Re: [PATCH v4 0/2] fix & prevent the missing preemption disabling

On 2021/10/26 上午10:42, Steven Rostedt wrote:

On Tue, 26 Oct 2021 10:09:12 +0800
王贇 [off-list ref] wrote:

quoted

Just a ping, to see if there are any more comments :-P

I guess you missed what went into mainline (and your name found a bug
in my perl script for importing patches ;-)

  https://lore.kernel.org/all/20211019091344.65629198@gandalf.local.home/ (local)

Cool~ Missing some chinese font maybe, that's fine :-)

Which means patch 1 needs to change:

quoted

+	/*
+	 * Disable preemption to fulfill the promise.
+	 *
+	 * Don't worry about the bit 0 cases, they indicate
+	 * the disabling behaviour has already been done by
+	 * internal call previously.
+	 */
+	preempt_disable_notrace();
+
 	return bit + 1;
 }

+/*
+ * Preemption will be enabled (if it was previously enabled).
+ */
 static __always_inline void trace_clear_recursion(int bit)
 {
 	if (!bit)
 		return;

+	if (bit > 0)
+		preempt_enable_notrace();
+

Where this wont work anymore.

Need to preempt disable and enable always.

Yup, will rebase on the latest changes~

Regards,
Michael Wang

-- Steve

quoted

 	barrier();
 	bit--;
 	trace_recursion_clear(bit);

@@ -209,7 +227,7 @@ static __always_inline void trace_clear_recursion(int bit)
  * tracing recursed in the same context (normal vs interrupt),
  *