mm: emit tracepoint when rss watermark is hit

Useful to track how rss is changing per tgid. Required for the
memory visibility work being done for Android.

Original patch by Tim Murray:
https://partner-android-review.googlesource.com/c/kernel/private/msm-google/+/1081280

Changes from original patch:
- don't bloat mm_struct
- add some noise reduction to rss tracking

Mot-CRs-fixed: (CR)

Change-Id: Ief904334235ff4380244e5803d7853579e70d202
Signed-off-by: Joel Fernandes <joelaf@google.com>
Reviewed-on: https://gerrit.mot.com/1453726
SME-Granted: SME Approvals Granted
Tested-by: Jira Key
SLTApproved: Slta Waiver
Reviewed-by: Xiangpo Zhao <zhaoxp3@motorola.com>
Submit-Approved: Jira Key

diff --git a/include/linux/mm.h b/include/linux/mm.h
index b14ca733a2b9b7929ab139cdcd6153cccdd737f8..131536db63dbdcb046c2fe7da0dd607e408ddf41 100644 (file)
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -1504,19 +1504,27 @@ static inline unsigned long get_mm_counter(struct mm_struct *mm, int member)
        return (unsigned long)val;
 }
 
+void mm_trace_rss_stat(int member, long count, long value);
+
 static inline void add_mm_counter(struct mm_struct *mm, int member, long value)
 {
-       atomic_long_add(value, &mm->rss_stat.count[member]);
+       long count = atomic_long_add_return(value, &mm->rss_stat.count[member]);
+
+       mm_trace_rss_stat(member, count, value);
 }
 
 static inline void inc_mm_counter(struct mm_struct *mm, int member)
 {
-       atomic_long_inc(&mm->rss_stat.count[member]);
+       long count = atomic_long_inc_return(&mm->rss_stat.count[member]);
+
+       mm_trace_rss_stat(member, count, 1);
 }
 
 static inline void dec_mm_counter(struct mm_struct *mm, int member)
 {
-       atomic_long_dec(&mm->rss_stat.count[member]);
+       long count = atomic_long_dec_return(&mm->rss_stat.count[member]);
+
+       mm_trace_rss_stat(member, count, -1);
 }
 
 /* Optimized variant when page is already known not to be PageAnon */

diff --git a/include/trace/events/kmem.h b/include/trace/events/kmem.h
index 285feeadac39fb7729d33c6118421d1a340e71ab..b5529c0d4e019bff24348fbf9e4d57ffacf43d9e 100644 (file)
--- a/include/trace/events/kmem.h
+++ b/include/trace/events/kmem.h
@@ -318,6 +318,27 @@ TRACE_EVENT(mm_page_alloc_extfrag,
                __entry->change_ownership)
 );
 
+TRACE_EVENT(rss_stat,
+
+       TP_PROTO(int member,
+               long count),
+
+       TP_ARGS(member, count),
+
+       TP_STRUCT__entry(
+               __field(int, member)
+               __field(long, size)
+       ),
+
+       TP_fast_assign(
+               __entry->member = member;
+               __entry->size = (count << PAGE_SHIFT);
+       ),
+
+       TP_printk("member=%d size=%ldB",
+               __entry->member,
+               __entry->size)
+       );
 #endif /* _TRACE_KMEM_H */
 
 /* This part must be outside protection */

diff --git a/mm/memory.c b/mm/memory.c
index 16956844e5535450ea80926db381c4b98c34ff52..a92eba1f28217e866fae36ec18b0597b1511900f 100644 (file)
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -71,6 +71,8 @@
 #include <linux/dax.h>
 #include <linux/oom.h>
 
+#include <trace/events/kmem.h>
+
 #include <asm/io.h>
 #include <asm/mmu_context.h>
 #include <asm/pgalloc.h>
@@ -139,6 +141,21 @@ static int __init init_zero_pfn(void)
 }
 core_initcall(init_zero_pfn);
 
+/*
+ * This threshold is the boundary in the value space, that the counter has to
+ * advance before we trace it. Should be a power of 2. It is to reduce unwanted
+ * trace overhead. The counter is number of pages.
+ */
+#define TRACE_MM_COUNTER_THRESHOLD 128
+
+void mm_trace_rss_stat(int member, long count, long value)
+{
+       long thresh_mask = ~(TRACE_MM_COUNTER_THRESHOLD - 1);
+
+       /* Threshold roll-over, trace it */
+       if ((count & thresh_mask) != ((count - value) & thresh_mask))
+               trace_rss_stat(member, count);
+}
 
 #if defined(SPLIT_RSS_COUNTING)