Donate to e Foundation | Murena handsets with /e/OS | Own a part of Murena! Learn more

Commit 1836ac85 authored by Ingo Molnar's avatar Ingo Molnar
Browse files

Merge tag 'perf-core-for-mingo-3' of... 

Merge tag 'perf-core-for-mingo-3' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux

 into perf/core

Pull perf/core improvements and fixes from Arnaldo Carvalho de Melo:

User visible changes:

  - Improve --filter support for 'perf probe', allowing using its arguments
    on other commands, as --add, --del, etc (Masami Hiramatsu)

  - Show warning when running 'perf kmem stat' on a unsuitable perf.data file,
    i.e. one with events that are not the ones required for the stat variant
    used (Namhyung Kim).

Infrastructure changes:

  - Auxtrace support patches, paving the way to support Intel PT and BTS (Adrian Hunter)

  - hists browser (top, report) refactorings (Namhyung Kim)

Signed-off-by: default avatarArnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: default avatarIngo Molnar <mingo@kernel.org>
parents 5ebe6afa 3698dab1

Makefile

+2 −3
Original line number Diff line number Diff line
@@ -215,7 +215,6 @@ VPATH := $(srctree)$(if $(KBUILD_EXTMOD),:$(KBUILD_EXTMOD)) 


export srctree objtree VPATH





# SUBARCH tells the usermode build what the underlying arch is.  That is set

# first, and if a usermode build is happening, the "ARCH=um" on the command

# line overrides the setting of ARCH below.  If a native build is happening,
@@ -1497,11 +1496,11 @@ image_name: 
# Clear a bunch of variables before executing the submake

tools/: FORCE

	$(Q)mkdir -p $(objtree)/tools

	$(Q)$(MAKE) LDFLAGS= MAKEFLAGS="$(filter --j% -j,$(MAKEFLAGS))" O=$(objtree) subdir=tools -C $(src)/tools/

	$(Q)$(MAKE) LDFLAGS= MAKEFLAGS="$(filter --j% -j,$(MAKEFLAGS))" O=$(O) subdir=tools -C $(src)/tools/



tools/%: FORCE

	$(Q)mkdir -p $(objtree)/tools

	$(Q)$(MAKE) LDFLAGS= MAKEFLAGS="$(filter --j% -j,$(MAKEFLAGS))" O=$(objtree) subdir=tools -C $(src)/tools/ $*

	$(Q)$(MAKE) LDFLAGS= MAKEFLAGS="$(filter --j% -j,$(MAKEFLAGS))" O=$(O) subdir=tools -C $(src)/tools/ $*



# Single targets

# ---------------------------------------------------------------------------

tools/Makefile

+15 −5
Original line number Diff line number Diff line 
# Some of the tools (perf) use same make variables

# as in kernel build.

export srctree=

export objtree=



include scripts/Makefile.include



help:
@@ -47,11 +52,16 @@ cgroup firewire hv guest usb virtio vm net: FORCE 
liblockdep: FORCE

	$(call descend,lib/lockdep)



libapikfs: FORCE

libapi: FORCE

	$(call descend,lib/api)



perf: libapikfs FORCE

	$(call descend,$@)

# The perf build does not follow the descend function setup,

# invoking it via it's own make rule.

PERF_O   = $(if $(O),$(O)/tools/perf,)



perf: FORCE

	$(Q)mkdir -p $(PERF_O) .

	$(Q)$(MAKE) --no-print-directory -C perf O=$(PERF_O) subdir=



selftests: FORCE

	$(call descend,testing/$@)
@@ -97,10 +107,10 @@ cgroup_clean hv_clean firewire_clean lguest_clean usb_clean virtio_clean vm_clea 
liblockdep_clean:

	$(call descend,lib/lockdep,clean)



libapikfs_clean:

libapi_clean:

	$(call descend,lib/api,clean)



perf_clean: libapikfs_clean

perf_clean:

	$(call descend,$(@:_clean=),clean)



selftests_clean:

tools/lib/traceevent/event-parse.c

+4 −2
Original line number Diff line number Diff line
@@ -1387,7 +1387,7 @@ static int event_read_fields(struct event_format *event, struct format_field **f 
			do_warning_event(event, "%s: no type found", __func__);

			goto fail;

		}

		field->name = last_token;

		field->name = field->alias = last_token;



		if (test_type(type, EVENT_OP))

			goto fail;
@@ -1469,7 +1469,7 @@ static int event_read_fields(struct event_format *event, struct format_field **f 
				size_dynamic = type_size(field->name);

				free_token(field->name);

				strcat(field->type, brackets);

				field->name = token;

				field->name = field->alias = token;

				type = read_token(&token);

			} else {

				char *new_type;
@@ -6444,6 +6444,8 @@ void pevent_ref(struct pevent *pevent) 
void pevent_free_format_field(struct format_field *field)

{

	free(field->type);

	if (field->alias != field->name)

		free(field->alias);

	free(field->name);

	free(field);

}

tools/lib/traceevent/event-parse.h

+1 −0
Original line number Diff line number Diff line
@@ -191,6 +191,7 @@ struct format_field { 
	struct event_format	*event;

	char			*type;

	char			*name;

	char			*alias;

	int			offset;

	int			size;

	unsigned int		arraylen;

tools/perf/Documentation/callchain-overhead-calculation.txt

0 → 100644
+108 −0
Original line number Diff line number Diff line 
Overhead calculation

--------------------

The overhead can be shown in two columns as 'Children' and 'Self' when

perf collects callchains.  The 'self' overhead is simply calculated by

adding all period values of the entry - usually a function (symbol).

This is the value that perf shows traditionally and sum of all the

'self' overhead values should be 100%.



The 'children' overhead is calculated by adding all period values of

the child functions so that it can show the total overhead of the

higher level functions even if they don't directly execute much.

'Children' here means functions that are called from another (parent)

function.



It might be confusing that the sum of all the 'children' overhead

values exceeds 100% since each of them is already an accumulation of

'self' overhead of its child functions.  But with this enabled, users

can find which function has the most overhead even if samples are

spread over the children.



Consider the following example; there are three functions like below.



-----------------------

void foo(void) {

    /* do something */

}



void bar(void) {

    /* do something */

    foo();

}



int main(void) {

    bar()

    return 0;

}

-----------------------



In this case 'foo' is a child of 'bar', and 'bar' is an immediate

child of 'main' so 'foo' also is a child of 'main'.  In other words,

'main' is a parent of 'foo' and 'bar', and 'bar' is a parent of 'foo'.



Suppose all samples are recorded in 'foo' and 'bar' only.  When it's

recorded with callchains the output will show something like below

in the usual (self-overhead-only) output of perf report:



----------------------------------

Overhead  Symbol

........  .....................

  60.00%  foo

          |

          --- foo

              bar

              main

              __libc_start_main



  40.00%  bar

          |

          --- bar

              main

              __libc_start_main

----------------------------------



When the --children option is enabled, the 'self' overhead values of

child functions (i.e. 'foo' and 'bar') are added to the parents to

calculate the 'children' overhead.  In this case the report could be

displayed as:



-------------------------------------------

Children      Self  Symbol

........  ........  ....................

 100.00%     0.00%  __libc_start_main

          |

          --- __libc_start_main



 100.00%     0.00%  main

          |

          --- main

              __libc_start_main



 100.00%    40.00%  bar

          |

          --- bar

              main

              __libc_start_main



  60.00%    60.00%  foo

          |

          --- foo

              bar

              main

              __libc_start_main

-------------------------------------------



In the above output, the 'self' overhead of 'foo' (60%) was add to the

'children' overhead of 'bar', 'main' and '\_\_libc_start_main'.

Likewise, the 'self' overhead of 'bar' (40%) was added to the

'children' overhead of 'main' and '\_\_libc_start_main'.



So '\_\_libc_start_main' and 'main' are shown first since they have

same (100%) 'children' overhead (even though they have zero 'self'

overhead) and they are the parents of 'foo' and 'bar'.



Since v3.16 the 'children' overhead is shown by default and the output

is sorted by its values. The 'children' overhead is disabled by

specifying --no-children option on the command line or by adding

'report.children = false' or 'top.children = false' in the perf config

file.