From: "Nilawar, Badal" <badal.nilawar@intel.com>
To: Mallesh Koujalagi <mallesh.koujalagi@intel.com>,
<intel-xe@lists.freedesktop.org>, <rodrigo.vivi@intel.com>,
<matthew.brost@intel.com>
Cc: <anshuman.gupta@intel.com>, <vinay.belgaumkar@intel.com>,
<riana.tauro@intel.com>, <karthik.poosa@intel.com>,
<sk.anirban@intel.com>, <raag.jadav@intel.com>,
<aravind.iddamsetty@linux.intel.com>,
<jani.nikula@linux.intel.com>
Subject: Re: [PATCH v4 2/6] drm/xe/cper: Prepare CPER headers for software errors
Date: Tue, 17 Nov 2026 11:19:12 +0530 [thread overview]
Message-ID: <15eee851-2dc4-4ee4-8e7f-12eceb43f54e@intel.com> (raw)
In-Reply-To: <20261112134636.1860446-10-mallesh.koujalagi@intel.com>
On 12-11-2026 19:16, Mallesh Koujalagi wrote:
> Add software error classification support and reuse the
> common CPER header initialization path.
>
> Cc: Badal Nilawar <badal.nilawar@intel.com>
> Signed-off-by: Mallesh Koujalagi <mallesh.koujalagi@intel.com>
> ---
> v2:
> - Move union inside xe_cper_intel_hdr structure. (Badal)
>
> v3:
> - Introduce structure where it's used. (Badal)
> - Move error-class union inside cper_init_intel_hdr. (Badal)
>
> v4:
> - Update commit message. (Badal)
> - Use ktime_get_real_ns.
> - Add BUILD_BUG_ON. (Raag)
> ---
> drivers/gpu/drm/xe/xe_cper.c | 24 ++++++++++++++++++---
> drivers/gpu/drm/xe/xe_cper_types.h | 34 ++++++++++++++++++++++++++++--
> 2 files changed, 53 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/gpu/drm/xe/xe_cper.c b/drivers/gpu/drm/xe/xe_cper.c
> index dd6d4251dd1d..b23f8bc9f993 100644
> --- a/drivers/gpu/drm/xe/xe_cper.c
> +++ b/drivers/gpu/drm/xe/xe_cper.c
> @@ -14,6 +14,8 @@
> #include <linux/time.h>
> #include <linux/timekeeping.h>
>
> +#include "abi/xe_log_abi.h"
> +
> #include "xe_cper.h"
> #include "xe_cper_types.h"
> #include "xe_device.h"
> @@ -92,11 +94,14 @@ static u64 cper_timestamp_now(void)
> return ts;
> }
>
> -static void cper_init_intel_hdr(struct xe_device *xe, const struct xe_ras_error_class *counter,
> +static void cper_init_intel_hdr(struct xe_device *xe, const union xe_cper_error_class *error_class,
> u64 first_timestamp, u32 sig_id, u32 error_count,
> struct xe_cper_intel_hdr *ihdr)
> {
> - ihdr->counter = *counter;
> + BUILD_BUG_ON(sizeof(struct xe_ras_error_class) !=
> + sizeof(struct xe_cper_driver_error_class));
> +
> + ihdr->error_class = *error_class;
> ihdr->validation_bits |= XE_CPER_VALID_LOCATION;
> ihdr->sig_id = U32_MAX;
>
> @@ -219,11 +224,12 @@ void xe_cper_emit_counter(struct xe_device *xe, int cper_sev, enum xe_sigid sigi
> const struct xe_ras_error_class *counter,
> const struct xe_ras_get_counter_response *response)
> {
> + union xe_cper_error_class error_class = { .counter = *counter };
> struct xe_cper_intel_hdr ihdr = {};
>
> xe_assert(xe, is_hw_sigid(sigid));
>
> - cper_init_intel_hdr(xe, counter, response->timestamp,
> + cper_init_intel_hdr(xe, &error_class, response->timestamp,
> sigid, response->value, &ihdr);
This can go in single line.
> cper_record_emit(xe, cper_sev, &INTEL_CPER_NOTIFY_GPU_ERROR, &ihdr, NULL, 0);
> }
> @@ -249,6 +255,18 @@ static void cper_emit_sw_err(struct xe_device *xe, int cper_sev, enum xe_sigid s
> u32 component, u32 location, const void *data,
> struct va_format *vaf)
> {
> + union xe_cper_error_class error_class = {};
> + struct xe_cper_intel_hdr ihdr = {};
> + struct xe_cper_driver_error_class *driver = &error_class.driver;
> + u64 timestamp = ktime_get_real_ns();
Same as rev3: Initialize timestamp closer to its use. Please do the same
for driver and maintain reverse Christmas tree ordering.
> +
> + driver->error_type = cper_sev;
> + driver->error_component = component;
> + driver->tile = location;
> + if (IS_ERR(data))
> + driver->cause = -PTR_ERR(data);
> +
> + cper_init_intel_hdr(xe, &error_class, timestamp, sigid, 1, &ihdr);
> }
>
> /**
> diff --git a/drivers/gpu/drm/xe/xe_cper_types.h b/drivers/gpu/drm/xe/xe_cper_types.h
> index 68e30df0440e..8e90a3220919 100644
> --- a/drivers/gpu/drm/xe/xe_cper_types.h
> +++ b/drivers/gpu/drm/xe/xe_cper_types.h
> @@ -45,6 +45,29 @@
> GUID_INIT(0xea9d8f84, 0x4258, 0x5227, \
> 0x80, 0x28, 0xb9, 0xb1, 0x3e, 0x6d, 0x58, 0xb0)
>
> +/**
> + * struct xe_cper_driver_error_class - Intel GPU CPER driver-detected error classification
> + */
> +struct xe_cper_driver_error_class {
> + /** @error_type: CPER severity: CPER_SEV_* */
> + u8 error_type;
> + /** @reserved0: Reserved, must be zero */
> + u8 reserved0;
> + /** @error_component: Driver component id: CLASS[7:0].TYPE[15:8] */
> + u16 error_component;
> + /** @tile: Driver location id: TYPE[7:0].ID[15:8] ('tile' and 'GT' field region) */
> + u16 tile;
> + /** @reserved1: Reserved, must be zero */
> + u16 reserved1;
> + /** @cause: errno of the failing operation, or 0 if none */
> + u32 cause;
> +};
> +
> +/* Same size as the counter-sourced error_class */
> +#define XE_CPER_ERROR_CLASS_SIZE 12
> +
> +static_assert(sizeof(struct xe_cper_driver_error_class) == XE_CPER_ERROR_CLASS_SIZE);
> +
> /**
> * struct xe_cper_intel_hdr - Intel-specific CPER header
> *
> @@ -52,8 +75,15 @@
> * All multi-byte fields are little-endian.
> */
> struct xe_cper_intel_hdr {
> - /** @counter: Error class (severity, component, location, cause) */
> - struct xe_ras_error_class counter;
> + /** @error_class: Error classification */
> + union xe_cper_error_class {
> + /** @error_class.counter: Error class (severity, component,
> + * location, cause)
> + */
Keep this in single line.
Thanks,
Badal
> + struct xe_ras_error_class counter;
> + /** @error_class.driver: Driver-detected error classification */
> + struct xe_cper_driver_error_class driver;
> + } error_class;
> /** @first_timestamp: First occurrence of this error class, in ticks */
> u64 first_timestamp;
> /** @sig_id: Aggregated error class SIG ID; set to U32_MAX if unknown */
next prev parent reply other threads:[~2026-11-17 5:50 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-12 13:46 [PATCH v4 0/6] drm/xe/cper: Software (driver-detected) CPER error reporting Mallesh Koujalagi
2026-11-12 13:46 ` [PATCH v4 1/6] drm/xe/cper: Add software CPER emission entry point Mallesh Koujalagi
2026-11-17 7:30 ` Nilawar, Badal
2026-11-17 8:48 ` Mallesh, Koujalagi
2026-11-12 13:46 ` [PATCH v4 2/6] drm/xe/cper: Prepare CPER headers for software errors Mallesh Koujalagi
2026-11-17 5:49 ` Nilawar, Badal [this message]
2026-11-17 6:38 ` Mallesh, Koujalagi
2026-11-12 13:46 ` [PATCH v4 3/6] drm/xe/cper: Emit software CPER records Mallesh Koujalagi
2026-11-17 7:34 ` Nilawar, Badal
2026-11-17 12:38 ` Mallesh, Koujalagi
2026-11-12 13:46 ` [PATCH v4 4/6] drm/xe/cper: Make software CPER emission atomic-safe Mallesh Koujalagi
2026-11-12 13:46 ` [PATCH v4 5/6] drm/xe/log: Add software CPER information to KUnit output Mallesh Koujalagi
2026-11-18 7:33 ` Nilawar, Badal
2026-11-18 8:11 ` Mallesh, Koujalagi
2026-11-12 13:46 ` [PATCH v4 6/6] drm/xe/log: Test software CPER emission parameters Mallesh Koujalagi
2026-11-17 14:05 ` Nilawar, Badal
2026-11-12 14:21 ` ✓ CI.KUnit: success for drm/xe/cper: Software (driver-detected) CPER error reporting (rev4) Patchwork
2026-11-12 14:57 ` ✓ Xe.CI.BAT: " Patchwork
2026-11-12 23:17 ` ✓ Xe.CI.FULL: " Patchwork
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=15eee851-2dc4-4ee4-8e7f-12eceb43f54e@intel.com \
--to=badal.nilawar@intel.com \
--cc=anshuman.gupta@intel.com \
--cc=aravind.iddamsetty@linux.intel.com \
--cc=intel-xe@lists.freedesktop.org \
--cc=jani.nikula@linux.intel.com \
--cc=karthik.poosa@intel.com \
--cc=mallesh.koujalagi@intel.com \
--cc=matthew.brost@intel.com \
--cc=raag.jadav@intel.com \
--cc=riana.tauro@intel.com \
--cc=rodrigo.vivi@intel.com \
--cc=sk.anirban@intel.com \
--cc=vinay.belgaumkar@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