Re: [PATCH v4 04/16] perf kvm: Use subtraction for comparison metrics

From: Arnaldo Carvalho de Melo
Date: Mon Mar 13 2023 - 13:58:33 EST


Em Mon, Mar 13, 2023 at 07:40:06PM +0800, Leo Yan escreveu:
> Currently the metrics comparison uses greater operator (>), it returns
> the boolean value (0 or 1).
>
> This patch changes to use subtraction as comparison result, which can
> be used by histograms sorting. Since the subtraction result is u64
> type, we change the function return type from int to int64_t to avoid
> overflow.

Not building at his point in the series:

LD /tmp/build/perf-tools-next/util/perf-in.o
builtin-kvm.c:80:18: error: initialization of ‘int (*)(struct kvm_event *, struct kvm_event *, int)’ from incompatible pointer type ‘int64_t (*)(struct kvm_event *, struct kvm_event *, int)’ {aka ‘long int (*)(struct kvm_event *, struct kvm_event *, int)’} [-Werror=incompatible-pointer-types]
80 | { #name, cmp_event_ ## compare_key }
| ^~~~~~~~~~
builtin-kvm.c:83:9: note: in expansion of macro ‘DEF_SORT_NAME_KEY’
83 | DEF_SORT_NAME_KEY(sample, count),
| ^~~~~~~~~~~~~~~~~
builtin-kvm.c:80:18: note: (near initialization for ‘keys[0].key’)
80 | { #name, cmp_event_ ## compare_key }
| ^~~~~~~~~~
builtin-kvm.c:83:9: note: in expansion of macro ‘DEF_SORT_NAME_KEY’
83 | DEF_SORT_NAME_KEY(sample, count),
| ^~~~~~~~~~~~~~~~~
builtin-kvm.c:80:18: error: initialization of ‘int (*)(struct kvm_event *, struct kvm_event *, int)’ from incompatible pointer type ‘int64_t (*)(struct kvm_event *, struct kvm_event *, int)’ {aka ‘long int (*)(struct kvm_event *, struct kvm_event *, int)’} [-Werror=incompatible-pointer-types]
80 | { #name, cmp_event_ ## compare_key }
| ^~~~~~~~~~
builtin-kvm.c:84:9: note: in expansion of macro ‘DEF_SORT_NAME_KEY’
84 | DEF_SORT_NAME_KEY(time, mean),
| ^~~~~~~~~~~~~~~~~
builtin-kvm.c:80:18: note: (near initialization for ‘keys[1].key’)
80 | { #name, cmp_event_ ## compare_key }
| ^~~~~~~~~~
builtin-kvm.c:84:9: note: in expansion of macro ‘DEF_SORT_NAME_KEY’
84 | DEF_SORT_NAME_KEY(time, mean),
| ^~~~~~~~~~~~~~~~~
cc1: all warnings being treated as errors
make[3]: *** [/var/home/acme/git/perf-tools-next/tools/build/Makefile.build:97: /tmp/build/perf-tools-next/builtin-kvm.o] Error 1
make[3]: *** Waiting for unfinished jobs....
make[2]: *** [Makefile.perf:673: /tmp/build/perf-tools-next/perf-in.o] Error 2
make[2]: *** Waiting for unfinished jobs....
CC /tmp/build/perf-tools-next/pmu-events/pmu-events.o
LD /tmp/build/perf-tools-next/pmu-events/pmu-events-in.o
make[1]: *** [Makefile.perf:236: sub-make] Error 2
make: *** [Makefile:113: install-bin] Error 2
make: Leaving directory '/var/home/acme/git/perf-tools-next/tools/perf'

Performance counter stats for 'make -k BUILD_BPF_SKEL=1 CORESIGHT=1 O=/tmp/build/perf-tools-next -C tools/perf install-bin':

46519453561 cycles:u
73484239007 instructions:u # 1.58 insn per cycle

10.053592077 seconds time elapsed

10.925077000 seconds user
8.045887000 seconds sys


⬢[acme@toolbox perf-tools-next]$ git log --oneline -4
40843f667620dcf3 (HEAD) perf kvm: Use subtraction for comparison metrics
c6a5a089864171f2 perf kvm: Move up metrics helpers
1d75bc17ffc933aa perf kvm: Add pointer to 'perf_kvm_stat' in kvm event
4568393639938020 perf kvm: Refactor overall statistics
⬢[acme@toolbox perf-tools-next]$

⬢[acme@toolbox perf-tools-next]$ gcc --version
gcc (GCC) 12.2.1 20221121 (Red Hat 12.2.1-4)

> Signed-off-by: Leo Yan <leo.yan@xxxxxxxxxx>
> ---
> tools/perf/builtin-kvm.c | 6 +++---
> 1 file changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/tools/perf/builtin-kvm.c b/tools/perf/builtin-kvm.c
> index 62c097a37da9..4e9519390da6 100644
> --- a/tools/perf/builtin-kvm.c
> +++ b/tools/perf/builtin-kvm.c
> @@ -63,10 +63,10 @@ static u64 get_event_ ##func(struct kvm_event *event, int vcpu) \
>
> #define COMPARE_EVENT_KEY(func, field) \
> GET_EVENT_KEY(func, field) \
> -static int cmp_event_ ## func(struct kvm_event *one, \
> +static int64_t cmp_event_ ## func(struct kvm_event *one, \
> struct kvm_event *two, int vcpu) \
> { \
> - return get_event_ ##func(one, vcpu) > \
> + return get_event_ ##func(one, vcpu) - \
> get_event_ ##func(two, vcpu); \
> }
>
> @@ -523,7 +523,7 @@ static void insert_to_result(struct rb_root *result, struct kvm_event *event,
> p = container_of(*rb, struct kvm_event, rb);
> parent = *rb;
>
> - if (bigger(event, p, vcpu))
> + if (bigger(event, p, vcpu) > 0)
> rb = &(*rb)->rb_left;
> else
> rb = &(*rb)->rb_right;
> --
> 2.34.1
>

--

- Arnaldo