Thread (8 messages) 8 messages, 3 authors, 2021-10-13

Re: [RESEND PATCH v2 1/2] ftrace: disable preemption between ftrace_test_recursion_trylock/unlock()

From: Miroslav Benes <mbenes@suse.cz>
Date: 2021-10-13 07:55:24
Also in: linux-riscv, linuxppc-dev, lkml

quoted hunk ↗ jump to hunk
diff --git a/include/linux/trace_recursion.h b/include/linux/trace_recursion.h
index a9f9c57..101e1fb 100644
--- a/include/linux/trace_recursion.h
+++ b/include/linux/trace_recursion.h
@@ -208,13 +208,29 @@ static __always_inline void trace_clear_recursion(int bit)
  * Use this for ftrace callbacks. This will detect if the function
  * tracing recursed in the same context (normal vs interrupt),
  *
+ * The ftrace_test_recursion_trylock() will disable preemption,
+ * which is required for the variant of synchronize_rcu() that is
+ * used to allow patching functions where RCU is not watching.
+ * See klp_synchronize_transition() for more details.
+ *
I think that you misunderstood. Steven proposed to put the comment before 
ftrace_test_recursion_trylock() call site in klp_ftrace_handler().
  * Returns: -1 if a recursion happened.
  *           >= 0 if no recursion
  */
 static __always_inline int ftrace_test_recursion_trylock(unsigned long ip,
 							 unsigned long parent_ip)
 {
-	return trace_test_and_set_recursion(ip, parent_ip, TRACE_FTRACE_START, TRACE_FTRACE_MAX);
+	int bit;
+
+	bit = trace_test_and_set_recursion(ip, parent_ip, TRACE_FTRACE_START, TRACE_FTRACE_MAX);
+	/*
+	 * The zero bit indicate we are nested
+	 * in another trylock(), which means the
+	 * preemption already disabled.
+	 */
+	if (bit > 0)
+		preempt_disable_notrace();
+
+	return bit;
 }
[...]
quoted hunk ↗ jump to hunk
diff --git a/kernel/livepatch/patch.c b/kernel/livepatch/patch.c
index e8029ae..6e66ccd 100644
--- a/kernel/livepatch/patch.c
+++ b/kernel/livepatch/patch.c
@@ -52,11 +52,6 @@ static void notrace klp_ftrace_handler(unsigned long ip,
Here
quoted hunk ↗ jump to hunk
 	bit = ftrace_test_recursion_trylock(ip, parent_ip);
 	if (WARN_ON_ONCE(bit < 0))
 		return;
-	/*
-	 * A variant of synchronize_rcu() is used to allow patching functions
-	 * where RCU is not watching, see klp_synchronize_transition().
-	 */
-	preempt_disable_notrace();

 	func = list_first_or_null_rcu(&ops->func_stack, struct klp_func,
 				      stack_node);
@@ -120,7 +115,6 @@ static void notrace klp_ftrace_handler(unsigned long ip,
 	klp_arch_set_pc(fregs, (unsigned long)func->new_func);

 unlock:
-	preempt_enable_notrace();
 	ftrace_test_recursion_unlock(bit);
 }
Side note... the comment will eventually conflict with peterz's 
https://lore.kernel.org/all/20210929152429.125997206@infradead.org/ (local).

Miroslav
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help