Re: [PATCH v2 3/4] drm/xe/sysctrl: Add system controller event support

[Raag Jadav <raag.jadav@intel.com>]

On Tue, Mar 10, 2026 at 11:51:56AM +0530, Mallesh, Koujalagi wrote:
> On 13-02-2026 01:46 pm, Raag Jadav wrote:
> > System controller reports different types of events to GFX endpoint for
> > different usecases, add initial support for them. This will be further
> > extended to service those usecases.
> > 
> > v2: Handle unexpected response length (Mallesh)
> > 
> > Signed-off-by: Raag Jadav <raag.jadav@intel.com>
> > ---
> >   drivers/gpu/drm/xe/Makefile                 |  1 +
> >   drivers/gpu/drm/xe/xe_sysctrl.c             |  5 ++
> >   drivers/gpu/drm/xe/xe_sysctrl.h             |  1 +
> >   drivers/gpu/drm/xe/xe_sysctrl_event.c       | 76 +++++++++++++++++++++
> >   drivers/gpu/drm/xe/xe_sysctrl_event_types.h | 49 +++++++++++++
> >   drivers/gpu/drm/xe/xe_sysctrl_mailbox.h     | 10 +++
> >   6 files changed, 142 insertions(+)
> >   create mode 100644 drivers/gpu/drm/xe/xe_sysctrl_event.c
> >   create mode 100644 drivers/gpu/drm/xe/xe_sysctrl_event_types.h
> > 
> > diff --git a/drivers/gpu/drm/xe/Makefile b/drivers/gpu/drm/xe/Makefile
> > index 8005293dc30f..59e083f90d7e 100644
> > --- a/drivers/gpu/drm/xe/Makefile
> > +++ b/drivers/gpu/drm/xe/Makefile
> > @@ -123,6 +123,7 @@ xe-y += xe_bb.o \
> >   	xe_survivability_mode.o \
> >   	xe_sync.o \
> >   	xe_sysctrl.o \
> > +	xe_sysctrl_event.o \
> >   	xe_sysctrl_mailbox.o \
> >   	xe_tile.o \
> >   	xe_tile_sysfs.o \
> > diff --git a/drivers/gpu/drm/xe/xe_sysctrl.c b/drivers/gpu/drm/xe/xe_sysctrl.c
> > index aba2166650aa..bbfb737efc88 100644
> > --- a/drivers/gpu/drm/xe/xe_sysctrl.c
> > +++ b/drivers/gpu/drm/xe/xe_sysctrl.c
> > @@ -31,6 +31,11 @@
> >   static void xe_sysctrl_work(struct work_struct *work)
> >   {
> > +	struct xe_sysctrl *sc = container_of(work, struct xe_sysctrl, work);
> > +	struct xe_device *xe = container_of(sc, struct xe_device, sc);
> > +
> > +	guard(mutex)(&sc->work_lock);
> > +	xe_sysctrl_event(xe);
> >   }
> >   static void xe_sysctrl_fini(void *arg)
> > diff --git a/drivers/gpu/drm/xe/xe_sysctrl.h b/drivers/gpu/drm/xe/xe_sysctrl.h
> > index 5919310b9db9..bd9acf575d14 100644
> > --- a/drivers/gpu/drm/xe/xe_sysctrl.h
> > +++ b/drivers/gpu/drm/xe/xe_sysctrl.h
> > @@ -12,5 +12,6 @@ struct xe_device;
> >   int xe_sysctrl_init(struct xe_device *xe);
> >   void xe_sysctrl_irq_handler(struct xe_device *xe, u32 master_ctl);
> > +void xe_sysctrl_event(struct xe_device *xe);
> >   #endif /* _XE_SYSCTRL_H_ */
> > diff --git a/drivers/gpu/drm/xe/xe_sysctrl_event.c b/drivers/gpu/drm/xe/xe_sysctrl_event.c
> > new file mode 100644
> > index 000000000000..7c3041f4196a
> > --- /dev/null
> > +++ b/drivers/gpu/drm/xe/xe_sysctrl_event.c
> > @@ -0,0 +1,76 @@
> > +// SPDX-License-Identifier: MIT
> > +/*
> > + * Copyright © 2026 Intel Corporation
> > + */
> > +
> > +#include "xe_assert.h"
> > +#include "xe_device.h"
> > +#include "xe_irq.h"
> > +#include "xe_printk.h"
> > +#include "xe_sysctrl.h"
> > +#include "xe_sysctrl_event_types.h"
> > +#include "xe_sysctrl_mailbox.h"
> > +#include "xe_sysctrl_mailbox_types.h"
> > +
> > +static void xe_sysctrl_get_pending_event(struct xe_device *xe,
> > +					 struct xe_sysctrl_mailbox_command *command)
> > +{
> > +	struct xe_sysctrl_event_response response;
> > +	size_t len;
> > +	int ret;
> > +
> > +	command->data_out = &response;
> > +	command->data_out_len = sizeof(response);
> > +
> > +	do {
> > +		memset(&response, 0, sizeof(response));
> > +
> > +		ret = xe_sysctrl_send_command(xe, command, &len);
> > +		if (ret) {
> > +			xe_err(xe, "sysctrl: failed to get pending event %d\n", ret);
> > +			return;
> > +		}
> > +
> > +		if (len != sizeof(response)) {
> > +			xe_err(xe, "sysctrl: unexpected response length %ld\n", len);
> > +			return;
> > +		}
> > +
> > +		if (response.event == XE_SYSCTRL_EVENT_THRESHOLD_CROSSED) {
> > +			xe_warn(xe, "[RAS]: error counter threshold crossed\n");
> > +		} else {
> > +			xe_err(xe, "sysctrl: unexpected event %#x\n", response.event);
> What about remaining events in response.count?

We tream them as firmware bugs, similar to above cases.

> > +			return;
> > +		}
> > +
> > +		xe_dbg(xe, "sysctrl: %u events pending\n", response.count);
> 
> What happen when sysctrl continuously reports pending events, this could
> loop forever by monopolizing the work queue thread?

I already have it locally but thanks for pointing it out.

Raag

> > +	} while (response.count);
> > +}
> > +
> > +static void xe_sysctrl_event_request_prep(struct xe_device *xe,
> > +					  struct xe_sysctrl_mailbox_app_msg_hdr *header,
> > +					  struct xe_sysctrl_event_request *request)
> > +{
> > +	struct pci_dev *pdev = to_pci_dev(xe->drm.dev);
> > +
> > +	header->data = REG_FIELD_PREP(APP_HDR_GROUP_ID_MASK, XE_SYSCTRL_GROUP_GFSP) |
> > +		       REG_FIELD_PREP(APP_HDR_COMMAND_MASK, XE_SYSCTRL_CMD_GET_PENDING_EVENT);
> > +
> > +	request->vector = xe_device_has_msix(xe) ? XE_IRQ_DEFAULT_MSIX : 0;
> > +	request->fn = PCI_FUNC(pdev->devfn);
> > +}
> > +
> > +void xe_sysctrl_event(struct xe_device *xe)
> > +{
> > +	struct xe_sysctrl_mailbox_app_msg_hdr header = {};
> > +	struct xe_sysctrl_mailbox_command command = {};
> > +	struct xe_sysctrl_event_request request = {};
> > +
> > +	xe_sysctrl_event_request_prep(xe, &header, &request);
> > +
> > +	command.header = header;
> > +	command.data_in = &request;
> > +	command.data_in_len = sizeof(request);
> > +
> > +	xe_sysctrl_get_pending_event(xe, &command);
> > +}
> > diff --git a/drivers/gpu/drm/xe/xe_sysctrl_event_types.h b/drivers/gpu/drm/xe/xe_sysctrl_event_types.h
> > new file mode 100644
> > index 000000000000..9c5fb95c58f7
> > --- /dev/null
> > +++ b/drivers/gpu/drm/xe/xe_sysctrl_event_types.h
> > @@ -0,0 +1,49 @@
> > +/* SPDX-License-Identifier: MIT */
> > +/*
> > + * Copyright © 2026 Intel Corporation
> > + */
> > +
> > +#ifndef _XE_SYSCTRL_EVENT_TYPES_H_
> > +#define _XE_SYSCTRL_EVENT_TYPES_H_
> > +
> > +#include <linux/types.h>
> > +
> > +#define XE_SYSCTRL_EVENT_DATA_LEN		68
> > +
> > +enum xe_sysctrl_event {
> > +	XE_SYSCTRL_EVENT_THRESHOLD_CROSSED	= 0x01,
> > +};
> > +
> > +/**
> > + * struct xe_sysctrl_event_request - Request structure for pending event
> > + */
> > +struct xe_sysctrl_event_request {
> > +	/** @vector: MSI-X vector that was triggered */
> > +	u32 vector;
> > +	/** @fn: Function index (0-7) of PCIe device */
> > +	u8 fn;
> > +	/** @reserved: Reserved for future use */
> > +	u16 reserved;
> > +	/** @reserved2: Reserved for future use */
> > +	u32 reserved2[2];
> > +} __packed;
> > +
> > +/**
> > + * struct xe_sysctrl_event_response - Response structure for pending event
> > + */
> > +struct xe_sysctrl_event_response {
> > +	/** @count: Number of pending events */
> > +	u32 count;
> > +	/** @event: Pending event */
> > +	enum xe_sysctrl_event event;
> > +	/** @timestamp: Timestamp of most recent event */
> > +	u64 timestamp;
> > +	/** @extended: Event has extended payload */
> > +	u8 extended:1;
> > +	/** @reserved: Reserved for future use */
> > +	u32 reserved:23;
> > +	/** @data: Generic event data */
> > +	u32 data[XE_SYSCTRL_EVENT_DATA_LEN];
> > +} __packed;
> > +
> > +#endif /* _XE_SYSCTRL_EVENT_TYPES_H_ */
> > diff --git a/drivers/gpu/drm/xe/xe_sysctrl_mailbox.h b/drivers/gpu/drm/xe/xe_sysctrl_mailbox.h
> > index 2b64165c8e76..f060be5124f2 100644
> > --- a/drivers/gpu/drm/xe/xe_sysctrl_mailbox.h
> > +++ b/drivers/gpu/drm/xe/xe_sysctrl_mailbox.h
> > @@ -27,6 +27,16 @@ struct xe_sysctrl_mailbox_command;
> >   #define XE_SYSCTRL_APP_HDR_VERSION(hdr) \
> >   	FIELD_GET(APP_HDR_VERSION_MASK, le32_to_cpu((hdr)->data))
> > +/* Command groups */
> > +enum xe_sysctrl_group {
> > +	XE_SYSCTRL_GROUP_GFSP			= 0x01,
> > +};
> > +
> > +/* Commands supported by GFSP group */
> > +enum xe_sysctrl_gfsp_cmd {
> > +	XE_SYSCTRL_CMD_GET_PENDING_EVENT	= 0x07,
> > +};
> > +
> >   void xe_sysctrl_mailbox_init(struct xe_sysctrl *sc);
> >   int xe_sysctrl_send_command(struct xe_device *xe,
> >   			    struct xe_sysctrl_mailbox_command *cmd,