resctrl has one mutex that is taken by the architecture specific code,
and the filesystem parts. The two interact via cpuhp, where the
architecture code updates the domain list. Filesystem handlers that
walk the domains list should not run concurrently with the cpuhp
callback modifying the list.
Exposing a lock from the filesystem code means the interface is not
cleanly defined, and creates the possibility of cross-architecture
lock ordering headaches. The interaction only exists so that certain
filesystem paths are serialised against cpu hotplug. The cpu hotplug
code already has a mechanism to do this using cpus_read_lock().
MPAM's monitors have an overflow interrupt, so it needs to be possible
to walk the domains list in irq context. RCU is ideal for this,
but some paths need to be able to sleep to allocate memory.
Because resctrl_{on,off}line_cpu() take the rdtgroup_mutex as part
of a cpuhp callback, cpus_read_lock() must always be taken first.
rdtgroup_schemata_write() already does this.
All but one of the filesystem code's domain list walkers are
currently protected by the rdtgroup_mutex taken in
rdtgroup_kn_lock_live(). The exception is rdt_bit_usage_show()
which takes the lock directly.
Make the domain list protected by RCU. An architecture-specific
lock prevents concurrent writers. rdt_bit_usage_show() can
walk the domain list under rcu_read_lock().
The other filesystem list walkers need to be able to sleep.
Add cpus_read_lock() to rdtgroup_kn_lock_live() so that the
cpuhp callbacks can't be invoked when file system operations are
occurring.
Add lockdep_assert_cpus_held() in the cases where the
rdtgroup_kn_lock_live() call isn't obvious.
Resctrl's domain online/offline calls now need to take the
rdtgroup_mutex themselves.
Signed-off-by: James Morse <[email protected]>
---
arch/x86/kernel/cpu/resctrl/core.c | 33 ++++++++------
arch/x86/kernel/cpu/resctrl/ctrlmondata.c | 14 ++++--
arch/x86/kernel/cpu/resctrl/monitor.c | 3 ++
arch/x86/kernel/cpu/resctrl/pseudo_lock.c | 3 ++
arch/x86/kernel/cpu/resctrl/rdtgroup.c | 54 ++++++++++++++++++++---
include/linux/resctrl.h | 2 +-
6 files changed, 84 insertions(+), 25 deletions(-)
diff --git a/arch/x86/kernel/cpu/resctrl/core.c b/arch/x86/kernel/cpu/resctrl/core.c
index 00b098b56e42..c4e4ce5bf87f 100644
--- a/arch/x86/kernel/cpu/resctrl/core.c
+++ b/arch/x86/kernel/cpu/resctrl/core.c
@@ -25,8 +25,14 @@
#include <asm/resctrl.h>
#include "internal.h"
-/* Mutex to protect rdtgroup access. */
-DEFINE_MUTEX(rdtgroup_mutex);
+/*
+ * rdt_domain structures are kfree()d when their last cpu goes offline,
+ * and allocated when the first cpu in a new domain comes online.
+ * The rdt_resource's domain list is updated when this happens. The domain
+ * list is protected by RCU, but callers can also take the cpus_read_lock()
+ * to prevent modification if they need to sleep. All writers take this mutex:
+ */
+static DEFINE_MUTEX(domain_list_lock);
/*
* The cached resctrl_pqr_state is strictly per CPU and can never be
@@ -489,6 +495,8 @@ static void domain_add_cpu(int cpu, struct rdt_resource *r)
struct rdt_domain *d;
int err;
+ lockdep_assert_held(&domain_list_lock);
+
d = rdt_find_domain(r, id, &add_pos);
if (IS_ERR(d)) {
pr_warn("Couldn't find cache id for CPU %d\n", cpu);
@@ -522,11 +530,12 @@ static void domain_add_cpu(int cpu, struct rdt_resource *r)
return;
}
- list_add_tail(&d->list, add_pos);
+ list_add_tail_rcu(&d->list, add_pos);
err = resctrl_online_domain(r, d);
if (err) {
- list_del(&d->list);
+ list_del_rcu(&d->list);
+ synchronize_rcu();
domain_free(hw_dom);
}
}
@@ -547,7 +556,8 @@ static void domain_remove_cpu(int cpu, struct rdt_resource *r)
cpumask_clear_cpu(cpu, &d->cpu_mask);
if (cpumask_empty(&d->cpu_mask)) {
resctrl_offline_domain(r, d);
- list_del(&d->list);
+ list_del_rcu(&d->list);
+ synchronize_rcu();
/*
* rdt_domain "d" is going to be freed below, so clear
@@ -575,30 +585,27 @@ static void clear_closid_rmid(int cpu)
static int resctrl_arch_online_cpu(unsigned int cpu)
{
struct rdt_resource *r;
- int err;
- mutex_lock(&rdtgroup_mutex);
+ mutex_lock(&domain_list_lock);
for_each_capable_rdt_resource(r)
domain_add_cpu(cpu, r);
clear_closid_rmid(cpu);
+ mutex_unlock(&domain_list_lock);
- err = resctrl_online_cpu(cpu);
- mutex_unlock(&rdtgroup_mutex);
-
- return err;
+ return resctrl_online_cpu(cpu);
}
static int resctrl_arch_offline_cpu(unsigned int cpu)
{
struct rdt_resource *r;
- mutex_lock(&rdtgroup_mutex);
resctrl_offline_cpu(cpu);
+ mutex_lock(&domain_list_lock);
for_each_capable_rdt_resource(r)
domain_remove_cpu(cpu, r);
clear_closid_rmid(cpu);
- mutex_unlock(&rdtgroup_mutex);
+ mutex_unlock(&domain_list_lock);
return 0;
}
diff --git a/arch/x86/kernel/cpu/resctrl/ctrlmondata.c b/arch/x86/kernel/cpu/resctrl/ctrlmondata.c
index 842266d1d148..dc700a09d661 100644
--- a/arch/x86/kernel/cpu/resctrl/ctrlmondata.c
+++ b/arch/x86/kernel/cpu/resctrl/ctrlmondata.c
@@ -209,6 +209,9 @@ static int parse_line(char *line, struct resctrl_schema *s,
struct rdt_domain *d;
unsigned long dom_id;
+ /* Walking r->domains, ensure it can't race with cpuhp */
+ lockdep_assert_cpus_held();
+
if (rdtgrp->mode == RDT_MODE_PSEUDO_LOCKSETUP &&
r->rid == RDT_RESOURCE_MBA) {
rdt_last_cmd_puts("Cannot pseudo-lock MBA resource\n");
@@ -314,6 +317,9 @@ int resctrl_arch_update_domains(struct rdt_resource *r, u32 closid)
int cpu;
u32 idx;
+ /* Walking r->domains, ensure it can't race with cpuhp */
+ lockdep_assert_cpus_held();
+
if (!zalloc_cpumask_var(&cpu_mask, GFP_KERNEL))
return -ENOMEM;
@@ -384,11 +390,9 @@ ssize_t rdtgroup_schemata_write(struct kernfs_open_file *of,
return -EINVAL;
buf[nbytes - 1] = '\0';
- cpus_read_lock();
rdtgrp = rdtgroup_kn_lock_live(of->kn);
if (!rdtgrp) {
rdtgroup_kn_unlock(of->kn);
- cpus_read_unlock();
return -ENOENT;
}
rdt_last_cmd_clear();
@@ -452,7 +456,6 @@ ssize_t rdtgroup_schemata_write(struct kernfs_open_file *of,
out:
rdtgroup_kn_unlock(of->kn);
- cpus_read_unlock();
return ret ?: nbytes;
}
@@ -472,6 +475,9 @@ static void show_doms(struct seq_file *s, struct resctrl_schema *schema, int clo
bool sep = false;
u32 ctrl_val;
+ /* Walking r->domains, ensure it can't race with cpuhp */
+ lockdep_assert_cpus_held();
+
seq_printf(s, "%*s:", max_name_width, schema->name);
list_for_each_entry(dom, &r->domains, list) {
if (sep)
@@ -534,7 +540,7 @@ void mon_event_read(struct rmid_read *rr, struct rdt_resource *r,
int evtid, int first)
{
/* When picking a cpu from cpu_mask, ensure it can't race with cpuhp */
- lockdep_assert_held(&rdtgroup_mutex);
+ lockdep_assert_cpus_held();
/*
* setup the parameters to pass to mon_event_count() to read the data.
diff --git a/arch/x86/kernel/cpu/resctrl/monitor.c b/arch/x86/kernel/cpu/resctrl/monitor.c
index 42a7fd6d56c9..777085df07cf 100644
--- a/arch/x86/kernel/cpu/resctrl/monitor.c
+++ b/arch/x86/kernel/cpu/resctrl/monitor.c
@@ -414,6 +414,9 @@ static void add_rmid_to_limbo(struct rmid_entry *entry)
u32 idx;
int err;
+ /* Walking r->domains, ensure it can't race with cpuhp */
+ lockdep_assert_cpus_held();
+
idx = resctrl_arch_rmid_idx_encode(entry->closid, entry->rmid);
arch_mon_ctx = resctrl_arch_mon_ctx_alloc(r, QOS_L3_OCCUP_EVENT_ID);
diff --git a/arch/x86/kernel/cpu/resctrl/pseudo_lock.c b/arch/x86/kernel/cpu/resctrl/pseudo_lock.c
index 48f196007e9c..690dc6c9f3d4 100644
--- a/arch/x86/kernel/cpu/resctrl/pseudo_lock.c
+++ b/arch/x86/kernel/cpu/resctrl/pseudo_lock.c
@@ -830,6 +830,9 @@ bool rdtgroup_pseudo_locked_in_hierarchy(struct rdt_domain *d)
struct rdt_domain *d_i;
bool ret = false;
+ /* Walking r->domains, ensure it can't race with cpuhp */
+ lockdep_assert_cpus_held();
+
if (!zalloc_cpumask_var(&cpu_with_psl, GFP_KERNEL))
return true;
diff --git a/arch/x86/kernel/cpu/resctrl/rdtgroup.c b/arch/x86/kernel/cpu/resctrl/rdtgroup.c
index f8ea04d95c42..5febb6dfaea0 100644
--- a/arch/x86/kernel/cpu/resctrl/rdtgroup.c
+++ b/arch/x86/kernel/cpu/resctrl/rdtgroup.c
@@ -35,6 +35,10 @@
DEFINE_STATIC_KEY_FALSE(rdt_enable_key);
DEFINE_STATIC_KEY_FALSE(rdt_mon_enable_key);
DEFINE_STATIC_KEY_FALSE(rdt_alloc_enable_key);
+
+/* Mutex to protect rdtgroup access. */
+DEFINE_MUTEX(rdtgroup_mutex);
+
static struct kernfs_root *rdt_root;
struct rdtgroup rdtgroup_default;
LIST_HEAD(rdt_all_groups);
@@ -930,7 +934,8 @@ static int rdt_bit_usage_show(struct kernfs_open_file *of,
mutex_lock(&rdtgroup_mutex);
hw_shareable = r->cache.shareable_bits;
- list_for_each_entry(dom, &r->domains, list) {
+ rcu_read_lock();
+ list_for_each_entry_rcu(dom, &r->domains, list) {
if (sep)
seq_putc(seq, ';');
sw_shareable = 0;
@@ -986,8 +991,10 @@ static int rdt_bit_usage_show(struct kernfs_open_file *of,
}
sep = true;
}
+ rcu_read_unlock();
seq_putc(seq, '\n');
mutex_unlock(&rdtgroup_mutex);
+
return 0;
}
@@ -1227,6 +1234,9 @@ static bool rdtgroup_mode_test_exclusive(struct rdtgroup *rdtgrp)
struct rdt_domain *d;
u32 ctrl;
+ /* Walking r->domains, ensure it can't race with cpuhp */
+ lockdep_assert_cpus_held();
+
list_for_each_entry(s, &resctrl_schema_all, list) {
r = s->res;
if (r->rid == RDT_RESOURCE_MBA)
@@ -1860,6 +1870,9 @@ static int set_cache_qos_cfg(int level, bool enable)
struct rdt_domain *d;
int cpu;
+ /* Walking r->domains, ensure it can't race with cpuhp */
+ lockdep_assert_cpus_held();
+
if (level == RDT_RESOURCE_L3)
update = l3_qos_cfg_update;
else if (level == RDT_RESOURCE_L2)
@@ -2052,6 +2065,7 @@ struct rdtgroup *rdtgroup_kn_lock_live(struct kernfs_node *kn)
atomic_inc(&rdtgrp->waitcount);
kernfs_break_active_protection(kn);
+ cpus_read_lock();
mutex_lock(&rdtgroup_mutex);
/* Was this group deleted while we waited? */
@@ -2069,6 +2083,7 @@ void rdtgroup_kn_unlock(struct kernfs_node *kn)
return;
mutex_unlock(&rdtgroup_mutex);
+ cpus_read_unlock();
if (atomic_dec_and_test(&rdtgrp->waitcount) &&
(rdtgrp->flags & RDT_DELETED)) {
@@ -2365,6 +2380,9 @@ static int reset_all_ctrls(struct rdt_resource *r)
struct rdt_domain *d;
int i, cpu;
+ /* Walking r->domains, ensure it can't race with cpuhp */
+ lockdep_assert_cpus_held();
+
if (!zalloc_cpumask_var(&cpu_mask, GFP_KERNEL))
return -ENOMEM;
@@ -2645,6 +2663,9 @@ static int mkdir_mondata_subdir_alldom(struct kernfs_node *parent_kn,
struct rdt_domain *dom;
int ret;
+ /* Walking r->domains, ensure it can't race with cpuhp */
+ lockdep_assert_cpus_held();
+
list_for_each_entry(dom, &r->domains, list) {
ret = mkdir_mondata_subdir(parent_kn, dom, r, prgrp);
if (ret)
@@ -3328,7 +3349,8 @@ static void domain_destroy_mon_state(struct rdt_domain *d)
kfree(d->mbm_local);
}
-void resctrl_offline_domain(struct rdt_resource *r, struct rdt_domain *d)
+static void _resctrl_offline_domain(struct rdt_resource *r,
+ struct rdt_domain *d)
{
lockdep_assert_held(&rdtgroup_mutex);
@@ -3363,6 +3385,13 @@ void resctrl_offline_domain(struct rdt_resource *r, struct rdt_domain *d)
domain_destroy_mon_state(d);
}
+void resctrl_offline_domain(struct rdt_resource *r, struct rdt_domain *d)
+{
+ mutex_lock(&rdtgroup_mutex);
+ _resctrl_offline_domain(r, d);
+ mutex_unlock(&rdtgroup_mutex);
+}
+
static int domain_setup_mon_state(struct rdt_resource *r, struct rdt_domain *d)
{
u32 idx_limit = resctrl_arch_system_num_rmid_idx();
@@ -3394,7 +3423,7 @@ static int domain_setup_mon_state(struct rdt_resource *r, struct rdt_domain *d)
return 0;
}
-int resctrl_online_domain(struct rdt_resource *r, struct rdt_domain *d)
+static int _resctrl_online_domain(struct rdt_resource *r, struct rdt_domain *d)
{
int err;
@@ -3425,12 +3454,23 @@ int resctrl_online_domain(struct rdt_resource *r, struct rdt_domain *d)
return 0;
}
+int resctrl_online_domain(struct rdt_resource *r, struct rdt_domain *d)
+{
+ int err;
+
+ mutex_lock(&rdtgroup_mutex);
+ err = _resctrl_online_domain(r, d);
+ mutex_unlock(&rdtgroup_mutex);
+
+ return err;
+}
+
int resctrl_online_cpu(unsigned int cpu)
{
- lockdep_assert_held(&rdtgroup_mutex);
-
+ mutex_lock(&rdtgroup_mutex);
/* The cpu is set in default rdtgroup after online. */
cpumask_set_cpu(cpu, &rdtgroup_default.cpu_mask);
+ mutex_unlock(&rdtgroup_mutex);
return 0;
}
@@ -3451,8 +3491,7 @@ void resctrl_offline_cpu(unsigned int cpu)
struct rdtgroup *rdtgrp;
struct rdt_resource *l3 = &rdt_resources_all[RDT_RESOURCE_L3].r_resctrl;
- lockdep_assert_held(&rdtgroup_mutex);
-
+ mutex_lock(&rdtgroup_mutex);
list_for_each_entry(rdtgrp, &rdt_all_groups, rdtgroup_list) {
if (cpumask_test_and_clear_cpu(cpu, &rdtgrp->cpu_mask)) {
clear_childcpus(rdtgrp, cpu);
@@ -3472,6 +3511,7 @@ void resctrl_offline_cpu(unsigned int cpu)
cqm_setup_limbo_handler(d, 0, cpu);
}
}
+ mutex_unlock(&rdtgroup_mutex);
}
/*
diff --git a/include/linux/resctrl.h b/include/linux/resctrl.h
index 530fb21cf9a5..c5af581790dc 100644
--- a/include/linux/resctrl.h
+++ b/include/linux/resctrl.h
@@ -153,7 +153,7 @@ struct resctrl_schema;
* @cache_level: Which cache level defines scope of this resource
* @cache: Cache allocation related data
* @membw: If the component has bandwidth controls, their properties.
- * @domains: All domains for this resource
+ * @domains: RCU list of all domains for this resource
* @name: Name to use in "schemata" file.
* @data_width: Character width of data when displaying
* @default_ctrl: Specifies default cache cbm or memory B/W percent.
--
2.30.2
Hi James,
On Fri, Oct 21, 2022 at 3:13 PM James Morse <[email protected]> wrote:
>
> MPAM's monitors have an overflow interrupt, so it needs to be possible
> to walk the domains list in irq context. RCU is ideal for this,
> but some paths need to be able to sleep to allocate memory.
I'm curious about this requirement. There are already counters which can
overflow on Intel, but we've been able to detect overflows soon enough
by checking at a reasonable interval. Are we expecting MSCs to have
counters that overflow so quickly that the overflows need to be handled
directly in IRQ context vs being able to run a threaded handler before
the second overflow?
It seems like MBM would be really intrusive if it could cause the system
to process overflow IRQs at a high rate.
Also is the overflow interrupt handler in one of your MPAM preview
branches? I was only able to find an error IRQ handler in
mpam/snapshot/v6.0:
https://git.kernel.org/pub/scm/linux/kernel/git/morse/linux.git/tree/drivers/platform/mpam/mpam_devices.c?h=mpam/snapshot/v6.0#n1813
-Peter
Hi Peter,
On 31/10/2022 14:21, Peter Newman wrote:
> On Fri, Oct 21, 2022 at 3:13 PM James Morse <[email protected]> wrote:
>>
>> MPAM's monitors have an overflow interrupt, so it needs to be possible
>> to walk the domains list in irq context. RCU is ideal for this,
>> but some paths need to be able to sleep to allocate memory.
> I'm curious about this requirement. There are already counters which can
> overflow on Intel, but we've been able to detect overflows soon enough
> by checking at a reasonable interval. Are we expecting MSCs to have
> counters that overflow so quickly that the overflows need to be handled
> directly in IRQ context vs being able to run a threaded handler before
> the second overflow?
Ultimately, I don't know. MPAM has three sizes of counter, 31, 44 and 63.
I think its entirely possible someone builds a system with an inconvenient size for the
way they use it - but this wasn't how I anticipated this getting used...
> It seems like MBM would be really intrusive if it could cause the system
> to process overflow IRQs at a high rate.
... I agree ..
> Also is the overflow interrupt handler in one of your MPAM preview
> branches? I was only able to find an error IRQ handler in
> mpam/snapshot/v6.0:
>
> https://git.kernel.org/pub/scm/linux/kernel/git/morse/linux.git/tree/drivers/platform/mpam/mpam_devices.c?h=mpam/snapshot/v6.0#n1813
Because there probably won't be enough monitors to expose the free-running resctrl files,
I anticipate that most use of the memory bandwidth counters will be via perf, which gives
MPAM the start/stop calls it needs to allocate and free a monitor.
The PMU driver gets asked to read the counters in IRQ context, see __perf_event_read()
called via smp_call_function_single(). (I'm sure there are others).
This is the first reason why the domain list needs to be protected by something like RCU.
Perf also has a sampling mode, where it sets the value to overflow after a specific number
of events, and times how long that takes to occur. (I haven't completely got my head round
it yet) In this mode, the MPAM driver would need to invoke the PMU driver to read the
counters in IRQ context.
I haven't written the overflow interrupt code yet because I've got no access to a
platform (virtual or otherwise) with working counters, so couldn't possibly test it. (See
also, the 44 and 63 bit counter support!)
I was being lazy with the commit message and only describing the last case. Is the
future-plot-arc important? It would result in a bit of a brain-dump/essay in the commit
message.
Thanks,
James