x86/intel_rdt: Disable PMU access
authorThomas Gleixner <tglx@linutronix.de>
Fri, 3 Aug 2018 11:06:35 +0000 (13:06 +0200)
committerThomas Gleixner <tglx@linutronix.de>
Fri, 3 Aug 2018 11:06:35 +0000 (13:06 +0200)
Peter is objecting to the direct PMU access in RDT. Right now the PMU usage
is broken anyway as it is not coordinated with perf.

Until this discussion settled, disable the PMU mechanics by simply
rejecting the type '2' measurement in the resctrl file.

Reported-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Reinette Chatre <reinette.chatre@intel.com>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: fenghua.yu@intel.com
Cc: tony.luck@intel.com
Cc: vikas.shivappa@linux.intel.com
CC: gavin.hindman@intel.com
Cc: jithu.joseph@intel.com
Cc: hpa@zytor.com
arch/x86/kernel/cpu/intel_rdt_pseudo_lock.c

index f80c58f8adc31e985f0fff32824371cd9f8b3d8f..40f3903ae5d98a9124efde23d164a65554fb24c8 100644 (file)
@@ -1171,7 +1171,7 @@ static ssize_t pseudo_lock_measure_trigger(struct file *file,
        buf[buf_size] = '\0';
        ret = kstrtoint(buf, 10, &sel);
        if (ret == 0) {
-               if (sel != 1 && sel != 2)
+               if (sel != 1)
                        return -EINVAL;
                ret = debugfs_file_get(file->f_path.dentry);
                if (ret)