[PATCH v2] tests/intel/xe_fault_injection: Ignore all errors while injecting fault

[Kamil Konieczny]

Hi Jonathan,
On 2025-06-03 at 19:36:39 +0000, Jonathan Cavitt wrote:
> From: Satyanarayana K V P <satyanarayana.k.v.p at intel.com>
> 
> Currently, numerous fault messages have been included in the dmesg
> ignore list, and this list continues to expand.  Each time a new fault
> injection point is introduced or a new feature is activated, additional
> fault messages appear, making it cumbersome to manage the dmesg ignore
> list.
> 
> However, we can safely assert that all dmesg reports that contain
> *ERROR* in their message can be ignored, so add them to the dmesg ignore
> list.  This unfortunately does not include the device probe error
> itself, so that must be added separately.
> 
> While we're here, we should also assert that any errors we see are only
> coming from the target PCI device.
> 
> v2:
> - Only ignore error-level dmesg reports (or, at least, reports with
>   *ERROR* in them), and device probe failues
> - Add PCI data to regex (Michal)
> 
> v3: (Michal)
> - Revert name change
> - Add change log
> - Remove fixes tag from commit
> - Rename ignore_faults_in_dmesg to igt_ignore_dmesg_errors_from_dut, and
>   move to lib/igt_core.c
> - Minor code fixes
> 
> Signed-off-by: Satyanarayana K V P <satyanarayana.k.v.p at intel.com>
> Signed-off-by: Jonathan Cavitt <jonathan.cavitt at intel.com>
> Suggested-by: Michal Wajdeczko <michal.wajdeczko at intel.com>
> Suggested-by: Daniele Ceraolo Spurio <daniele.ceraolospurio at intel.com>
> Suggested-by: Lucas De Marchi <lucas.demarchi at intel.com>
> Cc: Francois Dugast <francois.dugast at intel.com>
> Cc: Rodrigo Vivi <rodrigo.vivi at intel.com>
> Cc: John Harrison <john.c.harrison at intel.com>
> ---
>  lib/igt_core.c                   | 27 ++++++++++++++++++++++++
>  lib/igt_core.h                   |  1 +
>  tests/intel/xe_fault_injection.c | 36 +++++---------------------------
>  3 files changed, 33 insertions(+), 31 deletions(-)
> 
> diff --git a/lib/igt_core.c b/lib/igt_core.c
> index b06cdfd894..e666660ff3 100644
> --- a/lib/igt_core.c
> +++ b/lib/igt_core.c
> @@ -76,6 +76,7 @@
>  #include "igt_rc.h"
>  #include "igt_list.h"
>  #include "igt_map.h"
> +#include "igt_device.h"
>  #include "igt_device_scan.h"
>  #include "igt_thread.h"
>  #include "igt_vec.h"
> @@ -3581,6 +3582,32 @@ void igt_emit_ignore_dmesg_regex(const char *ignore_dmesg_regex)
>  	igt_kmsg(KMSG_INFO "%s%s\n", mark_ignore_dmesg, ignore_dmesg_regex);
>  }
>  
> +/**
> + * igt_ignore_dmesg_errors_from_dut(int fd)
> + * @fd: The file directory
> + *
> + * Submits an igt_emit_ignore_dmesg_regex request targeting all error-level
> + * dmesg reports (or, at least, reports that contain the string *ERROR*),
> + * as well as any device probe errors, that are reported from the pci
> + * slot of the given file directory @fd
> + */
> +void igt_ignore_dmesg_errors_from_dut(int fd)
> +{
> +	/*
> +	 * Catch any dmesg reports that contain the substring "*ERROR*", as well as any
> +	 * device probe errors.
> +	 */
> +	static const char *store = "probe with driver xe failed with error|\\*ERROR\\*";

I am not convinced you need a special function in lib for this,
why not just a function in xe fault test? You could think about
moving this into lib when there are more users from other tests.

Regards,
Kamil

> +	char pci_slot[NAME_MAX];
> +	char regex[1024];
> +
> +	/* Only block dmesg reports that target the pci slot of the given fd */
> +	igt_device_get_pci_slot_name(fd, pci_slot);
> +	snprintf(regex, sizeof(regex), "%s:.*(%s)", pci_slot, store);
> +
> +	igt_emit_ignore_dmesg_regex(regex);
> +}
> +
>  /**
>   * @igt_measured_usleep: Helper to model accurate sleep time for tests
>   * @usec: usec to sleep
> diff --git a/lib/igt_core.h b/lib/igt_core.h
> index 2db579423c..f6affbaf57 100644
> --- a/lib/igt_core.h
> +++ b/lib/igt_core.h
> @@ -1599,6 +1599,7 @@ int igt_pci_system_reinit(void);
>  void igt_pci_system_cleanup(void);
>  
>  void igt_emit_ignore_dmesg_regex(const char *ignore_dmesg_regex);
> +void igt_ignore_dmesg_errors_from_dut(int fd);
>  
>  unsigned int igt_measured_usleep(unsigned int usec);
>  #endif /* IGT_CORE_H */
> diff --git a/tests/intel/xe_fault_injection.c b/tests/intel/xe_fault_injection.c
> index 9fe6bfe351..594c1417a2 100644
> --- a/tests/intel/xe_fault_injection.c
> +++ b/tests/intel/xe_fault_injection.c
> @@ -64,32 +64,6 @@ static int fail_function_open(void)
>  	return debugfs_fail_function_dir_fd;
>  }
>  
> -static bool function_is_part_of_guc(const char function_name[])
> -{
> -	return strstr(function_name, "_guc_") != NULL ||
> -	       strstr(function_name, "_uc_") != NULL ||
> -	       strstr(function_name, "_wopcm_") != NULL;
> -}
> -
> -static void ignore_faults_in_dmesg(const char function_name[])
> -{
> -	/* Driver probe is expected to fail in all cases, so ignore in igt_runner */
> -	char regex[1024] = "probe with driver xe failed with error -12";
> -
> -	/*
> -	 * If GuC module fault is injected, GuC is expected to fail,
> -	 * so also ignore GuC init failures in igt_runner.
> -	 */
> -	if (function_is_part_of_guc(function_name)) {
> -		strcat(regex, "|GT[0-9a-fA-F]*: GuC init failed with -ENOMEM");
> -		strcat(regex, "|GT[0-9a-fA-F]*: Failed to initialize uC .-ENOMEM");
> -		strcat(regex, "|GT[0-9a-fA-F]*: Failed to enable GuC CT .-ENOMEM");
> -		strcat(regex, "|GT[0-9a-fA-F]*: GuC PC query task state failed: -ENOMEM");
> -	}
> -
> -	igt_emit_ignore_dmesg_regex(regex);
> -}
> -
>  /*
>   * The injectable file requires CONFIG_FUNCTION_ERROR_INJECTION in kernel.
>   */
> @@ -234,7 +208,7 @@ inject_fault_probe(int fd, char pci_slot[], const char function_name[])
>  	igt_info("Injecting error \"%s\" (%d) in function \"%s\"\n",
>  		 strerror(-INJECT_ERRNO), INJECT_ERRNO, function_name);
>  
> -	ignore_faults_in_dmesg(function_name);
> +	igt_ignore_dmesg_errors_from_dut(fd);
>  	injection_list_add(function_name);
>  	set_retval(function_name, INJECT_ERRNO);
>  
> @@ -299,7 +273,7 @@ exec_queue_create_fail(int fd, struct drm_xe_engine_class_instance *instance,
>  	igt_assert_eq(__xe_exec_queue_create(fd, vm, 1, 1, instance, 0, &exec_queue_id), 0);
>  	xe_exec_queue_destroy(fd, exec_queue_id);
>  
> -	ignore_faults_in_dmesg(function_name);
> +	igt_ignore_dmesg_errors_from_dut(fd);
>  	injection_list_add(function_name);
>  	set_retval(function_name, INJECT_ERRNO);
>  	igt_assert(__xe_exec_queue_create(fd, vm, 1, 1, instance, 0, &exec_queue_id) != 0);
> @@ -334,7 +308,7 @@ vm_create_fail(int fd, const char function_name[], unsigned int flags)
>  {
>  	igt_assert_eq(simple_vm_create(fd, flags), 0);
>  
> -	ignore_faults_in_dmesg(function_name);
> +	igt_ignore_dmesg_errors_from_dut(fd);
>  	injection_list_add(function_name);
>  	set_retval(function_name, INJECT_ERRNO);
>  	igt_assert(simple_vm_create(fd, flags) != 0);
> @@ -397,7 +371,7 @@ vm_bind_fail(int fd, const char function_name[])
>  
>  	igt_assert_eq(simple_vm_bind(fd, vm), 0);
>  
> -	ignore_faults_in_dmesg(function_name);
> +	igt_ignore_dmesg_errors_from_dut(fd);
>  	injection_list_add(function_name);
>  	set_retval(function_name, INJECT_ERRNO);
>  	igt_assert(simple_vm_bind(fd, vm) != 0);
> @@ -445,7 +419,7 @@ oa_add_config_fail(int fd, int sysfs, int devid, const char function_name[])
>  	igt_assert(igt_sysfs_scanf(sysfs, path, "%" PRIu64, &config_id) == 1);
>  	igt_assert_eq(intel_xe_perf_ioctl(fd, DRM_XE_OBSERVATION_OP_REMOVE_CONFIG, &config_id), 0);
>  
> -	ignore_faults_in_dmesg(function_name);
> +	igt_ignore_dmesg_errors_from_dut(fd);
>  	injection_list_add(function_name);
>  	set_retval(function_name, INJECT_ERRNO);
>  	igt_assert_lt(intel_xe_perf_ioctl(fd, DRM_XE_OBSERVATION_OP_ADD_CONFIG, &config), 0);
> -- 
> 2.43.0
>