lkml.org 
[lkml]   [2022]   [Oct]   [27]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
From
Subject[RFC PATCH 2/2] tracing/user_events: Fixup enable faults asyncly
Date
When events are enabled within the various tracing facilities, such as
ftrace/perf, the event_mutex is held. As events are enabled pages are
accessed. We do not want page faults to occur under this lock. Instead
queue the fault to a workqueue to be handled in a process context safe
way without the lock.

The enable address is disabled while the async fault-in occurs. This
ensures that we don't attempt fault-in more than is necessary. Once the
page has been faulted in, the address write is attempted again. If the
page couldn't fault-in, then we wait until the next time the event is
enabled to prevent any potential infinite loops.

Signed-off-by: Beau Belgrave <beaub@linux.microsoft.com>
---
kernel/trace/trace_events_user.c | 125 ++++++++++++++++++++++++++++++-
1 file changed, 121 insertions(+), 4 deletions(-)

diff --git a/kernel/trace/trace_events_user.c b/kernel/trace/trace_events_user.c
index 633f24c2a1ac..f1eb8101e053 100644
--- a/kernel/trace/trace_events_user.c
+++ b/kernel/trace/trace_events_user.c
@@ -81,11 +81,22 @@ struct user_event_enabler {
struct list_head link;
struct mm_struct *mm;
struct file *file;
+ refcount_t refcnt;
unsigned long enable_addr;
unsigned int enable_bit: 5,
- __reserved: 27;
+ __reserved: 26,
+ disabled: 1;
};

+/* Used for asynchronous faulting in of pages */
+struct user_event_enabler_fault {
+ struct work_struct work;
+ struct user_event_enabler *enabler;
+ struct user_event *event;
+};
+
+static struct kmem_cache *fault_cache;
+
/*
* Stores per-event properties, as users register events
* within a file a user_event might be created if it does not
@@ -236,6 +247,19 @@ static void user_event_enabler_destroy(struct user_event_enabler *enabler)
kfree(enabler);
}

+static __always_inline struct user_event_enabler
+*user_event_enabler_get(struct user_event_enabler *enabler)
+{
+ refcount_inc(&enabler->refcnt);
+ return enabler;
+}
+
+static void user_event_enabler_put(struct user_event_enabler *enabler)
+{
+ if (refcount_dec_and_test(&enabler->refcnt))
+ user_event_enabler_destroy(enabler);
+}
+
static void user_event_enabler_remove(struct file *file,
struct user_event *user)
{
@@ -249,13 +273,93 @@ static void user_event_enabler_remove(struct file *file,
if (enabler->file != file)
continue;

+ enabler->disabled = 0;
list_del(&enabler->link);
- user_event_enabler_destroy(enabler);
+ user_event_enabler_put(enabler);
}

mutex_unlock(&event_mutex);
}

+static void user_event_enabler_write(struct user_event_enabler *enabler,
+ struct user_event *user);
+
+static void user_event_enabler_fault_fixup(struct work_struct *work)
+{
+ struct user_event_enabler_fault *fault = container_of(
+ work, struct user_event_enabler_fault, work);
+ struct user_event_enabler *enabler = fault->enabler;
+ struct user_event *user = fault->event;
+ struct mm_struct *mm = enabler->mm;
+ unsigned long uaddr = enabler->enable_addr;
+ bool unlocked = false;
+ int ret;
+
+ might_sleep();
+
+ mmap_read_lock(mm);
+
+ ret = fixup_user_fault(mm, uaddr, FAULT_FLAG_WRITE | FAULT_FLAG_REMOTE,
+ &unlocked);
+
+ mmap_read_unlock(mm);
+
+ if (ret)
+ pr_warn("user_events: Fixup fault failed with %d "
+ "for mm: 0x%pK offset: 0x%llx event: %s\n", ret, mm,
+ (unsigned long long)uaddr, EVENT_NAME(user));
+
+ /* Prevent state changes from racing */
+ mutex_lock(&event_mutex);
+
+ /*
+ * If we managed to get the page, re-issue the write. We do not
+ * want to get into a possible infinite loop, which is why we only
+ * attempt again directly if the page came in. If we couldn't get
+ * the page here, then we will try again the next time the event is
+ * enabled/disabled.
+ */
+ enabler->disabled = 0;
+
+ if (!ret)
+ user_event_enabler_write(enabler, user);
+
+ mutex_unlock(&event_mutex);
+
+ refcount_dec(&user->refcnt);
+ user_event_enabler_put(enabler);
+ kmem_cache_free(fault_cache, fault);
+}
+
+static bool user_event_enabler_queue_fault(struct user_event_enabler *enabler,
+ struct user_event *user)
+{
+ struct user_event_enabler_fault *fault;
+
+ fault = kmem_cache_zalloc(fault_cache, GFP_NOWAIT | __GFP_NOWARN);
+
+ if (!fault)
+ return false;
+
+ INIT_WORK(&fault->work, user_event_enabler_fault_fixup);
+ fault->enabler = user_event_enabler_get(enabler);
+ fault->event = user;
+
+ refcount_inc(&user->refcnt);
+ enabler->disabled = 1;
+
+ if (!schedule_work(&fault->work)) {
+ enabler->disabled = 0;
+ refcount_dec(&user->refcnt);
+ user_event_enabler_put(enabler);
+ kmem_cache_free(fault_cache, fault);
+
+ return false;
+ }
+
+ return true;
+}
+
static void user_event_enabler_write(struct user_event_enabler *enabler,
struct user_event *user)
{
@@ -266,6 +370,11 @@ static void user_event_enabler_write(struct user_event_enabler *enabler,
void *kaddr;
int ret;

+ lockdep_assert_held(&event_mutex);
+
+ if (unlikely(enabler->disabled))
+ return;
+
mmap_read_lock(mm);

ret = pin_user_pages_remote(mm, uaddr, 1, FOLL_WRITE | FOLL_NOFAULT,
@@ -273,8 +382,10 @@ static void user_event_enabler_write(struct user_event_enabler *enabler,

mmap_read_unlock(mm);

- if (ret <= 0) {
- pr_warn("user_events: Enable write failed\n");
+ if (unlikely(ret <= 0)) {
+ if (!user_event_enabler_queue_fault(enabler, user))
+ pr_warn("user_events: Unable to queue fault handler\n");
+
return;
}

@@ -321,6 +432,7 @@ static struct user_event_enabler
enabler->file = file;
enabler->enable_addr = (unsigned long)reg->enable_addr;
enabler->enable_bit = reg->enable_bit;
+ refcount_set(&enabler->refcnt, 1);

/* Prevents state changes from racing with new enablers */
mutex_lock(&event_mutex);
@@ -1902,6 +2014,11 @@ static int __init trace_events_user_init(void)
{
int ret;

+ fault_cache = KMEM_CACHE(user_event_enabler_fault, 0);
+
+ if (!fault_cache)
+ return -ENOMEM;
+
init_group = user_event_group_create(&init_user_ns);

if (!init_group)
--
2.25.1
\
 
 \ /
  Last update: 2022-10-28 00:41    [W:0.121 / U:1.368 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site