* [PATCH v9 0/3] i2c: ma35d1: Add support for MA35D1 I2C controller
@ 2026-08-07 6:03 Zi-Yu Chen
2026-08-07 6:03 ` [PATCH v9 1/3] dt-bindings: i2c: nuvoton,ma35d1-i2c: Add " Zi-Yu Chen
` (2 more replies)
0 siblings, 3 replies; 6+ messages in thread
From: Zi-Yu Chen @ 2026-08-07 6:03 UTC (permalink / raw)
To: Andi Shyti, linux-i2c, devicetree, linux-arm-kernel, linux-kernel
Cc: Jacky Huang, Shan-Chun Hung, Rob Herring, Krzysztof Kozlowski,
Conor Dooley, Philipp Zabel, Andrew Jeffery, Zi-Yu Chen
This series adds support for the I2C controller found in the Nuvoton
MA35D1 SoC. The driver supports controller and optional target mode
and runtime power management.
The implementation has been tested on the Nuvoton MA35D1 SOM board.
Signed-off-by: Zi-Yu Chen <zychennvt@gmail.com>
---
Changes in v9:
- Centralize SI handling in the interrupt handler.
- Fix target mode comments.
- Initialize the target transmit byte.
- Add reset_control_deassert() during probe.
v8: https://lore.kernel.org/r/20260804084806.1846570-1-zychennvt@gmail.com/
Changes in v8:
- Remove rpm_put avoid duplicate processing.
- Add synchronize_irq wait for interrupt to finish before i2c->target = NULL.
- Add ma35d1_i2c_restore_idle() to handle bus errors or spurious interrupts.
- Change return types of TRX helper functions to irqreturn_t, returning
IRQ_NONE for unhandled status codes.
v7: https://lore.kernel.org/r/20260727081859.1737223-1-zychennvt@gmail.com/
Changes in v7:
- Disable IRQ before clearing i2c->msg.
- Support CONFIG_PM=n, and simplify runtime PM setup accordingly.
- Remove redundant register access in runtime suspend/resume callbacks.
- Defer clearing i2c->target until after pm_runtime_put_sync() in ma35d1_unreg_target().
- Enable interrupts alongside I2C enable in ma35d1_i2c_reset().
v6: https://lore.kernel.org/r/20260720030616.1067660-1-zychennvt@gmail.com/
Changes in v6:
- Patch 2 (driver):
- Removed dynamic IRQ toggling helpers for simpler interrupt management.
Based on Sashiko AI Feedback:
- Added zero-length read quirk.
- Set i2c->msg to NULL on timeout to avoid use-after-free.
- Cleared hardware registers on target unregistration to release the bus.
- Protected suspend-path register writes with a spinlock.
- Validated clock-frequency to prevent division-by-zero.
- Removed incorrect I2C_M_IGNORE_NAK handling.
- Corrected a comment typo in the receive data path.
- Patch 3 (dts):
- use a generic functional name audio-codec@1a
v5: https://lore.kernel.org/r/20260630012106.3616821-1-zychennvt@gmail.com/
Changes in v5:
- Patch 2 (driver):
- Add spinlock protection for MA35_CTL0 register access.
- Add NULL pointer check for i2c->msg in the ISR.
- Use devres-managed cleanup (devm_add_action_or_reset) for runtime PM.
- Use pm_runtime_force_suspend/resume() for Controller mode
system sleep.
- Control I2CEN bit enabling/disabling via runtime PM.
- Refactor arbitration lost handling in state machines.
- Remove unsupported I2C_FUNC_PROTOCOL_MANGLING flag.
- Clean up readl_poll_timeout() parameters, retry loop, and
unused macros.
v4: https://lore.kernel.org/r/20260610030208.2020275-1-zychennvt@gmail.com/
Changes in v4:
- Patch 1 (dt-bindings):
- Update example interrupt number.
- Patch 2 (driver):
- Remove redundant spinlocks and fix 'irq' type to signed.
- Fix target address recovery in ma35d1_i2c_reset().
- Refactor IRQ handler for multi-msg reads and NACK/IGNORE_NAK.
- Fix timeout UAF via explicit IRQ disabling during recovery.
- Drop IRQF_SHARED and fix PM paths in probe/unregistration.
- Limit suspend/resume wakeup logic to target mode only.
- Patch 3 (dts):
- Add missing i2c0 node to dtsi.
- Switch SoM configuration from i2c1 to i2c2 with updated pinctrl.
- Add nau8822 audio codec node under i2c2 on the SoM board.
- Drop redundant i2c0 alias and clean up dtsi property ordering.
v3: https://lore.kernel.org/r/20260512073953.564323-1-zychennvt@gmail.com/
Changes in v3:
- Fix minor DTS formatting issues (whitespace, missing newline)
v2: https://lore.kernel.org/r/20260316063726.41048-1-zychennvt@gmail.com
Changes in v2:
- Overall:
- Rebase on linux-i2c/i2c-next
- Switched terminology from "master/slave" to "controller/target".
- Patch 1 (dt-bindings):
- Simplified description and fixed 'reg' size in example.
- Patch 2 (driver):
- Modernized using devm_*, generic device properties, and FIELD_PREP/GENMASK.
- Optimized power management by moving clock control to runtime PM.
- Simplified code by removing redundant .remove(), .owner, and inlines.
- Added dev_err_probe() and default bus frequency handling.
- Patch 3 (dts):
- Moved i2c aliases to board dts and reordered nodes alphabetically.
v1: https://lore.kernel.org/r/20260302020822.13936-1-zychennvt@gmail.com
Zi-Yu Chen (3):
dt-bindings: i2c: nuvoton,ma35d1-i2c: Add MA35D1 I2C controller
i2c: ma35d1: Add Nuvoton MA35D1 I2C driver support
arm64: dts: nuvoton: Add I2C nodes for MA35D1 SoC
.../bindings/i2c/nuvoton,ma35d1-i2c.yaml | 63 ++
.../boot/dts/nuvoton/ma35d1-som-256m.dts | 20 +
arch/arm64/boot/dts/nuvoton/ma35d1.dtsi | 72 ++
drivers/i2c/busses/Kconfig | 13 +
drivers/i2c/busses/Makefile | 1 +
drivers/i2c/busses/i2c-ma35d1.c | 779 ++++++++++++++++++
6 files changed, 948 insertions(+)
create mode 100644 Documentation/devicetree/bindings/i2c/nuvoton,ma35d1-i2c.yaml
create mode 100644 drivers/i2c/busses/i2c-ma35d1.c
--
2.34.1
^ permalink raw reply [flat|nested] 6+ messages in thread* [PATCH v9 1/3] dt-bindings: i2c: nuvoton,ma35d1-i2c: Add MA35D1 I2C controller 2026-08-07 6:03 [PATCH v9 0/3] i2c: ma35d1: Add support for MA35D1 I2C controller Zi-Yu Chen @ 2026-08-07 6:03 ` Zi-Yu Chen 2026-08-07 6:03 ` [PATCH v9 2/3] i2c: ma35d1: Add Nuvoton MA35D1 I2C driver support Zi-Yu Chen 2026-08-07 6:04 ` [PATCH v9 3/3] arm64: dts: nuvoton: Add I2C nodes for MA35D1 SoC Zi-Yu Chen 2 siblings, 0 replies; 6+ messages in thread From: Zi-Yu Chen @ 2026-08-07 6:03 UTC (permalink / raw) To: Andi Shyti, linux-i2c, devicetree, linux-arm-kernel, linux-kernel Cc: Jacky Huang, Shan-Chun Hung, Rob Herring, Krzysztof Kozlowski, Conor Dooley, Philipp Zabel, Andrew Jeffery, Zi-Yu Chen, Krzysztof Kozlowski Add device tree binding documentation for the I2C controller found in the Nuvoton MA35D1 SoC. Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com> Signed-off-by: Zi-Yu Chen <zychennvt@gmail.com> --- .../bindings/i2c/nuvoton,ma35d1-i2c.yaml | 63 +++++++++++++++++++ 1 file changed, 63 insertions(+) create mode 100644 Documentation/devicetree/bindings/i2c/nuvoton,ma35d1-i2c.yaml diff --git a/Documentation/devicetree/bindings/i2c/nuvoton,ma35d1-i2c.yaml b/Documentation/devicetree/bindings/i2c/nuvoton,ma35d1-i2c.yaml new file mode 100644 index 000000000000..0713e1c1ecb8 --- /dev/null +++ b/Documentation/devicetree/bindings/i2c/nuvoton,ma35d1-i2c.yaml @@ -0,0 +1,63 @@ +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) +%YAML 1.2 +--- +$id: http://devicetree.org/schemas/i2c/nuvoton,ma35d1-i2c.yaml# +$schema: http://devicetree.org/meta-schemas/core.yaml# + +title: Nuvoton MA35D1 I2C Controller + +maintainers: + - Zi-Yu Chen <zychennvt@gmail.com> + +description: + The Nuvoton MA35D1 I2C controller supports controller and optional target mode. + +allOf: + - $ref: /schemas/i2c/i2c-controller.yaml# + +properties: + compatible: + const: nuvoton,ma35d1-i2c + + reg: + maxItems: 1 + + interrupts: + maxItems: 1 + + clocks: + maxItems: 1 + + clock-frequency: + description: + Desired I2C bus clock frequency in Hz. The absence of this property + indicates the default frequency 100 kHz. + + resets: + maxItems: 1 + +required: + - compatible + - reg + - interrupts + - clocks + - resets + +unevaluatedProperties: false + +examples: + - | + #include <dt-bindings/interrupt-controller/arm-gic.h> + #include <dt-bindings/clock/nuvoton,ma35d1-clk.h> + #include <dt-bindings/reset/nuvoton,ma35d1-reset.h> + + i2c0: i2c@40800000 { + compatible = "nuvoton,ma35d1-i2c"; + reg = <0x40800000 0x1000>; + interrupts = <GIC_SPI 13 IRQ_TYPE_LEVEL_HIGH>; + clocks = <&clk I2C0_GATE>; + clock-frequency = <100000>; + resets = <&sys MA35D1_RESET_I2C0>; + #address-cells = <1>; + #size-cells = <0>; + }; -- 2.34.1 ^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH v9 2/3] i2c: ma35d1: Add Nuvoton MA35D1 I2C driver support 2026-08-07 6:03 [PATCH v9 0/3] i2c: ma35d1: Add support for MA35D1 I2C controller Zi-Yu Chen 2026-08-07 6:03 ` [PATCH v9 1/3] dt-bindings: i2c: nuvoton,ma35d1-i2c: Add " Zi-Yu Chen @ 2026-08-07 6:03 ` Zi-Yu Chen 2026-08-07 6:18 ` sashiko-bot 2026-08-07 6:04 ` [PATCH v9 3/3] arm64: dts: nuvoton: Add I2C nodes for MA35D1 SoC Zi-Yu Chen 2 siblings, 1 reply; 6+ messages in thread From: Zi-Yu Chen @ 2026-08-07 6:03 UTC (permalink / raw) To: Andi Shyti, linux-i2c, devicetree, linux-arm-kernel, linux-kernel Cc: Jacky Huang, Shan-Chun Hung, Rob Herring, Krzysztof Kozlowski, Conor Dooley, Philipp Zabel, Andrew Jeffery, Zi-Yu Chen Add I2C support for Nuvoton MA35D1 SoC. The controller supports standard, fast and fast-plus modes, and provides controller/target functionality. Signed-off-by: Zi-Yu Chen <zychennvt@gmail.com> --- drivers/i2c/busses/Kconfig | 13 + drivers/i2c/busses/Makefile | 1 + drivers/i2c/busses/i2c-ma35d1.c | 779 ++++++++++++++++++++++++++++++++ 3 files changed, 793 insertions(+) create mode 100644 drivers/i2c/busses/i2c-ma35d1.c diff --git a/drivers/i2c/busses/Kconfig b/drivers/i2c/busses/Kconfig index d7b89508311f..e69120d46977 100644 --- a/drivers/i2c/busses/Kconfig +++ b/drivers/i2c/busses/Kconfig @@ -1602,4 +1602,17 @@ config I2C_VIRTIO This driver can also be built as a module. If so, the module will be called i2c-virtio. +config I2C_MA35D1 + tristate "Nuvoton MA35D1 I2C driver" + depends on ARCH_MA35 || COMPILE_TEST + select I2C_SLAVE + help + If you say yes to this option, support will be included for the + I2C controller in the Nuvoton MA35D1 SoC. This driver + supports the standard I2C bus protocols, including master and + slave modes. + + This driver can also be built as a module. If so, the module + will be called i2c-ma35d1. + endmenu diff --git a/drivers/i2c/busses/Makefile b/drivers/i2c/busses/Makefile index 3755c54b3d82..ca75dae4955c 100644 --- a/drivers/i2c/busses/Makefile +++ b/drivers/i2c/busses/Makefile @@ -131,6 +131,7 @@ obj-$(CONFIG_I2C_XILINX) += i2c-xiic.o obj-$(CONFIG_I2C_XLP9XX) += i2c-xlp9xx.o obj-$(CONFIG_I2C_RCAR) += i2c-rcar.o obj-$(CONFIG_I2C_GXP) += i2c-gxp.o +obj-$(CONFIG_I2C_MA35D1) += i2c-ma35d1.o # External I2C/SMBus adapter drivers obj-$(CONFIG_I2C_DIOLAN_U2C) += i2c-diolan-u2c.o diff --git a/drivers/i2c/busses/i2c-ma35d1.c b/drivers/i2c/busses/i2c-ma35d1.c new file mode 100644 index 000000000000..61d7605c3a05 --- /dev/null +++ b/drivers/i2c/busses/i2c-ma35d1.c @@ -0,0 +1,779 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * Copyright (c) 2026 Nuvoton technology corporation. + * + * Author: Zi-Yu Chen <zychennvt@gmail.com> + */ + +#include <linux/bitfield.h> +#include <linux/clk.h> +#include <linux/delay.h> +#include <linux/err.h> +#include <linux/errno.h> +#include <linux/i2c.h> +#include <linux/interrupt.h> +#include <linux/io.h> +#include <linux/iopoll.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/of.h> +#include <linux/platform_device.h> +#include <linux/pm_runtime.h> +#include <linux/property.h> +#include <linux/reset.h> +#include <linux/units.h> + +/* MA35D1 I2C registers offset */ +#define MA35_CTL0 0x00 +#define MA35_ADDR0 0x04 +#define MA35_DAT 0x08 +#define MA35_STATUS0 0x0c +#define MA35_CLKDIV 0x10 +#define MA35_TOCTL 0x14 +#define MA35_ADDR1 0x18 +#define MA35_ADDR2 0x1c +#define MA35_ADDR3 0x20 +#define MA35_ADDRMSK0 0x24 +#define MA35_ADDRMSK1 0x28 +#define MA35_ADDRMSK2 0x2c +#define MA35_ADDRMSK3 0x30 +#define MA35_WKCTL 0x3c +#define MA35_WKSTS 0x40 +#define MA35_CTL1 0x44 +#define MA35_STATUS1 0x48 +#define MA35_TMCTL 0x4c +#define MA35_BUSCTL 0x50 +#define MA35_BUSTCTL 0x54 +#define MA35_BUSSTS 0x58 +#define MA35_PKTSIZE 0x5c +#define MA35_PKTCRC 0x60 +#define MA35_BUSTOUT 0x64 +#define MA35_CLKTOUT 0x68 +#define MA35_AUTOCNT 0x78 + +/* MA35D1 I2C Status */ +/* Controller */ +#define MA35_M_START 0x08 /* Start */ +#define MA35_M_REPEAT_START 0x10 /* Controller Repeat Start */ +#define MA35_M_TRAN_ADDR_ACK 0x18 /* Controller Transmit Address ACK */ +#define MA35_M_TRAN_ADDR_NACK 0x20 /* Controller Transmit Address NACK */ +#define MA35_M_TRAN_DATA_ACK 0x28 /* Controller Transmit Data ACK */ +#define MA35_M_TRAN_DATA_NACK 0x30 /* Controller Transmit Data NACK */ +#define MA35_M_ARB_LOST 0x38 /* Controller Arbitration Lost */ +#define MA35_M_RECE_ADDR_ACK 0x40 /* Controller Receive Address ACK */ +#define MA35_M_RECE_ADDR_NACK 0x48 /* Controller Receive Address NACK */ +#define MA35_M_RECE_DATA_ACK 0x50 /* Controller Receive Data ACK */ +#define MA35_M_RECE_DATA_NACK 0x58 /* Controller Receive Data NACK */ +#define MA35_BUS_ERROR 0x00 /* Bus error */ + +/* Target */ +#define MA35_S_REPEAT_START_STOP 0xa0 /* Target Transmit Repeat Start or Stop */ +#define MA35_S_TRAN_ADDR_ACK 0xa8 /* Target Transmit Address ACK */ +#define MA35_S_TRAN_DATA_ACK 0xb8 /* Target Transmit Data ACK */ +#define MA35_S_TRAN_DATA_NACK 0xc0 /* Target Transmit Data NACK */ +#define MA35_S_TRAN_LAST_DATA_ACK 0xc8 /* Target Transmit Last Data ACK */ +#define MA35_S_RECE_ADDR_ACK 0x60 /* Target Receive Address ACK */ +#define MA35_S_RECE_ARB_LOST 0x68 /* Target Receive Arbitration Lost */ +#define MA35_S_RECE_DATA_ACK 0x80 /* Target Receive Data ACK */ +#define MA35_S_RECE_DATA_NACK 0x88 /* Target Receive Data NACK */ + +/* Other */ +#define MA35_ADDR_TRAN_ARB_LOST 0xb0 /* Address Transmit Arbitration Lost */ +#define MA35_BUS_RELEASED 0xf8 /* Bus Released */ + +/* I2C_CTL constant definitions. */ +#define MA35_CTL_AA BIT(2) +#define MA35_CTL_SI BIT(3) +#define MA35_CTL_STO BIT(4) +#define MA35_CTL_STA BIT(5) +#define MA35_CTL_I2CEN BIT(6) +#define MA35_CTL_INTEN BIT(7) +#define MA35_CTL_SI_AA (MA35_CTL_SI | MA35_CTL_AA) +#define MA35_CTL_STO_SI (MA35_CTL_STO | MA35_CTL_SI) +#define MA35_CTL_STA_SI (MA35_CTL_STA | MA35_CTL_SI) + +/* Constants */ +#define MA35_CLKDIV_MSK GENMASK(15, 0) +#define I2C_PM_TIMEOUT_MS 5000 +#define STOP_TIMEOUT_MS 50 + +struct ma35d1_i2c { + spinlock_t lock; + wait_queue_head_t wait; + struct i2c_msg *msg; + unsigned int msg_num; + unsigned int msg_idx; + unsigned int msg_ptr; + int err; + int irq; + void __iomem *regs; + struct clk *clk; + struct device *dev; + struct i2c_adapter adap; + struct i2c_client *target; + struct reset_control *rst; +}; + +static inline bool ma35d1_is_controller_status(unsigned int status) +{ + return status >= MA35_M_START && status <= MA35_M_RECE_DATA_NACK; +} + +/* + * ma35d1_i2c_write_ctl - Update the I2C control register + * @i2c: Pointer to the ma35d1 i2c instance + * @ctl: Control bits to set (e.g., MA35_CTL_STA, SI, AA) + * + * Read-modify-write the CTL0 register while preserving unrelated bits. + */ +static void ma35d1_i2c_write_ctl(struct ma35d1_i2c *i2c, unsigned int ctl) +{ + unsigned int val; + + guard(spinlock_irqsave)(&i2c->lock); + val = readl(i2c->regs + MA35_CTL0); + val &= ~(MA35_CTL_SI_AA); + val |= ctl; + writel(val, i2c->regs + MA35_CTL0); +} + +static void ma35d1_i2c_set_addr(struct ma35d1_i2c *i2c) +{ + unsigned int rw = i2c->msg->flags & I2C_M_RD; + + writel(((i2c->msg->addr & 0x7f) << 1) | rw, i2c->regs + MA35_DAT); +} + +static void ma35d1_i2c_controller_complete(struct ma35d1_i2c *i2c) +{ + i2c->msg_ptr = 0; + i2c->msg = NULL; + i2c->msg_idx++; + i2c->msg_num = 0; + + wake_up(&i2c->wait); +} + +static void ma35d1_i2c_reset(struct ma35d1_i2c *i2c) +{ + unsigned int clkdiv, slvaddr; + + clkdiv = readl(i2c->regs + MA35_CLKDIV); + slvaddr = readl(i2c->regs + MA35_ADDR0); + + disable_irq(i2c->irq); + i2c->msg = NULL; + reset_control_assert(i2c->rst); + usleep_range(10, 20); + reset_control_deassert(i2c->rst); + enable_irq(i2c->irq); + + writel(clkdiv, (i2c->regs + MA35_CLKDIV)); + ma35d1_i2c_write_ctl(i2c, MA35_CTL_I2CEN | MA35_CTL_INTEN); + + if (i2c->target) + writel(slvaddr, i2c->regs + MA35_ADDR0); +} + +static void ma35d1_i2c_stop(struct ma35d1_i2c *i2c, int ret) +{ + ma35d1_i2c_write_ctl(i2c, MA35_CTL_STO_SI); + + if (ret) + i2c->err = ret; + + ma35d1_i2c_controller_complete(i2c); +} + +/* + * ma35d1_i2c_restore_idle - restore controller state to idle + * @i2c: Pointer to the ma35d1 i2c instance + * + * Restore the controller to an idle state after an unexpected event. + * Abort an active controller transfer, otherwise re-arm target mode + * if enabled. + */ +static void ma35d1_i2c_restore_idle(struct ma35d1_i2c *i2c) +{ + if (i2c->msg) + ma35d1_i2c_stop(i2c, -EIO); + else if (i2c->target) + ma35d1_i2c_write_ctl(i2c, MA35_CTL_SI_AA); + else + ma35d1_i2c_write_ctl(i2c, MA35_CTL_SI); +} + +/* Check if this is the last message in the set */ +static inline bool is_last_msg(struct ma35d1_i2c *i2c) +{ + return i2c->msg_idx >= (i2c->msg_num - 1); +} + +/* Check if this is the last byte in the current message */ +static inline bool is_last_byte(struct ma35d1_i2c *i2c) +{ + return i2c->msg_ptr == i2c->msg->len - 1; +} + +/* Check if reached the end of the current message */ +static inline bool is_msgend(struct ma35d1_i2c *i2c) +{ + return i2c->msg_ptr >= i2c->msg->len; +} + +/* + * ma35d1_i2c_irq_target_trx - I2C Target state machine handler + * @i2c: ma35d1 i2c instance + * @i2c_status: hardware status code from MA35_STATUS0 + */ +static irqreturn_t ma35d1_i2c_irq_target_trx(struct ma35d1_i2c *i2c, + unsigned long i2c_status) +{ + unsigned char byte = 0; + + switch (i2c_status) { + case MA35_S_RECE_ARB_LOST: + /* + * Arbitration lost during address transmission phase. + * The hardware switches to Target Transmitter mode when + * our own SLA+W is detected on the bus. + */ + i2c->err = -EAGAIN; + ma35d1_i2c_controller_complete(i2c); + i2c_slave_event(i2c->target, I2C_SLAVE_WRITE_REQUESTED, &byte); + break; + + case MA35_S_RECE_ADDR_ACK: + /* Own SLA+W has been receive; ACK has been return */ + i2c_slave_event(i2c->target, I2C_SLAVE_WRITE_REQUESTED, &byte); + break; + + case MA35_S_TRAN_DATA_NACK: + case MA35_S_RECE_DATA_NACK: + /* + * Data byte or last data in I2CDAT has been transmitted and NACK received, + * or previously addressed with own SLA address and NACK returned. + */ + break; + + case MA35_S_RECE_DATA_ACK: + /* + * Previously address with own SLA address Data has been received; + * ACK has been returned + */ + byte = readb(i2c->regs + MA35_DAT); + i2c_slave_event(i2c->target, I2C_SLAVE_WRITE_RECEIVED, &byte); + break; + + case MA35_ADDR_TRAN_ARB_LOST: + /* + * Arbitration lost during address transmission phase. + * The hardware switches to Target Receiver mode when + * our own SLA+R is detected on the bus. + */ + i2c->err = -EAGAIN; + ma35d1_i2c_controller_complete(i2c); + i2c_slave_event(i2c->target, I2C_SLAVE_READ_REQUESTED, &byte); + writel(byte, i2c->regs + MA35_DAT); + break; + + case MA35_S_TRAN_ADDR_ACK: + /* Own SLA+R has been receive; ACK has been return */ + i2c_slave_event(i2c->target, I2C_SLAVE_READ_REQUESTED, &byte); + writel(byte, i2c->regs + MA35_DAT); + break; + + case MA35_S_TRAN_DATA_ACK: + i2c_slave_event(i2c->target, I2C_SLAVE_READ_PROCESSED, &byte); + writel(byte, i2c->regs + MA35_DAT); + break; + + case MA35_S_REPEAT_START_STOP: + /* + * A STOP or repeated START has been received + * while still addressed as Target/Receiver + */ + i2c_slave_event(i2c->target, I2C_SLAVE_STOP, &byte); + break; + + default: + dev_err(i2c->dev, "Status 0x%02lx is NOT processed\n", + i2c_status); + ma35d1_i2c_restore_idle(i2c); + return IRQ_NONE; + } + ma35d1_i2c_write_ctl(i2c, MA35_CTL_SI_AA); + return IRQ_HANDLED; +} + +/* + * ma35d1_i2c_irq_controller_trx - I2C Controller state machine handler + * @i2c: ma35d1 i2c instance + * @i2c_status: hardware status code from MA35_STATUS0 + */ +static irqreturn_t ma35d1_i2c_irq_controller_trx(struct ma35d1_i2c *i2c, + unsigned long i2c_status) +{ + unsigned char byte; + + /* Ensuring messages are absolutely valid in Controller mode. */ + if (unlikely(!i2c->msg)) { + dev_err(i2c->dev, "Spurious controller interrupt (status: 0x%02lx)\n", i2c_status); + ma35d1_i2c_restore_idle(i2c); + return IRQ_HANDLED; + } + + switch (i2c_status) { + case MA35_M_START: + case MA35_M_REPEAT_START: + ma35d1_i2c_set_addr(i2c); + ma35d1_i2c_write_ctl(i2c, MA35_CTL_SI); + break; + + case MA35_M_TRAN_ADDR_ACK: + case MA35_M_TRAN_DATA_ACK: + /* SLA+W has been transmitted and ACK has been received */ + if (i2c_status == MA35_M_TRAN_ADDR_ACK) { + if (is_last_msg(i2c) && i2c->msg->len == 0) { + ma35d1_i2c_stop(i2c, 0); + return IRQ_HANDLED; + } + } + + if (!is_msgend(i2c)) { + byte = i2c->msg->buf[i2c->msg_ptr++]; + writel(byte, i2c->regs + MA35_DAT); + ma35d1_i2c_write_ctl(i2c, MA35_CTL_SI); + } else if (!is_last_msg(i2c)) { + dev_dbg(i2c->dev, "WRITE: Next Message\n"); + + i2c->msg_ptr = 0; + i2c->msg_idx++; + i2c->msg++; + + ma35d1_i2c_write_ctl(i2c, MA35_CTL_STA | MA35_CTL_SI); + } else { + ma35d1_i2c_stop(i2c, 0); + } + break; + + case MA35_M_TRAN_DATA_NACK: + ma35d1_i2c_stop(i2c, -EIO); + break; + + case MA35_M_TRAN_ADDR_NACK: + case MA35_M_RECE_ADDR_NACK: + /* + * Controller Transmit Address NACK + * 0x20: SLA+W has been transmitted and NACK has been received + * 0x48: SLA+R has been transmitted and NACK has been received + */ + dev_dbg(i2c->dev, "\n i2c: ack was not received\n"); + ma35d1_i2c_stop(i2c, -ENXIO); + break; + + case MA35_M_RECE_ADDR_ACK: + if (is_msgend(i2c)) { + if (is_last_msg(i2c)) { + ma35d1_i2c_stop(i2c, 0); + } else { + dev_dbg(i2c->dev, "READ: Next Transfer\n"); + + i2c->msg_ptr = 0; + i2c->msg_idx++; + i2c->msg++; + + ma35d1_i2c_write_ctl(i2c, MA35_CTL_STA_SI); + } + } else if (i2c->msg->len == 1) { + ma35d1_i2c_write_ctl(i2c, MA35_CTL_SI); + } else { + ma35d1_i2c_write_ctl(i2c, MA35_CTL_SI_AA); + } + break; + + case MA35_M_RECE_DATA_ACK: + case MA35_M_RECE_DATA_NACK: + /* DATA has been received and ACK/NACK has been received */ + byte = readb(i2c->regs + MA35_DAT); + i2c->msg->buf[i2c->msg_ptr++] = byte; + + if (is_last_byte(i2c)) { + ma35d1_i2c_write_ctl(i2c, MA35_CTL_SI); + } else if (is_msgend(i2c)) { + if (is_last_msg(i2c)) { + dev_dbg(i2c->dev, "READ: Send Stop\n"); + + ma35d1_i2c_stop(i2c, 0); + } else { + dev_dbg(i2c->dev, "READ: Next Transfer\n"); + + i2c->msg_ptr = 0; + i2c->msg_idx++; + i2c->msg++; + + ma35d1_i2c_write_ctl(i2c, MA35_CTL_STA_SI); + } + } else { + ma35d1_i2c_write_ctl(i2c, MA35_CTL_SI_AA); + } + break; + + case MA35_M_ARB_LOST: + /* Arbitration Lost */ + i2c->err = -EAGAIN; + ma35d1_i2c_controller_complete(i2c); + ma35d1_i2c_write_ctl(i2c, MA35_CTL_SI); + break; + + default: + dev_err(i2c->dev, "Status 0x%02lx is NOT processed\n", + i2c_status); + ma35d1_i2c_restore_idle(i2c); + return IRQ_NONE; + } + return IRQ_HANDLED; +} + +static irqreturn_t ma35d1_i2c_irq(int irqno, void *dev_id) +{ + struct ma35d1_i2c *i2c = dev_id; + unsigned long status; + + status = readl(i2c->regs + MA35_STATUS0); + + if (status == MA35_BUS_ERROR) { + dev_err(i2c->dev, "bus error\n"); + ma35d1_i2c_restore_idle(i2c); + return IRQ_HANDLED; + } + + if (ma35d1_is_controller_status(status)) + return ma35d1_i2c_irq_controller_trx(i2c, status); + else + return ma35d1_i2c_irq_target_trx(i2c, status); +} + +static int ma35d1_i2c_doxfer(struct ma35d1_i2c *i2c, struct i2c_msg *msgs, + int num) +{ + unsigned long timeout; + unsigned int val; + int ret, err; + + i2c->msg = msgs; + i2c->msg_num = num; + i2c->msg_ptr = 0; + i2c->msg_idx = 0; + i2c->err = 0; + + ma35d1_i2c_write_ctl(i2c, MA35_CTL_STA); + + timeout = wait_event_timeout(i2c->wait, i2c->msg_num == 0, HZ * 5); + ret = i2c->msg_idx; + + if (timeout == 0) { + dev_dbg(i2c->dev, "xfer timeout\n"); + ret = -ETIMEDOUT; + goto reset; + } + + err = readl_poll_timeout(i2c->regs + MA35_CTL0, val, + !(val & MA35_CTL_STO), 100, + STOP_TIMEOUT_MS * 1000); + if (err) { + dev_err(i2c->dev, "bus idle timeout\n"); + ret = -EBUSY; + goto reset; + } else if (i2c->err) { + dev_dbg(i2c->dev, "xfer error %d\n", i2c->err); + ret = i2c->err; + } + return ret; + +reset: + ma35d1_i2c_reset(i2c); + return ret; +} + +static int ma35d1_i2c_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, + int num) +{ + struct ma35d1_i2c *i2c = i2c_get_adapdata(adap); + int ret; + + ret = pm_runtime_resume_and_get(i2c->dev); + if (ret) + return ret; + + ret = ma35d1_i2c_doxfer(i2c, msgs, num); + + if (i2c->target) + ma35d1_i2c_write_ctl(i2c, MA35_CTL_AA); + pm_runtime_put_autosuspend(i2c->dev); + + return ret; +} + +static int ma35d1_reg_target(struct i2c_client *target) +{ + struct ma35d1_i2c *i2c = i2c_get_adapdata(target->adapter); + unsigned int slvaddr; + int ret; + + if (i2c->target) + return -EBUSY; + + if (target->flags & I2C_CLIENT_TEN) + return -EAFNOSUPPORT; + + /* Keep device active for target mode listening. */ + ret = pm_runtime_resume_and_get(i2c->dev); + if (ret) { + dev_err(i2c->dev, "failed to resume i2c controller\n"); + return ret; + } + + i2c->target = target; + + slvaddr = target->addr << 1; + writel(slvaddr, i2c->regs + MA35_ADDR0); + + /* I2C enter SLV mode */ + ma35d1_i2c_write_ctl(i2c, MA35_CTL_AA); + + return 0; +} + +static int ma35d1_unreg_target(struct i2c_client *target) +{ + struct ma35d1_i2c *i2c = i2c_get_adapdata(target->adapter); + int ret; + + if (!i2c->target) + return -EINVAL; + + writel(0, i2c->regs + MA35_ADDR0); + /* Clear AA bit */ + ma35d1_i2c_write_ctl(i2c, 0); + + synchronize_irq(i2c->irq); + i2c->target = NULL; + + ret = pm_runtime_put_sync(i2c->dev); + if (ret < 0) + dev_err(i2c->dev, "failed to suspend i2c controller"); + + return 0; +} + +/* Declare Our I2C Functionality */ +static u32 ma35d1_i2c_func(struct i2c_adapter *adap) +{ + return I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL; +} + +/* I2C Bus Registration Info */ +static const struct i2c_algorithm ma35d1_i2c_algorithm = { + .xfer = ma35d1_i2c_xfer, + .functionality = ma35d1_i2c_func, + .reg_target = ma35d1_reg_target, + .unreg_target = ma35d1_unreg_target, +}; + +static const struct i2c_adapter_quirks ma35d1_i2c_quirks = { + .flags = I2C_AQ_NO_ZERO_LEN_READ, +}; + +static int ma35d1_i2c_suspend(struct device *dev) +{ + struct ma35d1_i2c *i2c = dev_get_drvdata(dev); + unsigned int val; + + /* Prepare for wake-up from I2C events if target mode is active */ + if (i2c->target) { + ma35d1_i2c_write_ctl(i2c, MA35_CTL_AA); + + /* Setup wake-up control */ + writel(0x1, i2c->regs + MA35_WKCTL); + + /* Clear pending wake-up flags */ + val = readl(i2c->regs + MA35_WKSTS); + writel(val, i2c->regs + MA35_WKSTS); + + enable_irq_wake(i2c->irq); + + return 0; + } + return pm_runtime_force_suspend(dev); +} + +static int ma35d1_i2c_resume(struct device *dev) +{ + struct ma35d1_i2c *i2c = dev_get_drvdata(dev); + unsigned int val; + + if (i2c->target) { + /* Disable wake-up */ + writel(0x0, i2c->regs + MA35_WKCTL); + + /* Clear pending wake-up flags */ + val = readl(i2c->regs + MA35_WKSTS); + writel(val, i2c->regs + MA35_WKSTS); + + disable_irq_wake(i2c->irq); + return 0; + } + return pm_runtime_force_resume(dev); +} + +static int ma35d1_i2c_runtime_suspend(struct device *dev) +{ + struct ma35d1_i2c *i2c = dev_get_drvdata(dev); + + clk_disable(i2c->clk); + + return 0; +} + +static int ma35d1_i2c_runtime_resume(struct device *dev) +{ + struct ma35d1_i2c *i2c = dev_get_drvdata(dev); + int ret; + + ret = clk_enable(i2c->clk); + if (ret) { + dev_err(dev, "failed to enable clock in resume\n"); + return ret; + } + + return 0; +} + +static void ma35d1_i2c_pm_cleanup(void *data) +{ + struct device *dev = data; + + pm_runtime_get_sync(dev); + pm_runtime_disable(dev); + pm_runtime_put_noidle(dev); +} + +static int ma35d1_i2c_probe(struct platform_device *pdev) +{ + unsigned int busfreq = I2C_MAX_STANDARD_MODE_FREQ; + struct device *dev = &pdev->dev; + struct ma35d1_i2c *i2c; + struct resource *res; + int ret, clkdiv; + u32 val; + + i2c = devm_kzalloc(dev, sizeof(*i2c), GFP_KERNEL); + if (!i2c) + return -ENOMEM; + + init_waitqueue_head(&i2c->wait); + spin_lock_init(&i2c->lock); + + i2c->dev = dev; + + i2c->clk = devm_clk_get_enabled(dev, NULL); + if (IS_ERR(i2c->clk)) + return dev_err_probe(dev, PTR_ERR(i2c->clk), + "failed to get core clk\n"); + + i2c->regs = devm_platform_get_and_ioremap_resource(pdev, 0, &res); + if (IS_ERR(i2c->regs)) + return PTR_ERR(i2c->regs); + + i2c->rst = devm_reset_control_get_exclusive(&pdev->dev, NULL); + if (IS_ERR(i2c->rst)) + return dev_err_probe(dev, PTR_ERR(i2c->rst), + "failed to get reset control\n"); + + ret = reset_control_deassert(i2c->rst); + if (ret) + return dev_err_probe(dev, ret, "failed to deassert reset line\n"); + + /* Setup info block for the I2C core */ + strscpy(i2c->adap.name, "ma35d1-i2c", sizeof(i2c->adap.name)); + i2c->adap.owner = THIS_MODULE; + i2c->adap.algo = &ma35d1_i2c_algorithm; + i2c->adap.quirks = &ma35d1_i2c_quirks; + i2c->adap.retries = 2; + i2c->adap.algo_data = i2c; + i2c->adap.dev.parent = &pdev->dev; + i2c->adap.dev.of_node = pdev->dev.of_node; + i2c_set_adapdata(&i2c->adap, i2c); + + if (!device_property_read_u32(dev, "clock-frequency", &val)) { + if (val != 0 && val <= MEGA) + busfreq = val; + } + /* Calculate divider based on the current peripheral clock rate */ + clkdiv = DIV_ROUND_CLOSEST(clk_get_rate(i2c->clk), busfreq * 4) - 1; + if (clkdiv < 0 || clkdiv > 0xffff) + return dev_err_probe(dev, -EINVAL, "invalid clkdiv value: %d\n", + clkdiv); + + i2c->irq = platform_get_irq(pdev, 0); + if (i2c->irq < 0) + return dev_err_probe(dev, i2c->irq, "failed to get irq\n"); + + platform_set_drvdata(pdev, i2c); + + pm_runtime_set_autosuspend_delay(dev, I2C_PM_TIMEOUT_MS); + pm_runtime_use_autosuspend(dev); + pm_runtime_set_active(dev); + pm_runtime_enable(dev); + + ret = devm_add_action_or_reset(dev, ma35d1_i2c_pm_cleanup, dev); + if (ret) + return ret; + + writel(MA35_CTL_I2CEN | MA35_CTL_INTEN, i2c->regs + MA35_CTL0); + writel(FIELD_PREP(MA35_CLKDIV_MSK, clkdiv), i2c->regs + MA35_CLKDIV); + + ret = devm_request_irq(dev, i2c->irq, ma35d1_i2c_irq, 0, dev_name(dev), + i2c); + if (ret) { + dev_err_probe(dev, ret, "cannot claim IRQ %d\n", i2c->irq); + return ret; + } + + ret = devm_i2c_add_adapter(dev, &i2c->adap); + if (ret) { + dev_err_probe(dev, ret, "failed to add bus to i2c core\n"); + return ret; + } + + dev_info(&i2c->adap.dev, "%pa MA35D1 I2C adapter registered\n", + &res->start); + return 0; +} + +static const struct dev_pm_ops ma35d1_i2c_pmops = { + SYSTEM_SLEEP_PM_OPS(ma35d1_i2c_suspend, ma35d1_i2c_resume) + RUNTIME_PM_OPS(ma35d1_i2c_runtime_suspend, + ma35d1_i2c_runtime_resume, NULL) +}; + +static const struct of_device_id ma35d1_i2c_of_match[] = { + { .compatible = "nuvoton,ma35d1-i2c" }, + {}, +}; +MODULE_DEVICE_TABLE(of, ma35d1_i2c_of_match); + +static struct platform_driver ma35d1_i2c_driver = { + .probe = ma35d1_i2c_probe, + .driver = { + .name = "ma35d1-i2c", + .of_match_table = ma35d1_i2c_of_match, + .pm = pm_ptr(&ma35d1_i2c_pmops), + }, +}; +module_platform_driver(ma35d1_i2c_driver); + +MODULE_AUTHOR("Zi-Yu Chen <zychennvt@gmail.com>"); +MODULE_DESCRIPTION("MA35D1 I2C Bus Driver"); +MODULE_LICENSE("GPL"); -- 2.34.1 ^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH v9 2/3] i2c: ma35d1: Add Nuvoton MA35D1 I2C driver support 2026-08-07 6:03 ` [PATCH v9 2/3] i2c: ma35d1: Add Nuvoton MA35D1 I2C driver support Zi-Yu Chen @ 2026-08-07 6:18 ` sashiko-bot 2026-08-07 7:18 ` zychen 0 siblings, 1 reply; 6+ messages in thread From: sashiko-bot @ 2026-08-07 6:18 UTC (permalink / raw) To: Zi-Yu Chen; +Cc: robh, conor+dt, devicetree Thank you for your contribution! Sashiko AI review found 6 potential issue(s) to consider: - [High] Unhandled `MA35_BUS_RELEASED` status causes spurious errors and randomly aborts ongoing I2C transfers. - [High] Missing NULL checks for `i2c->target` and race conditions during target unregistration lead to a NULL pointer dereference in the ISR. - [High] Unprotected initialization of shared transfer state in `ma35d1_i2c_doxfer` races with asynchronous I2C interrupts. - [Medium] Missing `pm_runtime_mark_last_busy()` defeats the autosuspend delay, causing immediate suspension. - [Medium] Missing reset assertion on driver remove or probe error leaves the hardware in a dirty state. - [Medium] Unconditional `disable_irq_wake()` during resume causes a kernel warning if `enable_irq_wake()` failed. -- --- Patch [2]: [PATCH v9 2/3] i2c: ma35d1: Add Nuvoton MA35D1 I2C driver support --- commit a1da9304aab06d4fdb61f6ee5807d7913cc199b8 Author: Zi-Yu Chen <zychennvt@gmail.com> Subject: i2c: ma35d1: Add Nuvoton MA35D1 I2C driver support This commit adds I2C controller driver support for the Nuvoton MA35D1 SoC. The driver provides both controller and target functionality across standard, fast, and fast-plus modes. > diff --git a/drivers/i2c/busses/i2c-ma35d1.c b/drivers/i2c/busses/i2c-ma35d1.c > new file mode 100644 > --- /dev/null > +++ b/drivers/i2c/busses/i2c-ma35d1.c [ ... ] > +static irqreturn_t ma35d1_i2c_irq_target_trx(struct ma35d1_i2c *i2c, > + unsigned long i2c_status) > +{ > + unsigned char byte = 0; > + > + switch (i2c_status) { > + case MA35_S_RECE_ARB_LOST: [ ... ] > + case MA35_S_RECE_ADDR_ACK: > + /* Own SLA+W has been receive; ACK has been return */ > + i2c_slave_event(i2c->target, I2C_SLAVE_WRITE_REQUESTED, &byte); [Severity: High] Can this result in a NULL pointer dereference if a target interrupt is pending during unregistration? When ma35d1_unreg_target() calls synchronize_irq() and sets i2c->target to NULL, pending interrupts in the irqchip might still fire immediately after. If that happens, i2c_slave_event() will dereference i2c->target without a NULL check. Could a defensive check be added here, or could the teardown synchronization be adjusted? > + break; [ ... ] > +static irqreturn_t ma35d1_i2c_irq(int irqno, void *dev_id) > +{ > + struct ma35d1_i2c *i2c = dev_id; > + unsigned long status; > + > + status = readl(i2c->regs + MA35_STATUS0); [Severity: High] Does this interrupt handler properly process the MA35_BUS_RELEASED (0xf8) status code? If a spurious interrupt occurs and the status is 0xf8, this falls through to the default cases in the transaction handlers, which log an error and call ma35d1_i2c_restore_idle(). This would unconditionally abort any active controller transfers via ma35d1_i2c_stop(i2c, -EIO). Could this be addressed by explicitly checking for 0xf8 and returning IRQ_NONE early? > + > + if (status == MA35_BUS_ERROR) { [ ... ] > +static int ma35d1_i2c_doxfer(struct ma35d1_i2c *i2c, struct i2c_msg *msgs, > + int num) > +{ > + unsigned long timeout; > + unsigned int val; > + int ret, err; > + > + i2c->msg = msgs; > + i2c->msg_num = num; > + i2c->msg_ptr = 0; > + i2c->msg_idx = 0; > + i2c->err = 0; [Severity: High] Could this lockless initialization race with asynchronous I2C interrupts? These transfer state variables are set in process context without holding i2c->lock or disabling interrupts. If a target mode event or spurious interrupt fires concurrently, the ISR could execute ma35d1_i2c_controller_complete() and clear the i2c->msg pointer. When doxfer() then sends the START command, the controller might see a null message pointer and incorrectly abort the transfer. Does this section need lock protection? > + > + ma35d1_i2c_write_ctl(i2c, MA35_CTL_STA); [ ... ] > +static int ma35d1_i2c_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, > + int num) > +{ > + struct ma35d1_i2c *i2c = i2c_get_adapdata(adap); > + int ret; > + > + ret = pm_runtime_resume_and_get(i2c->dev); > + if (ret) > + return ret; > + > + ret = ma35d1_i2c_doxfer(i2c, msgs, num); > + > + if (i2c->target) > + ma35d1_i2c_write_ctl(i2c, MA35_CTL_AA); > + pm_runtime_put_autosuspend(i2c->dev); [Severity: Medium] Does this bypass the autosuspend delay? Calling pm_runtime_put_autosuspend() without first updating the timestamp via pm_runtime_mark_last_busy() will cause the PM core to evaluate based on an outdated timestamp. This could result in the device suspending immediately rather than waiting for the 5000ms delay configured in the probe function. > + > + return ret; > +} [ ... ] > +static int ma35d1_i2c_suspend(struct device *dev) > +{ > + struct ma35d1_i2c *i2c = dev_get_drvdata(dev); > + unsigned int val; > + > + /* Prepare for wake-up from I2C events if target mode is active */ > + if (i2c->target) { > + ma35d1_i2c_write_ctl(i2c, MA35_CTL_AA); > + > + /* Setup wake-up control */ > + writel(0x1, i2c->regs + MA35_WKCTL); > + > + /* Clear pending wake-up flags */ > + val = readl(i2c->regs + MA35_WKSTS); > + writel(val, i2c->regs + MA35_WKSTS); > + > + enable_irq_wake(i2c->irq); > + > + return 0; > + } > + return pm_runtime_force_suspend(dev); > +} > + > +static int ma35d1_i2c_resume(struct device *dev) > +{ > + struct ma35d1_i2c *i2c = dev_get_drvdata(dev); > + unsigned int val; > + > + if (i2c->target) { > + /* Disable wake-up */ > + writel(0x0, i2c->regs + MA35_WKCTL); > + > + /* Clear pending wake-up flags */ > + val = readl(i2c->regs + MA35_WKSTS); > + writel(val, i2c->regs + MA35_WKSTS); > + > + disable_irq_wake(i2c->irq); [Severity: Medium] Can this cause a kernel warning if enable_irq_wake() fails during suspend? The return value of enable_irq_wake() is ignored in ma35d1_i2c_suspend(). If the irqchip does not support wakeup, the wake depth remains zero. Unconditionally calling disable_irq_wake() here will then trigger an "Unbalanced IRQ wake disable" warning. Should the driver record whether enabling wake succeeded before disabling it? > + return 0; > + } [ ... ] > +static int ma35d1_i2c_probe(struct platform_device *pdev) > +{ [ ... ] > + i2c->rst = devm_reset_control_get_exclusive(&pdev->dev, NULL); > + if (IS_ERR(i2c->rst)) > + return dev_err_probe(dev, PTR_ERR(i2c->rst), > + "failed to get reset control\n"); > + > + ret = reset_control_deassert(i2c->rst); > + if (ret) > + return dev_err_probe(dev, ret, "failed to deassert reset line\n"); [Severity: Medium] Is a cleanup action required to assert the reset line? The reset line is deasserted during initialization, but there appears to be no corresponding reset_control_assert() on the error path or driver unbind. If the module is unloaded and reloaded, the hardware could remain in a dirty state from the previous session, potentially causing unexpected behavior like spurious interrupts. Could a devm action be added to handle re-asserting the reset line? > + > + /* Setup info block for the I2C core */ > + strscpy(i2c->adap.name, "ma35d1-i2c", sizeof(i2c->adap.name)); -- Sashiko AI review · https://sashiko.dev/#/patchset/20260807060400.2047517-1-zychennvt@gmail.com?part=2 ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH v9 2/3] i2c: ma35d1: Add Nuvoton MA35D1 I2C driver support 2026-08-07 6:18 ` sashiko-bot @ 2026-08-07 7:18 ` zychen 0 siblings, 0 replies; 6+ messages in thread From: zychen @ 2026-08-07 7:18 UTC (permalink / raw) To: sashiko-reviews; +Cc: robh, conor+dt, devicetree sashiko-bot@kernel.org 於 2026/8/7 下午 02:18 寫道: > Thank you for your contribution! Sashiko AI review found 6 potential issue(s) to consider: > - [High] Unhandled `MA35_BUS_RELEASED` status causes spurious errors and randomly aborts ongoing I2C transfers. Not applicable. According to the MA35D1 TRM, MA35_BUS_RELEASED (0xf8) won't raise interrupt. > - [High] Missing NULL checks for `i2c->target` and race conditions during target unregistration lead to a NULL pointer dereference in the ISR. Not applicable. The target mode is disabled before synchronize_irq(), and i2c->target is cleared only after synchronize_irq() completes. Thus, any pending or running target interrupt is handled before the pointer is cleared, and a NULL check is not required. > - [High] Unprotected initialization of shared transfer state in `ma35d1_i2c_doxfer` races with asynchronous I2C interrupts. Not applicable. The target mode interrupt path does not access the controller transfer state variables. No additional locking is required. > - [Medium] Missing `pm_runtime_mark_last_busy()` defeats the autosuspend delay, causing immediate suspension. Not applicable. pm_runtime_put_autosuspend() already calls pm_runtime_mark_last_busy() internally. > - [Medium] Missing reset assertion on driver remove or probe error leaves the hardware in a dirty state. Not applicable. reset_control_deassert() is only used to ensure the controller is released from reset, as it may have been left asserted by the bootloader. The controller is fully reinitialized during probe, so reasserting the reset line on driver removal is unnecessary. > - [Medium] Unconditional `disable_irq_wake()` during resume causes a kernel warning if `enable_irq_wake()` failed. Not applicable. The MA35D1 I2C interrupt is a dedicated wake-capable IRQ, so enable_irq_wake() is expected to succeed on this platform. > -- > > --- Patch [2]: [PATCH v9 2/3] i2c: ma35d1: Add Nuvoton MA35D1 I2C driver support --- > commit a1da9304aab06d4fdb61f6ee5807d7913cc199b8 > Author: Zi-Yu Chen <zychennvt@gmail.com> > Subject: i2c: ma35d1: Add Nuvoton MA35D1 I2C driver support > > This commit adds I2C controller driver support for the Nuvoton MA35D1 SoC. > The driver provides both controller and target functionality across > standard, fast, and fast-plus modes. > >> diff --git a/drivers/i2c/busses/i2c-ma35d1.c b/drivers/i2c/busses/i2c-ma35d1.c >> new file mode 100644 >> --- /dev/null >> +++ b/drivers/i2c/busses/i2c-ma35d1.c > [ ... ] >> +static irqreturn_t ma35d1_i2c_irq_target_trx(struct ma35d1_i2c *i2c, >> + unsigned long i2c_status) >> +{ >> + unsigned char byte = 0; >> + >> + switch (i2c_status) { >> + case MA35_S_RECE_ARB_LOST: > [ ... ] >> + case MA35_S_RECE_ADDR_ACK: >> + /* Own SLA+W has been receive; ACK has been return */ >> + i2c_slave_event(i2c->target, I2C_SLAVE_WRITE_REQUESTED, &byte); > > [Severity: High] > Can this result in a NULL pointer dereference if a target interrupt is > pending during unregistration? > > When ma35d1_unreg_target() calls synchronize_irq() and sets i2c->target > to NULL, pending interrupts in the irqchip might still fire immediately > after. > > If that happens, i2c_slave_event() will dereference i2c->target without > a NULL check. Could a defensive check be added here, or could the teardown > synchronization be adjusted? > >> + break; > [ ... ] >> +static irqreturn_t ma35d1_i2c_irq(int irqno, void *dev_id) >> +{ >> + struct ma35d1_i2c *i2c = dev_id; >> + unsigned long status; >> + >> + status = readl(i2c->regs + MA35_STATUS0); > > [Severity: High] > Does this interrupt handler properly process the MA35_BUS_RELEASED (0xf8) > status code? > > If a spurious interrupt occurs and the status is 0xf8, this falls through > to the default cases in the transaction handlers, which log an error and > call ma35d1_i2c_restore_idle(). > > This would unconditionally abort any active controller transfers via > ma35d1_i2c_stop(i2c, -EIO). Could this be addressed by explicitly checking > for 0xf8 and returning IRQ_NONE early? > >> + >> + if (status == MA35_BUS_ERROR) { > [ ... ] >> +static int ma35d1_i2c_doxfer(struct ma35d1_i2c *i2c, struct i2c_msg *msgs, >> + int num) >> +{ >> + unsigned long timeout; >> + unsigned int val; >> + int ret, err; >> + >> + i2c->msg = msgs; >> + i2c->msg_num = num; >> + i2c->msg_ptr = 0; >> + i2c->msg_idx = 0; >> + i2c->err = 0; > > [Severity: High] > Could this lockless initialization race with asynchronous I2C interrupts? > > These transfer state variables are set in process context without holding > i2c->lock or disabling interrupts. If a target mode event or spurious > interrupt fires concurrently, the ISR could execute > ma35d1_i2c_controller_complete() and clear the i2c->msg pointer. > > When doxfer() then sends the START command, the controller might see a null > message pointer and incorrectly abort the transfer. Does this section need > lock protection? > >> + >> + ma35d1_i2c_write_ctl(i2c, MA35_CTL_STA); > [ ... ] >> +static int ma35d1_i2c_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, >> + int num) >> +{ >> + struct ma35d1_i2c *i2c = i2c_get_adapdata(adap); >> + int ret; >> + >> + ret = pm_runtime_resume_and_get(i2c->dev); >> + if (ret) >> + return ret; >> + >> + ret = ma35d1_i2c_doxfer(i2c, msgs, num); >> + >> + if (i2c->target) >> + ma35d1_i2c_write_ctl(i2c, MA35_CTL_AA); >> + pm_runtime_put_autosuspend(i2c->dev); > > [Severity: Medium] > Does this bypass the autosuspend delay? > > Calling pm_runtime_put_autosuspend() without first updating the timestamp > via pm_runtime_mark_last_busy() will cause the PM core to evaluate based > on an outdated timestamp. > > This could result in the device suspending immediately rather than waiting > for the 5000ms delay configured in the probe function. > >> + >> + return ret; >> +} > [ ... ] >> +static int ma35d1_i2c_suspend(struct device *dev) >> +{ >> + struct ma35d1_i2c *i2c = dev_get_drvdata(dev); >> + unsigned int val; >> + >> + /* Prepare for wake-up from I2C events if target mode is active */ >> + if (i2c->target) { >> + ma35d1_i2c_write_ctl(i2c, MA35_CTL_AA); >> + >> + /* Setup wake-up control */ >> + writel(0x1, i2c->regs + MA35_WKCTL); >> + >> + /* Clear pending wake-up flags */ >> + val = readl(i2c->regs + MA35_WKSTS); >> + writel(val, i2c->regs + MA35_WKSTS); >> + >> + enable_irq_wake(i2c->irq); >> + >> + return 0; >> + } >> + return pm_runtime_force_suspend(dev); >> +} >> + >> +static int ma35d1_i2c_resume(struct device *dev) >> +{ >> + struct ma35d1_i2c *i2c = dev_get_drvdata(dev); >> + unsigned int val; >> + >> + if (i2c->target) { >> + /* Disable wake-up */ >> + writel(0x0, i2c->regs + MA35_WKCTL); >> + >> + /* Clear pending wake-up flags */ >> + val = readl(i2c->regs + MA35_WKSTS); >> + writel(val, i2c->regs + MA35_WKSTS); >> + >> + disable_irq_wake(i2c->irq); > > [Severity: Medium] > Can this cause a kernel warning if enable_irq_wake() fails during suspend? > > The return value of enable_irq_wake() is ignored in ma35d1_i2c_suspend(). > If the irqchip does not support wakeup, the wake depth remains zero. > > Unconditionally calling disable_irq_wake() here will then trigger an > "Unbalanced IRQ wake disable" warning. Should the driver record whether > enabling wake succeeded before disabling it? > >> + return 0; >> + } > [ ... ] >> +static int ma35d1_i2c_probe(struct platform_device *pdev) >> +{ > [ ... ] >> + i2c->rst = devm_reset_control_get_exclusive(&pdev->dev, NULL); >> + if (IS_ERR(i2c->rst)) >> + return dev_err_probe(dev, PTR_ERR(i2c->rst), >> + "failed to get reset control\n"); >> + >> + ret = reset_control_deassert(i2c->rst); >> + if (ret) >> + return dev_err_probe(dev, ret, "failed to deassert reset line\n"); > > [Severity: Medium] > Is a cleanup action required to assert the reset line? > > The reset line is deasserted during initialization, but there appears to be > no corresponding reset_control_assert() on the error path or driver > unbind. > > If the module is unloaded and reloaded, the hardware could remain in a > dirty state from the previous session, potentially causing unexpected > behavior like spurious interrupts. Could a devm action be added to handle > re-asserting the reset line? > >> + >> + /* Setup info block for the I2C core */ >> + strscpy(i2c->adap.name, "ma35d1-i2c", sizeof(i2c->adap.name)); > Regards, Zi-Yu Chen ^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH v9 3/3] arm64: dts: nuvoton: Add I2C nodes for MA35D1 SoC 2026-08-07 6:03 [PATCH v9 0/3] i2c: ma35d1: Add support for MA35D1 I2C controller Zi-Yu Chen 2026-08-07 6:03 ` [PATCH v9 1/3] dt-bindings: i2c: nuvoton,ma35d1-i2c: Add " Zi-Yu Chen 2026-08-07 6:03 ` [PATCH v9 2/3] i2c: ma35d1: Add Nuvoton MA35D1 I2C driver support Zi-Yu Chen @ 2026-08-07 6:04 ` Zi-Yu Chen 2 siblings, 0 replies; 6+ messages in thread From: Zi-Yu Chen @ 2026-08-07 6:04 UTC (permalink / raw) To: Andi Shyti, linux-i2c, devicetree, linux-arm-kernel, linux-kernel Cc: Jacky Huang, Shan-Chun Hung, Rob Herring, Krzysztof Kozlowski, Conor Dooley, Philipp Zabel, Andrew Jeffery, Zi-Yu Chen Add I2C controller nodes to the MA35D1 SoC dtsi file. Also, enable the I2C2 interface on the MA35D1 SOM board and configure its pinctrl to allow communication with the onboard NAU8822 audio codec Signed-off-by: Zi-Yu Chen <zychennvt@gmail.com> --- .../boot/dts/nuvoton/ma35d1-som-256m.dts | 20 ++++++ arch/arm64/boot/dts/nuvoton/ma35d1.dtsi | 72 +++++++++++++++++++ 2 files changed, 92 insertions(+) diff --git a/arch/arm64/boot/dts/nuvoton/ma35d1-som-256m.dts b/arch/arm64/boot/dts/nuvoton/ma35d1-som-256m.dts index a029b660e8dc..422468dbc3bf 100644 --- a/arch/arm64/boot/dts/nuvoton/ma35d1-som-256m.dts +++ b/arch/arm64/boot/dts/nuvoton/ma35d1-som-256m.dts @@ -57,6 +57,18 @@ &clk { "integer"; }; +&i2c2 { + status = "okay"; + pinctrl-names = "default"; + pinctrl-0 = <&pinctrl_i2c2>; + + nau8822: audio-codec@1a { + compatible = "nuvoton,nau8822"; + reg = <0x1a> ; + nuvoton,spk-btl; + }; +}; + &pinctrl { uart-grp { pinctrl_uart0: uart0-pins { @@ -100,6 +112,14 @@ pinctrl_uart16: uart16-pins { power-source = <1>; }; }; + + i2c-grp { + pinctrl_i2c2: i2c2-pins { + nuvoton,pins = <1 8 4>, + <1 9 4>; + bias-disable; + }; + }; }; &uart0 { diff --git a/arch/arm64/boot/dts/nuvoton/ma35d1.dtsi b/arch/arm64/boot/dts/nuvoton/ma35d1.dtsi index 7228ad4735b5..28488889f564 100644 --- a/arch/arm64/boot/dts/nuvoton/ma35d1.dtsi +++ b/arch/arm64/boot/dts/nuvoton/ma35d1.dtsi @@ -372,6 +372,78 @@ uart15: serial@407f0000 { status = "disabled"; }; + i2c0: i2c@40800000 { + compatible = "nuvoton,ma35d1-i2c"; + reg = <0x0 0x40800000 0x0 0x1000>; + interrupts = <GIC_SPI 13 IRQ_TYPE_LEVEL_HIGH>; + clocks = <&clk I2C0_GATE>; + clock-frequency = <100000>; + resets = <&sys MA35D1_RESET_I2C0>; + status = "disabled"; + #address-cells = <1>; + #size-cells = <0>; + }; + + i2c1: i2c@40810000 { + compatible = "nuvoton,ma35d1-i2c"; + reg = <0x0 0x40810000 0x0 0x1000>; + interrupts = <GIC_SPI 69 IRQ_TYPE_LEVEL_HIGH>; + clocks = <&clk I2C1_GATE>; + clock-frequency = <100000>; + resets = <&sys MA35D1_RESET_I2C1>; + status = "disabled"; + #address-cells = <1>; + #size-cells = <0>; + }; + + i2c2: i2c@40820000 { + compatible = "nuvoton,ma35d1-i2c"; + reg = <0x0 0x40820000 0x0 0x1000>; + interrupts = <GIC_SPI 97 IRQ_TYPE_LEVEL_HIGH>; + clocks = <&clk I2C2_GATE>; + clock-frequency = <100000>; + resets = <&sys MA35D1_RESET_I2C2>; + status = "disabled"; + #address-cells = <1>; + #size-cells = <0>; + }; + + i2c3: i2c@40830000 { + compatible = "nuvoton,ma35d1-i2c"; + reg = <0x0 0x40830000 0x0 0x1000>; + interrupts = <GIC_SPI 98 IRQ_TYPE_LEVEL_HIGH>; + clocks = <&clk I2C3_GATE>; + clock-frequency = <100000>; + resets = <&sys MA35D1_RESET_I2C3>; + status = "disabled"; + #address-cells = <1>; + #size-cells = <0>; + }; + + i2c4: i2c@40840000 { + compatible = "nuvoton,ma35d1-i2c"; + reg = <0x0 0x40840000 0x0 0x1000>; + interrupts = <GIC_SPI 123 IRQ_TYPE_LEVEL_HIGH>; + clocks = <&clk I2C4_GATE>; + clock-frequency = <100000>; + resets = <&sys MA35D1_RESET_I2C4>; + status = "disabled"; + #address-cells = <1>; + #size-cells = <0>; + }; + + i2c5: i2c@40850000 { + compatible = "nuvoton,ma35d1-i2c"; + reg = <0x0 0x40850000 0x0 0x1000>; + interrupts = <GIC_SPI 124 IRQ_TYPE_LEVEL_HIGH>; + clocks = <&clk I2C5_GATE>; + clock-frequency = <100000>; + resets = <&sys MA35D1_RESET_I2C5>; + status = "disabled"; + #address-cells = <1>; + #size-cells = <0>; + }; + uart16: serial@40880000 { compatible = "nuvoton,ma35d1-uart"; reg = <0x0 0x40880000 0x0 0x100>; -- 2.34.1 ^ permalink raw reply related [flat|nested] 6+ messages in thread
end of thread, other threads:[~2026-08-07 7:18 UTC | newest] Thread overview: 6+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2026-08-07 6:03 [PATCH v9 0/3] i2c: ma35d1: Add support for MA35D1 I2C controller Zi-Yu Chen 2026-08-07 6:03 ` [PATCH v9 1/3] dt-bindings: i2c: nuvoton,ma35d1-i2c: Add " Zi-Yu Chen 2026-08-07 6:03 ` [PATCH v9 2/3] i2c: ma35d1: Add Nuvoton MA35D1 I2C driver support Zi-Yu Chen 2026-08-07 6:18 ` sashiko-bot 2026-08-07 7:18 ` zychen 2026-08-07 6:04 ` [PATCH v9 3/3] arm64: dts: nuvoton: Add I2C nodes for MA35D1 SoC Zi-Yu Chen
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox