From: "Summers, Stuart" <stuart.summers@intel.com>
To: "Summers, Stuart" <stuart.summers@intel.com>
Cc: "intel-xe@lists.freedesktop.org" <intel-xe@lists.freedesktop.org>,
"Lin, Shuicheng" <shuicheng.lin@intel.com>,
"Vivi, Rodrigo" <rodrigo.vivi@intel.com>,
"Roper, Matthew D" <matthew.d.roper@intel.com>,
"Brost, Matthew" <matthew.brost@intel.com>,
"Wajdeczko, Michal" <michal.wajdeczko@intel.com>,
"Nerlige Ramappa, Umesh" <umesh.nerlige.ramappa@intel.com>,
"Ceraolo Spurio, Daniele" <daniele.ceraolospurio@intel.com>
Subject: Re: [PATCH 06/10] drm/xe/guc: Add configfs support for guc_log_level
Date: Tue, 11 Aug 2026 19:19:42 +0000 [thread overview]
Message-ID: <e4ac79f952c7b8a7a655246c73076184ea1b1e36.camel@intel.com> (raw)
In-Reply-To: <20260807194518.9474-18-stuart.summers@intel.com>
On Fri, 2026-08-07 at 19:45 +0000, Stuart Summers wrote:
> Allow the GuC log level to be selected per-device via configfs in
> addition to the existing 'guc_log_level' module parameter. The
> configfs
> attribute lives under the new 'debug' configfs subdirectory:
>
> /sys/kernel/config/xe/<bdf>/debug/guc_log_level
>
> When the configfs attribute is set to a valid level (0-5), it
> overrides
> the module parameter for that device. The default value is -1
> (XE_GUC_LOG_LEVEL_UNSET), which means 'unset' and falls back to the
> 'guc_log_level' module parameter, ensuring existing users that rely
> solely on the module parameter are unaffected by this change.
>
> The module parameter implementation itself is untouched.
>
> Note that the expectation is a user will only set the log level when
> CONFIG_DRM_XE_DEBUG is set.
>
> Signed-off-by: Stuart Summers <stuart.summers@intel.com>
> Assisted-by: Copilot:claude-sonnet-4.6,claude-opus-4.7,claude-sonnet-
> 5,claude-opus-5
> ---
> drivers/gpu/drm/xe/xe_configfs.c | 2 +
> drivers/gpu/drm/xe/xe_configfs_debug.c | 84
> ++++++++++++++++++++++++++
> drivers/gpu/drm/xe/xe_configfs_debug.h | 7 +++
> drivers/gpu/drm/xe/xe_configfs_types.h | 1 +
> drivers/gpu/drm/xe/xe_defaults.h | 3 +
> drivers/gpu/drm/xe/xe_guc_log.c | 3 +-
> 6 files changed, 99 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/xe/xe_configfs.c
> b/drivers/gpu/drm/xe/xe_configfs.c
> index 111aacce8a84..ea09d5f9e925 100644
> --- a/drivers/gpu/drm/xe/xe_configfs.c
> +++ b/drivers/gpu/drm/xe/xe_configfs.c
> @@ -100,6 +100,7 @@ const struct xe_config_device
> xe_configfs_device_defaults = {
> .debug = {
> .engines_allowed = U64_MAX,
> .gt_types_allowed = U64_MAX,
> + .guc_log_level = XE_GUC_LOG_LEVEL_UNSET,
> .enable_multi_queue = true,
> .enable_psmi = false,
> .survivability_mode = false,
> @@ -418,6 +419,7 @@ static void dump_custom_dev_config(struct pci_dev
> *pdev,
> #if IS_ENABLED(CONFIG_DRM_XE_DEBUG)
> PRI_CUSTOM_ATTR("%llx", debug.engines_allowed);
> PRI_CUSTOM_ATTR("%llx", debug.gt_types_allowed);
> + PRI_CUSTOM_ATTR("%d", debug.guc_log_level);
> PRI_CUSTOM_ATTR("%d", debug.enable_multi_queue);
> PRI_CUSTOM_ATTR("%d", debug.enable_psmi);
> PRI_CUSTOM_ATTR("%d", debug.survivability_mode);
> diff --git a/drivers/gpu/drm/xe/xe_configfs_debug.c
> b/drivers/gpu/drm/xe/xe_configfs_debug.c
> index b2667573efbe..eb3c3bd9c33e 100644
> --- a/drivers/gpu/drm/xe/xe_configfs_debug.c
> +++ b/drivers/gpu/drm/xe/xe_configfs_debug.c
> @@ -17,6 +17,7 @@
> #include "xe_configfs_debug.h"
> #include "xe_configfs_types.h"
> #include "xe_gt_types.h"
> +#include "xe_guc_log.h"
> #include "xe_hw_engine_types.h"
> #include "xe_pci_types.h"
>
> @@ -44,6 +45,7 @@
> * ├── ctx_restore_post_bb
> * ├── engines_allowed
> * ├── gt_types_allowed
> + * ├── guc_log_level
> * ├── enable_multi_queue
> * ├── enable_psmi
> * └── survivability_mode
> @@ -167,6 +169,27 @@
> *
> * # echo '' >
> /sys/kernel/config/xe/0000:03:00.0/debug/gt_types_allowed
> *
> + * GuC log level:
> + * --------------
> + *
> + * Set the GuC firmware logging verbosity for this device. Accepted
> values
> + * match the ``guc_log_level`` module parameter:
> + *
> + * - 0: disable
> + * - 1: normal (non-verbose)
> + * - 2..%GUC_LOG_LEVEL_MAX: verbose levels
> + *
> + * Example::
> + *
> + * # echo 3 >
> /sys/kernel/config/xe/0000:03:00.0/debug/guc_log_level
> + *
> + * The default value is %XE_GUC_LOG_LEVEL_UNSET (-1), which means
> the value
> + * of the ``guc_log_level`` module parameter is used. Any value
> greater
> + * than -1 written to this attribute overrides the module parameter
> for
> + * this device.
> + *
> + * This attribute can only be set before binding to the device.
> + *
> * Enable multi-queue
> * ------------------
> *
> @@ -758,6 +781,65 @@ static ssize_t gt_types_allowed_store(struct
> config_item *item, const char *page
> return len;
> }
>
> +/**
> + * xe_configfs_get_guc_log_level - get configfs guc_log_level
> setting
> + * @pdev: pci device
> + *
> + * Returns the guc_log_level value configured via configfs. If the
> configfs
> + * value is negative (the default is %XE_GUC_LOG_LEVEL_UNSET, -1),
> the value
> + * of the ``guc_log_level`` module parameter is returned instead,
> allowing
> + * the configfs entry to override the module parameter without
> affecting
> + * users that rely solely on the module parameter.
> + *
> + * Return: GuC log level to use for this device.
> + */
> +int xe_configfs_get_guc_log_level(struct pci_dev *pdev)
> +{
> + struct xe_config_group_device *dev = find_device(pdev);
> + int level = xe_modparam.guc_log_level;
> +
> + if (!dev)
> + return level;
> +
> + scoped_guard(mutex, &dev->lock)
> + if (dev->config.debug.guc_log_level >= 0)
> + level = dev->config.debug.guc_log_level;
> +
> + config_group_put(&dev->group);
> +out:
Sashiko called out correctly that I forgot to drop the extra label
here... I'll post an update here shortly for this. I also need to post
IGT changes to xe_configfs to account for these new debug-specific
changes.
Thanks,
Stuart
> + return level;
> +}
> +
> +static ssize_t guc_log_level_show(struct config_item *item, char
> *page)
> +{
> + struct xe_config_device *dev =
> xe_configfs_subgroup_to_device(item);
> +
> + return sprintf(page, "%d\n", dev->debug.guc_log_level);
> +}
> +
> +static ssize_t guc_log_level_store(struct config_item *item, const
> char *page, size_t len)
> +{
> + struct xe_config_group_device *dev =
> xe_configfs_subgroup_to_group_device(item);
> + int val;
> + int ret;
> +
> + ret = kstrtoint(page, 0, &val);
> + if (ret)
> + return ret;
> +
> + if (val != XE_GUC_LOG_LEVEL_UNSET &&
> + (val < 0 || val > GUC_LOG_LEVEL_MAX))
> + return -EINVAL;
> +
> + guard(mutex)(&dev->lock);
> + if (xe_configfs_is_bound(dev))
> + return -EBUSY;
> +
> + dev->config.debug.guc_log_level = val;
> +
> + return len;
> +}
> +
> /**
> * xe_configfs_get_enable_multi_queue - get configfs
> enable_multi_queue setting
> * @pdev: pci device
> @@ -904,6 +986,7 @@ CONFIGFS_ATTR(, ctx_restore_mid_bb);
> CONFIGFS_ATTR(, ctx_restore_post_bb);
> CONFIGFS_ATTR(, engines_allowed);
> CONFIGFS_ATTR(, gt_types_allowed);
> +CONFIGFS_ATTR(, guc_log_level);
> CONFIGFS_ATTR(, enable_multi_queue);
> CONFIGFS_ATTR(, enable_psmi);
> CONFIGFS_ATTR(, survivability_mode);
> @@ -913,6 +996,7 @@ static struct configfs_attribute
> *xe_configfs_debug_attrs[] = {
> &attr_ctx_restore_post_bb,
> &attr_engines_allowed,
> &attr_gt_types_allowed,
> + &attr_guc_log_level,
> &attr_enable_multi_queue,
> &attr_enable_psmi,
> &attr_survivability_mode,
> diff --git a/drivers/gpu/drm/xe/xe_configfs_debug.h
> b/drivers/gpu/drm/xe/xe_configfs_debug.h
> index 4a027ea2082d..5fe39bb72f00 100644
> --- a/drivers/gpu/drm/xe/xe_configfs_debug.h
> +++ b/drivers/gpu/drm/xe/xe_configfs_debug.h
> @@ -7,7 +7,9 @@
>
> #include <linux/types.h>
>
> +#include "xe_defaults.h"
> #include "xe_hw_engine_types.h"
> +#include "xe_module.h"
>
> struct pci_dev;
> struct xe_config_device;
> @@ -24,6 +26,7 @@ u32 xe_configfs_get_ctx_restore_post_bb(struct
> pci_dev *pdev,
> u64 xe_configfs_get_engines_allowed(struct pci_dev *pdev);
> bool xe_configfs_primary_gt_allowed(struct pci_dev *pdev);
> bool xe_configfs_media_gt_allowed(struct pci_dev *pdev);
> +int xe_configfs_get_guc_log_level(struct pci_dev *pdev);
> bool xe_configfs_get_enable_multi_queue(struct pci_dev *pdev);
> bool xe_configfs_get_psmi_enabled(struct pci_dev *pdev);
> bool xe_configfs_get_survivability_mode(struct pci_dev *pdev);
> @@ -42,6 +45,10 @@ static inline u32
> xe_configfs_get_ctx_restore_post_bb(struct pci_dev *pdev,
> static inline u64 xe_configfs_get_engines_allowed(struct pci_dev
> *pdev) { return U64_MAX; }
> static inline bool xe_configfs_primary_gt_allowed(struct pci_dev
> *pdev) { return true; }
> static inline bool xe_configfs_media_gt_allowed(struct pci_dev
> *pdev) { return true; }
> +static inline int xe_configfs_get_guc_log_level(struct pci_dev
> *pdev)
> +{
> + return xe_modparam.guc_log_level;
> +}
> static inline bool xe_configfs_get_enable_multi_queue(struct pci_dev
> *pdev) { return true; }
> static inline bool xe_configfs_get_psmi_enabled(struct pci_dev
> *pdev) { return false; }
> static inline bool xe_configfs_get_survivability_mode(struct pci_dev
> *pdev) { return false; }
> diff --git a/drivers/gpu/drm/xe/xe_configfs_types.h
> b/drivers/gpu/drm/xe/xe_configfs_types.h
> index 6e8ba6f85fad..e9460c3de891 100644
> --- a/drivers/gpu/drm/xe/xe_configfs_types.h
> +++ b/drivers/gpu/drm/xe/xe_configfs_types.h
> @@ -38,6 +38,7 @@ struct xe_config_group_device {
> struct wa_bb
> ctx_restore_post_bb[XE_ENGINE_CLASS_MAX];
> u64 engines_allowed;
> u64 gt_types_allowed;
> + int guc_log_level;
> bool enable_multi_queue;
> bool enable_psmi;
> bool survivability_mode;
> diff --git a/drivers/gpu/drm/xe/xe_defaults.h
> b/drivers/gpu/drm/xe/xe_defaults.h
> index c8ae1d5f3d60..df88078e84b8 100644
> --- a/drivers/gpu/drm/xe/xe_defaults.h
> +++ b/drivers/gpu/drm/xe/xe_defaults.h
> @@ -13,6 +13,9 @@
> #define XE_DEFAULT_GUC_LOG_LEVEL 1
> #endif
>
> +/* Sentinel value for guc_log_level configfs: not set, fall back to
> module param */
> +#define XE_GUC_LOG_LEVEL_UNSET -1
> +
> #define
> XE_DEFAULT_PROBE_DISPLAY IS_ENABLED(CONFIG_DRM_XE_DISPL
> AY)
> #define XE_DEFAULT_VRAM_BAR_SIZE 0
> #define
> XE_DEFAULT_FORCE_PROBE CONFIG_DRM_XE_FORCE_PROBE
> diff --git a/drivers/gpu/drm/xe/xe_guc_log.c
> b/drivers/gpu/drm/xe/xe_guc_log.c
> index 538d4df0f7aa..531b9759d520 100644
> --- a/drivers/gpu/drm/xe/xe_guc_log.c
> +++ b/drivers/gpu/drm/xe/xe_guc_log.c
> @@ -13,6 +13,7 @@
> #include "abi/guc_lfd_abi.h"
> #include "regs/xe_guc_regs.h"
> #include "xe_bo.h"
> +#include "xe_configfs_debug.h"
> #include "xe_devcoredump.h"
> #include "xe_force_wake.h"
> #include "xe_gt_printk.h"
> @@ -637,7 +638,7 @@ int xe_guc_log_init(struct xe_guc_log *log)
>
> xe_map_memset(xe, &bo->vmap, 0, 0, xe_bo_size(bo));
> log->bo = bo;
> - log->level = xe_modparam.guc_log_level;
> + log->level = xe_configfs_get_guc_log_level(to_pci_dev(xe-
> >drm.dev));
>
> return 0;
> }
next prev parent reply other threads:[~2026-08-11 19:19 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-08-07 19:45 [PATCH 00/10] Add new debug infrastructure for configfs Stuart Summers
2026-08-07 19:45 ` [PATCH 01/10] drm/xe: Guard configfs attribute reads in getters Stuart Summers
2026-08-07 19:45 ` [PATCH 02/10] drm/xe: Sort xe_config_device fields Stuart Summers
2026-08-07 19:45 ` [PATCH 03/10] drm/xe: Split out configfs data structures Stuart Summers
2026-08-07 19:45 ` [PATCH 04/10] drm/xe: Add a new debug focused configfs group Stuart Summers
2026-08-07 19:45 ` [PATCH 05/10] drm/xe: Move debug configfs entries to xe_configfs_debug.c Stuart Summers
2026-08-07 19:45 ` [PATCH 06/10] drm/xe/guc: Add configfs support for guc_log_level Stuart Summers
2026-08-11 19:19 ` Summers, Stuart [this message]
2026-08-07 19:45 ` [PATCH 07/10] drm/xe/guc: Add support for NPK as a GuC log target Stuart Summers
2026-08-07 19:45 ` [PATCH 08/10] drm/xe: Add infrastructure for debug configfs parameters Stuart Summers
2026-08-07 19:45 ` [PATCH 09/10] drm/xe: Migrate existing debug configfs entries to params infrastructure Stuart Summers
2026-08-07 19:45 ` [PATCH 10/10] drm/xe: Taint kernel when debug configfs parameters are set Stuart Summers
2026-08-07 19:52 ` ✗ CI.checkpatch: warning for Add new debug infrastructure for configfs (rev5) Patchwork
2026-08-07 19:53 ` ✓ CI.KUnit: success " Patchwork
2026-08-07 20:44 ` ✓ Xe.CI.BAT: " Patchwork
2026-08-08 7:10 ` ✗ Xe.CI.FULL: failure " Patchwork
-- strict thread matches above, loose matches on Subject: below --
2026-08-20 21:06 [PATCH 00/10] Add new debug infrastructure for configfs Stuart Summers
2026-08-20 21:06 ` [PATCH 06/10] drm/xe/guc: Add configfs support for guc_log_level Stuart Summers
2026-08-20 20:07 [PATCH 00/10] Add new debug infrastructure for configfs Stuart Summers
2026-08-20 20:07 ` [PATCH 06/10] drm/xe/guc: Add configfs support for guc_log_level Stuart Summers
2026-08-05 23:20 [PATCH 00/10] Add new debug infrastructure for configfs Stuart Summers
2026-08-05 23:20 ` [PATCH 06/10] drm/xe/guc: Add configfs support for guc_log_level Stuart Summers
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=e4ac79f952c7b8a7a655246c73076184ea1b1e36.camel@intel.com \
--to=stuart.summers@intel.com \
--cc=daniele.ceraolospurio@intel.com \
--cc=intel-xe@lists.freedesktop.org \
--cc=matthew.brost@intel.com \
--cc=matthew.d.roper@intel.com \
--cc=michal.wajdeczko@intel.com \
--cc=rodrigo.vivi@intel.com \
--cc=shuicheng.lin@intel.com \
--cc=umesh.nerlige.ramappa@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox