Merge tag 'trace-v5.16' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace

Pull tracing updates from Steven Rostedt:

 - kprobes: Restructured stack unwinder to show properly on x86 when a
   stack dump happens from a kretprobe callback.

 - Fix to bootconfig parsing

 - Have tracefs allow owner and group permissions by default (only
   denying others). There's been pressure to allow non root to tracefs
   in a controlled fashion, and using groups is probably the safest.

 - Bootconfig memory managament updates.

 - Bootconfig clean up to have the tools directory be less dependent on
   changes in the kernel tree.

 - Allow perf to be traced by function tracer.

 - Rewrite of function graph tracer to be a callback from the function
   tracer instead of having its own trampoline (this change will happen
   on an arch by arch basis, and currently only x86_64 implements it).

 - Allow multiple direct trampolines (bpf hooks to functions) be batched
   together in one synchronization.

 - Allow histogram triggers to add variables that can perform
   calculations against the event's fields.

 - Use the linker to determine architecture callbacks from the ftrace
   trampoline to allow for proper parameter prototypes and prevent
   warnings from the compiler.

 - Extend histogram triggers to key off of variables.

 - Have trace recursion use bit magic to determine preempt context over
   if branches.

 - Have trace recursion disable preemption as all use cases do anyway.

 - Added testing for verification of tracing utilities.

 - Various small clean ups and fixes.

* tag 'trace-v5.16' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace: (101 commits)
  tracing/histogram: Fix semicolon.cocci warnings
  tracing/histogram: Fix documentation inline emphasis warning
  tracing: Increase PERF_MAX_TRACE_SIZE to handle Sentinel1 and docker together
  tracing: Show size of requested perf buffer
  bootconfig: Initialize ret in xbc_parse_tree()
  ftrace: do CPU checking after preemption disabled
  ftrace: disable preemption when recursion locked
  tracing/histogram: Document expression arithmetic and constants
  tracing/histogram: Optimize division by a power of 2
  tracing/histogram: Covert expr to const if both operands are constants
  tracing/histogram: Simplify handling of .sym-offset in expressions
  tracing: Fix operator precedence for hist triggers expression
  tracing: Add division and multiplication support for hist triggers
  tracing: Add support for creating hist trigger variables from literal
  selftests/ftrace: Stop tracing while reading the trace file by default
  MAINTAINERS: Update KPROBES and TRACING entries
  test_kprobes: Move it from kernel/ to lib/
  docs, kprobes: Remove invalid URL and add new reference
  samples/kretprobes: Fix return value if register_kretprobe() failed
  lib/bootconfig: Fix the xbc_get_info kerneldoc
  ...

1 files changed, 18 insertions, 9 deletions

diff --git a/include/linux/trace_recursion.h b/include/linux/trace_recursion.h
index fe95f0922526..c303f7a114e9 100644
--- a/include/linux/trace_recursion.h
+++ b/include/linux/trace_recursion.h
@@ -116,13 +116,9 @@ enum {
 
 static __always_inline int trace_get_context_bit(void)
 {
-	unsigned long pc = preempt_count();
+	unsigned char bit = interrupt_context_level();
 
-	if (!(pc & (NMI_MASK | HARDIRQ_MASK | SOFTIRQ_OFFSET)))
-		return TRACE_CTX_NORMAL;
-	else
-		return pc & NMI_MASK ? TRACE_CTX_NMI :
-			pc & HARDIRQ_MASK ? TRACE_CTX_IRQ : TRACE_CTX_SOFTIRQ;
+	return TRACE_CTX_NORMAL - bit;
 }
 
 #ifdef CONFIG_FTRACE_RECORD_RECURSION
@@ -139,6 +135,9 @@ extern void ftrace_record_recursion(unsigned long ip, unsigned long parent_ip);
 # define do_ftrace_record_recursion(ip, pip)	do { } while (0)
 #endif
 
+/*
+ * Preemption is promised to be disabled when return bit >= 0.
+ */
 static __always_inline int trace_test_and_set_recursion(unsigned long ip, unsigned long pip,
 							int start)
 {
@@ -148,8 +147,12 @@ static __always_inline int trace_test_and_set_recursion(unsigned long ip, unsign
 	bit = trace_get_context_bit() + start;
 	if (unlikely(val & (1 << bit))) {
 		/*
-		 * It could be that preempt_count has not been updated during
-		 * a switch between contexts. Allow for a single recursion.
+		 * If an interrupt occurs during a trace, and another trace
+		 * happens in that interrupt but before the preempt_count is
+		 * updated to reflect the new interrupt context, then this
+		 * will think a recursion occurred, and the event will be dropped.
+		 * Let a single instance happen via the TRANSITION_BIT to
+		 * not drop those events.
 		 */
 		bit = TRACE_CTX_TRANSITION + start;
 		if (val & (1 << bit)) {
@@ -162,11 +165,17 @@ static __always_inline int trace_test_and_set_recursion(unsigned long ip, unsign
 	current->trace_recursion = val;
 	barrier();
 
+	preempt_disable_notrace();
+
 	return bit;
 }
 
+/*
+ * Preemption will be enabled (if it was previously enabled).
+ */
 static __always_inline void trace_clear_recursion(int bit)
 {
+	preempt_enable_notrace();
 	barrier();
 	trace_recursion_clear(bit);
 }
@@ -178,7 +187,7 @@ static __always_inline void trace_clear_recursion(int bit)
  * tracing recursed in the same context (normal vs interrupt),
  *
  * Returns: -1 if a recursion happened.
- *           >= 0 if no recursion
+ *           >= 0 if no recursion.
  */
 static __always_inline int ftrace_test_recursion_trylock(unsigned long ip,
 							 unsigned long parent_ip)