[PATCH v6 0/5] drm/xe: i2c support

[PATCH v6 0/5] drm/xe: i2c support

[Heikki Krogerus]

Hi,

Changed since v5:
- The modification to the dw_i2c_plat_probe() is split into its own
  patch as proposed by Andi.
- I removed completely the comment from the last patch in the series
  ("drm/xe/xe_i2c: Add support for i2c in survivability mode").

Changed since v4:
- Cleanups requested by Andy.
- Casting the PCI power modes to make sparse happy - Raag.
- Limiting the use of this thing to Battlemage again. But I'm not sure
  if this is the correct thing to do.

Changed since v3:
- Cleanups as requested by Rodrigo.
- The licence is now changed, but still need confirmation for it!

Changed since v2:
- Added dependency on regmap when i2c is enabled.

Changed since v1:
- Now rebased on top of drm-tip.
- No longer ignoring errors from xe_i2c_probe().
- Cleanups pointed out by Lucas.

I've also included followup patches from Raag and Riana to this
series.

Original cover letter:

Some of the future GPUs will provide access to the on-board Synopsys
DesignWare I2C host adapter. The i2c is used to connect various
microcontrollers. The initially supported microcontroller unit is
called Add-In Management Controller (AMC).

Thanks,

Heikki Krogerus (3):
  i2c: designware: Use polling by default when there is no irq resource
  i2c: designware: Add quirk for Intel Xe
  drm/xe: Support for I2C attached MCUs

Raag Jadav (1):
  drm/xe/pm: Wire up suspend/resume for I2C controller

Riana Tauro (1):
  drm/xe/xe_i2c: Add support for i2c in survivability mode

 drivers/gpu/drm/xe/Kconfig                  |   1 +
 drivers/gpu/drm/xe/Makefile                 |   1 +
 drivers/gpu/drm/xe/regs/xe_i2c_regs.h       |  20 ++
 drivers/gpu/drm/xe/regs/xe_irq_regs.h       |   1 +
 drivers/gpu/drm/xe/regs/xe_pmt.h            |   2 +-
 drivers/gpu/drm/xe/regs/xe_regs.h           |   2 +
 drivers/gpu/drm/xe/xe_device.c              |   5 +
 drivers/gpu/drm/xe/xe_device_types.h        |   4 +
 drivers/gpu/drm/xe/xe_i2c.c                 | 329 ++++++++++++++++++++
 drivers/gpu/drm/xe/xe_i2c.h                 |  62 ++++
 drivers/gpu/drm/xe/xe_irq.c                 |   2 +
 drivers/gpu/drm/xe/xe_pm.c                  |   9 +
 drivers/gpu/drm/xe/xe_survivability_mode.c  |  19 +-
 drivers/i2c/busses/i2c-designware-platdrv.c |  18 +-
 14 files changed, 460 insertions(+), 15 deletions(-)
 create mode 100644 drivers/gpu/drm/xe/regs/xe_i2c_regs.h
 create mode 100644 drivers/gpu/drm/xe/xe_i2c.c
 create mode 100644 drivers/gpu/drm/xe/xe_i2c.h

-- 
2.47.2

[PATCH v6 1/5] i2c: designware: Use polling by default when there is no irq resource

[Heikki Krogerus]

The irq resource itself can be used as a generic way to
determine when polling is needed.

This not only removes the need for special additional device
properties that would soon be needed when the platform may
or may not have the irq, but it also removes the need to
check the platform in the first place in order to determine
is polling needed or not.

Signed-off-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
---
Hi guys,

I found the thread with Jarkko's comments from my archives. He wanted
the local flags variable to be added because he wants the order of the
calls to remain as it is now - the device is allocated only after the
irq is checked.

thanks,
---
 drivers/i2c/busses/i2c-designware-platdrv.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/drivers/i2c/busses/i2c-designware-platdrv.c b/drivers/i2c/busses/i2c-designware-platdrv.c
index 879719e91df2..3104f52e32be 100644
--- a/drivers/i2c/busses/i2c-designware-platdrv.c
+++ b/drivers/i2c/busses/i2c-designware-platdrv.c
@@ -205,25 +205,28 @@ static void i2c_dw_remove_lock_support(struct dw_i2c_dev *dev)
 
 static int dw_i2c_plat_probe(struct platform_device *pdev)
 {
+	u32 flags = (uintptr_t)device_get_match_data(&pdev->dev);
 	struct device *device = &pdev->dev;
 	struct i2c_adapter *adap;
 	struct dw_i2c_dev *dev;
 	int irq, ret;
 
-	irq = platform_get_irq(pdev, 0);
-	if (irq < 0)
+	irq = platform_get_irq_optional(pdev, 0);
+	if (irq == -ENXIO)
+		flags |= ACCESS_POLLING;
+	else if (irq < 0)
 		return irq;
 
 	dev = devm_kzalloc(device, sizeof(*dev), GFP_KERNEL);
 	if (!dev)
 		return -ENOMEM;
 
-	dev->flags = (uintptr_t)device_get_match_data(device);
 	if (device_property_present(device, "wx,i2c-snps-model"))
-		dev->flags = MODEL_WANGXUN_SP | ACCESS_POLLING;
+		flags = MODEL_WANGXUN_SP | ACCESS_POLLING;
 
 	dev->dev = device;
 	dev->irq = irq;
+	dev->flags = flags;
 	platform_set_drvdata(pdev, dev);
 
 	ret = dw_i2c_plat_request_regs(dev);
-- 
2.47.2

Re: [PATCH v6 1/5] i2c: designware: Use polling by default when there is no irq resource

[Andy Shevchenko]

On Tue, Jul 01, 2025 at 03:22:48PM +0300, Heikki Krogerus wrote:
> The irq resource itself can be used as a generic way to
> determine when polling is needed.
> 
> This not only removes the need for special additional device
> properties that would soon be needed when the platform may
> or may not have the irq, but it also removes the need to
> check the platform in the first place in order to determine
> is polling needed or not.

> Signed-off-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
> ---
> Hi guys,
> 
> I found the thread with Jarkko's comments from my archives. He wanted
> the local flags variable to be added because he wants the order of the
> calls to remain as it is now - the device is allocated only after the
> irq is checked.

Yes, thanks.

...

> +	u32 flags = (uintptr_t)device_get_match_data(&pdev->dev);

> +	irq = platform_get_irq_optional(pdev, 0);
> +	if (irq == -ENXIO)
> +		flags |= ACCESS_POLLING;
> +	else if (irq < 0)
>  		return irq;

>  	if (device_property_present(device, "wx,i2c-snps-model"))
> +		flags = MODEL_WANGXUN_SP | ACCESS_POLLING;

Now I'm a bit puzzled why do we need to add this flag explicitly here?
Does Wnagxun provides an IRQ and chooses at the same time to poll?
Shouldn't this patch rather fix that?

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH v6 1/5] i2c: designware: Use polling by default when there is no irq resource

[Heikki Krogerus]

On Tue, Jul 01, 2025 at 03:42:42PM +0300, Andy Shevchenko wrote:
> On Tue, Jul 01, 2025 at 03:22:48PM +0300, Heikki Krogerus wrote:
> > The irq resource itself can be used as a generic way to
> > determine when polling is needed.
> > 
> > This not only removes the need for special additional device
> > properties that would soon be needed when the platform may
> > or may not have the irq, but it also removes the need to
> > check the platform in the first place in order to determine
> > is polling needed or not.
> 
> > Signed-off-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
> > ---
> > Hi guys,
> > 
> > I found the thread with Jarkko's comments from my archives. He wanted
> > the local flags variable to be added because he wants the order of the
> > calls to remain as it is now - the device is allocated only after the
> > irq is checked.
> 
> Yes, thanks.
> 
> ...
> 
> > +	u32 flags = (uintptr_t)device_get_match_data(&pdev->dev);
> 
> > +	irq = platform_get_irq_optional(pdev, 0);
> > +	if (irq == -ENXIO)
> > +		flags |= ACCESS_POLLING;
> > +	else if (irq < 0)
> >  		return irq;
> 
> >  	if (device_property_present(device, "wx,i2c-snps-model"))
> > +		flags = MODEL_WANGXUN_SP | ACCESS_POLLING;
> 
> Now I'm a bit puzzled why do we need to add this flag explicitly here?
> Does Wnagxun provides an IRQ and chooses at the same time to poll?
> Shouldn't this patch rather fix that?

No. I do not want to touch the behavior here. The flags were
overwritten and continue to be overwritten.

I will propose an improvement for that together with some other
modifications to this file later, but those are out side the scope of
this series.

thanks,

-- 
heikki

Re: [PATCH v6 1/5] i2c: designware: Use polling by default when there is no irq resource

[Andy Shevchenko]

On Tue, Jul 01, 2025 at 03:55:08PM +0300, Heikki Krogerus wrote:
> On Tue, Jul 01, 2025 at 03:42:42PM +0300, Andy Shevchenko wrote:
> > On Tue, Jul 01, 2025 at 03:22:48PM +0300, Heikki Krogerus wrote:

...

> > > +	u32 flags = (uintptr_t)device_get_match_data(&pdev->dev);
> > 
> > > +	irq = platform_get_irq_optional(pdev, 0);
> > > +	if (irq == -ENXIO)
> > > +		flags |= ACCESS_POLLING;
> > > +	else if (irq < 0)
> > >  		return irq;
> > 
> > >  	if (device_property_present(device, "wx,i2c-snps-model"))
> > > +		flags = MODEL_WANGXUN_SP | ACCESS_POLLING;
> > 
> > Now I'm a bit puzzled why do we need to add this flag explicitly here?
> > Does Wnagxun provides an IRQ and chooses at the same time to poll?
> > Shouldn't this patch rather fix that?
> 
> No. I do not want to touch the behavior here. The flags were
> overwritten and continue to be overwritten.
> 
> I will propose an improvement for that together with some other
> modifications to this file later, but those are out side the scope of
> this series.

Sure.

-- 
With Best Regards,
Andy Shevchenko

Re: [PATCH v6 1/5] i2c: designware: Use polling by default when there is no irq resource

[Andi Shyti]

Hi Heikki,

On Tue, Jul 01, 2025 at 03:22:48PM +0300, Heikki Krogerus wrote:
> The irq resource itself can be used as a generic way to
> determine when polling is needed.
> 
> This not only removes the need for special additional device
> properties that would soon be needed when the platform may
> or may not have the irq, but it also removes the need to
> check the platform in the first place in order to determine
> is polling needed or not.
> 
> Signed-off-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>

Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>

Thanks,
Andi

[PATCH v6 2/5] i2c: designware: Add quirk for Intel Xe

[Heikki Krogerus]

The regmap is coming from the parent also in case of Xe
GPUs. Reusing the Wangxun quirk for that.

Originally-by: Michael J. Ruhl <michael.j.ruhl@intel.com>
Signed-off-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
---
 drivers/i2c/busses/i2c-designware-platdrv.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/drivers/i2c/busses/i2c-designware-platdrv.c b/drivers/i2c/busses/i2c-designware-platdrv.c
index 3104f52e32be..a35e4c64a1d4 100644
--- a/drivers/i2c/busses/i2c-designware-platdrv.c
+++ b/drivers/i2c/busses/i2c-designware-platdrv.c
@@ -101,7 +101,7 @@ static int bt1_i2c_request_regs(struct dw_i2c_dev *dev)
 }
 #endif
 
-static int txgbe_i2c_request_regs(struct dw_i2c_dev *dev)
+static int dw_i2c_get_parent_regmap(struct dw_i2c_dev *dev)
 {
 	dev->map = dev_get_regmap(dev->dev->parent, NULL);
 	if (!dev->map)
@@ -123,12 +123,15 @@ static int dw_i2c_plat_request_regs(struct dw_i2c_dev *dev)
 	struct platform_device *pdev = to_platform_device(dev->dev);
 	int ret;
 
+	if (device_is_compatible(dev->dev, "intel,xe-i2c"))
+		return dw_i2c_get_parent_regmap(dev);
+
 	switch (dev->flags & MODEL_MASK) {
 	case MODEL_BAIKAL_BT1:
 		ret = bt1_i2c_request_regs(dev);
 		break;
 	case MODEL_WANGXUN_SP:
-		ret = txgbe_i2c_request_regs(dev);
+		ret = dw_i2c_get_parent_regmap(dev);
 		break;
 	default:
 		dev->base = devm_platform_ioremap_resource(pdev, 0);
-- 
2.47.2

Re: [PATCH v6 2/5] i2c: designware: Add quirk for Intel Xe

[Andi Shyti]

Hi Heikki,

On Tue, Jul 01, 2025 at 03:22:49PM +0300, Heikki Krogerus wrote:
> The regmap is coming from the parent also in case of Xe
> GPUs. Reusing the Wangxun quirk for that.
> 
> Originally-by: Michael J. Ruhl <michael.j.ruhl@intel.com>
> Signed-off-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>

Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>

Thanks,
Andi

[PATCH v6 3/5] drm/xe: Support for I2C attached MCUs

[Heikki Krogerus]

Adding adaption/glue layer where the I2C host adapter
(Synopsys DesignWare I2C adapter) and the I2C clients (the
microcontroller units) are enumerated.

The microcontroller units (MCU) that are attached to the GPU
depend on the OEM. The initially supported MCU will be the
Add-In Management Controller (AMC).

Originally-by: Michael J. Ruhl <michael.j.ruhl@intel.com>
Signed-off-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
---
 drivers/gpu/drm/xe/Kconfig            |   1 +
 drivers/gpu/drm/xe/Makefile           |   1 +
 drivers/gpu/drm/xe/regs/xe_i2c_regs.h |  15 ++
 drivers/gpu/drm/xe/regs/xe_irq_regs.h |   1 +
 drivers/gpu/drm/xe/regs/xe_pmt.h      |   2 +-
 drivers/gpu/drm/xe/regs/xe_regs.h     |   2 +
 drivers/gpu/drm/xe/xe_device.c        |   5 +
 drivers/gpu/drm/xe/xe_device_types.h  |   4 +
 drivers/gpu/drm/xe/xe_i2c.c           | 300 ++++++++++++++++++++++++++
 drivers/gpu/drm/xe/xe_i2c.h           |  58 +++++
 drivers/gpu/drm/xe/xe_irq.c           |   2 +
 11 files changed, 390 insertions(+), 1 deletion(-)
 create mode 100644 drivers/gpu/drm/xe/regs/xe_i2c_regs.h
 create mode 100644 drivers/gpu/drm/xe/xe_i2c.c
 create mode 100644 drivers/gpu/drm/xe/xe_i2c.h

diff --git a/drivers/gpu/drm/xe/Kconfig b/drivers/gpu/drm/xe/Kconfig
index 827bc75b8b48..e78b3f46c846 100644
--- a/drivers/gpu/drm/xe/Kconfig
+++ b/drivers/gpu/drm/xe/Kconfig
@@ -45,6 +45,7 @@ config DRM_XE
 	select WANT_DEV_COREDUMP
 	select AUXILIARY_BUS
 	select HMM_MIRROR
+	select REGMAP if I2C
 	help
 	  Driver for Intel Xe2 series GPUs and later. Experimental support
 	  for Xe series is also available.
diff --git a/drivers/gpu/drm/xe/Makefile b/drivers/gpu/drm/xe/Makefile
index 1d97e5b63f4e..1201b8ab4702 100644
--- a/drivers/gpu/drm/xe/Makefile
+++ b/drivers/gpu/drm/xe/Makefile
@@ -125,6 +125,7 @@ xe-y += xe_bb.o \
 	xe_wait_user_fence.o \
 	xe_wopcm.o
 
+xe-$(CONFIG_I2C)	+= xe_i2c.o
 xe-$(CONFIG_HMM_MIRROR) += xe_hmm.o
 xe-$(CONFIG_DRM_XE_GPUSVM) += xe_svm.o
 
diff --git a/drivers/gpu/drm/xe/regs/xe_i2c_regs.h b/drivers/gpu/drm/xe/regs/xe_i2c_regs.h
new file mode 100644
index 000000000000..92dae4487614
--- /dev/null
+++ b/drivers/gpu/drm/xe/regs/xe_i2c_regs.h
@@ -0,0 +1,15 @@
+/* SPDX-License-Identifier: MIT */
+#ifndef _XE_I2C_REGS_H_
+#define _XE_I2C_REGS_H_
+
+#include "xe_reg_defs.h"
+#include "xe_regs.h"
+
+#define I2C_BRIDGE_OFFSET		(SOC_BASE + 0xd9000)
+#define I2C_CONFIG_SPACE_OFFSET		(SOC_BASE + 0xf6000)
+#define I2C_MEM_SPACE_OFFSET		(SOC_BASE + 0xf7400)
+
+#define REG_SG_REMAP_ADDR_PREFIX	XE_REG(SOC_BASE + 0x0164)
+#define REG_SG_REMAP_ADDR_POSTFIX	XE_REG(SOC_BASE + 0x0168)
+
+#endif /* _XE_I2C_REGS_H_ */
diff --git a/drivers/gpu/drm/xe/regs/xe_irq_regs.h b/drivers/gpu/drm/xe/regs/xe_irq_regs.h
index f0ecfcac4003..13635e4331d4 100644
--- a/drivers/gpu/drm/xe/regs/xe_irq_regs.h
+++ b/drivers/gpu/drm/xe/regs/xe_irq_regs.h
@@ -19,6 +19,7 @@
 #define   MASTER_IRQ				REG_BIT(31)
 #define   GU_MISC_IRQ				REG_BIT(29)
 #define   DISPLAY_IRQ				REG_BIT(16)
+#define   I2C_IRQ				REG_BIT(12)
 #define   GT_DW_IRQ(x)				REG_BIT(x)
 
 /*
diff --git a/drivers/gpu/drm/xe/regs/xe_pmt.h b/drivers/gpu/drm/xe/regs/xe_pmt.h
index b0efd9b48d1e..2995d72c3f78 100644
--- a/drivers/gpu/drm/xe/regs/xe_pmt.h
+++ b/drivers/gpu/drm/xe/regs/xe_pmt.h
@@ -5,7 +5,7 @@
 #ifndef _XE_PMT_H_
 #define _XE_PMT_H_
 
-#define SOC_BASE			0x280000
+#include "xe_regs.h"
 
 #define BMG_PMT_BASE_OFFSET		0xDB000
 #define BMG_DISCOVERY_OFFSET		(SOC_BASE + BMG_PMT_BASE_OFFSET)
diff --git a/drivers/gpu/drm/xe/regs/xe_regs.h b/drivers/gpu/drm/xe/regs/xe_regs.h
index 3abb17d2ca33..1926b4044314 100644
--- a/drivers/gpu/drm/xe/regs/xe_regs.h
+++ b/drivers/gpu/drm/xe/regs/xe_regs.h
@@ -7,6 +7,8 @@
 
 #include "regs/xe_reg_defs.h"
 
+#define SOC_BASE				0x280000
+
 #define GU_CNTL_PROTECTED			XE_REG(0x10100C)
 #define   DRIVERINT_FLR_DIS			REG_BIT(31)
 
diff --git a/drivers/gpu/drm/xe/xe_device.c b/drivers/gpu/drm/xe/xe_device.c
index 0b73cb72bad1..6db09cfc8eb8 100644
--- a/drivers/gpu/drm/xe/xe_device.c
+++ b/drivers/gpu/drm/xe/xe_device.c
@@ -43,6 +43,7 @@
 #include "xe_guc_pc.h"
 #include "xe_hw_engine_group.h"
 #include "xe_hwmon.h"
+#include "xe_i2c.h"
 #include "xe_irq.h"
 #include "xe_mmio.h"
 #include "xe_module.h"
@@ -902,6 +903,10 @@ int xe_device_probe(struct xe_device *xe)
 	if (err)
 		goto err_unregister_display;
 
+	err = xe_i2c_probe(xe);
+	if (err)
+		goto err_unregister_display;
+
 	for_each_gt(gt, xe, id)
 		xe_gt_sanitize_freq(gt);
 
diff --git a/drivers/gpu/drm/xe/xe_device_types.h b/drivers/gpu/drm/xe/xe_device_types.h
index 7e4f6d846af6..ce0c4ed834b8 100644
--- a/drivers/gpu/drm/xe/xe_device_types.h
+++ b/drivers/gpu/drm/xe/xe_device_types.h
@@ -34,6 +34,7 @@ struct dram_info;
 struct intel_display;
 struct intel_dg_nvm_dev;
 struct xe_ggtt;
+struct xe_i2c;
 struct xe_pat_ops;
 struct xe_pxp;
 
@@ -583,6 +584,9 @@ struct xe_device {
 	/** @pmu: performance monitoring unit */
 	struct xe_pmu pmu;
 
+	/** @i2c: I2C host controller */
+	struct xe_i2c *i2c;
+
 	/** @atomic_svm_timeslice_ms: Atomic SVM fault timeslice MS */
 	u32 atomic_svm_timeslice_ms;
 
diff --git a/drivers/gpu/drm/xe/xe_i2c.c b/drivers/gpu/drm/xe/xe_i2c.c
new file mode 100644
index 000000000000..ef91239879c0
--- /dev/null
+++ b/drivers/gpu/drm/xe/xe_i2c.c
@@ -0,0 +1,300 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT */
+/*
+ * Intel Xe I2C attached Microcontroller Units (MCU)
+ *
+ * Copyright (C) 2025 Intel Corporation.
+ */
+
+#include <linux/array_size.h>
+#include <linux/container_of.h>
+#include <linux/device.h>
+#include <linux/err.h>
+#include <linux/i2c.h>
+#include <linux/ioport.h>
+#include <linux/irq.h>
+#include <linux/irqdomain.h>
+#include <linux/notifier.h>
+#include <linux/pci.h>
+#include <linux/platform_device.h>
+#include <linux/property.h>
+#include <linux/regmap.h>
+#include <linux/sprintf.h>
+#include <linux/string.h>
+#include <linux/types.h>
+#include <linux/workqueue.h>
+
+#include "regs/xe_i2c_regs.h"
+#include "regs/xe_irq_regs.h"
+
+#include "xe_device.h"
+#include "xe_device_types.h"
+#include "xe_i2c.h"
+#include "xe_mmio.h"
+#include "xe_platform_types.h"
+
+/**
+ * DOC: Xe I2C devices
+ *
+ * Register a platform device for the I2C host controller (Synpsys DesignWare
+ * I2C) if the registers of that controller are mapped to the MMIO, and also the
+ * I2C client device for the Add-In Management Controller (the MCU) attached to
+ * the host controller.
+ *
+ * See drivers/i2c/busses/i2c-designware-* for more information on the I2C host
+ * controller.
+ */
+
+static const char adapter_name[] = "i2c_designware";
+
+static const struct property_entry xe_i2c_adapter_properties[] = {
+	PROPERTY_ENTRY_STRING("compatible", "intel,xe-i2c"),
+	PROPERTY_ENTRY_U32("clock-frequency", I2C_MAX_FAST_MODE_PLUS_FREQ),
+	{ }
+};
+
+static inline void xe_i2c_read_endpoint(struct xe_mmio *mmio, void *ep)
+{
+	u32 *val = ep;
+
+	val[0] = xe_mmio_read32(mmio, REG_SG_REMAP_ADDR_PREFIX);
+	val[1] = xe_mmio_read32(mmio, REG_SG_REMAP_ADDR_POSTFIX);
+}
+
+static void xe_i2c_client_work(struct work_struct *work)
+{
+	struct xe_i2c *i2c = container_of(work, struct xe_i2c, work);
+	struct i2c_board_info info = {
+		.type	= "amc",
+		.flags	= I2C_CLIENT_HOST_NOTIFY,
+		.addr	= i2c->ep.addr[1],
+	};
+
+	i2c->client[0] = i2c_new_client_device(i2c->adapter, &info);
+}
+
+static int xe_i2c_notifier(struct notifier_block *nb, unsigned long action, void *data)
+{
+	struct xe_i2c *i2c = container_of(nb, struct xe_i2c, bus_notifier);
+	struct i2c_adapter *adapter = i2c_verify_adapter(data);
+	struct device *dev = data;
+
+	if (action == BUS_NOTIFY_ADD_DEVICE &&
+	    adapter && dev->parent == &i2c->pdev->dev) {
+		i2c->adapter = adapter;
+		schedule_work(&i2c->work);
+		return NOTIFY_OK;
+	}
+
+	return NOTIFY_DONE;
+}
+
+static int xe_i2c_register_adapter(struct xe_i2c *i2c)
+{
+	struct pci_dev *pci = to_pci_dev(i2c->drm_dev);
+	struct platform_device *pdev;
+	struct fwnode_handle *fwnode;
+	int ret;
+
+	fwnode = fwnode_create_software_node(xe_i2c_adapter_properties, NULL);
+	if (!fwnode)
+		return -ENOMEM;
+
+	/*
+	 * Not using platform_device_register_full() here because we don't have
+	 * a handle to the platform_device before it returns. xe_i2c_notifier()
+	 * uses that handle, but it may be called before
+	 * platform_device_register_full() is done.
+	 */
+	pdev = platform_device_alloc(adapter_name, pci_dev_id(pci));
+	if (!pdev) {
+		ret = -ENOMEM;
+		goto err_fwnode_remove;
+	}
+
+	if (i2c->adapter_irq) {
+		struct resource res;
+
+		res = DEFINE_RES_IRQ_NAMED(i2c->adapter_irq, "xe_i2c");
+
+		ret = platform_device_add_resources(pdev, &res, 1);
+		if (ret)
+			goto err_pdev_put;
+	}
+
+	pdev->dev.parent = i2c->drm_dev;
+	pdev->dev.fwnode = fwnode;
+	i2c->adapter_node = fwnode;
+	i2c->pdev = pdev;
+
+	ret = platform_device_add(pdev);
+	if (ret)
+		goto err_pdev_put;
+
+	return 0;
+
+err_pdev_put:
+	platform_device_put(pdev);
+err_fwnode_remove:
+	fwnode_remove_software_node(fwnode);
+
+	return ret;
+}
+
+static void xe_i2c_unregister_adapter(struct xe_i2c *i2c)
+{
+	platform_device_unregister(i2c->pdev);
+	fwnode_remove_software_node(i2c->adapter_node);
+}
+
+/**
+ * xe_i2c_irq_handler: Handler for I2C interrupts
+ * @xe: xe device instance
+ * @master_ctl: interrupt register
+ *
+ * Forward interrupts generated by the I2C host adapter to the I2C host adapter
+ * driver.
+ */
+void xe_i2c_irq_handler(struct xe_device *xe, u32 master_ctl)
+{
+	if (!xe->i2c || !xe->i2c->adapter_irq)
+		return;
+
+	if (master_ctl & I2C_IRQ)
+		generic_handle_irq_safe(xe->i2c->adapter_irq);
+}
+
+static int xe_i2c_irq_map(struct irq_domain *h, unsigned int virq,
+			  irq_hw_number_t hw_irq_num)
+{
+	irq_set_chip_and_handler(virq, &dummy_irq_chip, handle_simple_irq);
+	return 0;
+}
+
+static const struct irq_domain_ops xe_i2c_irq_ops = {
+	.map = xe_i2c_irq_map,
+};
+
+static int xe_i2c_create_irq(struct xe_i2c *i2c)
+{
+	struct irq_domain *domain;
+
+	if (!(i2c->ep.capabilities & XE_I2C_EP_CAP_IRQ))
+		return 0;
+
+	domain = irq_domain_create_linear(dev_fwnode(i2c->drm_dev), 1, &xe_i2c_irq_ops, NULL);
+	if (!domain)
+		return -ENOMEM;
+
+	i2c->adapter_irq = irq_create_mapping(domain, 0);
+	i2c->irqdomain = domain;
+
+	return 0;
+}
+
+static void xe_i2c_remove_irq(struct xe_i2c *i2c)
+{
+	if (!i2c->irqdomain)
+		return;
+
+	irq_dispose_mapping(i2c->adapter_irq);
+	irq_domain_remove(i2c->irqdomain);
+}
+
+static int xe_i2c_read(void *context, unsigned int reg, unsigned int *val)
+{
+	struct xe_i2c *i2c = context;
+
+	*val = xe_mmio_read32(i2c->mmio, XE_REG(reg + I2C_MEM_SPACE_OFFSET));
+
+	return 0;
+}
+
+static int xe_i2c_write(void *context, unsigned int reg, unsigned int val)
+{
+	struct xe_i2c *i2c = context;
+
+	xe_mmio_write32(i2c->mmio, XE_REG(reg + I2C_MEM_SPACE_OFFSET), val);
+
+	return 0;
+}
+
+static const struct regmap_config i2c_regmap_config = {
+	.reg_bits = 32,
+	.val_bits = 32,
+	.reg_read = xe_i2c_read,
+	.reg_write = xe_i2c_write,
+	.fast_io = true,
+};
+
+static void xe_i2c_remove(void *data)
+{
+	struct xe_i2c *i2c = data;
+	unsigned int i;
+
+	for (i = 0; i < XE_I2C_MAX_CLIENTS; i++)
+		i2c_unregister_device(i2c->client[i]);
+
+	bus_unregister_notifier(&i2c_bus_type, &i2c->bus_notifier);
+	xe_i2c_unregister_adapter(i2c);
+	xe_i2c_remove_irq(i2c);
+}
+
+/**
+ * xe_i2c_probe: Probe the I2C host adapter and the I2C clients attached to it
+ * @xe: xe device instance
+ *
+ * Register all the I2C devices described in the I2C Endpoint data structure.
+ *
+ * Return: 0 on success, error code on failure
+ */
+int xe_i2c_probe(struct xe_device *xe)
+{
+	struct device *drm_dev = xe->drm.dev;
+	struct xe_i2c_endpoint ep;
+	struct regmap *regmap;
+	struct xe_i2c *i2c;
+	int ret;
+
+	if (xe->info.platform != XE_BATTLEMAGE)
+		return 0;
+
+	xe_i2c_read_endpoint(xe_root_tile_mmio(xe), &ep);
+	if (ep.cookie != XE_I2C_EP_COOKIE_DEVICE)
+		return 0;
+
+	i2c = devm_kzalloc(drm_dev, sizeof(*i2c), GFP_KERNEL);
+	if (!i2c)
+		return -ENOMEM;
+
+	INIT_WORK(&i2c->work, xe_i2c_client_work);
+	i2c->mmio = xe_root_tile_mmio(xe);
+	i2c->drm_dev = drm_dev;
+	i2c->ep = ep;
+
+	regmap = devm_regmap_init(drm_dev, NULL, i2c, &i2c_regmap_config);
+	if (IS_ERR(regmap))
+		return PTR_ERR(regmap);
+
+	i2c->bus_notifier.notifier_call = xe_i2c_notifier;
+	ret = bus_register_notifier(&i2c_bus_type, &i2c->bus_notifier);
+	if (ret)
+		return ret;
+
+	ret = xe_i2c_create_irq(i2c);
+	if (ret)
+		goto err_unregister_notifier;
+
+	ret = xe_i2c_register_adapter(i2c);
+	if (ret)
+		goto err_remove_irq;
+
+	return devm_add_action_or_reset(drm_dev, xe_i2c_remove, i2c);
+
+err_remove_irq:
+	xe_i2c_remove_irq(i2c);
+
+err_unregister_notifier:
+	bus_unregister_notifier(&i2c_bus_type, &i2c->bus_notifier);
+
+	return ret;
+}
diff --git a/drivers/gpu/drm/xe/xe_i2c.h b/drivers/gpu/drm/xe/xe_i2c.h
new file mode 100644
index 000000000000..7ea40f4e4aa4
--- /dev/null
+++ b/drivers/gpu/drm/xe/xe_i2c.h
@@ -0,0 +1,58 @@
+/* SPDX-License-Identifier: MIT */
+#ifndef _XE_I2C_H_
+#define _XE_I2C_H_
+
+#include <linux/bits.h>
+#include <linux/notifier.h>
+#include <linux/types.h>
+#include <linux/workqueue.h>
+
+struct device;
+struct fwnode_handle;
+struct i2c_adapter;
+struct i2c_client;
+struct irq_domain;
+struct platform_device;
+struct xe_device;
+struct xe_mmio;
+
+#define XE_I2C_MAX_CLIENTS		3
+
+#define XE_I2C_EP_COOKIE_DEVICE		0xde
+
+/* Endpoint Capabilities */
+#define XE_I2C_EP_CAP_IRQ		BIT(0)
+
+struct xe_i2c_endpoint {
+	u8 cookie;
+	u8 capabilities;
+	u16 addr[XE_I2C_MAX_CLIENTS];
+};
+
+struct xe_i2c {
+	struct fwnode_handle *adapter_node;
+	struct platform_device *pdev;
+	struct i2c_adapter *adapter;
+	struct i2c_client *client[XE_I2C_MAX_CLIENTS];
+
+	struct notifier_block bus_notifier;
+	struct work_struct work;
+
+	struct irq_domain *irqdomain;
+	int adapter_irq;
+
+	struct xe_i2c_endpoint ep;
+	struct device *drm_dev;
+
+	struct xe_mmio *mmio;
+};
+
+#if IS_ENABLED(CONFIG_I2C)
+int xe_i2c_probe(struct xe_device *xe);
+void xe_i2c_irq_handler(struct xe_device *xe, u32 master_ctl);
+#else
+static inline int xe_i2c_probe(struct xe_device *xe) { return 0; }
+static inline void xe_i2c_irq_handler(struct xe_device *xe, u32 master_ctl) { }
+#endif
+
+#endif
diff --git a/drivers/gpu/drm/xe/xe_irq.c b/drivers/gpu/drm/xe/xe_irq.c
index 5362d3174b06..c43e62dc692e 100644
--- a/drivers/gpu/drm/xe/xe_irq.c
+++ b/drivers/gpu/drm/xe/xe_irq.c
@@ -18,6 +18,7 @@
 #include "xe_gt.h"
 #include "xe_guc.h"
 #include "xe_hw_engine.h"
+#include "xe_i2c.h"
 #include "xe_memirq.h"
 #include "xe_mmio.h"
 #include "xe_pxp.h"
@@ -476,6 +477,7 @@ static irqreturn_t dg1_irq_handler(int irq, void *arg)
 			if (xe->info.has_heci_cscfi)
 				xe_heci_csc_irq_handler(xe, master_ctl);
 			xe_display_irq_handler(xe, master_ctl);
+			xe_i2c_irq_handler(xe, master_ctl);
 			gu_misc_iir = gu_misc_irq_ack(xe, master_ctl);
 		}
 	}
-- 
2.47.2

[PATCH v6 4/5] drm/xe/pm: Wire up suspend/resume for I2C controller

[Heikki Krogerus]

From: Raag Jadav <raag.jadav@intel.com>

Wire up suspend/resume handles for I2C controller to match its power
state with SGUnit.

Signed-off-by: Raag Jadav <raag.jadav@intel.com>
Signed-off-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Reviewed-by: Karthik Poosa <karthik.poosa@intel.com>
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
---
 drivers/gpu/drm/xe/regs/xe_i2c_regs.h |  5 +++++
 drivers/gpu/drm/xe/xe_i2c.c           | 29 +++++++++++++++++++++++++++
 drivers/gpu/drm/xe/xe_i2c.h           |  4 ++++
 drivers/gpu/drm/xe/xe_pm.c            |  9 +++++++++
 4 files changed, 47 insertions(+)

diff --git a/drivers/gpu/drm/xe/regs/xe_i2c_regs.h b/drivers/gpu/drm/xe/regs/xe_i2c_regs.h
index 92dae4487614..af781c8e4a80 100644
--- a/drivers/gpu/drm/xe/regs/xe_i2c_regs.h
+++ b/drivers/gpu/drm/xe/regs/xe_i2c_regs.h
@@ -2,6 +2,8 @@
 #ifndef _XE_I2C_REGS_H_
 #define _XE_I2C_REGS_H_
 
+#include <linux/pci_regs.h>
+
 #include "xe_reg_defs.h"
 #include "xe_regs.h"
 
@@ -12,4 +14,7 @@
 #define REG_SG_REMAP_ADDR_PREFIX	XE_REG(SOC_BASE + 0x0164)
 #define REG_SG_REMAP_ADDR_POSTFIX	XE_REG(SOC_BASE + 0x0168)
 
+#define I2C_CONFIG_CMD			XE_REG(I2C_CONFIG_SPACE_OFFSET + PCI_COMMAND)
+#define I2C_CONFIG_PMCSR		XE_REG(I2C_CONFIG_SPACE_OFFSET + 0x84)
+
 #endif /* _XE_I2C_REGS_H_ */
diff --git a/drivers/gpu/drm/xe/xe_i2c.c b/drivers/gpu/drm/xe/xe_i2c.c
index ef91239879c0..d395795d1e7e 100644
--- a/drivers/gpu/drm/xe/xe_i2c.c
+++ b/drivers/gpu/drm/xe/xe_i2c.c
@@ -226,6 +226,31 @@ static const struct regmap_config i2c_regmap_config = {
 	.fast_io = true,
 };
 
+void xe_i2c_pm_suspend(struct xe_device *xe)
+{
+	struct xe_mmio *mmio = xe_root_tile_mmio(xe);
+
+	if (!xe->i2c || xe->i2c->ep.cookie != XE_I2C_EP_COOKIE_DEVICE)
+		return;
+
+	xe_mmio_rmw32(mmio, I2C_CONFIG_PMCSR, PCI_PM_CTRL_STATE_MASK, (__force u32)PCI_D3hot);
+	drm_dbg(&xe->drm, "pmcsr: 0x%08x\n", xe_mmio_read32(mmio, I2C_CONFIG_PMCSR));
+}
+
+void xe_i2c_pm_resume(struct xe_device *xe, bool d3cold)
+{
+	struct xe_mmio *mmio = xe_root_tile_mmio(xe);
+
+	if (!xe->i2c || xe->i2c->ep.cookie != XE_I2C_EP_COOKIE_DEVICE)
+		return;
+
+	if (d3cold)
+		xe_mmio_rmw32(mmio, I2C_CONFIG_CMD, 0, PCI_COMMAND_MEMORY);
+
+	xe_mmio_rmw32(mmio, I2C_CONFIG_PMCSR, PCI_PM_CTRL_STATE_MASK, (__force u32)PCI_D0);
+	drm_dbg(&xe->drm, "pmcsr: 0x%08x\n", xe_mmio_read32(mmio, I2C_CONFIG_PMCSR));
+}
+
 static void xe_i2c_remove(void *data)
 {
 	struct xe_i2c *i2c = data;
@@ -270,6 +295,10 @@ int xe_i2c_probe(struct xe_device *xe)
 	i2c->mmio = xe_root_tile_mmio(xe);
 	i2c->drm_dev = drm_dev;
 	i2c->ep = ep;
+	xe->i2c = i2c;
+
+	/* PCI PM isn't aware of this device, bring it up and match it with SGUnit state. */
+	xe_i2c_pm_resume(xe, true);
 
 	regmap = devm_regmap_init(drm_dev, NULL, i2c, &i2c_regmap_config);
 	if (IS_ERR(regmap))
diff --git a/drivers/gpu/drm/xe/xe_i2c.h b/drivers/gpu/drm/xe/xe_i2c.h
index 7ea40f4e4aa4..b767ed8ce52b 100644
--- a/drivers/gpu/drm/xe/xe_i2c.h
+++ b/drivers/gpu/drm/xe/xe_i2c.h
@@ -50,9 +50,13 @@ struct xe_i2c {
 #if IS_ENABLED(CONFIG_I2C)
 int xe_i2c_probe(struct xe_device *xe);
 void xe_i2c_irq_handler(struct xe_device *xe, u32 master_ctl);
+void xe_i2c_pm_suspend(struct xe_device *xe);
+void xe_i2c_pm_resume(struct xe_device *xe, bool d3cold);
 #else
 static inline int xe_i2c_probe(struct xe_device *xe) { return 0; }
 static inline void xe_i2c_irq_handler(struct xe_device *xe, u32 master_ctl) { }
+static inline void xe_i2c_pm_suspend(struct xe_device *xe) { }
+static inline void xe_i2c_pm_resume(struct xe_device *xe, bool d3cold) { }
 #endif
 
 #endif
diff --git a/drivers/gpu/drm/xe/xe_pm.c b/drivers/gpu/drm/xe/xe_pm.c
index 26e95460af87..46471e166b96 100644
--- a/drivers/gpu/drm/xe/xe_pm.c
+++ b/drivers/gpu/drm/xe/xe_pm.c
@@ -19,6 +19,7 @@
 #include "xe_ggtt.h"
 #include "xe_gt.h"
 #include "xe_guc.h"
+#include "xe_i2c.h"
 #include "xe_irq.h"
 #include "xe_pcode.h"
 #include "xe_pxp.h"
@@ -146,6 +147,8 @@ int xe_pm_suspend(struct xe_device *xe)
 
 	xe_display_pm_suspend_late(xe);
 
+	xe_i2c_pm_suspend(xe);
+
 	drm_dbg(&xe->drm, "Device suspended\n");
 	return 0;
 
@@ -191,6 +194,8 @@ int xe_pm_resume(struct xe_device *xe)
 	if (err)
 		goto err;
 
+	xe_i2c_pm_resume(xe, xe->d3cold.allowed);
+
 	xe_irq_resume(xe);
 
 	for_each_gt(gt, xe, id)
@@ -484,6 +489,8 @@ int xe_pm_runtime_suspend(struct xe_device *xe)
 
 	xe_display_pm_runtime_suspend_late(xe);
 
+	xe_i2c_pm_suspend(xe);
+
 	xe_rpm_lockmap_release(xe);
 	xe_pm_write_callback_task(xe, NULL);
 	return 0;
@@ -531,6 +538,8 @@ int xe_pm_runtime_resume(struct xe_device *xe)
 			goto out;
 	}
 
+	xe_i2c_pm_resume(xe, xe->d3cold.allowed);
+
 	xe_irq_resume(xe);
 
 	for_each_gt(gt, xe, id)
-- 
2.47.2

[PATCH v6 5/5] drm/xe/xe_i2c: Add support for i2c in survivability mode

[Heikki Krogerus]

From: Riana Tauro <riana.tauro@intel.com>

Initialize i2c in survivability mode to allow firmware
update of Add-In Management Controller (AMC) in
survivability mode.

Signed-off-by: Riana Tauro <riana.tauro@intel.com>
Signed-off-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Reviewed-by: Raag Jadav <raag.jadav@intel.com>
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
---
 drivers/gpu/drm/xe/xe_survivability_mode.c | 19 +++++++++++--------
 1 file changed, 11 insertions(+), 8 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_survivability_mode.c b/drivers/gpu/drm/xe/xe_survivability_mode.c
index 1f710b3fc599..41705f5d52e3 100644
--- a/drivers/gpu/drm/xe/xe_survivability_mode.c
+++ b/drivers/gpu/drm/xe/xe_survivability_mode.c
@@ -14,6 +14,7 @@
 #include "xe_device.h"
 #include "xe_gt.h"
 #include "xe_heci_gsc.h"
+#include "xe_i2c.h"
 #include "xe_mmio.h"
 #include "xe_pcode_api.h"
 #include "xe_vsec.h"
@@ -173,20 +174,22 @@ static int enable_survivability_mode(struct pci_dev *pdev)
 	survivability->mode = true;
 
 	ret = xe_heci_gsc_init(xe);
-	if (ret) {
-		/*
-		 * But if it fails, device can't enter survivability
-		 * so move it back for correct error handling
-		 */
-		survivability->mode = false;
-		return ret;
-	}
+	if (ret)
+		goto err;
 
 	xe_vsec_init(xe);
 
+	ret = xe_i2c_probe(xe);
+	if (ret)
+		goto err;
+
 	dev_err(dev, "In Survivability Mode\n");
 
 	return 0;
+
+err:
+	survivability->mode = false;
+	return ret;
 }
 
 /**
-- 
2.47.2

Re: [PATCH v6 0/5] drm/xe: i2c support

[Andi Shyti]

Hi Heikki,

> Heikki Krogerus (3):
>   i2c: designware: Use polling by default when there is no irq resource
>   i2c: designware: Add quirk for Intel Xe

I think you could have kept Jarkko's ack here. Up to you.

Anyone against me taking the two above already?

Andi

>   drm/xe: Support for I2C attached MCUs
> 
> Raag Jadav (1):
>   drm/xe/pm: Wire up suspend/resume for I2C controller
> 
> Riana Tauro (1):
>   drm/xe/xe_i2c: Add support for i2c in survivability mode

Re: [PATCH v6 0/5] drm/xe: i2c support

[Rodrigo Vivi]

On Wed, Jul 02, 2025 at 07:18:50PM +0200, Andi Shyti wrote:
> Hi Heikki,
> 
> > Heikki Krogerus (3):
> >   i2c: designware: Use polling by default when there is no irq resource
> >   i2c: designware: Add quirk for Intel Xe
> 
> I think you could have kept Jarkko's ack here. Up to you.
> 
> Anyone against me taking the two above already?

I was thinking about keeping Jarkko's ack and get all merged at
once in drm-xe-next.

But first we need the SPDX headers in .c files fixed...

> 
> Andi
> 
> >   drm/xe: Support for I2C attached MCUs
> > 
> > Raag Jadav (1):
> >   drm/xe/pm: Wire up suspend/resume for I2C controller
> > 
> > Riana Tauro (1):
> >   drm/xe/xe_i2c: Add support for i2c in survivability mode

Re: [PATCH v6 0/5] drm/xe: i2c support

[Andi Shyti]

Hi Rodrigo,

On Wed, Jul 02, 2025 at 05:52:09PM -0400, Rodrigo Vivi wrote:
> On Wed, Jul 02, 2025 at 07:18:50PM +0200, Andi Shyti wrote:
> > > Heikki Krogerus (3):
> > >   i2c: designware: Use polling by default when there is no irq resource
> > >   i2c: designware: Add quirk for Intel Xe
> > 
> > I think you could have kept Jarkko's ack here. Up to you.
> > 
> > Anyone against me taking the two above already?
> 
> I was thinking about keeping Jarkko's ack and get all merged at
> once in drm-xe-next.

fine with me.

Andi

Re: [PATCH v6 0/5] drm/xe: i2c support

[Rodrigo Vivi]

On Thu, Jul 03, 2025 at 07:41:05PM +0200, Andi Shyti wrote:
> Hi Rodrigo,
> 
> On Wed, Jul 02, 2025 at 05:52:09PM -0400, Rodrigo Vivi wrote:
> > On Wed, Jul 02, 2025 at 07:18:50PM +0200, Andi Shyti wrote:
> > > > Heikki Krogerus (3):
> > > >   i2c: designware: Use polling by default when there is no irq resource
> > > >   i2c: designware: Add quirk for Intel Xe
> > > 
> > > I think you could have kept Jarkko's ack here. Up to you.
> > > 
> > > Anyone against me taking the two above already?
> > 
> > I was thinking about keeping Jarkko's ack and get all merged at
> > once in drm-xe-next.
> 
> fine with me.

pushed to drm-xe-next

I fixed the SPDX header and the tags while pushing it

thank you all!

> 
> Andi