From: sashiko-bot@kernel.org
To: "Xianwei Zhao" <xianwei.zhao@amlogic.com>
Cc: linux-amlogic@lists.infradead.org, devicetree@vger.kernel.org,
robh@kernel.org, conor+dt@kernel.org, neil.armstrong@linaro.org
Subject: Re: [PATCH v12 2/2] i2c: amlogic: Add Amlogic A9 I2C controller driver
Date: Wed, 18 Nov 2026 10:20:04 +0000 [thread overview]
Message-ID: <sashiko-outbox-220673@kernel.org> (raw)
In-Reply-To: <20261118-a9-i2c-v12-2-fc83c53121e6@amlogic.com>
Thank you for your contribution! Sashiko AI review found 2 potential
issues to consider:
- [Medium] i2c: amlogic-a9: transfer started before IRQ setup in
aml_i2c_do_start()
- [Medium] i2c: amlogic-a9: stale irq_state and edge IRQ miss in
aml_i2c_irq()
--
commit ce6e05692c16fd50e5f34df38eaf42209813110e
Author: Junyi Zhao <junyi.zhao@amlogic.com>
i2c: amlogic: Add Amlogic A9 I2C controller driver
This commit adds the I2C controller driver for Amlogic A9 SoCs along with its
Kconfig, Makefile, and MAINTAINERS entries.
> diff --git a/drivers/i2c/busses/i2c-amlogic-a9.c b/drivers/i2c/busses/i2c-amlogic-a9.c
> new file mode 100644
> index 0000000000000..d25bd50e78f6d
> --- /dev/null
> +++ b/drivers/i2c/busses/i2c-amlogic-a9.c
[ ... ]
> +static irqreturn_t aml_i2c_irq(int irqno, void *dev_id)
> +{
> + struct aml_i2c *i2c = dev_id;
> + u8 *buf = NULL;
> + u32 rd_len, i;
> + u32 rd_val;
> + u32 irq_state;
> +
> + guard(spinlock)(&i2c->lock);
> + irq_state = readl(i2c->regs + REG_CFG_IRQ_STATE);
> + if (i2c->state == A9_STATE_IDLE) {
> + dev_warn(i2c->dev, "i2c irq triggered abnormally, irq_state:0x%x,irq_en:0x%x\n",
> + irq_state, readl(i2c->regs + REG_CFG_IRQ_ENABLE));
> + goto irq_trans_out;
> + }
> +
> + if (irq_state & A9_TRANS_ERROR) {
> + i2c->error = -ENXIO;
> + goto complete_irq_out;
> + }
> +
> + switch (i2c->state) {
> + case A9_STATE_READ:
> + buf = i2c->msg->buf;
> + rd_len = min(i2c->msg->len - i2c->pos, A9_I2C_FIFO_DEPTH);
> + for (i = 0; i < rd_len; i++) {
> + rd_val = readl(i2c->regs + REG_CFG_RX);
> + buf[i2c->pos++] = rd_val & TX_RX_DATA;
> + }
> + /* Rx must clear thresh after read fifo */
> + clear_irq_status(i2c, A9_RX_THRESH_READ);
> + break;
> + case A9_STATE_WRITE:
> + aml_i2c_prepare_xfer(i2c, A9_FIFO_FILL_HALF);
> + if (i2c->pos == i2c->msg->len)
> + aml_i2c_set_mask(i2c, REG_CFG_IRQ_ENABLE, A9_TX_THRESH_WRITE, 0);
> + /* Tx must clear thresh after write fifo */
> + clear_irq_status(i2c, A9_TX_THRESH_WRITE);
> + break;
> + default:
> + break;
> + }
> +
> + if (!(irq_state & A9_TRANS_DONE) ||
> + (i2c->state == A9_STATE_READ && i2c->pos < i2c->msg->len))
> + return IRQ_HANDLED;
[Severity: Medium]
Can aml_i2c_irq() miss A9_TRANS_DONE on multi-chunk transfers and cause
wait_for_completion_timeout() to time out?
The binding specifies a rising-edge interrupt (IRQ_TYPE_EDGE_RISING), which
requires (REG_CFG_IRQ_STATE & REG_CFG_IRQ_ENABLE) to drop to 0 and transition
back to 1 to latch a new interrupt in the GIC:
Documentation/devicetree/bindings/i2c/amlogic,a9-i2c.yaml:
...
interrupts = <0 164 1>;
...
On a multi-chunk read whose length is a multiple of A9_I2C_FIFO_DEPTH (32),
such as 64 bytes, A9_RX_THRESH_READ remains enabled for the final 32-byte
chunk and fires as soon as the 32nd byte enters the RX FIFO, before the final
NACK and STOP finish on the bus:
CPU0 I2C HW
aml_i2c_irq()
irq_state = readl(REG_CFG_IRQ_STATE)
/* A9_RX_THRESH_READ only */
read 32 bytes from REG_CFG_RX
STOP finishes:
REG_CFG_IRQ_STATE |= A9_PHY_DONE
clear_irq_status(A9_RX_THRESH_READ)
/* A9_PHY_DONE keeps IRQ line high; no new rising edge */
!(irq_state & A9_TRANS_DONE)
return IRQ_HANDLED <-- completion missed, hangs 500 ms
Because irq_state is not re-read after draining or filling the FIFO, the same
miss happens if A9_PHY_DONE or A9_NCK_ERROR asserts while A9_STATE_WRITE is
filling REG_CFG_TX before clear_irq_status(i2c, A9_TX_THRESH_WRITE).
Similarly, if both A9_RX_THRESH_READ and A9_PHY_DONE are already set in
irq_state when more than 32 bytes remain to be read, aml_i2c_irq() reads only
32 bytes, clears only A9_RX_THRESH_READ, and returns early because
i2c->pos < i2c->msg->len, leaving A9_PHY_DONE continuously asserted with no
further rising edge to read the remaining bytes or complete the transfer.
> +
> +complete_irq_out:
> + i2c->state = A9_STATE_IDLE;
> + reset_fifo_pos(i2c);
> + aml_i2c_set_mask(i2c, REG_CFG_IRQ_ENABLE, A9_RX_THRESH_READ | A9_TX_THRESH_WRITE, 0);
> + complete(&i2c->done);
> +irq_trans_out:
> + clear_irq_status(i2c, IRQ_ALL_MASK);
> +
> + return IRQ_HANDLED;
> +}
> +
> +static void aml_i2c_do_start(struct aml_i2c *i2c, struct i2c_msg *msg)
> +{
> + unsigned int reg_val;
> + unsigned int trans_len;
> +
> + trans_len = msg->len;
> + reg_val = (unsigned int)(((msg->addr & 0x7f) << START_SADDR_SHIFT) |
> + ((msg->flags & I2C_M_RD) ? START_READ : 0) |
> + ((trans_len & 0xfff) << START_LEN_SHIFT) | START_START);
> + aml_i2c_set_mask(i2c, REG_CFG_START,
> + START_START | START_SLAVE_ADDR | START_READ | START_LEN, reg_val);
> + clear_irq_status(i2c, IRQ_ALL_MASK);
> + /* enable thresh mode */
> + if (trans_len > A9_I2C_FIFO_DEPTH) {
> + reg_val = (msg->flags & I2C_M_RD) ? A9_RX_THRESH_READ : A9_TX_THRESH_WRITE;
> + aml_i2c_set_mask(i2c, REG_CFG_IRQ_ENABLE,
> + A9_RX_THRESH_READ | A9_TX_THRESH_WRITE, reg_val);
> + }
> +}
[Severity: Medium]
Can clear_irq_status(i2c, IRQ_ALL_MASK) wipe out interrupt status bits for
the transfer that was just started by writing START_START to REG_CFG_START?
CPU0 I2C HW
aml_i2c_xfer_msg()
aml_i2c_do_start()
writel(..., REG_CFG_START)
address NACK or fast transfer finishes:
REG_CFG_IRQ_STATE |= A9_NCK_ERROR
GIC latches rising-edge IRQ
clear_irq_status(IRQ_ALL_MASK)
spin_unlock_irqrestore()
aml_i2c_irq()
irq_state = readl(REG_CFG_IRQ_STATE) /* 0 */
switch (i2c->state)
/* reads empty RX FIFO or writes extra TX FIFO bytes */
!(irq_state & A9_TRANS_DONE)
return IRQ_HANDLED <-- never completes, times out
If the CPU is delayed between writing REG_CFG_START and clearing
REG_CFG_IRQ_STATE (for example under CONFIG_PREEMPT_RT where spinlock_t is
preemptible, or on a vCPU), clearing IRQ_ALL_MASK after asserting START_START
discards the newly asserted status bits. When aml_i2c_irq() runs for the
already-latched edge interrupt, it sees irq_state == 0, misses A9_TRANS_ERROR,
unconditionally accesses the FIFO in the switch (i2c->state) block, and
returns without completing i2c->done.
Clearing REG_CFG_IRQ_STATE and enabling the threshold interrupts in
REG_CFG_IRQ_ENABLE before asserting START_START in REG_CFG_START would avoid
this race.
--
Sashiko AI review · https://sashiko.dev/#/patchset/20261118-a9-i2c-v12-0-fc83c53121e6@amlogic.com?part=2
_______________________________________________
linux-amlogic mailing list
linux-amlogic@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-amlogic
WARNING: multiple messages have this Message-ID (diff)
From: sashiko-bot@kernel.org
To: "Xianwei Zhao" <xianwei.zhao@amlogic.com>
Cc: linux-amlogic@lists.infradead.org, devicetree@vger.kernel.org,
robh@kernel.org, conor+dt@kernel.org, neil.armstrong@linaro.org
Subject: Re: [PATCH v12 2/2] i2c: amlogic: Add Amlogic A9 I2C controller driver
Date: Wed, 18 Nov 2026 10:20:04 +0000 [thread overview]
Message-ID: <sashiko-outbox-220673@kernel.org> (raw)
In-Reply-To: <20261118-a9-i2c-v12-2-fc83c53121e6@amlogic.com>
Thank you for your contribution! Sashiko AI review found 2 potential
issues to consider:
- [Medium] i2c: amlogic-a9: transfer started before IRQ setup in
aml_i2c_do_start()
- [Medium] i2c: amlogic-a9: stale irq_state and edge IRQ miss in
aml_i2c_irq()
--
commit ce6e05692c16fd50e5f34df38eaf42209813110e
Author: Junyi Zhao <junyi.zhao@amlogic.com>
i2c: amlogic: Add Amlogic A9 I2C controller driver
This commit adds the I2C controller driver for Amlogic A9 SoCs along with its
Kconfig, Makefile, and MAINTAINERS entries.
> diff --git a/drivers/i2c/busses/i2c-amlogic-a9.c b/drivers/i2c/busses/i2c-amlogic-a9.c
> new file mode 100644
> index 0000000000000..d25bd50e78f6d
> --- /dev/null
> +++ b/drivers/i2c/busses/i2c-amlogic-a9.c
[ ... ]
> +static irqreturn_t aml_i2c_irq(int irqno, void *dev_id)
> +{
> + struct aml_i2c *i2c = dev_id;
> + u8 *buf = NULL;
> + u32 rd_len, i;
> + u32 rd_val;
> + u32 irq_state;
> +
> + guard(spinlock)(&i2c->lock);
> + irq_state = readl(i2c->regs + REG_CFG_IRQ_STATE);
> + if (i2c->state == A9_STATE_IDLE) {
> + dev_warn(i2c->dev, "i2c irq triggered abnormally, irq_state:0x%x,irq_en:0x%x\n",
> + irq_state, readl(i2c->regs + REG_CFG_IRQ_ENABLE));
> + goto irq_trans_out;
> + }
> +
> + if (irq_state & A9_TRANS_ERROR) {
> + i2c->error = -ENXIO;
> + goto complete_irq_out;
> + }
> +
> + switch (i2c->state) {
> + case A9_STATE_READ:
> + buf = i2c->msg->buf;
> + rd_len = min(i2c->msg->len - i2c->pos, A9_I2C_FIFO_DEPTH);
> + for (i = 0; i < rd_len; i++) {
> + rd_val = readl(i2c->regs + REG_CFG_RX);
> + buf[i2c->pos++] = rd_val & TX_RX_DATA;
> + }
> + /* Rx must clear thresh after read fifo */
> + clear_irq_status(i2c, A9_RX_THRESH_READ);
> + break;
> + case A9_STATE_WRITE:
> + aml_i2c_prepare_xfer(i2c, A9_FIFO_FILL_HALF);
> + if (i2c->pos == i2c->msg->len)
> + aml_i2c_set_mask(i2c, REG_CFG_IRQ_ENABLE, A9_TX_THRESH_WRITE, 0);
> + /* Tx must clear thresh after write fifo */
> + clear_irq_status(i2c, A9_TX_THRESH_WRITE);
> + break;
> + default:
> + break;
> + }
> +
> + if (!(irq_state & A9_TRANS_DONE) ||
> + (i2c->state == A9_STATE_READ && i2c->pos < i2c->msg->len))
> + return IRQ_HANDLED;
[Severity: Medium]
Can aml_i2c_irq() miss A9_TRANS_DONE on multi-chunk transfers and cause
wait_for_completion_timeout() to time out?
The binding specifies a rising-edge interrupt (IRQ_TYPE_EDGE_RISING), which
requires (REG_CFG_IRQ_STATE & REG_CFG_IRQ_ENABLE) to drop to 0 and transition
back to 1 to latch a new interrupt in the GIC:
Documentation/devicetree/bindings/i2c/amlogic,a9-i2c.yaml:
...
interrupts = <0 164 1>;
...
On a multi-chunk read whose length is a multiple of A9_I2C_FIFO_DEPTH (32),
such as 64 bytes, A9_RX_THRESH_READ remains enabled for the final 32-byte
chunk and fires as soon as the 32nd byte enters the RX FIFO, before the final
NACK and STOP finish on the bus:
CPU0 I2C HW
aml_i2c_irq()
irq_state = readl(REG_CFG_IRQ_STATE)
/* A9_RX_THRESH_READ only */
read 32 bytes from REG_CFG_RX
STOP finishes:
REG_CFG_IRQ_STATE |= A9_PHY_DONE
clear_irq_status(A9_RX_THRESH_READ)
/* A9_PHY_DONE keeps IRQ line high; no new rising edge */
!(irq_state & A9_TRANS_DONE)
return IRQ_HANDLED <-- completion missed, hangs 500 ms
Because irq_state is not re-read after draining or filling the FIFO, the same
miss happens if A9_PHY_DONE or A9_NCK_ERROR asserts while A9_STATE_WRITE is
filling REG_CFG_TX before clear_irq_status(i2c, A9_TX_THRESH_WRITE).
Similarly, if both A9_RX_THRESH_READ and A9_PHY_DONE are already set in
irq_state when more than 32 bytes remain to be read, aml_i2c_irq() reads only
32 bytes, clears only A9_RX_THRESH_READ, and returns early because
i2c->pos < i2c->msg->len, leaving A9_PHY_DONE continuously asserted with no
further rising edge to read the remaining bytes or complete the transfer.
> +
> +complete_irq_out:
> + i2c->state = A9_STATE_IDLE;
> + reset_fifo_pos(i2c);
> + aml_i2c_set_mask(i2c, REG_CFG_IRQ_ENABLE, A9_RX_THRESH_READ | A9_TX_THRESH_WRITE, 0);
> + complete(&i2c->done);
> +irq_trans_out:
> + clear_irq_status(i2c, IRQ_ALL_MASK);
> +
> + return IRQ_HANDLED;
> +}
> +
> +static void aml_i2c_do_start(struct aml_i2c *i2c, struct i2c_msg *msg)
> +{
> + unsigned int reg_val;
> + unsigned int trans_len;
> +
> + trans_len = msg->len;
> + reg_val = (unsigned int)(((msg->addr & 0x7f) << START_SADDR_SHIFT) |
> + ((msg->flags & I2C_M_RD) ? START_READ : 0) |
> + ((trans_len & 0xfff) << START_LEN_SHIFT) | START_START);
> + aml_i2c_set_mask(i2c, REG_CFG_START,
> + START_START | START_SLAVE_ADDR | START_READ | START_LEN, reg_val);
> + clear_irq_status(i2c, IRQ_ALL_MASK);
> + /* enable thresh mode */
> + if (trans_len > A9_I2C_FIFO_DEPTH) {
> + reg_val = (msg->flags & I2C_M_RD) ? A9_RX_THRESH_READ : A9_TX_THRESH_WRITE;
> + aml_i2c_set_mask(i2c, REG_CFG_IRQ_ENABLE,
> + A9_RX_THRESH_READ | A9_TX_THRESH_WRITE, reg_val);
> + }
> +}
[Severity: Medium]
Can clear_irq_status(i2c, IRQ_ALL_MASK) wipe out interrupt status bits for
the transfer that was just started by writing START_START to REG_CFG_START?
CPU0 I2C HW
aml_i2c_xfer_msg()
aml_i2c_do_start()
writel(..., REG_CFG_START)
address NACK or fast transfer finishes:
REG_CFG_IRQ_STATE |= A9_NCK_ERROR
GIC latches rising-edge IRQ
clear_irq_status(IRQ_ALL_MASK)
spin_unlock_irqrestore()
aml_i2c_irq()
irq_state = readl(REG_CFG_IRQ_STATE) /* 0 */
switch (i2c->state)
/* reads empty RX FIFO or writes extra TX FIFO bytes */
!(irq_state & A9_TRANS_DONE)
return IRQ_HANDLED <-- never completes, times out
If the CPU is delayed between writing REG_CFG_START and clearing
REG_CFG_IRQ_STATE (for example under CONFIG_PREEMPT_RT where spinlock_t is
preemptible, or on a vCPU), clearing IRQ_ALL_MASK after asserting START_START
discards the newly asserted status bits. When aml_i2c_irq() runs for the
already-latched edge interrupt, it sees irq_state == 0, misses A9_TRANS_ERROR,
unconditionally accesses the FIFO in the switch (i2c->state) block, and
returns without completing i2c->done.
Clearing REG_CFG_IRQ_STATE and enabling the threshold interrupts in
REG_CFG_IRQ_ENABLE before asserting START_START in REG_CFG_START would avoid
this race.
--
Sashiko AI review · https://sashiko.dev/#/patchset/20261118-a9-i2c-v12-0-fc83c53121e6@amlogic.com?part=2
next prev parent reply other threads:[~2026-11-18 10:20 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-18 9:59 [PATCH v12 0/2] i2c: Add driver support Amlogic A9 I2C Xianwei Zhao
2026-11-18 9:59 ` Xianwei Zhao via B4 Relay
2026-11-18 9:59 ` Xianwei Zhao via B4 Relay
2026-11-18 9:59 ` [PATCH v12 1/2] dt-bindings: i2c: add Amlogic A9 I2C controller Xianwei Zhao
2026-11-18 9:59 ` Xianwei Zhao via B4 Relay
2026-11-18 9:59 ` Xianwei Zhao via B4 Relay
2026-11-18 9:59 ` [PATCH v12 2/2] i2c: amlogic: Add Amlogic A9 I2C controller driver Xianwei Zhao
2026-11-18 9:59 ` Xianwei Zhao via B4 Relay
2026-11-18 9:59 ` Xianwei Zhao via B4 Relay
2026-11-18 10:20 ` sashiko-bot [this message]
2026-11-18 10:20 ` sashiko-bot
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=sashiko-outbox-220673@kernel.org \
--to=sashiko-bot@kernel.org \
--cc=conor+dt@kernel.org \
--cc=devicetree@vger.kernel.org \
--cc=linux-amlogic@lists.infradead.org \
--cc=neil.armstrong@linaro.org \
--cc=robh@kernel.org \
--cc=sashiko-reviews@lists.linux.dev \
--cc=xianwei.zhao@amlogic.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.