All of lore.kernel.org
 help / color / mirror / Atom feed
From: Josef Bacik <jbacik@fb.com>
To: Steven Rostedt <rostedt@goodmis.org>
Cc: <linux-kernel@vger.kernel.org>, <kernel-team@fb.com>
Subject: Re: [PATCH 3/5] trace-cmd: lookup syscall names in profile
Date: Wed, 20 May 2015 15:06:58 -0400	[thread overview]
Message-ID: <555CDB52.9000901@fb.com> (raw)
In-Reply-To: <20150520150133.76c2a08c@gandalf.local.home>

On 05/20/2015 03:01 PM, Steven Rostedt wrote:
> On Wed, 20 May 2015 14:56:17 -0400
> Steven Rostedt <rostedt@goodmis.org> wrote:
>
>> On Wed, 20 May 2015 14:22:31 -0400
>> Josef Bacik <jbacik@fb.com> wrote:
>>
>>>> I'll keep this as is, but I'm going to add code to check for this
>>>> library, and go back to the default raw numbers if it doesn't exist.
>>>>
>>>
>>> Maybe we should think about autoconf at some point?  Thanks,
>>
>> Egad no! I hate autoconf. And so do many others!
>>
>> A simple test works fine. I don't want to require libaudit. Just warn
>> users that compile without it.
>>
>> I like to install trace-cmd on minimal setups (like embedded devices).
>> autoconf will kill that.
>>
>
> Here's what I did to fix it:
>
> -- Steve
>
> diff --git a/Makefile b/Makefile
> index 402f7114ab42..a8e63ad0c223 100644
> --- a/Makefile
> +++ b/Makefile
> @@ -160,7 +160,7 @@ bindir_relative_SQ = $(subst ','\'',$(bindir_relative))
>   plugin_dir_SQ = $(subst ','\'',$(plugin_dir))
>   python_dir_SQ = $(subst ','\'',$(python_dir))
>
> -LIBS = -L. -ltracecmd -ldl -laudit
> +LIBS = -L. -ltracecmd -ldl
>   LIB_FILE = libtracecmd.a
>
>   PACKAGES= gtk+-2.0 libxml-2.0 gthread-2.0
> @@ -231,12 +231,25 @@ override CFLAGS += -D_GNU_SOURCE
>   ifndef NO_PTRACE
>   ifneq ($(call try-cc,$(SOURCE_PTRACE),),y)
>   	NO_PTRACE = 1
> -	CFLAGS += -DWARN_NO_PTRACE
> +	override CFLAGS += -DWARN_NO_PTRACE
>   endif
>   endif
>
>   ifdef NO_PTRACE
> -CFLAGS += -DNO_PTRACE
> +override CFLAGS += -DNO_PTRACE
> +endif
> +
> +ifndef NO_AUDIT
> +ifneq ($(call try-cc,$(SOURCE_AUDIT),-laudit),y)
> +	NO_AUDIT = 1
> +	override CFLAGS += -DWARN_NO_AUDIT
> +endif
> +endif
> +
> +ifdef NO_AUDIT
> +override CFLAGS += -DNO_AUDIT
> +else
> +LIBS += -laudit
>   endif
>
>   # Append required CFLAGS
> diff --git a/features.mk b/features.mk
> index cb2e8bd897be..9c9d63c13a23 100644
> --- a/features.mk
> +++ b/features.mk
> @@ -34,3 +34,21 @@ int main (void)
>   	return ret;
>   }
>   endef
> +
> +define SOURCE_AUDIT
> +#include <stdio.h>
> +#include <libaudit.h>
> +
> +int main (void)
> +{
> +	char *name;
> +	int ret;
> +	ret = audit_detect_machine();
> +	if (ret < 0)
> +		return ret;
> +	name = audit_syscall_to_name(1, ret);
> +	if (!name)
> +		return -1;
> +	return ret;
> +}
> +endef
> diff --git a/trace-profile.c b/trace-profile.c
> index 24f989715f05..640dca822d46 100644
> --- a/trace-profile.c
> +++ b/trace-profile.c
> @@ -23,11 +23,17 @@
>   #include <stdio.h>
>   #include <stdlib.h>
>   #include <string.h>
> +#ifndef NO_AUDIT
>   #include <libaudit.h>
> -
> +#endif
>   #include "trace-local.h"
>   #include "trace-hash.h"
>
> +#ifdef WARN_NO_AUDIT
> +# warning "lib audit not found, using raw syscalls "	\
> +	"(install libaudit-devel and try again)"
> +#endif
> +
>   #define TASK_STATE_TO_CHAR_STR "RSDTtXZxKWP"
>   #define TASK_STATE_MAX		1024
>
> @@ -848,6 +854,7 @@ static void func_print(struct trace_seq *s, struct event_hash *event_hash)
>
>   static void syscall_print(struct trace_seq *s, struct event_hash *event_hash)
>   {
> +#ifndef NO_AUDIT
>   	const char *name = NULL;
>   	int machine;
>
> @@ -860,6 +867,7 @@ static void syscall_print(struct trace_seq *s, struct event_hash *event_hash)
>   	trace_seq_printf(s, "syscall:%s", name);
>   	return;
>   fail:
> +#endif
>   	trace_seq_printf(s, "%s:%d", event_hash->event_data->event->name,
>   			 (int)event_hash->val);
>   }
>

Looks good to me, thanks,

Josef

  reply	other threads:[~2015-05-20 19:07 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-05-08 19:28 [PATCH 0/5] trace-cmd profile enhancements Josef Bacik
2015-05-08 19:29 ` [PATCH 1/5] trace-cmd: add option to group like comms for profile Josef Bacik
2015-05-20 17:31   ` Steven Rostedt
2015-05-20 17:51     ` Josef Bacik
2015-05-08 19:29 ` [PATCH 2/5] trace-cmd: group schedule stacktraces together with --merge Josef Bacik
2015-05-20 17:36   ` Steven Rostedt
2015-05-20 17:54     ` Josef Bacik
2015-05-08 19:29 ` [PATCH 3/5] trace-cmd: lookup syscall names in profile Josef Bacik
2015-05-20 18:00   ` Steven Rostedt
2015-05-20 18:22     ` Josef Bacik
2015-05-20 18:56       ` Steven Rostedt
2015-05-20 19:01         ` Steven Rostedt
2015-05-20 19:06           ` Josef Bacik [this message]
2015-05-08 19:29 ` [PATCH 4/5] trace-cmd: sort the events in the profile output Josef Bacik
2015-05-08 19:29 ` [PATCH 5/5] trace-cmd: make trace-hist work with instances Josef Bacik
2015-05-08 20:13 ` [PATCH 0/5] trace-cmd profile enhancements Steven Rostedt

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=555CDB52.9000901@fb.com \
    --to=jbacik@fb.com \
    --cc=kernel-team@fb.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=rostedt@goodmis.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.