The patch titled
Subject: percpu: improve percpu_alloc_percpu event trace
has been added to the -mm mm-unstable branch. Its filename is
percpu-improve-percpu_alloc_percpu-event-trace.patch
This patch should soon appear in the mm-unstable branch at
git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
Before you just go and hit "reply", please:
a) Consider who else should be cc'ed
b) Prefer to cc a suitable mailing list as well
c) Ideally: find the original patch on the mailing list and do a
reply-to-all to that, adding suitable additional cc's
*** Remember to use Documentation/process/submit-checklist.rst when testing your code ***
The -mm tree is included into linux-next via the mm-everything
branch at git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
and is updated there every 2-3 working days
------------------------------------------------------
From: Vasily Averin <vvs@xxxxxxxxxx>
Subject: percpu: improve percpu_alloc_percpu event trace
Add call_site, bytes_alloc and gfp_flags fields to the output of the
percpu_alloc_percpu ftrace event:
mkdir-4393 [001] 169.334788: percpu_alloc_percpu:
call_site=mem_cgroup_css_alloc+0xa6 reserved=0 is_atomic=0 size=2408 align=8
base_addr=0xffffc7117fc00000 off=402176 ptr=0x3dc867a62300 bytes_alloc=14448
gfp_flags=GFP_KERNEL_ACCOUNT
This is required to track memcg-accounted percpu allocations.
Link: https://lkml.kernel.org/r/a07be858-c8a3-7851-9086-e3262cbcf707@xxxxxxxxxx
Signed-off-by: Vasily Averin <vvs@xxxxxxxxxx>
Cc: Shakeel Butt <shakeelb@xxxxxxxxxx>
Cc: Steven Rostedt <rostedt@xxxxxxxxxxx>
Cc: Ingo Molnar <mingo@xxxxxxxxxx>
Cc: Roman Gushchin <roman.gushchin@xxxxxxxxx>
Cc: Vlastimil Babka <vbabka@xxxxxxx>
Cc: Michal Hocko <mhocko@xxxxxxxx>
Cc: Dennis Zhou <dennis@xxxxxxxxxx>
Cc: Tejun Heo <tj@xxxxxxxxxx>
Cc: Christoph Lameter <cl@xxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
include/trace/events/percpu.h | 23 +++++++++++++++++------
mm/percpu-internal.h | 8 ++++----
mm/percpu.c | 5 +++--
3 files changed, 24 insertions(+), 12 deletions(-)
--- a/include/trace/events/percpu.h~percpu-improve-percpu_alloc_percpu-event-trace
+++ a/include/trace/events/percpu.h
@@ -6,15 +6,20 @@
#define _TRACE_PERCPU_H
#include <linux/tracepoint.h>
+#include <trace/events/mmflags.h>
TRACE_EVENT(percpu_alloc_percpu,
- TP_PROTO(bool reserved, bool is_atomic, size_t size,
- size_t align, void *base_addr, int off, void __percpu *ptr),
+ TP_PROTO(unsigned long call_site,
+ bool reserved, bool is_atomic, size_t size,
+ size_t align, void *base_addr, int off,
+ void __percpu *ptr, size_t bytes_alloc, gfp_t gfp_flags),
- TP_ARGS(reserved, is_atomic, size, align, base_addr, off, ptr),
+ TP_ARGS(call_site, reserved, is_atomic, size, align, base_addr, off,
+ ptr, bytes_alloc, gfp_flags),
TP_STRUCT__entry(
+ __field( unsigned long, call_site )
__field( bool, reserved )
__field( bool, is_atomic )
__field( size_t, size )
@@ -22,9 +27,11 @@ TRACE_EVENT(percpu_alloc_percpu,
__field( void *, base_addr )
__field( int, off )
__field( void __percpu *, ptr )
+ __field( size_t, bytes_alloc )
+ __field( gfp_t, gfp_flags )
),
-
TP_fast_assign(
+ __entry->call_site = call_site;
__entry->reserved = reserved;
__entry->is_atomic = is_atomic;
__entry->size = size;
@@ -32,12 +39,16 @@ TRACE_EVENT(percpu_alloc_percpu,
__entry->base_addr = base_addr;
__entry->off = off;
__entry->ptr = ptr;
+ __entry->bytes_alloc = bytes_alloc;
+ __entry->gfp_flags = gfp_flags;
),
- TP_printk("reserved=%d is_atomic=%d size=%zu align=%zu base_addr=%p off=%d ptr=%p",
+ TP_printk("call_site=%pS reserved=%d is_atomic=%d size=%zu align=%zu base_addr=%p off=%d ptr=%p bytes_alloc=%zu gfp_flags=%s",
+ (void *)__entry->call_site,
__entry->reserved, __entry->is_atomic,
__entry->size, __entry->align,
- __entry->base_addr, __entry->off, __entry->ptr)
+ __entry->base_addr, __entry->off, __entry->ptr,
+ __entry->bytes_alloc, show_gfp_flags(__entry->gfp_flags))
);
TRACE_EVENT(percpu_free_percpu,
--- a/mm/percpu.c~percpu-improve-percpu_alloc_percpu-event-trace
+++ a/mm/percpu.c
@@ -1884,8 +1884,9 @@ area_found:
ptr = __addr_to_pcpu_ptr(chunk->base_addr + off);
kmemleak_alloc_percpu(ptr, size, gfp);
- trace_percpu_alloc_percpu(reserved, is_atomic, size, align,
- chunk->base_addr, off, ptr);
+ trace_percpu_alloc_percpu(_RET_IP_, reserved, is_atomic, size, align,
+ chunk->base_addr, off, ptr,
+ pcpu_obj_full_size(size), gfp);
pcpu_memcg_post_alloc_hook(objcg, chunk, off, size);
--- a/mm/percpu-internal.h~percpu-improve-percpu_alloc_percpu-event-trace
+++ a/mm/percpu-internal.h
@@ -113,7 +113,6 @@ static inline int pcpu_chunk_map_bits(st
return pcpu_nr_pages_to_map_bits(chunk->nr_pages);
}
-#ifdef CONFIG_MEMCG_KMEM
/**
* pcpu_obj_full_size - helper to calculate size of each accounted object
* @size: size of area to allocate in bytes
@@ -123,13 +122,14 @@ static inline int pcpu_chunk_map_bits(st
*/
static inline size_t pcpu_obj_full_size(size_t size)
{
- size_t extra_size;
+ size_t extra_size = 0;
- extra_size = size / PCPU_MIN_ALLOC_SIZE * sizeof(struct obj_cgroup *);
+#ifdef CONFIG_MEMCG_KMEM
+ extra_size += size / PCPU_MIN_ALLOC_SIZE * sizeof(struct obj_cgroup *);
+#endif
return size * num_possible_cpus() + extra_size;
}
-#endif /* CONFIG_MEMCG_KMEM */
#ifdef CONFIG_PERCPU_STATS
_
Patches currently in -mm which might be from vvs@xxxxxxxxxx are
percpu-improve-percpu_alloc_percpu-event-trace.patch
