[PATCH v3 0/2] Add trace events for Qualcomm GENI SPI drivers

[PATCH v3 0/2] Add trace events for Qualcomm GENI SPI drivers

[Praveen Talari]

Add tracepoints to the Qualcomm GENI (Generic Interface) SPI driver.
These trace events enable runtime debugging and performance analysis
of SPI operations.

The trace events capture SPI clock configuration, setup parameters,
transfer details, interrupt status.

Usage examples:

Enable all SPI traces:
  echo 1 > /sys/kernel/tracing/events/spi/enable  
  echo 1 > /sys/kernel/debug/tracing/events/qcom_geni_spi/enable
  cat /sys/kernel/debug/tracing/trace_pipe

Example trace output:

1003.956560: spi_message_submit: spi16.0 000000001b20b93c
1003.956642: spi_controller_busy: spi16
1003.956643: spi_message_start: spi16.0 000000001b20b93c
1003.956646: geni_spi_setup_params: 888000.spi: cs=0 mode=0x00000020
     mode_changed=0x00000007 cs_changed=0
1003.956647: spi_set_cs: spi16.0 activate
1003.956648: spi_transfer_start: spi16.0 00000000ea1cf8b6 len=16
     tx=[4c-80-e4-ca-68-4d-95-aa-ee-99-ae-d7-69-e9-5f-39]
rx=[00-00-00-00-00-00-00-00-00-00-00-00-00-00-00-00]
1003.956653: geni_spi_clk_cfg: 888000.spi: req_hz=20000000
     sclk_hz=100000000 clk_idx=5 clk_div=5 bpw=8
1003.956691: geni_spi_transfer: 888000.spi: len=16 m_cmd=0x00000003
1003.956708: geni_spi_irq: 888000.spi: m_irq=0x08000081
     dma_tx=0x00000000 dma_rx=0x00000000
1003.956717: spi_transfer_stop: spi16.0 00000000ea1cf8b6 len=16
     tx=[4c-80-e4-ca-68-4d-95-aa-ee-99-ae-d7-69-e9-5f-39]
rx=[4c-80-e4-ca-68-4d-95-aa-ee-99-ae-d7-69-e9-5f-39]
1003.956717: spi_set_cs: spi16.0 deactivate
1003.956718: spi_message_done: spi16.0 000000001b20b93c len=16/16

Signed-off-by: Praveen Talari <praveen.talari@oss.qualcomm.com>
---
Changes in v3:
- Replaced geni_spi_fifo_params with geni_spi_setup_params trace event.
- Updated commit text.
- Link to v2: https://lore.kernel.org/r/20260512-add-tracepoints-for-qcom-geni-spi-v2-0-3b184068ecf9@oss.qualcomm.com

Changes in v2:
- Removed tx/rx data capture since spi core had already support.
- Updated commit text in patches and cover letter.
- Link to v1: https://lore.kernel.org/r/20260506-add-tracepoints-for-qcom-geni-spi-v1-0-c957cfe712d1@oss.qualcomm.com

---
Praveen Talari (2):
      spi: qcom-geni: trace: Add trace events for Qualcomm GENI SPI
      spi: qcom-geni: Add trace events for Qualcomm GENI SPI driver

 drivers/spi/spi-geni-qcom.c          |  13 +++++
 include/trace/events/qcom_geni_spi.h | 103 +++++++++++++++++++++++++++++++++++
 2 files changed, 116 insertions(+)
---
base-commit: 1f5ffc672165ff851063a5fd044b727ab2517ae3
change-id: 20260506-add-tracepoints-for-qcom-geni-spi-e31457c2267c

Best regards,
-- 
Praveen Talari <praveen.talari@oss.qualcomm.com>

[PATCH v3 1/2] spi: qcom-geni: trace: Add trace events for Qualcomm GENI SPI

[Praveen Talari]

Add tracepoint support to the Qualcomm GENI SPI driver to provide
runtime visibility into driver behavior without requiring invasive debug
patches.

The trace events cover clock and setup parameter configuration,
transfer metadata, interrupt status to be making it easier to diagnose
communication issues in the field..

Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Signed-off-by: Praveen Talari <praveen.talari@oss.qualcomm.com>
---
v2->v3:
- Renamed geni_spi_fifo_params to geni_spi_setup_params trace event.
- Updated commit text.

v1->v2:
- Removed TX/RX data tracepoints.
- Updated commit text.
---
 include/trace/events/qcom_geni_spi.h | 103 +++++++++++++++++++++++++++++++++++
 1 file changed, 103 insertions(+)

diff --git a/include/trace/events/qcom_geni_spi.h b/include/trace/events/qcom_geni_spi.h
new file mode 100644
index 000000000000..6d027adf2e1d
--- /dev/null
+++ b/include/trace/events/qcom_geni_spi.h
@@ -0,0 +1,103 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM qcom_geni_spi
+
+#if !defined(_TRACE_QCOM_GENI_SPI_H) || defined(TRACE_HEADER_MULTI_READ)
+#define _TRACE_QCOM_GENI_SPI_H
+
+#include <linux/tracepoint.h>
+
+TRACE_EVENT(geni_spi_setup_params,
+	    TP_PROTO(struct device *dev, u8 cs, u32 mode,
+		     u32 mode_changed, bool cs_changed),
+	    TP_ARGS(dev, cs, mode, mode_changed, cs_changed),
+
+	    TP_STRUCT__entry(__string(name, dev_name(dev))
+			     __field(u8, cs)
+			     __field(u32, mode)
+			     __field(u32, mode_changed)
+			     __field(bool, cs_changed)
+	    ),
+
+	    TP_fast_assign(__assign_str(name);
+			   __entry->cs = cs;
+			   __entry->mode = mode;
+			   __entry->mode_changed = mode_changed;
+			   __entry->cs_changed = cs_changed;
+	    ),
+
+	    TP_printk("%s: cs=%u mode=0x%08x mode_changed=0x%08x cs_changed=%d",
+		      __get_str(name), __entry->cs, __entry->mode,
+		      __entry->mode_changed, __entry->cs_changed)
+);
+
+TRACE_EVENT(geni_spi_clk_cfg,
+	    TP_PROTO(struct device *dev, unsigned long req_hz,
+		     unsigned long sclk_hz, unsigned int clk_idx,
+		     unsigned int clk_div, unsigned int bpw),
+	    TP_ARGS(dev, req_hz, sclk_hz, clk_idx, clk_div, bpw),
+
+	    TP_STRUCT__entry(__string(name, dev_name(dev))
+			     __field(unsigned long, req_hz)
+			     __field(unsigned long, sclk_hz)
+			     __field(unsigned int, clk_idx)
+			     __field(unsigned int, clk_div)
+			     __field(unsigned int, bpw)
+	    ),
+
+	    TP_fast_assign(__assign_str(name);
+			   __entry->req_hz = req_hz;
+			   __entry->sclk_hz = sclk_hz;
+			   __entry->clk_idx = clk_idx;
+			   __entry->clk_div = clk_div;
+			   __entry->bpw = bpw;
+	    ),
+
+	    TP_printk("%s: req_hz=%lu sclk_hz=%lu clk_idx=%u clk_div=%u bpw=%u",
+		      __get_str(name), __entry->req_hz, __entry->sclk_hz,
+		      __entry->clk_idx, __entry->clk_div, __entry->bpw)
+);
+
+TRACE_EVENT(geni_spi_transfer,
+	    TP_PROTO(struct device *dev, unsigned int len, u32 m_cmd),
+	    TP_ARGS(dev, len, m_cmd),
+
+	    TP_STRUCT__entry(__string(name, dev_name(dev))
+			     __field(unsigned int, len)
+			     __field(u32, m_cmd)
+	    ),
+
+	    TP_fast_assign(__assign_str(name);
+			   __entry->len = len;
+			   __entry->m_cmd = m_cmd;
+	    ),
+
+	    TP_printk("%s: len=%u m_cmd=0x%08x",
+		      __get_str(name), __entry->len, __entry->m_cmd)
+);
+
+TRACE_EVENT(geni_spi_irq,
+	    TP_PROTO(struct device *dev, u32 m_irq, u32 dma_tx, u32 dma_rx),
+	    TP_ARGS(dev, m_irq, dma_tx, dma_rx),
+
+	    TP_STRUCT__entry(__string(name, dev_name(dev))
+			     __field(u32, m_irq)
+			     __field(u32, dma_tx)
+			     __field(u32, dma_rx)
+	    ),
+
+	    TP_fast_assign(__assign_str(name);
+			   __entry->m_irq = m_irq;
+			   __entry->dma_tx = dma_tx;
+			   __entry->dma_rx = dma_rx;
+	    ),
+
+	    TP_printk("%s: m_irq=0x%08x dma_tx=0x%08x dma_rx=0x%08x",
+		      __get_str(name), __entry->m_irq, __entry->dma_tx,
+		      __entry->dma_rx)
+);
+
+#endif /* _TRACE_QCOM_GENI_SPI_H */
+
+/* This part must be outside protection */
+#include <trace/define_trace.h>

-- 
2.34.1

Re: [PATCH v3 1/2] spi: qcom-geni: trace: Add trace events for Qualcomm GENI SPI

[Mukesh Savaliya]

On 5/18/2026 10:30 PM, Praveen Talari wrote:
> Add tracepoint support to the Qualcomm GENI SPI driver to provide
> runtime visibility into driver behavior without requiring invasive debug
> patches.
> 
> The trace events cover clock and setup parameter configuration,
> transfer metadata, interrupt status to be making it easier to diagnose
> communication issues in the field..
> 
> Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
> Signed-off-by: Praveen Talari <praveen.talari@oss.qualcomm.com>
> ---
> v2->v3:
> - Renamed geni_spi_fifo_params to geni_spi_setup_params trace event.
> - Updated commit text.
> 
Reviewed-by: Mukesh Kumar Savaliya <mukesh.savaliya@oss.qualcomm.com>

Re: [PATCH v3 1/2] spi: qcom-geni: trace: Add trace events for Qualcomm GENI SPI

[Steven Rostedt]

On Mon, 18 May 2026 22:30:51 +0530
Praveen Talari <praveen.talari@oss.qualcomm.com> wrote:

> @@ -0,0 +1,103 @@
> +/* SPDX-License-Identifier: GPL-2.0 */
> +#undef TRACE_SYSTEM
> +#define TRACE_SYSTEM qcom_geni_spi
> +
> +#if !defined(_TRACE_QCOM_GENI_SPI_H) || defined(TRACE_HEADER_MULTI_READ)
> +#define _TRACE_QCOM_GENI_SPI_H
> +
> +#include <linux/tracepoint.h>
> +
> +TRACE_EVENT(geni_spi_setup_params,
> +	    TP_PROTO(struct device *dev, u8 cs, u32 mode,
> +		     u32 mode_changed, bool cs_changed),
> +	    TP_ARGS(dev, cs, mode, mode_changed, cs_changed),
> +
> +	    TP_STRUCT__entry(__string(name, dev_name(dev))
> +			     __field(u8, cs)

A u8 followed by a u32 will create a 3 byte hole in the structure
layout that gets recorded onto the ring buffer. Best to move that field
to after the bool cs_changed, for better compaction.

> +			     __field(u32, mode)
> +			     __field(u32, mode_changed)
> +			     __field(bool, cs_changed)
> +	    ),
> +
> +	    TP_fast_assign(__assign_str(name);
> +			   __entry->cs = cs;
> +			   __entry->mode = mode;
> +			   __entry->mode_changed = mode_changed;
> +			   __entry->cs_changed = cs_changed;
> +	    ),
> +
> +	    TP_printk("%s: cs=%u mode=0x%08x mode_changed=0x%08x cs_changed=%d",
> +		      __get_str(name), __entry->cs, __entry->mode,
> +		      __entry->mode_changed, __entry->cs_changed)
> +);
> +
> +TRACE_EVENT(geni_spi_clk_cfg,
> +	    TP_PROTO(struct device *dev, unsigned long req_hz,
> +		     unsigned long sclk_hz, unsigned int clk_idx,
> +		     unsigned int clk_div, unsigned int bpw),
> +	    TP_ARGS(dev, req_hz, sclk_hz, clk_idx, clk_div, bpw),
> +
> +	    TP_STRUCT__entry(__string(name, dev_name(dev))

__string items inject a 4 byte meta data so they are basically the same
as a u32 item on the structure. Move this to the end or after the long
words so that on 64 bit architectures you don't create a 4 byte hole
here.

-- Steve

> +			     __field(unsigned long, req_hz)
> +			     __field(unsigned long, sclk_hz)
> +			     __field(unsigned int, clk_idx)
> +			     __field(unsigned int, clk_div)
> +			     __field(unsigned int, bpw)
> +	    ),
> +
> +	    TP_fast_assign(__assign_str(name);
> +			   __entry->req_hz = req_hz;
> +			   __entry->sclk_hz = sclk_hz;
> +			   __entry->clk_idx = clk_idx;
> +			   __entry->clk_div = clk_div;
> +			   __entry->bpw = bpw;
> +	    ),
> +
> +	    TP_printk("%s: req_hz=%lu sclk_hz=%lu clk_idx=%u clk_div=%u bpw=%u",
> +		      __get_str(name), __entry->req_hz, __entry->sclk_hz,
> +		      __entry->clk_idx, __entry->clk_div, __entry->bpw)
> +);

[PATCH v3 2/2] spi: qcom-geni: Add trace events for Qualcomm GENI SPI driver

[Praveen Talari]

Add tracepoints to the Qualcomm GENI (Generic Interface) SPI driver.
These trace events enable runtime debugging and performance analysis
of SPI operations.

The trace events capture SPI clock configuration, setup parameters,
transfer details, interrupt status.

Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
Signed-off-by: Praveen Talari <praveen.talari@oss.qualcomm.com>
---
v2->v3:
- Replaced geni_spi_fifo_params with geni_spi_setup_params trace event.
- Updated commit text.

v1->v2:
- Removed tx/rx data capture since spi core had already support.
- Updated commit text.
---
 drivers/spi/spi-geni-qcom.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)

diff --git a/drivers/spi/spi-geni-qcom.c b/drivers/spi/spi-geni-qcom.c
index d5fb0edc8e0c..a04cdc1e5ad4 100644
--- a/drivers/spi/spi-geni-qcom.c
+++ b/drivers/spi/spi-geni-qcom.c
@@ -1,6 +1,9 @@
 // SPDX-License-Identifier: GPL-2.0
 // Copyright (c) 2017-2018, The Linux foundation. All rights reserved.
 
+#define CREATE_TRACE_POINTS
+#include <trace/events/qcom_geni_spi.h>
+
 #include <linux/clk.h>
 #include <linux/dmaengine.h>
 #include <linux/dma-mapping.h>
@@ -332,6 +335,9 @@ static int geni_spi_set_clock_and_bw(struct spi_geni_master *mas,
 	writel(clk_sel, se->base + SE_GENI_CLK_SEL);
 	writel(m_clk_cfg, se->base + GENI_SER_M_CLK_CFG);
 
+	trace_geni_spi_clk_cfg(mas->dev, clk_hz, mas->cur_sclk_hz, idx, div,
+			       mas->cur_bits_per_word);
+
 	/* Set BW quota for CPU as driver supports FIFO mode only. */
 	se->icc_paths[CPU_TO_GENI].avg_bw = Bps_to_icc(mas->cur_speed_hz);
 	ret = geni_icc_set_bw(se);
@@ -366,6 +372,9 @@ static int setup_fifo_params(struct spi_device *spi_slv,
 	if ((mode_changed & SPI_CS_HIGH) || (cs_changed && (spi_slv->mode & SPI_CS_HIGH)))
 		writel((spi_slv->mode & SPI_CS_HIGH) ? BIT(chipselect) : 0, se->base + SE_SPI_DEMUX_OUTPUT_INV);
 
+	trace_geni_spi_setup_params(mas->dev, chipselect, spi_slv->mode,
+				    mode_changed, cs_changed);
+
 	return 0;
 }
 
@@ -861,6 +870,8 @@ static int setup_se_xfer(struct spi_transfer *xfer,
 	spin_lock_irq(&mas->lock);
 	geni_se_setup_m_cmd(se, m_cmd, m_params);
 
+	trace_geni_spi_transfer(mas->dev, len, m_cmd);
+
 	if (mas->cur_xfer_mode == GENI_SE_DMA) {
 		if (m_cmd & SPI_RX_ONLY)
 			geni_se_rx_init_dma(se, sg_dma_address(xfer->rx_sg.sgl),
@@ -915,6 +926,8 @@ static irqreturn_t geni_spi_isr(int irq, void *data)
 	if (!m_irq && !dma_tx_status && !dma_rx_status)
 		return IRQ_NONE;
 
+	trace_geni_spi_irq(mas->dev, m_irq, dma_tx_status, dma_rx_status);
+
 	if (m_irq & (M_CMD_OVERRUN_EN | M_ILLEGAL_CMD_EN | M_CMD_FAILURE_EN |
 		     M_RX_FIFO_RD_ERR_EN | M_RX_FIFO_WR_ERR_EN |
 		     M_TX_FIFO_RD_ERR_EN | M_TX_FIFO_WR_ERR_EN))

-- 
2.34.1

Re: [PATCH v3 2/2] spi: qcom-geni: Add trace events for Qualcomm GENI SPI driver

[Mukesh Savaliya]

Hi Praveen, one question below.

On 5/18/2026 10:30 PM, Praveen Talari wrote:
> Add tracepoints to the Qualcomm GENI (Generic Interface) SPI driver.
> These trace events enable runtime debugging and performance analysis
> of SPI operations.
> 
> The trace events capture SPI clock configuration, setup parameters,
> transfer details, interrupt status.
Don't you need trace logs around PM operations ? ie. runtime and system PM ?
> 
> Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
> Signed-off-by: Praveen Talari <praveen.talari@oss.qualcomm.com>
> ---

[...]

Re: [PATCH v3 2/2] spi: qcom-geni: Add trace events for Qualcomm GENI SPI driver

[Konrad Dybcio]

On 5/19/26 12:59 PM, Mukesh Savaliya wrote:
> Hi Praveen, one question below.
> 
> On 5/18/2026 10:30 PM, Praveen Talari wrote:
>> Add tracepoints to the Qualcomm GENI (Generic Interface) SPI driver.
>> These trace events enable runtime debugging and performance analysis
>> of SPI operations.
>>
>> The trace events capture SPI clock configuration, setup parameters,
>> transfer details, interrupt status.
> Don't you need trace logs around PM operations ? ie. runtime and system PM ?

The PM core provides a couple, are those enough?

Konrad