linux-kernel - [RFC patch 00/32] TRACE_EVENT: cleanup for code-size reduction

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-Id: <20110502211123.163877033@efficios.com>
Date:	Mon, 02 May 2011 17:11:23 -0400
From:	Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
To:	LKML <linux-kernel@...r.kernel.org>,
	Steven Rostedt <rostedt@...dmis.org>
Cc:	Ingo Molnar <mingo@...e.hu>, Thomas Gleixner <tglx@...utronix.de>,
	Frederic Weisbecker <fweisbec@...il.com>
Subject: [RFC patch 00/32] TRACE_EVENT: cleanup for code-size reduction

Hi,

I implemented the TRACE_EVENT macros as arrays of fields and arrays of
events in LTTng, and realised that it could benefit Ftrace too by
cutting down its code size. As I know that code size footprint is an
issue Steven has been wanting to tackle for a while, I am sending this
"semicolon removal" patchset, which is a prerequisite to be able to
generate arrays of events.

The main difference between the approaches Ftrace and LTTng have taken
is that Ftrace creates a "ftrace_define_fields_##call" function for each
event which calls trace_define_field for each field of each event. In
LTTng, I did the following (code snippet from [1]):

struct lttng_event_field {
        const char *name;
        const struct lttng_type type;
};

struct lttng_event_desc {
        const struct lttng_event_field *fields;
        const char *name;
        unsigned int nr_fields;
};

#undef DEFINE_EVENT
#define DEFINE_EVENT(_template, _name, _proto, _args)                          \
                {                                                              \
                        .fields = __event_fields___##_template,                \
                        .name = #_name,                                        \
                        .nr_fields = ARRAY_SIZE(__event_fields___##_template), \
                },

#define TP_ID1(_token, _system) _token##_system
#define TP_ID(_token, _system)  TP_ID1(_token, _system)

static const struct lttng_event_desc TP_ID(__event_desc___, TRACE_SYSTEM)[] = {
#include TRACE_INCLUDE(TRACE_INCLUDE_FILE)
};

And for creation of __event_fields___##_template, which allows to keep
all fields in another array (which should be defined in a phase prior to
the event array definition), I did:

#undef __field
#define __field(_type, _item)                                   \
        { .name = #_item, .type = { .atype = atype_integer, .name = #_type} },

(similar for __field_ext, __array, __dynamic_array, __string)

#undef TP_STRUCT__entry
#define TP_STRUCT__entry(args...) args  /* Only one used in this phase */

#undef DECLARE_EVENT_CLASS
#define DECLARE_EVENT_CLASS(_name, _proto, _args, _tstruct, _assign, _print) \
        static const struct lttng_event_field __event_fields___##_name[] = { \
                _tstruct \
        };

For the declaration of struct lttng_type, which contains the field type
information in a union, see [2].

Best regards,

Mathieu

[1] http://git.lttng.org/?p=lttng-modules.git;a=blob;f=probes/lttng-events.h;h=a05c29d56771b89807ffef035b7cd6df34a0b97f;hb=HEAD
[2] http://git.lttng.org/?p=lttng-modules.git;a=blob;f=probes/lttng-types.h;h=0192bffaf2f321f446f71a13901e2ab09fd69663;hb=HEAD
-- 
Mathieu Desnoyers
Operating System Efficiency R&D Consultant
EfficiOS Inc.
http://www.efficios.com
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/