Skip to contentSkip to conversation

Implement Shutdown callback for geni-i2c

PATCH Applied Roja Rani Yarubandi 3 participants 9 messages Dec 21, 2020 – Jan 8, 2021

Subsystems (4)

v7 of 11 Applied Latest is v11 →

lore
Share & tools
lore
#4 lore raw

PATCH v7 2/2 i2c: i2c-qcom-geni: Add shutdown callback for i2c

In reply to Roja Rani Yarubandi

If the hardware is still accessing memory after SMMU translation is disabled (as part of smmu shutdown callback), then the IOVAs (I/O virtual address) which it was using will go on the bus as the physical addresses which will result in unknown crashes like NoC/interconnect errors.

So, implement shutdown callback to i2c driver to stop on-going transfer and unmap DMA mappings during system "reboot" or "shutdown".

Fixes 37692de5d523 ("i2c: i2c-qcom-geni: Add bus driver for the Qualcomm GENI I2C controller")

---
Changes in V2:
- As per Stephen's comments added seperate function for stop transfer, fixed minor nitpicks.
- As per Stephen's comments, changed commit text.

Changes in V3:
- As per Stephen's comments, squashed patch 1 into patch 2, added Fixes tag.
- As per Akash's comments, included FIFO case in stop_xfer, fixed minor nitpicks.

Changes in V4:
- As per Stephen's comments cleaned up geni_i2c_stop_xfer function, added dma_buf in geni_i2c_dev struct to call i2c_put_dma_safe_msg_buf() from other functions, removed "iova" check in geni_se_rx_dma_unprep() and geni_se_tx_dma_unprep() functions.
- Added two helper functions geni_i2c_rx_one_msg_done() and geni_i2c_tx_one_msg_done() to unwrap the things after rx/tx FIFO/DMA transfers, so that the same can be used in geni_i2c_stop_xfer() function during shutdown callback. Updated commit text accordingly.
- Checking whether it is tx/rx transfer using I2C_M_RD which is valid for both FIFO and DMA cases, so dropped DMA_RX_ACTIVE and DMA_TX_ACTIVE bit checking

Changes in V5:
- As per Stephen's comments, added spin_lock_irqsave & spin_unlock_irqsave in geni_i2c_stop_xfer() function.

Changes in V6:
- As per Stephen's comments, taken care of unsafe lock order in geni_i2c_stop_xfer().
- Moved spin_lock/unlock to geni_i2c_rx_msg_cleanup() and geni_i2c_tx_msg_cleanup() functions.

Changes in V7:
- No changes

 drivers/i2c/busses/i2c-qcom-geni.c | 35 ++++++++++++++++++++++++++++++
 1 file changed, 35 insertions(+)
drivers/i2c/busses/i2c-qcom-geni.c +35 -0
index 214b4c913a13..7903e6145543 100644
@@ -375,6 +375,33 @@ static void geni_i2c_tx_msg_cleanup(struct geni_i2c_dev *gi2c,
 	}
 }
 
+static void geni_i2c_stop_xfer(struct geni_i2c_dev *gi2c)
+{
+	int ret;
+	u32 geni_status;
+	struct i2c_msg *cur;
+
+	/* Resume device, as runtime suspend can happen anytime during transfer */
+	ret = pm_runtime_get_sync(gi2c->se.dev);
+	if (ret < 0) {
+		dev_err(gi2c->se.dev, "Failed to resume device: %d\n", ret);
+		return;
+	}
+
+	geni_status = readl_relaxed(gi2c->se.base + SE_GENI_STATUS);
+	if (!(geni_status & M_GENI_CMD_ACTIVE))
+		goto out;
+
+	cur = gi2c->cur;
+	geni_i2c_abort_xfer(gi2c);
+	if (cur->flags & I2C_M_RD)
+		geni_i2c_rx_msg_cleanup(gi2c, cur);
+	else
+		geni_i2c_tx_msg_cleanup(gi2c, cur);
+out:
+	pm_runtime_put_sync_suspend(gi2c->se.dev);
+}
+
 static int geni_i2c_rx_one_msg(struct geni_i2c_dev *gi2c, struct i2c_msg *msg,
 				u32 m_param)
 {
@@ -650,6 +677,13 @@ static int geni_i2c_remove(struct platform_device *pdev)
 	return 0;
 }
 
+static void  geni_i2c_shutdown(struct platform_device *pdev)
+{
+	struct geni_i2c_dev *gi2c = platform_get_drvdata(pdev);
+
+	geni_i2c_stop_xfer(gi2c);
+}
+
 static int __maybe_unused geni_i2c_runtime_suspend(struct device *dev)
 {
 	int ret;
@@ -714,6 +748,7 @@ MODULE_DEVICE_TABLE(of, geni_i2c_dt_match);
 static struct platform_driver geni_i2c_driver = {
 	.probe  = geni_i2c_probe,
 	.remove = geni_i2c_remove,
+	.shutdown = geni_i2c_shutdown,
 	.driver = {
 		.name = "geni_i2c",
 		.pm = &geni_i2c_pm_ops,
Signature · 2 lines
QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a member 
of Code Aurora Forum, hosted by The Linux Foundation
#0 lore raw

v7 cover Implement Shutdown callback for geni-i2c

Hi,

Posting V7 with comments answering.

Roja Rani Yarubandi (2):
i2c: i2c-qcom-geni: Store DMA mapping data in geni_i2c_dev struct
i2c: i2c-qcom-geni: Add shutdown callback for i2c

 drivers/i2c/busses/i2c-qcom-geni.c | 94 +++++++++++++++++++++++++-----
 1 file changed, 78 insertions(+), 16 deletions(-)
Signature · 2 lines
QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a member 
of Code Aurora Forum, hosted by The Linux Foundation
#1 lore raw

PATCH v7 1/2 i2c: i2c-qcom-geni: Store DMA mapping data in geni_i2c_dev struct

Store DMA mapping data in geni_i2c_dev struct to enhance DMA mapping data scope. For example during shutdown callback to unmap DMA mapping, this stored DMA mapping data can be used to call geni_se_tx_dma_unprep and geni_se_rx_dma_unprep functions.

Add two helper functions geni_i2c_rx_msg_cleanup and geni_i2c_tx_msg_cleanup to unwrap the things after rx/tx FIFO/DMA transfers, so that the same can be used in geni_i2c_stop_xfer() function during shutdown callback.

---
Changes in V5:
- As per Stephen's comments separated this patch from shutdown callback patch, gi2c->cur = NULL is not removed from geni_i2c_abort_xfer(), and made a copy of gi2c->cur and passed to cleanup functions.

Changes in V6:
- Added spin_lock/unlock in geni_i2c_rx_msg_cleanup() and geni_i2c_tx_msg_cleanup() functions.

Changes in V7:
- As per Akash's comments analysed the race chances and removed spin lock in cleanup functions.

 drivers/i2c/busses/i2c-qcom-geni.c | 59 ++++++++++++++++++++++--------
 1 file changed, 43 insertions(+), 16 deletions(-)
drivers/i2c/busses/i2c-qcom-geni.c +43 -16
index 046d241183c5..214b4c913a13 100644
@@ -86,6 +86,9 @@ struct geni_i2c_dev {
 	u32 clk_freq_out;
 	const struct geni_i2c_clk_fld *clk_fld;
 	int suspended;
+	void *dma_buf;
+	size_t xfer_len;
+	dma_addr_t dma_addr;
 };
 
 struct geni_i2c_err_log {
@@ -348,14 +351,39 @@ static void geni_i2c_tx_fsm_rst(struct geni_i2c_dev *gi2c)
 		dev_err(gi2c->se.dev, "Timeout resetting TX_FSM\n");
 }
 
+static void geni_i2c_rx_msg_cleanup(struct geni_i2c_dev *gi2c,
+				     struct i2c_msg *cur)
+{
+	gi2c->cur_rd = 0;
+	if (gi2c->dma_buf) {
+		if (gi2c->err)
+			geni_i2c_rx_fsm_rst(gi2c);
+		geni_se_rx_dma_unprep(&gi2c->se, gi2c->dma_addr, gi2c->xfer_len);
+		i2c_put_dma_safe_msg_buf(gi2c->dma_buf, cur, !gi2c->err);
+	}
+}
+
+static void geni_i2c_tx_msg_cleanup(struct geni_i2c_dev *gi2c,
+				     struct i2c_msg *cur)
+{
+	gi2c->cur_wr = 0;
+	if (gi2c->dma_buf) {
+		if (gi2c->err)
+			geni_i2c_tx_fsm_rst(gi2c);
+		geni_se_tx_dma_unprep(&gi2c->se, gi2c->dma_addr, gi2c->xfer_len);
+		i2c_put_dma_safe_msg_buf(gi2c->dma_buf, cur, !gi2c->err);
+	}
+}
+
 static int geni_i2c_rx_one_msg(struct geni_i2c_dev *gi2c, struct i2c_msg *msg,
 				u32 m_param)
 {
-	dma_addr_t rx_dma;
+	dma_addr_t rx_dma = 0;
 	unsigned long time_left;
 	void *dma_buf;
 	struct geni_se *se = &gi2c->se;
 	size_t len = msg->len;
+	struct i2c_msg *cur;
 
 	dma_buf = i2c_get_dma_safe_msg_buf(msg, 32);
 	if (dma_buf)
@@ -370,19 +398,18 @@ static int geni_i2c_rx_one_msg(struct geni_i2c_dev *gi2c, struct i2c_msg *msg,
 		geni_se_select_mode(se, GENI_SE_FIFO);
 		i2c_put_dma_safe_msg_buf(dma_buf, msg, false);
 		dma_buf = NULL;
+	} else {
+		gi2c->xfer_len = len;
+		gi2c->dma_addr = rx_dma;
+		gi2c->dma_buf = dma_buf;
 	}
 
+	cur = gi2c->cur;
 	time_left = wait_for_completion_timeout(&gi2c->done, XFER_TIMEOUT);
 	if (!time_left)
 		geni_i2c_abort_xfer(gi2c);
 
-	gi2c->cur_rd = 0;
-	if (dma_buf) {
-		if (gi2c->err)
-			geni_i2c_rx_fsm_rst(gi2c);
-		geni_se_rx_dma_unprep(se, rx_dma, len);
-		i2c_put_dma_safe_msg_buf(dma_buf, msg, !gi2c->err);
-	}
+	geni_i2c_rx_msg_cleanup(gi2c, cur);
 
 	return gi2c->err;
 }
@@ -390,11 +417,12 @@ static int geni_i2c_rx_one_msg(struct geni_i2c_dev *gi2c, struct i2c_msg *msg,
 static int geni_i2c_tx_one_msg(struct geni_i2c_dev *gi2c, struct i2c_msg *msg,
 				u32 m_param)
 {
-	dma_addr_t tx_dma;
+	dma_addr_t tx_dma = 0;
 	unsigned long time_left;
 	void *dma_buf;
 	struct geni_se *se = &gi2c->se;
 	size_t len = msg->len;
+	struct i2c_msg *cur;
 
 	dma_buf = i2c_get_dma_safe_msg_buf(msg, 32);
 	if (dma_buf)
@@ -409,22 +437,21 @@ static int geni_i2c_tx_one_msg(struct geni_i2c_dev *gi2c, struct i2c_msg *msg,
 		geni_se_select_mode(se, GENI_SE_FIFO);
 		i2c_put_dma_safe_msg_buf(dma_buf, msg, false);
 		dma_buf = NULL;
+	} else {
+		gi2c->xfer_len = len;
+		gi2c->dma_addr = tx_dma;
+		gi2c->dma_buf = dma_buf;
 	}
 
 	if (!dma_buf) /* Get FIFO IRQ */
 		writel_relaxed(1, se->base + SE_GENI_TX_WATERMARK_REG);
 
+	cur = gi2c->cur;
 	time_left = wait_for_completion_timeout(&gi2c->done, XFER_TIMEOUT);
 	if (!time_left)
 		geni_i2c_abort_xfer(gi2c);
 
-	gi2c->cur_wr = 0;
-	if (dma_buf) {
-		if (gi2c->err)
-			geni_i2c_tx_fsm_rst(gi2c);
-		geni_se_tx_dma_unprep(se, tx_dma, len);
-		i2c_put_dma_safe_msg_buf(dma_buf, msg, !gi2c->err);
-	}
+	geni_i2c_tx_msg_cleanup(gi2c, cur);
 
 	return gi2c->err;
 }
Signature · 2 lines
QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a member 
of Code Aurora Forum, hosted by The Linux Foundation
Reviewed-by on v7 1/2
#2 lore raw

Re: [PATCH V7 1/2] i2c: i2c-qcom-geni: Store DMA mapping data in geni_i2c_dev struct

Quoted 11 lines · Roja Rani Yarubandi

Store DMA mapping data in geni_i2c_dev struct to enhance DMA mapping data scope. For example during shutdown callback to unmap DMA mapping, this stored DMA mapping data can be used to call geni_se_tx_dma_unprep and geni_se_rx_dma_unprep functions.

Add two helper functions geni_i2c_rx_msg_cleanup and geni_i2c_tx_msg_cleanup to unwrap the things after rx/tx FIFO/DMA transfers, so that the same can be used in geni_i2c_stop_xfer() function during shutdown callback.

Signature · 1 line
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,\na Linux Foundation Collaborative Project
Reviewed-by on v7 2/2
#5 lore raw

Re: [PATCH V7 2/2] i2c: i2c-qcom-geni: Add shutdown callback for i2c

In reply to Roja Rani Yarubandi
Quoted 11 lines · Roja Rani Yarubandi

If the hardware is still accessing memory after SMMU translation is disabled (as part of smmu shutdown callback), then the IOVAs (I/O virtual address) which it was using will go on the bus as the physical addresses which will result in unknown crashes like NoC/interconnect errors.

So, implement shutdown callback to i2c driver to stop on-going transfer and unmap DMA mappings during system "reboot" or "shutdown".

Fixes 37692de5d523 ("i2c: i2c-qcom-geni: Add bus driver for the Qualcomm GENI I2C controller")
Signature · 1 line
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,\na Linux Foundation Collaborative Project
Applied
#3 lore raw

Re: [PATCH V7 1/2] i2c: i2c-qcom-geni: Store DMA mapping data in geni_i2c_dev struct

In reply to Roja Rani Yarubandi
Quoted 11 lines · Roja Rani Yarubandi

Store DMA mapping data in geni_i2c_dev struct to enhance DMA mapping data scope. For example during shutdown callback to unmap DMA mapping, this stored DMA mapping data can be used to call geni_se_tx_dma_unprep and geni_se_rx_dma_unprep functions.

Add two helper functions geni_i2c_rx_msg_cleanup and geni_i2c_tx_msg_cleanup to unwrap the things after rx/tx FIFO/DMA transfers, so that the same can be used in geni_i2c_stop_xfer() function during shutdown callback.

Applied to for-next, thanks!

#6 lore raw

Re: [PATCH V7 2/2] i2c: i2c-qcom-geni: Add shutdown callback for i2c

In reply to Roja Rani Yarubandi
Quoted 4 lines
drivers/i2c/busses/i2c-qcom-geni.c +3 -0
+	geni_status = readl_relaxed(gi2c->se.base + SE_GENI_STATUS);
+	if (!(geni_status & M_GENI_CMD_ACTIVE))
+		goto out;

drivers/i2c/busses/i2c-qcom-geni.c static void geni_i2c_tx_msg_cleanup(struct geni_i2c_dev *gi2c, · line 403

Quoted 10 lines
drivers/i2c/busses/i2c-qcom-geni.c +10 -0
+
+	cur = gi2c->cur;
+	geni_i2c_abort_xfer(gi2c);
+	if (cur->flags & I2C_M_RD)
+		geni_i2c_rx_msg_cleanup(gi2c, cur);
+	else
+		geni_i2c_tx_msg_cleanup(gi2c, cur);
+out:
+	pm_runtime_put_sync_suspend(gi2c->se.dev);
+}

The use of 'goto' is not needed here IMHO. I think:

	if (geni_status & M_GENI_CMD_ACTIVE) {
		do_the_stuff
	}

	pm_runtime_put_sync_suspend(...);

is more readable, in fact. Also, I don't think we really need the 'cur' variable and just use 'gi2c->cur' but that's very minor and you can keep it if you like it.

Reset looks good!

#7 lore raw

Re: [PATCH V7 2/2] i2c: i2c-qcom-geni: Add shutdown callback for i2c

Hi Wolfram,

Quoted 16 lines · Wolfram Sang

drivers/i2c/busses/i2c-qcom-geni.c static void geni_i2c_tx_msg_cleanup(struct geni_i2c_dev *gi2c, · line 403

Quoted 13 lines
drivers/i2c/busses/i2c-qcom-geni.c +13 -0
+	geni_status = readl_relaxed(gi2c->se.base + SE_GENI_STATUS);
+	if (!(geni_status & M_GENI_CMD_ACTIVE))
+		goto out;
+
+	cur = gi2c->cur;
+	geni_i2c_abort_xfer(gi2c);
+	if (cur->flags & I2C_M_RD)
+		geni_i2c_rx_msg_cleanup(gi2c, cur);
+	else
+		geni_i2c_tx_msg_cleanup(gi2c, cur);
+out:
+	pm_runtime_put_sync_suspend(gi2c->se.dev);
+}

The use of 'goto' is not needed here IMHO. I think:

Quoted 8 lines
	if (geni_status & M_GENI_CMD_ACTIVE) {
		do_the_stuff
	}

	pm_runtime_put_sync_suspend(...);

is more readable, in fact.

In context to the previous comment [1], I have implemented this way. But, yeah anything is fine for me.

Quoted 4 lines

Also, I don't think we really need the 'cur'
variable and just use 'gi2c->cur' but that's very minor and you can keep it if you like it.

In geni_i2c_abort_xfer() function gi2c->cur will be made NULL, so copying it before to "cur" is needed here.

Quoted 1 line

Reset looks good!

[1] https://patchwork.kernel.org/project/linux-arm-msm/patch/[email protected]/#23560541

Thanks,
Roja

#8 lore raw

Re: [PATCH V7 2/2] i2c: i2c-qcom-geni: Add shutdown callback for i2c

Quoted 2 lines
Quoted 2 lines

The use of 'goto' is not needed here IMHO. I think:

...

Quoted 2 lines

In context to the previous comment [1], I have implemented this way. But, yeah anything is fine for me.

Thanks, I really think it is better.

Quoted 2 lines

In geni_i2c_abort_xfer() function gi2c->cur will be made NULL, so copying it before to "cur" is needed here.

Okay then, thanks for the heads up!