fs/resctrl: Eliminate false positive lockdep warning when reading SNC counters

Running resctrl_tests on an SNC-2 system with lockdep debugging enabled
triggers several warnings with following trace:

  WARNING: CPU: 0 PID: 1914 at kernel/cpu.c:528 lockdep_assert_cpus_held
  ...
  Call Trace:
  __mon_event_count
  ? __lock_acquire
  ? __pfx___mon_event_count
  mon_event_count
  ? __pfx_smp_mon_event_count
  smp_mon_event_count
  smp_call_on_cpu_callback

get_cpu_cacheinfo_level() called from __mon_event_count() requires CPU hotplug
lock to be held. The hotplug lock is indeed held during this time, as
confirmed by the lockdep_assert_cpus_held() within mon_event_read() that calls
mon_event_count() via IPI, but the lockdep tracking is not able to follow the
IPI.

Fresh CPU cache information via get_cpu_cacheinfo_level() from
__mon_event_count() was added to support the fix for the issue where resctrl
inappropriately maintained links to L3 cache information that will be stale in
the case when the associated CPU goes offline.

Keep the cacheinfo ID in struct rdt_mon_domain to ensure that resctrl does not
maintain stale cache information while CPUs can go offline. Return to using
a pointer to the L3 cache information (struct cacheinfo) in struct rmid_read,
rmid_read::ci. Initialize rmid_read::ci before the IPI where it is used. CPU
hotplug lock is held across rmid_read::ci initialization and use to ensure
that it points to accurate cache information.

Fixes: 594902c ("x86,fs/resctrl: Remove inappropriate references to cacheinfo in the resctrl subsystem")
Signed-off-by: Reinette Chatre <[email protected]>
Signed-off-by: Borislav Petkov (AMD) <[email protected]>

Author: rchatre

fs/resctrl/ctrlmondata.c

@@ -625,11 +625,11 @@ int rdtgroup_mondata_show(struct seq_file *m, void *arg)
 		 */
 		list_for_each_entry(d, &r->mon_domains, hdr.list) {
 			if (d->ci_id == domid) {
-				rr.ci_id = d->ci_id;
 				cpu = cpumask_any(&d->hdr.cpu_mask);
 				ci = get_cpu_cacheinfo_level(cpu, RESCTRL_L3_CACHE);
 				if (!ci)
 					continue;
+				rr.ci = ci;
 				mon_event_read(&rr, r, NULL, rdtgrp,
 					       &ci->shared_cpu_map, evtid, false);
 				goto checkresult;

fs/resctrl/internal.h

@@ -98,7 +98,7 @@ struct mon_data {
  *	   domains in @r sharing L3 @ci.id
  * @evtid: Which monitor event to read.
  * @first: Initialize MBM counter when true.
- * @ci_id: Cacheinfo id for L3. Only set when @d is NULL. Used when summing domains.
+ * @ci:    Cacheinfo for L3. Only set when @d is NULL. Used when summing domains.
  * @err:   Error encountered when reading counter.
  * @val:   Returned value of event counter. If @rgrp is a parent resource group,
  *	   @val includes the sum of event counts from its child resource groups.
@@ -112,7 +112,7 @@ struct rmid_read {
 	struct rdt_mon_domain	*d;
 	enum resctrl_event_id	evtid;
 	bool			first;
-	unsigned int		ci_id;
+	struct cacheinfo	*ci;
 	int			err;
 	u64			val;
 	void			*arch_mon_ctx;

fs/resctrl/monitor.c

@@ -361,7 +361,6 @@ static int __mon_event_count(u32 closid, u32 rmid, struct rmid_read *rr)
 {
 	int cpu = smp_processor_id();
 	struct rdt_mon_domain *d;
-	struct cacheinfo *ci;
 	struct mbm_state *m;
 	int err, ret;
 	u64 tval = 0;
@@ -389,8 +388,7 @@ static int __mon_event_count(u32 closid, u32 rmid, struct rmid_read *rr)
 	}
 
 	/* Summing domains that share a cache, must be on a CPU for that cache. */
-	ci = get_cpu_cacheinfo_level(cpu, RESCTRL_L3_CACHE);
-	if (!ci || ci->id != rr->ci_id)
+	if (!cpumask_test_cpu(cpu, &rr->ci->shared_cpu_map))
 		return -EINVAL;
 
 	/*
@@ -402,7 +400,7 @@ static int __mon_event_count(u32 closid, u32 rmid, struct rmid_read *rr)
 	 */
 	ret = -EINVAL;
 	list_for_each_entry(d, &rr->r->mon_domains, hdr.list) {
-		if (d->ci_id != rr->ci_id)
+		if (d->ci_id != rr->ci->id)
 			continue;
 		err = resctrl_arch_rmid_read(rr->r, d, closid, rmid,
 					     rr->evtid, &tval, rr->arch_mon_ctx);