Re: [Intel-xe] [PATCH 02/14] drm/xe: Introduce the dev_coredump infrastructure.

["Thomas Hellström" <thomas.hellstrom@linux.intel.com>]

On 4/26/23 22:57, Rodrigo Vivi wrote:
> The goal is to use devcoredump infrastructure to report error states
> captured at the crash time.
>
> The error state will contain useful information for GPU hang debug, such
> as INSTDONE registers and the current buffers getting executed, as well
> as any other information that helps user space and allow later replays of
> the error.
>
> The proposal here is to avoid a Xe only error_state like i915 and use
> a standard dev_coredump infrastructure to expose the error state.
>
> For our own case, the data is only useful if it is a snapshot of the
> time when the GPU crash has happened, since we reset the GPU immediately
> after and the registers might have changed. So the proposal here is to
> have an internal snapshot to be printed out later.
>
> Also, usually a subsequent GPU hang can be only a cause of the initial
> one. So we only save the 'first' hang. The dev_coredump has a delayed
> work queue where it remove the coredump and free all the data withing a
> few moments of the error. When that happens we also reset our capture
> state and allow further snapshots.
>
> Right now this infra only print out the time of the hang. More information
> will be migrated here on subsequent work. Also, in order to organize the
> dump better, the goal is to propose dev_coredump changes itself to allow
> multiple files and different controls. But for now we start Xe usage of
> it without any dependency on dev_coredump core changes.
>
> Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
> Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
> ---
>   drivers/gpu/drm/xe/Kconfig                |   1 +
>   drivers/gpu/drm/xe/Makefile               |   1 +
>   drivers/gpu/drm/xe/xe_devcoredump.c       | 144 ++++++++++++++++++++++
>   drivers/gpu/drm/xe/xe_devcoredump.h       |  22 ++++
>   drivers/gpu/drm/xe/xe_devcoredump_types.h |  47 +++++++
>   drivers/gpu/drm/xe/xe_device_types.h      |   4 +
>   drivers/gpu/drm/xe/xe_guc_submit.c        |   2 +
>   drivers/gpu/drm/xe/xe_pci.c               |   2 +
>   8 files changed, 223 insertions(+)
>   create mode 100644 drivers/gpu/drm/xe/xe_devcoredump.c
>   create mode 100644 drivers/gpu/drm/xe/xe_devcoredump.h
>   create mode 100644 drivers/gpu/drm/xe/xe_devcoredump_types.h
>
> diff --git a/drivers/gpu/drm/xe/Kconfig b/drivers/gpu/drm/xe/Kconfig
> index f6f3b491d162..d44794f99338 100644
> --- a/drivers/gpu/drm/xe/Kconfig
> +++ b/drivers/gpu/drm/xe/Kconfig
> @@ -35,6 +35,7 @@ config DRM_XE
>   	select DRM_TTM_HELPER
>   	select DRM_SCHED
>   	select MMU_NOTIFIER
> +	select WANT_DEV_COREDUMP
>   	help
>   	  Experimental driver for Intel Xe series GPUs
>   
> diff --git a/drivers/gpu/drm/xe/Makefile b/drivers/gpu/drm/xe/Makefile
> index ee4a95beec20..9d675f7c77aa 100644
> --- a/drivers/gpu/drm/xe/Makefile
> +++ b/drivers/gpu/drm/xe/Makefile
> @@ -34,6 +34,7 @@ xe-y += xe_bb.o \
>   	xe_bo.o \
>   	xe_bo_evict.o \
>   	xe_debugfs.o \
> +	xe_devcoredump.o \
>   	xe_device.o \
>   	xe_dma_buf.o \
>   	xe_engine.o \
> diff --git a/drivers/gpu/drm/xe/xe_devcoredump.c b/drivers/gpu/drm/xe/xe_devcoredump.c
> new file mode 100644
> index 000000000000..d9531183f03a
> --- /dev/null
> +++ b/drivers/gpu/drm/xe/xe_devcoredump.c
> @@ -0,0 +1,144 @@
> +// SPDX-License-Identifier: MIT
> +/*
> + * Copyright © 2023 Intel Corporation
> + */
> +
> +#include "xe_devcoredump.h"
> +#include "xe_devcoredump_types.h"
> +
> +#include <linux/devcoredump.h>
> +#include <generated/utsrelease.h>
> +
> +#include "xe_engine.h"
> +#include "xe_gt.h"
> +
> +/**
> + * DOC: Xe device coredump
> + *
> + * Devices overview:
> + * Xe uses dev_coredump infrastructure for exposing the crash errors in a
> + * standardized way.
> + * devcoredump exposes a temporary device under /sys/class/devcoredump/
> + * which is linked with our card device directly.
> + * The core dump can be accessed either from
> + * /sys/class/drm/card<n>/device/devcoredump/ or from
> + * /sys/class/devcoredump/devcd<m> where
> + * /sys/class/devcoredump/devcd<m>/failing_device is a link to
> + * /sys/class/drm/card<n>/device/.
> + *
> + * Snapshot at hang:
> + * The 'data' file is printed with a drm_printer pointer at devcoredump read
> + * time. For this reason, we need to take snapshots from when the hang has
> + * happened, and not only when the user is reading the file. Otherwise the
> + * information is outdated since the resets might have happened in between.
> + *
> + * 'First' failure snapshot:
> + * In general, the first hang is the most critical one since the following hangs
> + * can be a consequence of the initial hang. For this reason we only take the
> + * snapshot of the 'first' failure and ignore subsequent calls of this function,
> + * at least while the coredump device is alive. Dev_coredump has a delayed work
> + * queue that will eventually delete the device and free all the dump
> + * information. At this time we also clear the faulty_engine and allow the next
> + * hang capture.
> + */
> +
> +static ssize_t xe_devcoredump_read(char *buffer, loff_t offset,
> +				   size_t count, void *data, size_t datalen)
> +{
> +	struct xe_devcoredump *coredump = data;
> +	struct xe_devcoredump_snapshot *ss;
> +	struct drm_printer p;
> +	struct drm_print_iterator iter;
> +	struct timespec64 ts;
> +
> +	iter.data = buffer;
> +	iter.offset = 0;
> +	iter.start = offset;
> +	iter.remain = count;
> +
> +	mutex_lock(&coredump->lock);
> +
> +	ss = &coredump->snapshot;
> +	p = drm_coredump_printer(&iter);
> +
> +	drm_printf(&p, "**** Xe Device Coredump ****\n");
> +	drm_printf(&p, "kernel: " UTS_RELEASE "\n");
> +	drm_printf(&p, "module: " KBUILD_MODNAME "\n");
> +
> +	ts = ktime_to_timespec64(ss->snapshot_time);
> +	drm_printf(&p, "Snapshot time: %lld.%09ld\n", ts.tv_sec, ts.tv_nsec);
> +	ts = ktime_to_timespec64(ss->boot_time);
> +	drm_printf(&p, "Boot time: %lld.%09ld\n", ts.tv_sec, ts.tv_nsec);
> +	ts = ktime_to_timespec64(ktime_sub(ss->snapshot_time, ss->boot_time));
> +	drm_printf(&p, "Uptime: %lld.%09ld\n", ts.tv_sec, ts.tv_nsec);
> +
> +	mutex_unlock(&coredump->lock);
> +
> +	return count - iter.remain;
> +}
> +
> +static void xe_devcoredump_free(void *data)
> +{
> +	struct xe_devcoredump *coredump = data;
> +	struct xe_device *xe = container_of(coredump, struct xe_device,
> +					    devcoredump);
> +	mutex_lock(&coredump->lock);
> +
> +	coredump->faulty_engine = NULL;
> +	drm_info(&xe->drm, "Xe device coredump has been deleted.\n");
> +
> +	mutex_unlock(&coredump->lock);
> +}
> +
> +static void devcoredump_snapshot(struct xe_devcoredump *coredump)
> +{
> +	struct xe_devcoredump_snapshot *ss = &coredump->snapshot;
> +
> +	lockdep_assert_held(&coredump->lock);
> +	ss->snapshot_time = ktime_get_real();
> +	ss->boot_time = ktime_get_boottime();
> +}
> +
> +/**
> + * xe_devcoredump - Take the required snapshots and initialize coredump device.
> + * @e: The faulty xe_engine, where the issue was detected.
> + *
> + * This function should be called at the crash time. It is skipped if we still
> + * have the core dump device available with the information of the 'first'
> + * snapshot.
> + */
> +void xe_devcoredump(struct xe_engine *e)
> +{
> +	struct xe_device *xe = gt_to_xe(e->gt);
> +	struct xe_devcoredump *coredump = &xe->devcoredump;

For !long running engines, this is the dma-fence signalling critical 
path, and since the drm_scheduler has not yet been properly annotated, 
we should probably annotate that here, to avoid seeing strange deadlocks 
during coredumps....

/Thomas



> +
> +	mutex_lock(&coredump->lock);
> +	if (coredump->faulty_engine) {
> +		drm_dbg(&xe->drm, "Multiple hangs are occuring, but only the first snapshot was taken\n");
> +		mutex_unlock(&coredump->lock);
> +		return;
> +	}
> +	coredump->faulty_engine = e;
> +	devcoredump_snapshot(coredump);
> +	mutex_unlock(&coredump->lock);
> +
> +	drm_info(&xe->drm, "Xe device coredump has been created\n");
> +	drm_info(&xe->drm, "Check your /sys/class/drm/card<n>/device/devcoredump/data\n");
> +
> +	dev_coredumpm(xe->drm.dev, THIS_MODULE, coredump, 0, GFP_KERNEL,
> +		      xe_devcoredump_read, xe_devcoredump_free);
> +}
> +
> +/**
> + * xe_devcoredump_init - Initialize xe_devcoredump.
> + * @xe: Xe device.
> + *
> + * This function should be called at the probe so the mutex lock can be
> + * initialized.
> + */
> +void xe_devcoredump_init(struct xe_device *xe)
> +{
> +	struct xe_devcoredump *coredump = &xe->devcoredump;
> +
> +	mutex_init(&coredump->lock);
> +}
> diff --git a/drivers/gpu/drm/xe/xe_devcoredump.h b/drivers/gpu/drm/xe/xe_devcoredump.h
> new file mode 100644
> index 000000000000..30941d2e554b
> --- /dev/null
> +++ b/drivers/gpu/drm/xe/xe_devcoredump.h
> @@ -0,0 +1,22 @@
> +/* SPDX-License-Identifier: MIT */
> +/*
> + * Copyright © 2023 Intel Corporation
> + */
> +
> +#ifndef _XE_DEVCOREDUMP_H_
> +#define _XE_DEVCOREDUMP_H_
> +
> +struct xe_device;
> +struct xe_engine;
> +
> +void xe_devcoredump_init(struct xe_device *xe);
> +
> +#ifdef CONFIG_DEV_COREDUMP
> +void xe_devcoredump(struct xe_engine *e);
> +#else
> +static inline void xe_devcoredump(struct xe_engine *e)
> +{
> +}
> +#endif
> +
> +#endif
> diff --git a/drivers/gpu/drm/xe/xe_devcoredump_types.h b/drivers/gpu/drm/xe/xe_devcoredump_types.h
> new file mode 100644
> index 000000000000..3f395fa9104e
> --- /dev/null
> +++ b/drivers/gpu/drm/xe/xe_devcoredump_types.h
> @@ -0,0 +1,47 @@
> +/* SPDX-License-Identifier: MIT */
> +/*
> + * Copyright © 2023 Intel Corporation
> + */
> +
> +#ifndef _XE_DEVCOREDUMP_TYPES_H_
> +#define _XE_DEVCOREDUMP_TYPES_H_
> +
> +#include <linux/ktime.h>
> +#include <linux/mutex.h>
> +
> +struct xe_device;
> +
> +/**
> + * struct xe_devcoredump_snapshot - Crash snapshot
> + *
> + * This struct contains all the useful information quickly captured at the time
> + * of the crash. So, any subsequent reads of the coredump points to a data that
> + * shows the state of the GPU of when the issue has happened.
> + */
> +struct xe_devcoredump_snapshot {
> +	/** @snapshot_time:  Time of this capture. */
> +	ktime_t snapshot_time;
> +	/** @boot_time:  Relative boot time so the uptime can be calculated. */
> +	ktime_t boot_time;
> +};
> +
> +/**
> + * struct xe_devcoredump - Xe devcoredump main structure
> + *
> + * This struct represents the live and active dev_coredump node.
> + * It is created/populated at the time of a crash/error. Then it
> + * is read later when user access the device coredump data file
> + * for reading the information.
> + */
> +struct xe_devcoredump {
> +	/** @xe: Xe device. */
> +	struct xe_device *xe;
> +	/** @falty_engine: Engine where the crash/error happened. */
> +	struct xe_engine *faulty_engine;
> +	/** @lock: Protects data from races between capture and read out. */
> +	struct mutex lock;
> +	/** @snapshot: Snapshot is captured at time of the first crash */
> +	struct xe_devcoredump_snapshot snapshot;
> +};
> +
> +#endif
> diff --git a/drivers/gpu/drm/xe/xe_device_types.h b/drivers/gpu/drm/xe/xe_device_types.h
> index 1cb404e48aaa..2a0995824692 100644
> --- a/drivers/gpu/drm/xe/xe_device_types.h
> +++ b/drivers/gpu/drm/xe/xe_device_types.h
> @@ -12,6 +12,7 @@
>   #include <drm/drm_file.h>
>   #include <drm/ttm/ttm_device.h>
>   
> +#include "xe_devcoredump_types.h"
>   #include "xe_gt_types.h"
>   #include "xe_platform_types.h"
>   #include "xe_step_types.h"
> @@ -55,6 +56,9 @@ struct xe_device {
>   	/** @drm: drm device */
>   	struct drm_device drm;
>   
> +	/** @devcoredump: device coredump */
> +	struct xe_devcoredump devcoredump;
> +
>   	/** @info: device info */
>   	struct intel_device_info {
>   		/** @graphics_name: graphics IP name */
> diff --git a/drivers/gpu/drm/xe/xe_guc_submit.c b/drivers/gpu/drm/xe/xe_guc_submit.c
> index e857013070b9..231fb4145297 100644
> --- a/drivers/gpu/drm/xe/xe_guc_submit.c
> +++ b/drivers/gpu/drm/xe/xe_guc_submit.c
> @@ -14,6 +14,7 @@
>   #include <drm/drm_managed.h>
>   
>   #include "regs/xe_lrc_layout.h"
> +#include "xe_devcoredump.h"
>   #include "xe_device.h"
>   #include "xe_engine.h"
>   #include "xe_force_wake.h"
> @@ -800,6 +801,7 @@ guc_engine_timedout_job(struct drm_sched_job *drm_job)
>   		drm_warn(&xe->drm, "Timedout job: seqno=%u, guc_id=%d, flags=0x%lx",
>   			 xe_sched_job_seqno(job), e->guc->id, e->flags);
>   		simple_error_capture(e);
> +		xe_devcoredump(e);
>   	} else {
>   		drm_dbg(&xe->drm, "Timedout signaled job: seqno=%u, guc_id=%d, flags=0x%lx",
>   			 xe_sched_job_seqno(job), e->guc->id, e->flags);
> diff --git a/drivers/gpu/drm/xe/xe_pci.c b/drivers/gpu/drm/xe/xe_pci.c
> index e512e8b69831..1d496210b580 100644
> --- a/drivers/gpu/drm/xe/xe_pci.c
> +++ b/drivers/gpu/drm/xe/xe_pci.c
> @@ -16,6 +16,7 @@
>   
>   #include "regs/xe_regs.h"
>   #include "regs/xe_gt_regs.h"
> +#include "xe_devcoredump.h"
>   #include "xe_device.h"
>   #include "xe_display.h"
>   #include "xe_drv.h"
> @@ -657,6 +658,7 @@ static int xe_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
>   		return err;
>   	}
>   
> +	xe_devcoredump_init(xe);
>   	xe_pm_runtime_init(xe);
>   
>   	return 0;