s=%04d)", pkt_cnt, "vsi", i);
+ }
+ }
+
+ seq_putc(seq, '\n');
+
+ return 0;
+}
+
+/* The number of packets transmitted or dropped on the egress port. */
+static int ppe_port_tx_counter_get(struct ppe_device *ppe_dev,
+ struct seq_file *seq)
+{
+ u32 reg, pkt_cnt = 0, drop_cnt = 0;
+ int ret, i, tag;
+
+ seq_printf(seq, "%-24s", "VPORT TX/TX_DROP:");
+ tag = 0;
+ for (i = 0; i < PPE_VPORT_TX_COUNTER_TBL_ENTRIES; i++) {
+ reg = PPE_VPORT_TX_COUNTER_TBL_ADDR + PPE_VPORT_TX_COUNTER_TBL_INC * i;
+ ret = ppe_pkt_cnt_get(ppe_dev, reg, PPE_PKT_CNT_SIZE_3WORD,
+ &pkt_cnt, NULL);
+ if (ret) {
+ dev_err(ppe_dev->dev, "CNT ERROR %d\n", ret);
+ return ret;
+ }
+
+ reg = PPE_VPORT_TX_DROP_CNT_TBL_ADDR + PPE_VPORT_TX_DROP_CNT_TBL_INC * i;
+ ret = ppe_pkt_cnt_get(ppe_dev, reg, PPE_PKT_CNT_SIZE_3WORD,
+ &drop_cnt, NULL);
+ if (ret) {
+ dev_err(ppe_dev->dev, "CNT ERROR %d\n", ret);
+ return ret;
+ }
+
+ if (pkt_cnt > 0 || drop_cnt > 0) {
+ if (!((++tag) % 4))
+ seq_printf(seq, "\n%-24s", "");
+
+ seq_printf(seq, "%10u/%u(%s=%04d)", pkt_cnt, drop_cnt,
+ "port", i);
+ }
+ }
+
+ seq_putc(seq, '\n');
+
+ seq_printf(seq, "%-24s", "PORT TX/TX_DROP:");
+ tag = 0;
+ for (i = 0; i < PPE_PORT_TX_COUNTER_TBL_ENTRIES; i++) {
+ reg = PPE_PORT_TX_COUNTER_TBL_ADDR + PPE_PORT_TX_COUNTER_TBL_INC * i;
+ ret = ppe_pkt_cnt_get(ppe_dev, reg, PPE_PKT_CNT_SIZE_3WORD,
+ &pkt_cnt, NULL);
+ if (ret) {
+ dev_err(ppe_dev->dev, "CNT ERROR %d\n", ret);
+ return ret;
+ }
+
+ reg = PPE_PORT_TX_DROP_CNT_TBL_ADDR + PPE_PORT_TX_DROP_CNT_TBL_INC * i;
+ ret = ppe_pkt_cnt_get(ppe_dev, reg, PPE_PKT_CNT_SIZE_3WORD,
+ &drop_cnt, NULL);
+ if (ret) {
+ dev_err(ppe_dev->dev, "CNT ERROR %d\n", ret);
+ return ret;
+ }
+
+ if (pkt_cnt > 0 || drop_cnt > 0) {
+ if (!((++tag) % 4))
+ seq_printf(seq, "\n%-24s", "");
+
+ seq_printf(seq, "%10u/%u(%s=%04d)", pkt_cnt, drop_cnt,
+ "port", i);
+ }
+ }
+
+ seq_putc(seq, '\n');
+
+ return 0;
+}
+
+/* The number of packets transmitted or pending by the PPE queue. */
+static int ppe_queue_counter_get(struct ppe_device *ppe_dev,
+ struct seq_file *seq)
+{
+ u32 reg, val, pkt_cnt = 0, pend_cnt = 0, drop_cnt = 0;
+ int ret, i, tag = 0;
+
+ seq_printf(seq, "%-24s", "QUEUE TX/PEND/DROP:");
+ for (i = 0; i < PPE_QUEUE_TX_COUNTER_TBL_ENTRIES; i++) {
+ reg = PPE_QUEUE_TX_COUNTER_TBL_ADDR + PPE_QUEUE_TX_COUNTER_TBL_INC * i;
+ ret = ppe_pkt_cnt_get(ppe_dev, reg, PPE_PKT_CNT_SIZE_3WORD,
+ &pkt_cnt, NULL);
+ if (ret) {
+ dev_err(ppe_dev->dev, "CNT ERROR %d\n", ret);
+ return ret;
+ }
+
+ if (i < PPE_AC_UNICAST_QUEUE_CFG_TBL_ENTRIES) {
+ reg = PPE_AC_UNICAST_QUEUE_CNT_TBL_ADDR +
+ PPE_AC_UNICAST_QUEUE_CNT_TBL_INC * i;
+ ret = regmap_read(ppe_dev->regmap, reg, &val);
+ if (ret) {
+ dev_err(ppe_dev->dev, "CNT ERROR %d\n", ret);
+ return ret;
+ }
+
+ pend_cnt = FIELD_GET(PPE_AC_UNICAST_QUEUE_CNT_TBL_PEND_CNT, val);
+
+ reg = PPE_UNICAST_DROP_CNT_TBL_ADDR +
+ PPE_AC_UNICAST_QUEUE_CNT_TBL_INC *
+ (i * PPE_UNICAST_DROP_TYPES + PPE_UNICAST_DROP_FORCE_OFFSET);
+
+ ret = ppe_pkt_cnt_get(ppe_dev, reg, PPE_PKT_CNT_SIZE_3WORD,
+ &drop_cnt, NULL);
+ if (ret) {
+ dev_err(ppe_dev->dev, "CNT ERROR %d\n", ret);
+ return ret;
+ }
+ } else {
+ int mq_offset = i - PPE_AC_UNICAST_QUEUE_CFG_TBL_ENTRIES;
+
+ reg = PPE_AC_MULTICAST_QUEUE_CNT_TBL_ADDR +
+ PPE_AC_MULTICAST_QUEUE_CNT_TBL_INC * mq_offset;
+ ret = regmap_read(ppe_dev->regmap, reg, &val);
+ if (ret) {
+ dev_err(ppe_dev->dev, "CNT ERROR %d\n", ret);
+ return ret;
+ }
+
+ pend_cnt = FIELD_GET(PPE_AC_MULTICAST_QUEUE_CNT_TBL_PEND_CNT, val);
+
+ if (mq_offset < PPE_P0_MULTICAST_QUEUE_NUM) {
+ reg = PPE_CPU_PORT_MULTICAST_FORCE_DROP_CNT_TBL_ADDR(mq_offset);
+ } else {
+ mq_offset -= PPE_P0_MULTICAST_QUEUE_NUM;
+
+ reg = PPE_P1_MULTICAST_DROP_CNT_TBL_ADDR;
+ reg += (mq_offset / PPE_MULTICAST_QUEUE_NUM) *
+ PPE_MULTICAST_QUEUE_PORT_ADDR_INC;
+ reg += (mq_offset % PPE_MULTICAST_QUEUE_NUM) *
+ PPE_MULTICAST_DROP_CNT_TBL_INC *
+ PPE_MULTICAST_DROP_TYPES;
+ }
+
+ ret = ppe_pkt_cnt_get(ppe_dev, reg, PPE_PKT_CNT_SIZE_3WORD,
+ &drop_cnt, NULL);
+ if (ret) {
+ dev_err(ppe_dev->dev, "CNT ERROR %d\n", ret);
+ return ret;
+ }
+ }
+
+ if (pkt_cnt > 0 || pend_cnt > 0 || drop_cnt > 0) {
+ if (!((++tag) % 4))
+ seq_printf(seq, "\n%-24s", "");
+
+ seq_printf(seq, "%10u/%u/%u(%s=%04d)",
+ pkt_cnt, pend_cnt, drop_cnt, "queue", i);
+ }
+ }
+
+ seq_putc(seq, '\n');
+
+ return 0;
+}
+
+/* Display the various packet counters of PPE. */
+static int ppe_packet_counter_show(struct seq_file *seq, void *v)
+{
+ struct ppe_debugfs_entry *entry = seq->private;
+ struct ppe_device *ppe_dev = entry->ppe;
+ int ret;
+
+ switch (entry->counter_type) {
+ case PPE_CNT_BM:
+ ret = ppe_bm_counter_get(ppe_dev, seq);
+ break;
+ case PPE_CNT_PARSE:
+ ret = ppe_parse_pkt_counter_get(ppe_dev, seq);
+ break;
+ case PPE_CNT_PORT_RX:
+ ret = ppe_port_rx_counter_get(ppe_dev, seq);
+ break;
+ case PPE_CNT_VLAN_RX:
+ ret = ppe_vlan_rx_counter_get(ppe_dev, seq);
+ break;
+ case PPE_CNT_L2_FWD:
+ ret = ppe_l2_counter_get(ppe_dev, seq);
+ break;
+ case PPE_CNT_CPU_CODE:
+ ret = ppe_cpu_code_counter_get(ppe_dev, seq);
+ break;
+ case PPE_CNT_VLAN_TX:
+ ret = ppe_vlan_tx_counter_get(ppe_dev, seq);
+ break;
+ case PPE_CNT_PORT_TX:
+ ret = ppe_port_tx_counter_get(ppe_dev, seq);
+ break;
+ case PPE_CNT_QM:
+ ret = ppe_queue_counter_get(ppe_dev, seq);
+ break;
+ default:
+ ret = -EINVAL;
+ break;
+ }
+
+ return ret;
+}
+
+/* Flush the various packet counters of PPE. */
+static ssize_t ppe_packet_counter_write(struct file *file,
+ const char __user *buf,
+ size_t count, loff_t *pos)
+{
+ struct ppe_debugfs_entry *entry = file_inode(file)->i_private;
+ struct ppe_device *ppe_dev = entry->ppe;
+ u32 reg;
+ int i;
+
+ switch (entry->counter_type) {
+ case PPE_CNT_BM:
+ for (i = 0; i < PPE_DROP_CNT_TBL_ENTRIES; i++) {
+ reg = PPE_DROP_CNT_TBL_ADDR + i * PPE_DROP_CNT_TBL_INC;
+ ppe_tbl_pkt_cnt_clear(ppe_dev, reg, PPE_PKT_CNT_SIZE_1WORD);
+ }
+
+ for (i = 0; i < PPE_DROP_STAT_TBL_ENTRIES; i++) {
+ reg = PPE_DROP_STAT_TBL_ADDR + PPE_DROP_STAT_TBL_INC * i;
+ ppe_tbl_pkt_cnt_clear(ppe_dev, reg, PPE_PKT_CNT_SIZE_3WORD);
+ }
+
+ break;
+ case PPE_CNT_PARSE:
+ for (i = 0; i < PPE_IPR_PKT_CNT_TBL_ENTRIES; i++) {
+ reg = PPE_IPR_PKT_CNT_TBL_ADDR + i * PPE_IPR_PKT_CNT_TBL_INC;
+ ppe_tbl_pkt_cnt_clear(ppe_dev, reg, PPE_PKT_CNT_SIZE_1WORD);
+
+ reg = PPE_TPR_PKT_CNT_TBL_ADDR + i * PPE_TPR_PKT_CNT_TBL_INC;
+ ppe_tbl_pkt_cnt_clear(ppe_dev, reg, PPE_PKT_CNT_SIZE_1WORD);
+ }
+
+ break;
+ case PPE_CNT_PORT_RX:
+ for (i = 0; i < PPE_PORT_RX_CNT_TBL_ENTRIES; i++) {
+ reg = PPE_PORT_RX_CNT_TBL_ADDR + PPE_PORT_RX_CNT_TBL_INC * i;
+ ppe_tbl_pkt_cnt_clear(ppe_dev, reg, PPE_PKT_CNT_SIZE_5WORD);
+ }
+
+ for (i = 0; i < PPE_PHY_PORT_RX_CNT_TBL_ENTRIES; i++) {
+ reg = PPE_PHY_PORT_RX_CNT_TBL_ADDR + PPE_PHY_PORT_RX_CNT_TBL_INC * i;
+ ppe_tbl_pkt_cnt_clear(ppe_dev, reg, PPE_PKT_CNT_SIZE_5WORD);
+ }
+
+ break;
+ case PPE_CNT_VLAN_RX:
+ for (i = 0; i < PPE_VLAN_CNT_TBL_ENTRIES; i++) {
+ reg = PPE_VLAN_CNT_TBL_ADDR + PPE_VLAN_CNT_TBL_INC * i;
+ ppe_tbl_pkt_cnt_clear(ppe_dev, reg, PPE_PKT_CNT_SIZE_3WORD);
+ }
+
+ break;
+ case PPE_CNT_L2_FWD:
+ for (i = 0; i < PPE_PRE_L2_CNT_TBL_ENTRIES; i++) {
+ reg = PPE_PRE_L2_CNT_TBL_ADDR + PPE_PRE_L2_CNT_TBL_INC * i;
+ ppe_tbl_pkt_cnt_clear(ppe_dev, reg, PPE_PKT_CNT_SIZE_5WORD);
+ }
+
+ break;
+ case PPE_CNT_CPU_CODE:
+ for (i = 0; i < PPE_DROP_CPU_CNT_TBL_ENTRIES; i++) {
+ reg = PPE_DROP_CPU_CNT_TBL_ADDR + PPE_DROP_CPU_CNT_TBL_INC * i;
+ ppe_tbl_pkt_cnt_clear(ppe_dev, reg, PPE_PKT_CNT_SIZE_3WORD);
+ }
+
+ break;
+ case PPE_CNT_VLAN_TX:
+ for (i = 0; i < PPE_EG_VSI_COUNTER_TBL_ENTRIES; i++) {
+ reg = PPE_EG_VSI_COUNTER_TBL_ADDR + PPE_EG_VSI_COUNTER_TBL_INC * i;
+ ppe_tbl_pkt_cnt_clear(ppe_dev, reg, PPE_PKT_CNT_SIZE_3WORD);
+ }
+
+ break;
+ case PPE_CNT_PORT_TX:
+ for (i = 0; i < PPE_PORT_TX_COUNTER_TBL_ENTRIES; i++) {
+ reg = PPE_PORT_TX_DROP_CNT_TBL_ADDR + PPE_PORT_TX_DROP_CNT_TBL_INC * i;
+ ppe_tbl_pkt_cnt_clear(ppe_dev, reg, PPE_PKT_CNT_SIZE_3WORD);
+
+ reg = PPE_PORT_TX_COUNTER_TBL_ADDR + PPE_PORT_TX_COUNTER_TBL_INC * i;
+ ppe_tbl_pkt_cnt_clear(ppe_dev, reg, PPE_PKT_CNT_SIZE_3WORD);
+ }
+
+ for (i = 0; i < PPE_VPORT_TX_COUNTER_TBL_ENTRIES; i++) {
+ reg = PPE_VPORT_TX_COUNTER_TBL_ADDR + PPE_VPORT_TX_COUNTER_TBL_INC * i;
+ ppe_tbl_pkt_cnt_clear(ppe_dev, reg, PPE_PKT_CNT_SIZE_3WORD);
+
+ reg = PPE_VPORT_TX_DROP_CNT_TBL_ADDR + PPE_VPORT_TX_DROP_CNT_TBL_INC * i;
+ ppe_tbl_pkt_cnt_clear(ppe_dev, reg, PPE_PKT_CNT_SIZE_3WORD);
+ }
+
+ break;
+ case PPE_CNT_QM:
+ for (i = 0; i < PPE_QUEUE_TX_COUNTER_TBL_ENTRIES; i++) {
+ reg = PPE_QUEUE_TX_COUNTER_TBL_ADDR + PPE_QUEUE_TX_COUNTER_TBL_INC * i;
+ ppe_tbl_pkt_cnt_clear(ppe_dev, reg, PPE_PKT_CNT_SIZE_3WORD);
+ }
+
+ break;
+ default:
+ break;
+ }
+
+ return count;
+}
+DEFINE_SHOW_STORE_ATTRIBUTE(ppe_packet_counter);
+
+void ppe_debugfs_setup(struct ppe_device *ppe_dev)
+{
+ struct ppe_debugfs_entry *entry;
+ int i;
+
+ ppe_dev->debugfs_root = debugfs_create_dir("ppe", NULL);
+ if (IS_ERR(ppe_dev->debugfs_root))
+ return;
+
+ for (i = 0; i < ARRAY_SIZE(debugfs_files); i++) {
+ entry = devm_kzalloc(ppe_dev->dev, sizeof(*entry), GFP_KERNEL);
+ if (!entry)
+ return;
+
+ entry->ppe = ppe_dev;
+ entry->counter_type = debugfs_files[i].counter_type;
+
+ debugfs_create_file(debugfs_files[i].name, 0444,
+ ppe_dev->debugfs_root, entry,
+ &ppe_packet_counter_fops);
+ }
+}
+
+void ppe_debugfs_teardown(struct ppe_device *ppe_dev)
+{
+ debugfs_remove_recursive(ppe_dev->debugfs_root);
+ ppe_dev->debugfs_root = NULL;
+}
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_debugfs.h b/drivers/net/ethernet/qualcomm/ppe/ppe_debugfs.h
new file mode 100644
index 000000000000..81f49a709123
--- /dev/null
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_debugfs.h
@@ -0,0 +1,16 @@
+/* SPDX-License-Identifier: GPL-2.0-only
+ *
+ * Copyright (c) Qualcomm Technologies, Inc. and/or its subsidiaries.
+ */
+
+/* PPE debugfs counters setup. */
+
+#ifndef __PPE_DEBUGFS_H__
+#define __PPE_DEBUGFS_H__
+
+#include "ppe.h"
+
+void ppe_debugfs_setup(struct ppe_device *ppe_dev);
+void ppe_debugfs_teardown(struct ppe_device *ppe_dev);
+
+#endif
diff --git a/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h b/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
index 6fc63f82ee80..746dfbb5a682 100644
--- a/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
+++ b/drivers/net/ethernet/qualcomm/ppe/ppe_regs.h
@@ -16,6 +16,36 @@
#define PPE_BM_SCH_CTRL_SCH_OFFSET GENMASK(14, 8)
#define PPE_BM_SCH_CTRL_SCH_EN BIT(31)
+/* PPE drop counters. */
+#define PPE_DROP_CNT_TBL_ADDR 0xb024
+#define PPE_DROP_CNT_TBL_ENTRIES 8
+#define PPE_DROP_CNT_TBL_INC 4
+
+/* BM port drop counters. */
+#define PPE_DROP_STAT_TBL_ADDR 0xe000
+#define PPE_DROP_STAT_TBL_ENTRIES 30
+#define PPE_DROP_STAT_TBL_INC 0x10
+
+/* Egress VLAN counters. */
+#define PPE_EG_VSI_COUNTER_TBL_ADDR 0x41000
+#define PPE_EG_VSI_COUNTER_TBL_ENTRIES 64
+#define PPE_EG_VSI_COUNTER_TBL_INC 0x10
+
+/* Port TX counters. */
+#define PPE_PORT_TX_COUNTER_TBL_ADDR 0x45000
+#define PPE_PORT_TX_COUNTER_TBL_ENTRIES 8
+#define PPE_PORT_TX_COUNTER_TBL_INC 0x10
+
+/* Virtual port TX counters. */
+#define PPE_VPORT_TX_COUNTER_TBL_ADDR 0x47000
+#define PPE_VPORT_TX_COUNTER_TBL_ENTRIES 256
+#define PPE_VPORT_TX_COUNTER_TBL_INC 0x10
+
+/* Queue counters. */
+#define PPE_QUEUE_TX_COUNTER_TBL_ADDR 0x4a000
+#define PPE_QUEUE_TX_COUNTER_TBL_ENTRIES 300
+#define PPE_QUEUE_TX_COUNTER_TBL_INC 0x10
+
/* RSS settings are to calculate the random RSS hash value generated during
* packet receive to ARM cores. This hash is then used to generate the queue
* offset used to determine the queue used to transmit the packet to ARM cores.
@@ -213,6 +243,51 @@
#define PPE_L2_PORT_SET_DST_INFO(tbl_cfg, value) \
FIELD_MODIFY(PPE_L2_VP_PORT_W0_DST_INFO, tbl_cfg, value)
+/* Port RX and RX drop counters. */
+#define PPE_PORT_RX_CNT_TBL_ADDR 0x150000
+#define PPE_PORT_RX_CNT_TBL_ENTRIES 256
+#define PPE_PORT_RX_CNT_TBL_INC 0x20
+
+/* Physical port RX and RX drop counters. */
+#define PPE_PHY_PORT_RX_CNT_TBL_ADDR 0x156000
+#define PPE_PHY_PORT_RX_CNT_TBL_ENTRIES 8
+#define PPE_PHY_PORT_RX_CNT_TBL_INC 0x20
+
+/* Counters for the packet to CPU port. */
+#define PPE_DROP_CPU_CNT_TBL_ADDR 0x160000
+#define PPE_DROP_CPU_CNT_TBL_ENTRIES 1280
+#define PPE_DROP_CPU_CNT_TBL_INC 0x10
+
+/* VLAN counters. */
+#define PPE_VLAN_CNT_TBL_ADDR 0x178000
+#define PPE_VLAN_CNT_TBL_ENTRIES 64
+#define PPE_VLAN_CNT_TBL_INC 0x10
+
+/* PPE L2 counters. */
+#define PPE_PRE_L2_CNT_TBL_ADDR 0x17c000
+#define PPE_PRE_L2_CNT_TBL_ENTRIES 64
+#define PPE_PRE_L2_CNT_TBL_INC 0x20
+
+/* Port TX drop counters. */
+#define PPE_PORT_TX_DROP_CNT_TBL_ADDR 0x17d000
+#define PPE_PORT_TX_DROP_CNT_TBL_ENTRIES 8
+#define PPE_PORT_TX_DROP_CNT_TBL_INC 0x10
+
+/* Virtual port TX counters. */
+#define PPE_VPORT_TX_DROP_CNT_TBL_ADDR 0x17e000
+#define PPE_VPORT_TX_DROP_CNT_TBL_ENTRIES 256
+#define PPE_VPORT_TX_DROP_CNT_TBL_INC 0x10
+
+/* Counters for the tunnel packet. */
+#define PPE_TPR_PKT_CNT_TBL_ADDR 0x1d0080
+#define PPE_TPR_PKT_CNT_TBL_ENTRIES 8
+#define PPE_TPR_PKT_CNT_TBL_INC 4
+
+/* Counters for the all packet received. */
+#define PPE_IPR_PKT_CNT_TBL_ADDR 0x1e0080
+#define PPE_IPR_PKT_CNT_TBL_ENTRIES 8
+#define PPE_IPR_PKT_CNT_TBL_INC 4
+
/* PPE service code configuration for the tunnel packet. */
#define PPE_TL_SERVICE_TBL_ADDR 0x306000
#define PPE_TL_SERVICE_TBL_ENTRIES 256
@@ -325,6 +400,18 @@
#define PPE_BM_PORT_GROUP_ID_INC 0x4
#define PPE_BM_PORT_GROUP_ID_SHARED_GROUP_ID GENMASK(1, 0)
+/* Counters for PPE buffers used for packets cached. */
+#define PPE_BM_USED_CNT_TBL_ADDR 0x6001c0
+#define PPE_BM_USED_CNT_TBL_ENTRIES 15
+#define PPE_BM_USED_CNT_TBL_INC 0x4
+#define PPE_BM_USED_CNT_VAL GENMASK(10, 0)
+
+/* Counters for PPE buffers used for packets received after pause frame sent. */
+#define PPE_BM_REACT_CNT_TBL_ADDR 0x600240
+#define PPE_BM_REACT_CNT_TBL_ENTRIES 15
+#define PPE_BM_REACT_CNT_TBL_INC 0x4
+#define PPE_BM_REACT_CNT_VAL GENMASK(8, 0)
+
#define PPE_BM_SHARED_GROUP_CFG_ADDR 0x600290
#define PPE_BM_SHARED_GROUP_CFG_ENTRIES 4
#define PPE_BM_SHARED_GROUP_CFG_INC 0x4
@@ -449,9 +536,56 @@
#define PPE_AC_GRP_SET_BUF_LIMIT(tbl_cfg, value) \
FIELD_MODIFY(PPE_AC_GRP_W1_BUF_LIMIT, (tbl_cfg) + 0x1, value)
+/* Counters for packets handled by unicast queues (0-255). */
+#define PPE_AC_UNICAST_QUEUE_CNT_TBL_ADDR 0x84e000
+#define PPE_AC_UNICAST_QUEUE_CNT_TBL_ENTRIES 256
+#define PPE_AC_UNICAST_QUEUE_CNT_TBL_INC 0x10
+#define PPE_AC_UNICAST_QUEUE_CNT_TBL_PEND_CNT GENMASK(12, 0)
+
+/* Counters for packets handled by multicast queues (256-299). */
+#define PPE_AC_MULTICAST_QUEUE_CNT_TBL_ADDR 0x852000
+#define PPE_AC_MULTICAST_QUEUE_CNT_TBL_ENTRIES 44
+#define PPE_AC_MULTICAST_QUEUE_CNT_TBL_INC 0x10
+#define PPE_AC_MULTICAST_QUEUE_CNT_TBL_PEND_CNT GENMASK(12, 0)
+
/* Table addresses for per-queue enqueue setting. */
#define PPE_ENQ_OPR_TBL_ADDR 0x85c000
#define PPE_ENQ_OPR_TBL_ENTRIES 300
#define PPE_ENQ_OPR_TBL_INC 0x10
#define PPE_ENQ_OPR_TBL_ENQ_DISABLE BIT(0)
+
+/* Unicast drop count includes the possible drops with WRED for the green,
+ * yellow and red categories.
+ */
+#define PPE_UNICAST_DROP_CNT_TBL_ADDR 0x9e0000
+#define PPE_UNICAST_DROP_CNT_TBL_ENTRIES 1536
+#define PPE_UNICAST_DROP_CNT_TBL_INC 0x10
+#define PPE_UNICAST_DROP_TYPES 6
+#define PPE_UNICAST_DROP_FORCE_OFFSET 3
+
+/* There are 16 multicast queues dedicated to CPU port 0. Multicast drop
+ * count includes the force drop for green, yellow and red category packets.
+ */
+#define PPE_P0_MULTICAST_DROP_CNT_TBL_ADDR 0x9f0000
+#define PPE_P0_MULTICAST_DROP_CNT_TBL_ENTRIES 48
+#define PPE_P0_MULTICAST_DROP_CNT_TBL_INC 0x10
+#define PPE_P0_MULTICAST_QUEUE_NUM 16
+
+/* Each PPE physical port has four dedicated multicast queues, providing
+ * a total of 12 entries per port. The multicast drop count includes forced
+ * drops for green, yellow, and red category packets.
+ */
+#define PPE_MULTICAST_QUEUE_PORT_ADDR_INC 0x1000
+#define PPE_MULTICAST_DROP_CNT_TBL_INC 0x10
+#define PPE_MULTICAST_DROP_TYPES 3
+#define PPE_MULTICAST_QUEUE_NUM 4
+#define PPE_MULTICAST_DROP_CNT_TBL_ENTRIES 12
+
+#define PPE_CPU_PORT_MULTICAST_FORCE_DROP_CNT_TBL_ADDR(mq_offset) \
+ (PPE_P0_MULTICAST_DROP_CNT_TBL_ADDR + \
+ (mq_offset) * PPE_P0_MULTICAST_DROP_CNT_TBL_INC * \
+ PPE_MULTICAST_DROP_TYPES)
+
+#define PPE_P1_MULTICAST_DROP_CNT_TBL_ADDR \
+ (PPE_P0_MULTICAST_DROP_CNT_TBL_ADDR + PPE_MULTICAST_QUEUE_PORT_ADDR_INC)
#endif
--
2.34.1
^ permalink raw reply related [flat|nested] 17+ messages in thread
* [PATCH net-next v6 14/14] MAINTAINERS: Add maintainer for Qualcomm PPE driver
2025-07-20 10:57 [PATCH net-next v6 00/14] Add PPE driver for Qualcomm IPQ9574 SoC Luo Jie
` (12 preceding siblings ...)
2025-07-20 10:57 ` [PATCH net-next v6 13/14] net: ethernet: qualcomm: Add PPE debugfs support for PPE counters Luo Jie
@ 2025-07-20 10:57 ` Luo Jie
13 siblings, 0 replies; 17+ messages in thread
From: Luo Jie @ 2025-07-20 10:57 UTC (permalink / raw)
To: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Rob Herring, Krzysztof Kozlowski, Conor Dooley,
Lei Wei, Suruchi Agarwal, Pavithra R, Simon Horman,
Jonathan Corbet, Kees Cook, Gustavo A. R. Silva, Philipp Zabel
Cc: linux-arm-msm, netdev, devicetree, linux-kernel, linux-doc,
linux-hardening, quic_kkumarcs, quic_linchen, Luo Jie
Add maintainer entry for PPE (Packet Process Engine) driver supported
for Qualcomm IPQ SoCs.
Signed-off-by: Luo Jie <quic_luoj@quicinc.com>
---
MAINTAINERS | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/MAINTAINERS b/MAINTAINERS
index 1bc1698bc5ae..1caca0fa90c5 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -20577,6 +20577,14 @@ S: Maintained
F: Documentation/devicetree/bindings/mtd/qcom,nandc.yaml
F: drivers/mtd/nand/raw/qcom_nandc.c
+QUALCOMM PPE DRIVER
+M: Luo Jie <quic_luoj@quicinc.com>
+L: netdev@vger.kernel.org
+S: Supported
+F: Documentation/devicetree/bindings/net/qcom,ipq9574-ppe.yaml
+F: Documentation/networking/device_drivers/ethernet/qualcomm/ppe/ppe.rst
+F: drivers/net/ethernet/qualcomm/ppe/
+
QUALCOMM QSEECOM DRIVER
M: Maximilian Luz <luzmaximilian@gmail.com>
L: linux-arm-msm@vger.kernel.org
--
2.34.1
^ permalink raw reply related [flat|nested] 17+ messages in thread
* Re: [PATCH net-next v6 01/14] dt-bindings: net: Add PPE for Qualcomm IPQ9574 SoC
2025-07-20 10:57 ` [PATCH net-next v6 01/14] dt-bindings: net: Add PPE " Luo Jie
@ 2025-07-21 19:35 ` Rob Herring
2025-07-23 5:00 ` Luo Jie
0 siblings, 1 reply; 17+ messages in thread
From: Rob Herring @ 2025-07-21 19:35 UTC (permalink / raw)
To: Luo Jie
Cc: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Krzysztof Kozlowski, Conor Dooley, Lei Wei,
Suruchi Agarwal, Pavithra R, Simon Horman, Jonathan Corbet,
Kees Cook, Gustavo A. R. Silva, Philipp Zabel, linux-arm-msm,
netdev, devicetree, linux-kernel, linux-doc, linux-hardening,
quic_kkumarcs, quic_linchen
On Sun, Jul 20, 2025 at 06:57:08PM +0800, Luo Jie wrote:
> The PPE (packet process engine) hardware block is available in Qualcomm
> IPQ chipsets that support PPE architecture, such as IPQ9574. The PPE in
> the IPQ9574 SoC includes six ethernet ports (6 GMAC and 6 XGMAC), which
> are used to connect with external PHY devices by PCS. It includes an L2
> switch function for bridging packets among the 6 ethernet ports and the
> CPU port. The CPU port enables packet transfer between the ethernet ports
> and the ARM cores in the SoC, using the ethernet DMA.
>
> The PPE also includes packet processing offload capabilities for various
> networking functions such as route and bridge flows, VLANs, different
> tunnel protocols and VPN.
>
> The PPE switch is modeled according to the ethernet switch schema, with
> additional properties defined for the switch node for interrupts, clocks,
> resets, interconnects and Ethernet DMA. The switch port node is extended
> with additional properties for clocks and resets.
>
> Signed-off-by: Luo Jie <quic_luoj@quicinc.com>
> ---
> .../devicetree/bindings/net/qcom,ipq9574-ppe.yaml | 529 +++++++++++++++++++++
> 1 file changed, 529 insertions(+)
>
> diff --git a/Documentation/devicetree/bindings/net/qcom,ipq9574-ppe.yaml b/Documentation/devicetree/bindings/net/qcom,ipq9574-ppe.yaml
> new file mode 100644
> index 000000000000..d48169a8ba7c
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/net/qcom,ipq9574-ppe.yaml
> @@ -0,0 +1,529 @@
> +# SPDX-License-Identifier: GPL-2.0 OR BSD-2-Clause
> +%YAML 1.2
> +---
> +$id: http://devicetree.org/schemas/net/qcom,ipq9574-ppe.yaml#
> +$schema: http://devicetree.org/meta-schemas/core.yaml#
> +
> +title: Qualcomm IPQ packet process engine (PPE)
> +
> +maintainers:
> + - Luo Jie <quic_luoj@quicinc.com>
> + - Lei Wei <quic_leiwei@quicinc.com>
> + - Suruchi Agarwal <quic_suruchia@quicinc.com>
> + - Pavithra R <quic_pavir@quicinc.com>
> +
> +description: |
> + The Ethernet functionality in the PPE (Packet Process Engine) is comprised
> + of three components, the switch core, port wrapper and Ethernet DMA.
> +
> + The Switch core in the IPQ9574 PPE has maximum of 6 front panel ports and
> + two FIFO interfaces. One of the two FIFO interfaces is used for Ethernet
> + port to host CPU communication using Ethernet DMA. The other is used
> + communicating to the EIP engine which is used for IPsec offload. On the
> + IPQ9574, the PPE includes 6 GMAC/XGMACs that can be connected with external
> + Ethernet PHY. Switch core also includes BM (Buffer Management), QM (Queue
> + Management) and SCH (Scheduler) modules for supporting the packet processing.
> +
> + The port wrapper provides connections from the 6 GMAC/XGMACS to UNIPHY (PCS)
> + supporting various modes such as SGMII/QSGMII/PSGMII/USXGMII/10G-BASER. There
> + are 3 UNIPHY (PCS) instances supported on the IPQ9574.
> +
> + Ethernet DMA is used to transmit and receive packets between the six Ethernet
> + ports and ARM host CPU.
> +
> + The follow diagram shows the PPE hardware block along with its connectivity
> + to the external hardware blocks such clock hardware blocks (CMNPLL, GCC,
> + NSS clock controller) and ethernet PCS/PHY blocks. For depicting the PHY
> + connectivity, one 4x1 Gbps PHY (QCA8075) and two 10 GBps PHYs are used as an
> + example.
> +
> + +---------+
> + | 48 MHZ |
> + +----+----+
> + |(clock)
> + v
> + +----+----+
> + +------| CMN PLL |
> + | +----+----+
> + | |(clock)
> + | v
> + | +----+----+ +----+----+ (clock) +----+----+
> + | +---| NSSCC | | GCC |--------->| MDIO |
> + | | +----+----+ +----+----+ +----+----+
> + | | |(clock & reset) |(clock)
> + | | v v
> + | | +----+---------------------+--+----------+----------+---------+
> + | | | +-----+ |EDMA FIFO | | EIP FIFO|
> + | | | | SCH | +----------+ +---------+
> + | | | +-----+ | | |
> + | | | +------+ +------+ +-------------------+ |
> + | | | | BM | | QM | IPQ9574-PPE | L2/L3 Process | |
> + | | | +------+ +------+ +-------------------+ |
> + | | | | |
> + | | | +-------+ +-------+ +-------+ +-------+ +-------+ +-------+ |
> + | | | | MAC0 | | MAC1 | | MAC2 | | MAC3 | | XGMAC4| |XGMAC5 | |
> + | | | +---+---+ +---+---+ +---+---+ +---+---+ +---+---+ +---+---+ |
> + | | | | | | | | | |
> + | | +-----+---------+---------+---------+---------+---------+-----+
> + | | | | | | | |
> + | | +---+---------+---------+---------+---+ +---+---+ +---+---+
> + +--+---->| PCS0 | | PCS1 | | PCS2 |
> + |(clock) +---+---------+---------+---------+---+ +---+---+ +---+---+
> + | | | | | | |
> + | +---+---------+---------+---------+---+ +---+---+ +---+---+
> + +------->| QCA8075 PHY | | PHY4 | | PHY5 |
> + (clock) +-------------------------------------+ +-------+ +-------+
> +
> +properties:
> + compatible:
> + enum:
> + - qcom,ipq9574-ppe
> +
> + reg:
> + maxItems: 1
> +
> + clocks:
> + items:
> + - description: PPE core clock from NSS clock controller
> + - description: PPE APB (Advanced Peripheral Bus) clock from NSS clock controller
> + - description: PPE ingress process engine clock from NSS clock controller
> + - description: PPE BM, QM and scheduler clock from NSS clock controller
Drop 'from NSS clock controller'. That's outside the scope of this
binding.
> +
> + clock-names:
> + items:
> + - const: ppe
> + - const: apb
> + - const: ipe
> + - const: btq
> +
> + resets:
> + maxItems: 1
> + description: PPE reset, which is necessary before configuring PPE hardware
> +
> + interrupts:
> + maxItems: 1
> + description: PPE switch miscellaneous interrupt
> +
> + interconnects:
> + items:
> + - description: Clock path leading to PPE switch core function
> + - description: Clock path leading to PPE register access
> + - description: Clock path leading to QoS generation
> + - description: Clock path leading to timeout reference
> + - description: Clock path leading to NSS NOC from memory NOC
> + - description: Clock path leading to memory NOC from NSS NOC
> + - description: Clock path leading to enhanced memory NOC from NSS NOC
Clock path? This should be bus interconnect paths.
> +
> + interconnect-names:
> + items:
> + - const: ppe
> + - const: ppe_cfg
> + - const: qos_gen
> + - const: timeout_ref
> + - const: nssnoc_memnoc
> + - const: memnoc_nssnoc
> + - const: memnoc_nssnoc_1
> +
> + ethernet-dma:
> + type: object
> + additionalProperties: false
> + description:
> + EDMA (Ethernet DMA) is used to transmit packets between PPE and ARM
> + host CPU. There are 32 TX descriptor rings, 32 TX completion rings,
> + 24 RX descriptor rings and 8 RX fill rings supported.
> +
> + properties:
> + clocks:
> + items:
> + - description: EDMA system clock from NSS Clock Controller
> + - description: EDMA APB (Advanced Peripheral Bus) clock from
> + NSS Clock Controller
> +
> + clock-names:
> + items:
> + - const: sys
> + - const: apb
> +
> + resets:
> + maxItems: 1
> + description: EDMA reset from NSS clock controller
> +
> + interrupts:
> + minItems: 65
> + maxItems: 65
> +
> + interrupt-names:
> + minItems: 65
> + maxItems: 65
> + description:
> + Interrupts "txcmpl_[0-31]" are the Ethernet DMA TX completion ring interrupts.
> + Interrupts "rxfill_[0-7]" are the Ethernet DMA RX fill ring interrupts.
> + Interrupts "rxdesc_[0-23]" are the Ethernet DMA RX Descriptor ring interrupts.
> + Interrupt "misc" is the Ethernet DMA miscellaneous error interrupt.
items:
oneOf:
- pattern: '^txcmpl_([1-2]?[0-9]|3[01])$'
- pattern: '^rxfill_[0-7]$'
- pattern: '^rxdesc_(1?[0-9]|2[0-3])$'
- const: misc
> +
> + required:
> + - clocks
> + - clock-names
> + - resets
> + - interrupts
> + - interrupt-names
> +
> +patternProperties:
> + "^(ethernet-)?ports$":
New binding, does 'ethernet-' part need to be optional? No.
> + patternProperties:
> + "^ethernet-port@[1-6]+$":
> + type: object
> + unevaluatedProperties: false
> + $ref: ethernet-switch-port.yaml#
> +
> + properties:
> + reg:
> + minimum: 1
> + maximum: 6
> + description: PPE Ethernet port ID
> +
> + clocks:
> + items:
> + - description: Port MAC clock from NSS clock controller
> + - description: Port RX clock from NSS clock controller
> + - description: Port TX clock from NSS clock controller
> +
> + clock-names:
> + items:
> + - const: mac
> + - const: rx
> + - const: tx
> +
> + resets:
> + items:
> + - description: Port MAC reset from NSS clock controller
> + - description: Port RX reset from NSS clock controller
> + - description: Port TX reset from NSS clock controller
> +
> + reset-names:
> + items:
> + - const: mac
> + - const: rx
> + - const: tx
> +
> + required:
> + - reg
> + - clocks
> + - clock-names
> + - resets
> + - reset-names
> +
> +required:
> + - compatible
> + - reg
> + - clocks
> + - clock-names
> + - resets
> + - interconnects
> + - interconnect-names
> + - ethernet-dma
> +
> +allOf:
> + - $ref: ethernet-switch.yaml
> +
> +unevaluatedProperties: false
^ permalink raw reply [flat|nested] 17+ messages in thread
* Re: [PATCH net-next v6 01/14] dt-bindings: net: Add PPE for Qualcomm IPQ9574 SoC
2025-07-21 19:35 ` Rob Herring
@ 2025-07-23 5:00 ` Luo Jie
0 siblings, 0 replies; 17+ messages in thread
From: Luo Jie @ 2025-07-23 5:00 UTC (permalink / raw)
To: Rob Herring
Cc: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski,
Paolo Abeni, Krzysztof Kozlowski, Conor Dooley, Lei Wei,
Suruchi Agarwal, Pavithra R, Simon Horman, Jonathan Corbet,
Kees Cook, Gustavo A. R. Silva, Philipp Zabel, linux-arm-msm,
netdev, devicetree, linux-kernel, linux-doc, linux-hardening,
quic_kkumarcs, quic_linchen
On 7/22/2025 3:35 AM, Rob Herring wrote:
> On Sun, Jul 20, 2025 at 06:57:08PM +0800, Luo Jie wrote:
>> The PPE (packet process engine) hardware block is available in Qualcomm
>> IPQ chipsets that support PPE architecture, such as IPQ9574. The PPE in
>> the IPQ9574 SoC includes six ethernet ports (6 GMAC and 6 XGMAC), which
>> are used to connect with external PHY devices by PCS. It includes an L2
>> switch function for bridging packets among the 6 ethernet ports and the
>> CPU port. The CPU port enables packet transfer between the ethernet ports
>> and the ARM cores in the SoC, using the ethernet DMA.
>>
>> The PPE also includes packet processing offload capabilities for various
>> networking functions such as route and bridge flows, VLANs, different
>> tunnel protocols and VPN.
>>
>> The PPE switch is modeled according to the ethernet switch schema, with
>> additional properties defined for the switch node for interrupts, clocks,
>> resets, interconnects and Ethernet DMA. The switch port node is extended
>> with additional properties for clocks and resets.
>>
>> Signed-off-by: Luo Jie <quic_luoj@quicinc.com>
>> ---
>> .../devicetree/bindings/net/qcom,ipq9574-ppe.yaml | 529 +++++++++++++++++++++
>> 1 file changed, 529 insertions(+)
>>
>> diff --git a/Documentation/devicetree/bindings/net/qcom,ipq9574-ppe.yaml b/Documentation/devicetree/bindings/net/qcom,ipq9574-ppe.yaml
>> new file mode 100644
>> index 000000000000..d48169a8ba7c
>> --- /dev/null
>> +++ b/Documentation/devicetree/bindings/net/qcom,ipq9574-ppe.yaml
>> @@ -0,0 +1,529 @@
>> +# SPDX-License-Identifier: GPL-2.0 OR BSD-2-Clause
>> +%YAML 1.2
>> +---
>> +$id: http://devicetree.org/schemas/net/qcom,ipq9574-ppe.yaml#
>> +$schema: http://devicetree.org/meta-schemas/core.yaml#
>> +
>> +title: Qualcomm IPQ packet process engine (PPE)
>> +
>> +maintainers:
>> + - Luo Jie <quic_luoj@quicinc.com>
>> + - Lei Wei <quic_leiwei@quicinc.com>
>> + - Suruchi Agarwal <quic_suruchia@quicinc.com>
>> + - Pavithra R <quic_pavir@quicinc.com>
>> +
>> +description: |
>> + The Ethernet functionality in the PPE (Packet Process Engine) is comprised
>> + of three components, the switch core, port wrapper and Ethernet DMA.
>> +
>> + The Switch core in the IPQ9574 PPE has maximum of 6 front panel ports and
>> + two FIFO interfaces. One of the two FIFO interfaces is used for Ethernet
>> + port to host CPU communication using Ethernet DMA. The other is used
>> + communicating to the EIP engine which is used for IPsec offload. On the
>> + IPQ9574, the PPE includes 6 GMAC/XGMACs that can be connected with external
>> + Ethernet PHY. Switch core also includes BM (Buffer Management), QM (Queue
>> + Management) and SCH (Scheduler) modules for supporting the packet processing.
>> +
>> + The port wrapper provides connections from the 6 GMAC/XGMACS to UNIPHY (PCS)
>> + supporting various modes such as SGMII/QSGMII/PSGMII/USXGMII/10G-BASER. There
>> + are 3 UNIPHY (PCS) instances supported on the IPQ9574.
>> +
>> + Ethernet DMA is used to transmit and receive packets between the six Ethernet
>> + ports and ARM host CPU.
>> +
>> + The follow diagram shows the PPE hardware block along with its connectivity
>> + to the external hardware blocks such clock hardware blocks (CMNPLL, GCC,
>> + NSS clock controller) and ethernet PCS/PHY blocks. For depicting the PHY
>> + connectivity, one 4x1 Gbps PHY (QCA8075) and two 10 GBps PHYs are used as an
>> + example.
>> +
>> + +---------+
>> + | 48 MHZ |
>> + +----+----+
>> + |(clock)
>> + v
>> + +----+----+
>> + +------| CMN PLL |
>> + | +----+----+
>> + | |(clock)
>> + | v
>> + | +----+----+ +----+----+ (clock) +----+----+
>> + | +---| NSSCC | | GCC |--------->| MDIO |
>> + | | +----+----+ +----+----+ +----+----+
>> + | | |(clock & reset) |(clock)
>> + | | v v
>> + | | +----+---------------------+--+----------+----------+---------+
>> + | | | +-----+ |EDMA FIFO | | EIP FIFO|
>> + | | | | SCH | +----------+ +---------+
>> + | | | +-----+ | | |
>> + | | | +------+ +------+ +-------------------+ |
>> + | | | | BM | | QM | IPQ9574-PPE | L2/L3 Process | |
>> + | | | +------+ +------+ +-------------------+ |
>> + | | | | |
>> + | | | +-------+ +-------+ +-------+ +-------+ +-------+ +-------+ |
>> + | | | | MAC0 | | MAC1 | | MAC2 | | MAC3 | | XGMAC4| |XGMAC5 | |
>> + | | | +---+---+ +---+---+ +---+---+ +---+---+ +---+---+ +---+---+ |
>> + | | | | | | | | | |
>> + | | +-----+---------+---------+---------+---------+---------+-----+
>> + | | | | | | | |
>> + | | +---+---------+---------+---------+---+ +---+---+ +---+---+
>> + +--+---->| PCS0 | | PCS1 | | PCS2 |
>> + |(clock) +---+---------+---------+---------+---+ +---+---+ +---+---+
>> + | | | | | | |
>> + | +---+---------+---------+---------+---+ +---+---+ +---+---+
>> + +------->| QCA8075 PHY | | PHY4 | | PHY5 |
>> + (clock) +-------------------------------------+ +-------+ +-------+
>> +
>> +properties:
>> + compatible:
>> + enum:
>> + - qcom,ipq9574-ppe
>> +
>> + reg:
>> + maxItems: 1
>> +
>> + clocks:
>> + items:
>> + - description: PPE core clock from NSS clock controller
>> + - description: PPE APB (Advanced Peripheral Bus) clock from NSS clock controller
>> + - description: PPE ingress process engine clock from NSS clock controller
>> + - description: PPE BM, QM and scheduler clock from NSS clock controller
>
> Drop 'from NSS clock controller'. That's outside the scope of this
> binding.
Ok, will update.
>
>> +
>> + clock-names:
>> + items:
>> + - const: ppe
>> + - const: apb
>> + - const: ipe
>> + - const: btq
>> +
>> + resets:
>> + maxItems: 1
>> + description: PPE reset, which is necessary before configuring PPE hardware
>> +
>> + interrupts:
>> + maxItems: 1
>> + description: PPE switch miscellaneous interrupt
>> +
>> + interconnects:
>> + items:
>> + - description: Clock path leading to PPE switch core function
>> + - description: Clock path leading to PPE register access
>> + - description: Clock path leading to QoS generation
>> + - description: Clock path leading to timeout reference
>> + - description: Clock path leading to NSS NOC from memory NOC
>> + - description: Clock path leading to memory NOC from NSS NOC
>> + - description: Clock path leading to enhanced memory NOC from NSS NOC
>
> Clock path? This should be bus interconnect paths.
Sure, I will use "bus interconnect path" for clarity.
>
>> +
>> + interconnect-names:
>> + items:
>> + - const: ppe
>> + - const: ppe_cfg
>> + - const: qos_gen
>> + - const: timeout_ref
>> + - const: nssnoc_memnoc
>> + - const: memnoc_nssnoc
>> + - const: memnoc_nssnoc_1
>> +
>> + ethernet-dma:
>> + type: object
>> + additionalProperties: false
>> + description:
>> + EDMA (Ethernet DMA) is used to transmit packets between PPE and ARM
>> + host CPU. There are 32 TX descriptor rings, 32 TX completion rings,
>> + 24 RX descriptor rings and 8 RX fill rings supported.
>> +
>> + properties:
>> + clocks:
>> + items:
>> + - description: EDMA system clock from NSS Clock Controller
>> + - description: EDMA APB (Advanced Peripheral Bus) clock from
>> + NSS Clock Controller
>> +
>> + clock-names:
>> + items:
>> + - const: sys
>> + - const: apb
>> +
>> + resets:
>> + maxItems: 1
>> + description: EDMA reset from NSS clock controller
>> +
>> + interrupts:
>> + minItems: 65
>> + maxItems: 65
>> +
>> + interrupt-names:
>> + minItems: 65
>> + maxItems: 65
>> + description:
>> + Interrupts "txcmpl_[0-31]" are the Ethernet DMA TX completion ring interrupts.
>> + Interrupts "rxfill_[0-7]" are the Ethernet DMA RX fill ring interrupts.
>> + Interrupts "rxdesc_[0-23]" are the Ethernet DMA RX Descriptor ring interrupts.
>> + Interrupt "misc" is the Ethernet DMA miscellaneous error interrupt.
>
> items:
> oneOf:
> - pattern: '^txcmpl_([1-2]?[0-9]|3[01])$'
> - pattern: '^rxfill_[0-7]$'
> - pattern: '^rxdesc_(1?[0-9]|2[0-3])$'
> - const: misc
>
Thanks for the suggestion, I will add the constraint.
>> +
>> + required:
>> + - clocks
>> + - clock-names
>> + - resets
>> + - interrupts
>> + - interrupt-names
>> +
>> +patternProperties:
>> + "^(ethernet-)?ports$":
>
> New binding, does 'ethernet-' part need to be optional? No.
OK, understand. Will update to make 'ethernet-' mandatory.
>
>> + patternProperties:
>> + "^ethernet-port@[1-6]+$":
>> + type: object
>> + unevaluatedProperties: false
>> + $ref: ethernet-switch-port.yaml#
>> +
>> + properties:
>> + reg:
>> + minimum: 1
>> + maximum: 6
>> + description: PPE Ethernet port ID
>> +
>> + clocks:
>> + items:
>> + - description: Port MAC clock from NSS clock controller
>> + - description: Port RX clock from NSS clock controller
>> + - description: Port TX clock from NSS clock controller
>> +
>> + clock-names:
>> + items:
>> + - const: mac
>> + - const: rx
>> + - const: tx
>> +
>> + resets:
>> + items:
>> + - description: Port MAC reset from NSS clock controller
>> + - description: Port RX reset from NSS clock controller
>> + - description: Port TX reset from NSS clock controller
>> +
>> + reset-names:
>> + items:
>> + - const: mac
>> + - const: rx
>> + - const: tx
>> +
>> + required:
>> + - reg
>> + - clocks
>> + - clock-names
>> + - resets
>> + - reset-names
>> +
>> +required:
>> + - compatible
>> + - reg
>> + - clocks
>> + - clock-names
>> + - resets
>> + - interconnects
>> + - interconnect-names
>> + - ethernet-dma
>> +
>> +allOf:
>> + - $ref: ethernet-switch.yaml
>> +
>> +unevaluatedProperties: false
^ permalink raw reply [flat|nested] 17+ messages in thread
end of thread, other threads:[~2025-07-23 5:01 UTC | newest]
Thread overview: 17+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2025-07-20 10:57 [PATCH net-next v6 00/14] Add PPE driver for Qualcomm IPQ9574 SoC Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 01/14] dt-bindings: net: Add PPE " Luo Jie
2025-07-21 19:35 ` Rob Herring
2025-07-23 5:00 ` Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 02/14] docs: networking: Add PPE driver documentation " Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 03/14] net: ethernet: qualcomm: Add PPE driver for " Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 04/14] net: ethernet: qualcomm: Initialize PPE buffer management for IPQ9574 Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 05/14] net: ethernet: qualcomm: Initialize PPE queue " Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 06/14] net: ethernet: qualcomm: Initialize the PPE scheduler settings Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 07/14] net: ethernet: qualcomm: Initialize PPE queue settings Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 08/14] net: ethernet: qualcomm: Initialize PPE service code settings Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 09/14] net: ethernet: qualcomm: Initialize PPE port control settings Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 10/14] net: ethernet: qualcomm: Initialize PPE RSS hash settings Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 11/14] net: ethernet: qualcomm: Initialize PPE queue to Ethernet DMA ring mapping Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 12/14] net: ethernet: qualcomm: Initialize PPE L2 bridge settings Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 13/14] net: ethernet: qualcomm: Add PPE debugfs support for PPE counters Luo Jie
2025-07-20 10:57 ` [PATCH net-next v6 14/14] MAINTAINERS: Add maintainer for Qualcomm PPE driver Luo Jie
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).