Message ID | 6c7d1e3005dbec5483bdb9b7b60071175bf7bf70.1688811201.git.christophe.jaillet@wanadoo.fr |
---|---|
State | New |
Headers |
Return-Path: <linux-kernel-owner@vger.kernel.org> Delivered-To: ouuuleilei@gmail.com Received: by 2002:a59:9f45:0:b0:3ea:f831:8777 with SMTP id v5csp3841662vqx; Sat, 8 Jul 2023 03:49:17 -0700 (PDT) X-Google-Smtp-Source: APBJJlFeRX7hK7A3FB4dZQFv8e2cMEN7Db4Vh/legMaJtiFx2qjeqPCdVpJjbN0sFfIsw7tclDeU X-Received: by 2002:a17:906:778c:b0:993:d5bd:a765 with SMTP id s12-20020a170906778c00b00993d5bda765mr3050588ejm.20.1688813357058; Sat, 08 Jul 2023 03:49:17 -0700 (PDT) ARC-Seal: i=1; a=rsa-sha256; t=1688813357; cv=none; d=google.com; s=arc-20160816; b=INYwFnLFBRyq6Z8SfGfJbvFzoWdtuqMI+BGmVdVbfxRLEnFIQWEYFu/9ouWEE8aruY JVB61oosmtL9wtDG5AsgbUGje90cSMkyoGsD3McHQRPsoGA+lHGwfXU6vtnH6/VcLSJW egk4AiLTZx7sn2opm4B0aGDb+/tKVcIfbx3+BxlNjbssVTeJ+aSW7Qp8CLVXLMHZRSt9 8v40ni52+6GzaRGJAXLeaCAGRm9sjkp1stm9s8gyPktY3lWQgEwjJvGUHLrXMgkluTrf z31EFMK6McoBrR0vN3Trtcz1yC/3neXfWl6waY4VO/RHyh6J7KJR9Zm5XJd+IrcebphG mkPg== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=list-id:precedence:content-transfer-encoding:mime-version :message-id:date:subject:cc:to:from:dkim-signature; bh=9O9yEyZRJBIiLE4+IqXvqAN9ixKeBIx6+ibfzyadF5o=; fh=0WOQaqOU7etC0AFM88lvsAWAYktCJoK6wy1iyKHV33E=; b=HYKki7f9Fo1r5QKQ/ZaFt79L2mAJeVumS9px4u0uV7B9ss08pfD4LdomimCcQ2xCd7 7I61V5u+0qku4fJoSZYYOe1wmSwt4R9o0+iETmilL1fwVs9E7atUJbKJJ1NOjWUSYQ2+ iVqtITiuZ+HA83Ar4GKIIIFq4p6/bKPukUKiEYYOzAdDnPlpv9dvUQySVcTwVfTgcoWl DQ3ggwhjExsrf3LcWhj5yaKezwzgBcJait1azMxB5+QwCdOs27qXe4CsQqnE7jojrExR uLSuULdtWWNsB2LBpizl4bC2h2BX8GAFdY89FFnY6PRmcmxVPHo186O9hElnFzBzd4vb TP9A== ARC-Authentication-Results: i=1; mx.google.com; dkim=pass header.i=@wanadoo.fr header.s=t20230301 header.b=cU+NAgl8; spf=pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) smtp.mailfrom=linux-kernel-owner@vger.kernel.org Received: from out1.vger.email (out1.vger.email. [2620:137:e000::1:20]) by mx.google.com with ESMTP id t9-20020a1709064f0900b00993664a9970si3496654eju.873.2023.07.08.03.48.53; Sat, 08 Jul 2023 03:49:17 -0700 (PDT) Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20; Authentication-Results: mx.google.com; dkim=pass header.i=@wanadoo.fr header.s=t20230301 header.b=cU+NAgl8; spf=pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) smtp.mailfrom=linux-kernel-owner@vger.kernel.org Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S229682AbjGHKOD (ORCPT <rfc822;tebrre53rla2o@gmail.com> + 99 others); Sat, 8 Jul 2023 06:14:03 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:52002 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S229627AbjGHKOB (ORCPT <rfc822;linux-kernel@vger.kernel.org>); Sat, 8 Jul 2023 06:14:01 -0400 Received: from smtp.smtpout.orange.fr (smtp-14.smtpout.orange.fr [80.12.242.14]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 96B7A1FD9 for <linux-kernel@vger.kernel.org>; Sat, 8 Jul 2023 03:13:55 -0700 (PDT) Received: from pop-os.home ([86.243.2.178]) by smtp.orange.fr with ESMTPA id I4wlqeROQBC88I4wlqne8P; Sat, 08 Jul 2023 12:13:48 +0200 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=wanadoo.fr; s=t20230301; t=1688811228; bh=9O9yEyZRJBIiLE4+IqXvqAN9ixKeBIx6+ibfzyadF5o=; h=From:To:Cc:Subject:Date; b=cU+NAgl8Ayr2k12EeZ8TTODNngu8q5zX2SPR1YT3RykLA+r9UeqMtCf6A4HyDuXJp IpZORkJ3jcErhBjaVeAKGB38tD2U6rCOTtM0AJWGVJ15xvcNgdVeQrkaKX5o4/Cznh FrJjWb5pQpoGTGqTaroNV2YS33AaGblLSnC0ZKQUYdEY7vYEc+xBnket4EgS7S8beP MX8AFtX2ziZFeaENzmNZeXaVd6ntNeeuZ7c8pfeiRmb+NgcuUYtaAIDXoRfgu2UQeT 4uI+qwa995A5cdWn8eQip6eVrJrfecV4CstgyCZS4fiXObURj6VPinxgfO/vMidKH4 wRdS9LJ4v5SAg== X-ME-Helo: pop-os.home X-ME-Auth: Y2hyaXN0b3BoZS5qYWlsbGV0QHdhbmFkb28uZnI= X-ME-Date: Sat, 08 Jul 2023 12:13:48 +0200 X-ME-IP: 86.243.2.178 From: Christophe JAILLET <christophe.jaillet@wanadoo.fr> To: Greg Kroah-Hartman <gregkh@linuxfoundation.org>, "Rafael J. Wysocki" <rafael@kernel.org> Cc: linux-kernel@vger.kernel.org, kernel-janitors@vger.kernel.org, Christophe JAILLET <christophe.jaillet@wanadoo.fr> Subject: [PATCH] kobject: Reorder fields in 'struct kobject' Date: Sat, 8 Jul 2023 12:13:45 +0200 Message-Id: <6c7d1e3005dbec5483bdb9b7b60071175bf7bf70.1688811201.git.christophe.jaillet@wanadoo.fr> X-Mailer: git-send-email 2.34.1 MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Spam-Status: No, score=-2.1 required=5.0 tests=BAYES_00,DKIM_SIGNED, DKIM_VALID,DKIM_VALID_AU,DKIM_VALID_EF,RCVD_IN_DNSWL_NONE, RCVD_IN_MSPIKE_H3,RCVD_IN_MSPIKE_WL,SPF_HELO_PASS,SPF_PASS, T_SCC_BODY_TEXT_LINE autolearn=ham autolearn_force=no version=3.4.6 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on lindbergh.monkeyblade.net Precedence: bulk List-ID: <linux-kernel.vger.kernel.org> X-Mailing-List: linux-kernel@vger.kernel.org X-getmail-retrieved-from-mailbox: =?utf-8?q?INBOX?= X-GMAIL-THRID: =?utf-8?q?1770849154321679013?= X-GMAIL-MSGID: =?utf-8?q?1770849154321679013?= |
Series |
kobject: Reorder fields in 'struct kobject'
|
|
Commit Message
Christophe JAILLET
July 8, 2023, 10:13 a.m. UTC
Group some variables based on their sizes to reduce hole and avoid padding.
On x86_64, this shrinks the size of 'struct kobject' from 256 to 244 bytes.
This structure is often included in some other structures. So these other
structures will also benefit from this 8 bytes saving.
This is especially nice for structure like 'cma_kobject' or 'class_dir'
that are now 256 bytes long. When they are kzalloc()'ed, 256 bytes are
allocated, instead of 512.
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
---
Try to see how often this is included in another struct. Certainly not
perfect, but gives an idea.
git grep -P \\tstruct\ kobj[^*]*$ | wc -l
163
Using pahole
Before:
======
struct kobject {
const char * name; /* 0 8 */
struct list_head entry; /* 8 16 */
struct kobject * parent; /* 24 8 */
struct kset * kset; /* 32 8 */
const struct kobj_type * ktype; /* 40 8 */
struct kernfs_node * sd; /* 48 8 */
struct kref kref; /* 56 4 */
/* XXX 4 bytes hole, try to pack */
/* --- cacheline 1 boundary (64 bytes) --- */
struct delayed_work release; /* 64 184 */
/* XXX last struct has 4 bytes of padding */
/* --- cacheline 3 boundary (192 bytes) was 56 bytes ago --- */
unsigned int state_initialized:1; /* 248: 0 4 */
unsigned int state_in_sysfs:1; /* 248: 1 4 */
unsigned int state_add_uevent_sent:1; /* 248: 2 4 */
unsigned int state_remove_uevent_sent:1; /* 248: 3 4 */
unsigned int uevent_suppress:1; /* 248: 4 4 */
/* size: 256, cachelines: 4, members: 13 */
/* sum members: 244, holes: 1, sum holes: 4 */
/* sum bitfield members: 5 bits (0 bytes) */
/* padding: 4 */
/* paddings: 1, sum paddings: 4 */
/* bit_padding: 27 bits */
};
After:
=====
struct kobject {
const char * name; /* 0 8 */
struct list_head entry; /* 8 16 */
struct kobject * parent; /* 24 8 */
struct kset * kset; /* 32 8 */
const struct kobj_type * ktype; /* 40 8 */
struct kernfs_node * sd; /* 48 8 */
struct kref kref; /* 56 4 */
unsigned int state_initialized:1; /* 60: 0 4 */
unsigned int state_in_sysfs:1; /* 60: 1 4 */
unsigned int state_add_uevent_sent:1; /* 60: 2 4 */
unsigned int state_remove_uevent_sent:1; /* 60: 3 4 */
unsigned int uevent_suppress:1; /* 60: 4 4 */
/* XXX 27 bits hole, try to pack */
/* --- cacheline 1 boundary (64 bytes) --- */
struct delayed_work release; /* 64 184 */
/* XXX last struct has 4 bytes of padding */
/* size: 248, cachelines: 4, members: 13 */
/* sum members: 244 */
/* sum bitfield members: 5 bits, bit holes: 1, sum bit holes: 27 bits */
/* paddings: 1, sum paddings: 4 */
/* last cacheline: 56 bytes */
};
---
include/linux/kobject.h | 8 +++++---
1 file changed, 5 insertions(+), 3 deletions(-)
Comments
On Sat, Jul 08, 2023 at 12:13:45PM +0200, Christophe JAILLET wrote: > Group some variables based on their sizes to reduce hole and avoid padding. > On x86_64, this shrinks the size of 'struct kobject' from 256 to 244 bytes. > > This structure is often included in some other structures. So these other > structures will also benefit from this 8 bytes saving. > > This is especially nice for structure like 'cma_kobject' or 'class_dir' > that are now 256 bytes long. When they are kzalloc()'ed, 256 bytes are > allocated, instead of 512. > > Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr> > --- > Try to see how often this is included in another struct. Certainly not > perfect, but gives an idea. > git grep -P \\tstruct\ kobj[^*]*$ | wc -l > 163 > > > Using pahole > > Before: > ====== > struct kobject { > const char * name; /* 0 8 */ > struct list_head entry; /* 8 16 */ > struct kobject * parent; /* 24 8 */ > struct kset * kset; /* 32 8 */ > const struct kobj_type * ktype; /* 40 8 */ > struct kernfs_node * sd; /* 48 8 */ > struct kref kref; /* 56 4 */ > > /* XXX 4 bytes hole, try to pack */ > > /* --- cacheline 1 boundary (64 bytes) --- */ > struct delayed_work release; /* 64 184 */ > > /* XXX last struct has 4 bytes of padding */ > > /* --- cacheline 3 boundary (192 bytes) was 56 bytes ago --- */ > unsigned int state_initialized:1; /* 248: 0 4 */ > unsigned int state_in_sysfs:1; /* 248: 1 4 */ > unsigned int state_add_uevent_sent:1; /* 248: 2 4 */ > unsigned int state_remove_uevent_sent:1; /* 248: 3 4 */ > unsigned int uevent_suppress:1; /* 248: 4 4 */ > > /* size: 256, cachelines: 4, members: 13 */ > /* sum members: 244, holes: 1, sum holes: 4 */ > /* sum bitfield members: 5 bits (0 bytes) */ > /* padding: 4 */ > /* paddings: 1, sum paddings: 4 */ > /* bit_padding: 27 bits */ > }; > > > After: > ===== > struct kobject { > const char * name; /* 0 8 */ > struct list_head entry; /* 8 16 */ > struct kobject * parent; /* 24 8 */ > struct kset * kset; /* 32 8 */ > const struct kobj_type * ktype; /* 40 8 */ > struct kernfs_node * sd; /* 48 8 */ > struct kref kref; /* 56 4 */ > unsigned int state_initialized:1; /* 60: 0 4 */ > unsigned int state_in_sysfs:1; /* 60: 1 4 */ > unsigned int state_add_uevent_sent:1; /* 60: 2 4 */ > unsigned int state_remove_uevent_sent:1; /* 60: 3 4 */ > unsigned int uevent_suppress:1; /* 60: 4 4 */ > > /* XXX 27 bits hole, try to pack */ > > /* --- cacheline 1 boundary (64 bytes) --- */ > struct delayed_work release; /* 64 184 */ > > /* XXX last struct has 4 bytes of padding */ > > /* size: 248, cachelines: 4, members: 13 */ > /* sum members: 244 */ > /* sum bitfield members: 5 bits, bit holes: 1, sum bit holes: 27 bits */ > /* paddings: 1, sum paddings: 4 */ > /* last cacheline: 56 bytes */ > }; > --- > include/linux/kobject.h | 8 +++++--- > 1 file changed, 5 insertions(+), 3 deletions(-) > > diff --git a/include/linux/kobject.h b/include/linux/kobject.h > index c392c811d9ad..c30affcc43b4 100644 > --- a/include/linux/kobject.h > +++ b/include/linux/kobject.h > @@ -69,14 +69,16 @@ struct kobject { > const struct kobj_type *ktype; > struct kernfs_node *sd; /* sysfs directory entry */ > struct kref kref; > -#ifdef CONFIG_DEBUG_KOBJECT_RELEASE > - struct delayed_work release; > -#endif > + > unsigned int state_initialized:1; > unsigned int state_in_sysfs:1; > unsigned int state_add_uevent_sent:1; > unsigned int state_remove_uevent_sent:1; > unsigned int uevent_suppress:1; > + > +#ifdef CONFIG_DEBUG_KOBJECT_RELEASE > + struct delayed_work release; > +#endif > }; Very nice, thanks for doing this! For some reason I thought I did this already, but it turned out to be in an old branch of mine that never saw the light of day as I went down some other path trying to do something else. So many thanks for this, I'll queue it up after 6.5-rc1 is out. greg k-h
diff --git a/include/linux/kobject.h b/include/linux/kobject.h index c392c811d9ad..c30affcc43b4 100644 --- a/include/linux/kobject.h +++ b/include/linux/kobject.h @@ -69,14 +69,16 @@ struct kobject { const struct kobj_type *ktype; struct kernfs_node *sd; /* sysfs directory entry */ struct kref kref; -#ifdef CONFIG_DEBUG_KOBJECT_RELEASE - struct delayed_work release; -#endif + unsigned int state_initialized:1; unsigned int state_in_sysfs:1; unsigned int state_add_uevent_sent:1; unsigned int state_remove_uevent_sent:1; unsigned int uevent_suppress:1; + +#ifdef CONFIG_DEBUG_KOBJECT_RELEASE + struct delayed_work release; +#endif }; __printf(2, 3) int kobject_set_name(struct kobject *kobj, const char *name, ...);