tracing: Some more updates for 4.15

PATCH New Steven Rostedt 1 participant 5 messages Oct 16, 2017

Subsystems (2)

Tracing Printk

lore

Share & tools

lore

Steven Rostedt

author maintainer

2017-10-16 22:02 UTC #0 lore raw

cover tracing: Some more updates for 4.15

git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace.git
for-next

Head SHA1: f40a37cb4916f17806b8a89d8eb76f6943c69189

Arnd Bergmann (1):
tracing: mark trace_test_buffer as __maybe_unused

Steven Rostedt (VMware) (3):
printk: Remove superfluous memory barriers from printk_safe tracing/xen: Hide events that are not used when X86_PAE is not defined tracing, memcg, vmscan: Hide trace events when not in use

----

 include/trace/events/vmscan.h |  4 ++++
 include/trace/events/xen.h    | 35 +++++++++++++++++++----------------
 kernel/printk/printk_safe.c   | 15 ++++++++-------
 kernel/trace/trace_selftest.c |  2 +-
 4 files changed, 32 insertions(+), 24 deletions(-)

Steven Rostedt

author maintainer

2017-10-16 22:02 UTC #1 lore raw

PATCH 1/4 printk: Remove superfluous memory barriers from printk_safe

From: "Steven Rostedt (VMware)" <[email protected]>

The variable printk_safe_irq_ready is set and never cleared at system boot up, when there's only one CPU active. It is set before other CPUs come on line. Also, it is extremely unlikely that an NMI would trigger this early in boot up (which I wonder why we even have this variable at all).

Also mark the printk_safe_irq_ready as read mostly, as it is set at system boot up, and never touched again.

Link http://lkml.kernel.org/r/[email protected] open here

Reviewed-by Petr Mladek <[email protected]>

Signed-off-by Steven Rostedt (VMware) <[email protected]>

---
 kernel/printk/printk_safe.c | 15 ++++++++-------
 1 file changed, 8 insertions(+), 7 deletions(-)

kernel/printk/printk_safe.c +8 -7

index 3cdaeaef9ce1..724d9292d4b9 100644
@@ -39,7 +39,7 @@
  * There are situations when we want to make sure that all buffers
  * were handled or when IRQs are blocked.
  */
-static int printk_safe_irq_ready;
+static int printk_safe_irq_ready __read_mostly;
 
 #define SAFE_LOG_BUF_LEN ((1 << CONFIG_PRINTK_SAFE_LOG_BUF_SHIFT) -	\
 				sizeof(atomic_t) -			\

@@ -63,11 +63,8 @@ static DEFINE_PER_CPU(struct printk_safe_seq_buf, nmi_print_seq);
 /* Get flushed in a more safe context. */
 static void queue_flush_work(struct printk_safe_seq_buf *s)
 {
-	if (printk_safe_irq_ready) {
-		/* Make sure that IRQ work is really initialized. */
-		smp_rmb();
+	if (printk_safe_irq_ready)
 		irq_work_queue(&s->work);
-	}
 }
 
 /*

@@ -398,8 +395,12 @@ void __init printk_safe_init(void)
 #endif
 	}
 
-	/* Make sure that IRQ works are initialized before enabling. */
-	smp_wmb();
+	/*
+	 * In the highly unlikely event that a NMI were to trigger at
+	 * this moment. Make sure IRQ work is set up before this
+	 * variable is set.
+	 */
+	barrier();
 	printk_safe_irq_ready = 1;
 
 	/* Flush pending messages that did not have scheduled IRQ works. */

Signature · 1 line

2.13.2

Steven Rostedt

author maintainer

2017-10-16 22:02 UTC #2 lore raw

PATCH 2/4 tracing: mark trace_test_buffer as __maybe_unused

In reply to Steven Rostedt

From: Arnd Bergmann <[email protected]>

After trace_selftest_startup_sched_switch is removed, trace_test_buffer() is only used sometimes, leading to this warning:

kernel/trace/trace_selftest.c:62:12: error: 'trace_test_buffer' defined but not used [-Werror=unused-function]

There is no simple #ifdef condition that captures well whether the function is in fact used or not, so marking it as __maybe_unused is probably the best way to shut up the warning. The function will then be silently dropped when there is no user.

Link http://lkml.kernel.org/r/[email protected] open here

Fixes d8c4deee6dc6 ("tracing: Remove obsolete sched_switch tracer selftest")

Signed-off-by Arnd Bergmann <[email protected]>

Signed-off-by Steven Rostedt (VMware) <[email protected]>

---
 kernel/trace/trace_selftest.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

kernel/trace/trace_selftest.c +1 -1

index 364f78abdf47..eb9ba5c1ba40 100644
@@ -59,7 +59,7 @@ static int trace_test_buffer_cpu(struct trace_buffer *buf, int cpu)
  * Test the trace buffer to see if all the elements
  * are still sane.
  */
-static int trace_test_buffer(struct trace_buffer *buf, unsigned long *count)
+static int __maybe_unused trace_test_buffer(struct trace_buffer *buf, unsigned long *count)
 {
 	unsigned long flags, cnt = 0;
 	int cpu, ret = 0;

Signature · 1 line

2.13.2

Steven Rostedt

author maintainer

2017-10-16 22:03 UTC #3 lore raw

PATCH 3/4 tracing/xen: Hide events that are not used when X86_PAE is not defined

In reply to Steven Rostedt

From: "Steven Rostedt (VMware)" <[email protected]>

TRACE_EVENTS() take up memory. If they are defined but not used, then they simply waste space. If their use case is behind a define, then the trace events should be as well.

The trace events xen_mmu_set_pte_atomic, xen_mmu_pte_clear, and xen_mmu_pmd_clear are not used when CONFIG_X86_PAE is not defined.

Link http://lkml.kernel.org/r/[email protected] open here

Reviewed-by Boris Ostrovsky <[email protected]>

Reviewed-by Jeremy Linton <[email protected]>

Signed-off-by Steven Rostedt (VMware) <[email protected]>

---
 include/trace/events/xen.h | 35 +++++++++++++++++++----------------
 1 file changed, 19 insertions(+), 16 deletions(-)

include/trace/events/xen.h +19 ~~-16~~

index 1b4fed72f573..6118d82334ec 100644
@@ -147,7 +147,6 @@ DECLARE_EVENT_CLASS(xen_mmu__set_pte,
 		     TP_ARGS(ptep, pteval))
 
 DEFINE_XEN_MMU_SET_PTE(xen_mmu_set_pte);
-DEFINE_XEN_MMU_SET_PTE(xen_mmu_set_pte_atomic);
 
 TRACE_EVENT(xen_mmu_set_pte_at,
 	    TP_PROTO(struct mm_struct *mm, unsigned long addr,

@@ -169,21 +168,6 @@ TRACE_EVENT(xen_mmu_set_pte_at,
 		      (int)sizeof(pteval_t) * 2, (unsigned long long)__entry->pteval)
 	);
 
-TRACE_EVENT(xen_mmu_pte_clear,
-	    TP_PROTO(struct mm_struct *mm, unsigned long addr, pte_t *ptep),
-	    TP_ARGS(mm, addr, ptep),
-	    TP_STRUCT__entry(
-		    __field(struct mm_struct *, mm)
-		    __field(unsigned long, addr)
-		    __field(pte_t *, ptep)
-		    ),
-	    TP_fast_assign(__entry->mm = mm;
-			   __entry->addr = addr;
-			   __entry->ptep = ptep),
-	    TP_printk("mm %p addr %lx ptep %p",
-		      __entry->mm, __entry->addr, __entry->ptep)
-	);
-
 TRACE_DEFINE_SIZEOF(pmdval_t);
 
 TRACE_EVENT(xen_mmu_set_pmd,

@@ -201,6 +185,24 @@ TRACE_EVENT(xen_mmu_set_pmd,
 		      (int)sizeof(pmdval_t) * 2, (unsigned long long)__entry->pmdval)
 	);
 
+#ifdef CONFIG_X86_PAE
+DEFINE_XEN_MMU_SET_PTE(xen_mmu_set_pte_atomic);
+
+TRACE_EVENT(xen_mmu_pte_clear,
+	    TP_PROTO(struct mm_struct *mm, unsigned long addr, pte_t *ptep),
+	    TP_ARGS(mm, addr, ptep),
+	    TP_STRUCT__entry(
+		    __field(struct mm_struct *, mm)
+		    __field(unsigned long, addr)
+		    __field(pte_t *, ptep)
+		    ),
+	    TP_fast_assign(__entry->mm = mm;
+			   __entry->addr = addr;
+			   __entry->ptep = ptep),
+	    TP_printk("mm %p addr %lx ptep %p",
+		      __entry->mm, __entry->addr, __entry->ptep)
+	);
+
 TRACE_EVENT(xen_mmu_pmd_clear,
 	    TP_PROTO(pmd_t *pmdp),
 	    TP_ARGS(pmdp),

@@ -210,6 +212,7 @@ TRACE_EVENT(xen_mmu_pmd_clear,
 	    TP_fast_assign(__entry->pmdp = pmdp),
 	    TP_printk("pmdp %p", __entry->pmdp)
 	);
+#endif
 
 #if CONFIG_PGTABLE_LEVELS >= 4

Signature · 1 line

2.13.2

Steven Rostedt

author maintainer

2017-10-16 22:03 UTC #4 lore raw

PATCH 4/4 tracing, memcg, vmscan: Hide trace events when not in use

In reply to Steven Rostedt

From: "Steven Rostedt (VMware)" <[email protected]>

When trace events are defined but not used they still create data structures and functions for their use, even though nothing may be using them.

The trace events mm_vmscan_memcg_reclaim_begin, mm_vmscan_memcg_softlimit_reclaim_begin, mm_vmscan_memcg_reclaim_end, and mm_vmscan_memcg_softlimit_reclaim_end are not used if CONFIG_MEMCG is not defined. Do not create these trace events unless CONFIG_MEMCG is defined.

Link http://lkml.kernel.org/r/[email protected] open here

Acked-by Michal Hocko <[email protected]>

Signed-off-by Steven Rostedt (VMware) <[email protected]>

---
 include/trace/events/vmscan.h | 4 ++++
 1 file changed, 4 insertions(+)

include/trace/events/vmscan.h +4 -0

index 27e8a5c77579..cde5fea0a179 100644
@@ -133,6 +133,7 @@ DEFINE_EVENT(mm_vmscan_direct_reclaim_begin_template, mm_vmscan_direct_reclaim_b
 	TP_ARGS(order, may_writepage, gfp_flags, classzone_idx)
 );
 
+#ifdef CONFIG_MEMCG
 DEFINE_EVENT(mm_vmscan_direct_reclaim_begin_template, mm_vmscan_memcg_reclaim_begin,
 
 	TP_PROTO(int order, int may_writepage, gfp_t gfp_flags, int classzone_idx),

@@ -146,6 +147,7 @@ DEFINE_EVENT(mm_vmscan_direct_reclaim_begin_template, mm_vmscan_memcg_softlimit_
 
 	TP_ARGS(order, may_writepage, gfp_flags, classzone_idx)
 );
+#endif /* CONFIG_MEMCG */
 
 DECLARE_EVENT_CLASS(mm_vmscan_direct_reclaim_end_template,

@@ -171,6 +173,7 @@ DEFINE_EVENT(mm_vmscan_direct_reclaim_end_template, mm_vmscan_direct_reclaim_end
 	TP_ARGS(nr_reclaimed)
 );
 
+#ifdef CONFIG_MEMCG
 DEFINE_EVENT(mm_vmscan_direct_reclaim_end_template, mm_vmscan_memcg_reclaim_end,
 
 	TP_PROTO(unsigned long nr_reclaimed),

@@ -184,6 +187,7 @@ DEFINE_EVENT(mm_vmscan_direct_reclaim_end_template, mm_vmscan_memcg_softlimit_re
 
 	TP_ARGS(nr_reclaimed)
 );
+#endif /* CONFIG_MEMCG */
 
 TRACE_EVENT(mm_shrink_slab_start,
 	TP_PROTO(struct shrinker *shr, struct shrink_control *sc,

Signature · 1 line

2.13.2