From: "Martin Liška" <mliska@suse.cz>
To: Ingo Molnar <mingo@kernel.org>,
Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: linux-kernel@vger.kernel.org, Ingo Molnar <mingo@redhat.com>,
Paul Mackerras <paulus@samba.org>,
Peter Zijlstra <a.p.zijlstra@chello.nl>
Subject: Re: [PATCH] perf: fix wrong DEBUG configuration
Date: Mon, 25 May 2015 10:04:24 +0200 [thread overview]
Message-ID: <5562D788.9000201@suse.cz> (raw)
In-Reply-To: <20150522070250.GA30715@gmail.com>
[-- Attachment #1: Type: text/plain, Size: 1998 bytes --]
On 05/22/2015 09:02 AM, Ingo Molnar wrote:
>
> * Arnaldo Carvalho de Melo <acme@kernel.org> wrote:
>
>> Em Wed, May 20, 2015 at 06:16:51PM +0200, Martin Liška escreveu:
>>> On 05/20/2015 04:55 PM, Ingo Molnar wrote:
>>>> * Arnaldo Carvalho de Melo <acme@kernel.org> wrote:
>>>>> So the rule has been: What are the kernel requirements for the
>>>>> toolchain? tools/perf/ should build with that.
>>>>
>>>> So we could use -Og if it works, like Kbuild does it:
>>>>
>>>> KBUILD_CFLAGS += $(call cc-option,-mno-fp-ret-in-387)
>>>>
>>>> the 'cc-option' Make function does some magic of silently calling GCC
>>>> with that option and observing the result.
>>
>>>> See:
>>
>>>> scripts/Kbuild.include:cc-option = $(call try-run,\
>>
>>> I am sorry, I did mistake in understanding of DEBUG variable.
>>> Following patch should be fixed, except missing auto-detection
>>> of -Og option.
>>>
>>> Unfortunately, following hunk does not work, no -Ox is added to CFLAGS?
>>>
>>> -- CFLAGS += -Og
>>> ++ CFLAGS += $(call cc-option,-Og,-O0)
>>
>> I don't know if we have this cc-option, perhaps Ingo is suggesting
>> we get it in tools/build/? Or include scripts/Kbuild.include and
>> then use it?
>>
>> I.e. we have checks to see if we can use, for instance
>> -fstack-protector-all, see tools/build/feature/Makefile, using this
>> cc-option thing, importing it from Kbuild would solve the issue at
>> hand in a definitive way and in line with what we have been
>> pursuing: to make the tools/ build system be based on Kbuild.
>
> So I'd suggest copying any necessary functions instead of outright
> including all of Kbuild in the tooling build system which creates
> non-trivial dependencies that is not necessarily tested as thoroughly
> on the kbuild side as on the tooling side.
>
> Thanks,
>
> Ingo
>
Hi.
Thanks for advice, I copied necessary functions to include/utilities.mak.
Following patch works for, tested both with a compiler capable of -Og and
a not capable one.
Thanks,
Martin
[-- Attachment #2: 0001-perf-fix-wrong-DEBUG-configuration-v4.patch --]
[-- Type: text/x-patch, Size: 3160 bytes --]
Currently, GCC optimizes -O6 same as -O3 level, thus change the value
to -O6.
Right optimize debugging experience is given by passing -Og to compiler.
Assign default value for pointers that are identified by compiler as
non-initialized.
Signed-off-by: Martin Liska <mliska@suse.cz>
---
tools/perf/arch/common.c | 2 +-
tools/perf/config/Makefile | 4 +++-
tools/perf/config/utilities.mak | 19 +++++++++++++++++++
tools/perf/util/symbol.c | 2 +-
tools/perf/util/trace-event-parse.c | 2 +-
5 files changed, 25 insertions(+), 4 deletions(-)
diff --git a/tools/perf/arch/common.c b/tools/perf/arch/common.c
index 49776f1..b7bb42c 100644
--- a/tools/perf/arch/common.c
+++ b/tools/perf/arch/common.c
@@ -61,7 +61,7 @@ const char *const mips_triplets[] = {
static bool lookup_path(char *name)
{
bool found = false;
- char *path, *tmp;
+ char *path, *tmp = NULL;
char buf[PATH_MAX];
char *env = getenv("PATH");
diff --git a/tools/perf/config/Makefile b/tools/perf/config/Makefile
index e3b3724..47e4ae1 100644
--- a/tools/perf/config/Makefile
+++ b/tools/perf/config/Makefile
@@ -129,7 +129,9 @@ ifndef DEBUG
endif
ifeq ($(DEBUG),0)
- CFLAGS += -O6
+ CFLAGS += -O3
+else
+ CFLAGS += $(call cc-option,-Og,-O0)
endif
ifdef PARSER_DEBUG
diff --git a/tools/perf/config/utilities.mak b/tools/perf/config/utilities.mak
index c16ce83..0ebef09 100644
--- a/tools/perf/config/utilities.mak
+++ b/tools/perf/config/utilities.mak
@@ -177,3 +177,22 @@ $(if $($(1)),$(call _ge_attempt,$($(1)),$(1)),$(call _ge_attempt,$(2)))
endef
_ge_attempt = $(if $(get-executable),$(get-executable),$(call _gea_err,$(2)))
_gea_err = $(if $(1),$(error Please set '$(1)' appropriately))
+
+# try-run
+# Usage: option = $(call try-run, $(CC)...-o "$$TMP",option-ok,otherwise)
+# Exit code chooses option. "$$TMP" is can be used as temporary file and
+# is automatically cleaned up.
+try-run = $(shell set -e; \
+ TMP="$(TMPOUT).$$$$.tmp"; \
+ TMPO="$(TMPOUT).$$$$.o"; \
+ if ($(1)) >/dev/null 2>&1; \
+ then echo "$(2)"; \
+ else echo "$(3)"; \
+ fi; \
+ rm -f "$$TMP" "$$TMPO")
+
+# cc-option
+# Usage: cflags-y += $(call cc-option,-march=winchip-c6,-march=i586)
+
+cc-option = $(call try-run,\
+ $(CC) $(KBUILD_CPPFLAGS) $(KBUILD_CFLAGS) $(1) -c -x c /dev/null -o "$$TMP",$(1),$(2))
diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
index 82a31fd..a19fbd4 100644
--- a/tools/perf/util/symbol.c
+++ b/tools/perf/util/symbol.c
@@ -400,7 +400,7 @@ static struct symbol *symbols__find_by_name(struct rb_root *symbols,
const char *name)
{
struct rb_node *n;
- struct symbol_name_rb_node *s;
+ struct symbol_name_rb_node *s = NULL;
if (symbols == NULL)
return NULL;
diff --git a/tools/perf/util/trace-event-parse.c b/tools/perf/util/trace-event-parse.c
index 25d6c73..d495741 100644
--- a/tools/perf/util/trace-event-parse.c
+++ b/tools/perf/util/trace-event-parse.c
@@ -173,7 +173,7 @@ void parse_ftrace_printk(struct pevent *pevent,
char *line;
char *next = NULL;
char *addr_str;
- char *fmt;
+ char *fmt = NULL;
line = strtok_r(file, "\n", &next);
while (line) {
--
2.1.4
next prev parent reply other threads:[~2015-05-25 8:04 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-05-19 13:41 [PATCH] perf: fix wrong DEBUG configuration Martin Liška
2015-05-19 14:04 ` Arnaldo Carvalho de Melo
2015-05-20 13:07 ` Martin Liška
2015-05-20 13:17 ` Arnaldo Carvalho de Melo
2015-05-20 13:29 ` Martin Liška
2015-05-20 13:53 ` Arnaldo Carvalho de Melo
2015-05-20 14:55 ` Ingo Molnar
2015-05-20 16:16 ` Martin Liška
2015-05-21 15:05 ` Arnaldo Carvalho de Melo
2015-05-22 7:02 ` Ingo Molnar
2015-05-22 13:50 ` Martin Liška
2015-05-25 8:04 ` Martin Liška [this message]
2015-05-25 10:47 ` Ingo Molnar
2015-05-25 11:04 ` Martin Liška
2015-05-25 11:09 ` Ingo Molnar
2015-05-25 11:12 ` Martin Liška
2015-05-25 13:52 ` Arnaldo Carvalho de Melo
2015-05-25 18:32 ` Ingo Molnar
2015-05-26 9:13 ` Martin Liška
2015-05-26 14:36 ` Arnaldo Carvalho de Melo
2015-05-26 15:22 ` Arnaldo Carvalho de Melo
2015-05-26 15:39 ` Martin Liška
2015-05-26 15:48 ` Martin Liška
2015-05-26 15:53 ` Arnaldo Carvalho de Melo
2015-05-27 16:51 ` [tip:perf/core] perf tools: Assign default value for some pointers tip-bot for Martin Liška
2015-05-27 16:51 ` [tip:perf/core] perf tools: Improve setting of gcc debug option tip-bot for Martin Liska
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=5562D788.9000201@suse.cz \
--to=mliska@suse.cz \
--cc=a.p.zijlstra@chello.nl \
--cc=acme@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@kernel.org \
--cc=mingo@redhat.com \
--cc=paulus@samba.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.