linux.git/tools/testing/selftests/resctrl/cache.c, branch v7.2-rc1

selftests/resctrl: Simplify perf usage in CAT test

2026-05-05T00:40:02+00:00

The CAT test relies on the PERF_COUNT_HW_CACHE_MISSES event to determine if
modifying a cache portion size is successful. This event is configured to
report the data as part of an event group, but no other events are added to
the group.

Remove the unnecessary PERF_FORMAT_GROUP format setting. This eliminates
the need for struct perf_event_read and results in read() of the associated
file descriptor to return just one value associated with the
PERF_COUNT_HW_CACHE_MISSES event of interest.

Link: https://lore.kernel.org/r/fb69325eba5031b735fa79effaaacd797c9c6040.1775266384.git.reinette.chatre@intel.com
Signed-off-by: Reinette Chatre 
Tested-by: Chen Yu 
Reviewed-by: Ilpo Järvinen 
Signed-off-by: Shuah Khan

selftests/resctrl: Reduce interference from L2 occupancy during cache occupancy test

2026-05-05T00:40:02+00:00

The CMT test creates a new control group that is also capable of monitoring
and assigns the workload to it. The workload allocates a buffer that by
default fills a portion of the L3 and keeps reading from the buffer,
measuring the L3 occupancy at intervals. The test passes if the workload's
L3 occupancy is within 15% of the buffer size.

The CMT test does not take into account that some of the workload's data
may land in L2/L1. Matching L3 occupancy to the size of the buffer while
a portion of the buffer can be allocated into L2 is not accurate.

Take the L2 cache into account to improve test accuracy:
 - Reduce the workload's L2 cache allocation to the minimum on systems that
   support L2 cache allocation. Do so with a new utility in preparation for
   all L3 cache allocation tests needing the same capability.
 - Increase the buffer size to accommodate data that may be allocated into
   the L2 cache. Use a buffer size double the L3 portion to keep using the
   L3 portion size as goal for L3 occupancy while taking into account that
   some of the data may be in L2.

Running the CMT test on a sample system while introducing significant
cache misses using "stress-ng --matrix-3d 0 --matrix-3d-zyx" shows
significant improvement in L3 cache occupancy:

Before:

    # Starting CMT test ...
    # Mounting resctrl to "/sys/fs/resctrl"
    # Cache size :335544320
    # Writing benchmark parameters to resctrl FS
    # Write schema "L3:0=fffe0" to resctrl FS
    # Write schema "L3:0=1f" to resctrl FS
    # Benchmark PID: 7089
    # Checking for pass/fail
    # Pass: Check cache miss rate within 15%
    # Percent diff=12
    # Number of bits: 5
    # Average LLC val: 73269248
    # Cache span (bytes): 83886080
    ok 1 CMT: test

After:
    # Starting CMT test ...
    # Mounting resctrl to "/sys/fs/resctrl"
    # Cache size :335544320
    # Writing benchmark parameters to resctrl FS
    # Write schema "L3:0=fffe0" to resctrl FS
    # Write schema "L3:0=1f" to resctrl FS
    # Write schema "L2:1=0x1" to resctrl FS
    # Benchmark PID: 7171
    # Checking for pass/fail
    # Pass: Check cache miss rate within 15%
    # Percent diff=0
    # Number of bits: 5
    # Average LLC val: 83755008
    # Cache span (bytes): 83886080
    ok 1 CMT: test

Link: https://lore.kernel.org/r/00445fa64c251b86b86023f87220ee1ad8561460.1775266384.git.reinette.chatre@intel.com
Reported-by: Dave Martin 
Signed-off-by: Reinette Chatre 
Tested-by: Chen Yu 
Reviewed-by: Ilpo Järvinen 
Link: https://lore.kernel.org/lkml/aO+7MeSMV29VdbQs@e133380.arm.com/
Signed-off-by: Shuah Khan

selftests/resctrl: Use correct type for pids

2024-07-11T17:23:53+00:00

A few functions receive PIDs through int arguments. PIDs variables
should be of type pid_t, not int.

Convert pid arguments from int to pid_t.

Before printing PID, match the type to %d by casting to int which is
enough for Linux (standard would allow using a longer integer type but
generalizing for that would complicate the code unnecessarily, the
selftest code does not need to be portable).

Signed-off-by: Ilpo Järvinen 
Tested-by: Babu Moger 
Reviewed-by: Reinette Chatre 
Signed-off-by: Shuah Khan

selftests/resctrl: Open perf fd before start & add error handling

2024-02-13T20:56:44+00:00

Perf fd (pe_fd) is opened, reset, and enabled during every test the CAT
selftest runs. Also, ioctl(pe_fd, ...) calls are not error checked even
if ioctl() could return an error.

Open perf fd only once before the tests and only reset and enable the
counter within the test loop. Add error checking to pe_fd ioctl()
calls.

Signed-off-by: Ilpo Järvinen 
Reviewed-by: Reinette Chatre 
Signed-off-by: Shuah Khan

selftests/resctrl: Move cat_val() to cat_test.c and rename to cat_test()

2024-02-13T20:56:44+00:00

The main CAT test function is called cat_val() and resides in cache.c
which is illogical.

Rename the function to cat_test() and move it into cat_test.c.

Signed-off-by: Ilpo Järvinen 
Reviewed-by: Reinette Chatre 
Signed-off-by: Shuah Khan

selftests/resctrl: Convert perf related globals to locals

2024-02-13T20:56:44+00:00

Perf related variables pea_llc_miss, pe_read, and pe_fd are globals in
cache.c.

Convert them to locals for better scoping and make pea_llc_miss simpler
by renaming it to pea. Make close(pe_fd) handling easier to understand
by doing it inside cat_val().

Make also sizeof()s use safer way to determine the right struct.

Signed-off-by: Ilpo Järvinen 
Reviewed-by: Reinette Chatre 
Signed-off-by: Shuah Khan

selftests/resctrl: Improve perf init

2024-02-13T20:56:44+00:00

struct perf_event_attr initialization is spread into
perf_event_initialize() and perf_event_attr_initialize() and setting
->config is hardcoded by the deepest level.

perf_event_attr init belongs to perf_event_attr_initialize() so move it
entirely there. Rename the other function
perf_event_initialized_read_format().

Call each init function directly from the test as they will take
different parameters (especially true after the perf related global
variables are moved to local variables).

Signed-off-by: Ilpo Järvinen 
Reviewed-by: Reinette Chatre 
Signed-off-by: Shuah Khan

selftests/resctrl: Consolidate naming of perf event related things

2024-02-13T20:56:44+00:00

Naming for perf event related functions, types, and variables is
inconsistent.

Make struct read_format and all functions related to perf events start
with "perf_". Adjust variable names towards the same direction but use
shorter names for variables where appropriate (pe prefix).

Signed-off-by: Ilpo Järvinen 
Reviewed-by: Reinette Chatre 
Signed-off-by: Shuah Khan

selftests/resctrl: Remove nested calls in perf event handling

2024-02-13T20:56:44+00:00

Perf event handling has functions that are the sole caller of another
perf event handling related function:
  - reset_enable_llc_perf() calls perf_event_open_llc_miss()
  - perf_event_measure() calls get_llc_perf()

Remove the extra layer of calls to make the code easier to follow by
moving the code into the calling function.

Signed-off-by: Ilpo Järvinen 
Reviewed-by: Reinette Chatre 
Signed-off-by: Shuah Khan

selftests/resctrl: Remove unnecessary __u64 -> unsigned long conversion

2024-02-13T20:56:44+00:00

Perf counters are __u64 but the code converts them to unsigned long
before printing them out.

Remove unnecessary type conversion and retain the perf originating
value as __u64.

Signed-off-by: Ilpo Järvinen 
Reviewed-by: Reinette Chatre 
Signed-off-by: Shuah Khan