Re: [PATCH v3 08/10] selftests/resctrl: Remove requirement on cache miss rate

[Reinette Chatre <reinette.chatre@intel.com>]

Hi Ilpo,

On 3/27/26 10:45 AM, Ilpo Järvinen wrote:
> On Fri, 13 Mar 2026, Reinette Chatre wrote:
>> -/*
>> - * Minimum difference in LLC misses between a test with n+1 bits CBM to the
>> - * test with n bits is MIN_DIFF_PERCENT_PER_BIT * (n - 1). With e.g. 5 vs 4
>> - * bits in the CBM mask, the minimum difference must be at least
>> - * MIN_DIFF_PERCENT_PER_BIT * (4 - 1) = 3 percent.
>> - *
>> - * The relationship between number of used CBM bits and difference in LLC
>> - * misses is not expected to be linear. With a small number of bits, the
>> - * margin is smaller than with larger number of bits. For selftest purposes,
>> - * however, linear approach is enough because ultimately only pass/fail
>> - * decision has to be made and distinction between strong and stronger
>> - * signal is irrelevant.
>> - */
>> -#define MIN_DIFF_PERCENT_PER_BIT	1UL
>> -
>>  static int show_results_info(__u64 sum_llc_val, int no_of_bits,
>>  			     unsigned long cache_span,
>> -			     unsigned long min_diff_percent,
>>  			     unsigned long num_of_runs, bool platform,
>>  			     __s64 *prev_avg_llc_val)
>>  {
>>  	__u64 avg_llc_val = 0;
>> -	float avg_diff;
>>  	int ret = 0;
>>  
>>  	avg_llc_val = sum_llc_val / num_of_runs;
>>  	if (*prev_avg_llc_val) {
>> -		float delta = (__s64)(avg_llc_val - *prev_avg_llc_val);
>> -
>> -		avg_diff = delta / *prev_avg_llc_val;
>> -		ret = platform && (avg_diff * 100) < (float)min_diff_percent;
>> -
>> -		ksft_print_msg("%s Check cache miss rate changed more than %.1f%%\n",
>> -			       ret ? "Fail:" : "Pass:", (float)min_diff_percent);
>> +		ret = platform && (avg_llc_val < *prev_avg_llc_val);
>>  
>> -		ksft_print_msg("Percent diff=%.1f\n", avg_diff * 100);
>> +		ksft_print_msg("%s Check cache miss rate increased\n",
>> +			       ret ? "Fail:" : "Pass:");
> 
> While I'm fine with removing the amount of change check, this no longer 
> shows any numbers which would be a bit annoying if/when there's a failure.
> 

This snippet only removes display of the number that is no longer computed ("avg_diff").
The values that are compared now, avg_llc_val and it previous value, are printed
in the call to show_cache_info() that follows this snippet but is not visible in the diff.
 
Below is an example of what a user running the CAT test will see after these changes.
Since show_cache_info() always prints avg_llc_val the user can obtain insight into failure
by considering it and its previous measurement.

# Starting L3_CAT test ...
# Mounting resctrl to "/sys/fs/resctrl"
# Cache size :117964800
# Writing benchmark parameters to resctrl FS
# Write schema "L2:1=0x1" to resctrl FS
# Write schema "L3:0=1fc0" to resctrl FS
# Write schema "L3:0=3f" to resctrl FS
# Write schema "L3:0=1ff0" to resctrl FS
# Write schema "L3:0=f" to resctrl FS
# Write schema "L3:0=1ffc" to resctrl FS
# Write schema "L3:0=3" to resctrl FS
# Checking for pass/fail
# Number of bits: 6
# Average LLC val: 445092
# Cache span (lines): 737280
# Pass: Check cache miss rate increased
# Number of bits: 4
# Average LLC val: 724472
# Cache span (lines): 491520
# Pass: Check cache miss rate increased
# Number of bits: 2
# Average LLC val: 1085470
# Cache span (lines): 245760
ok 4 L3_CAT: test

Reinette