From: Eric Sandeen <sandeen@sandeen.net>
To: xfs-oss <xfs@oss.sgi.com>
Subject: [PATCH] optimize dmapi event tests w/o dmapi config
Date: Sun, 19 Aug 2007 01:05:10 -0500 [thread overview]
Message-ID: <46C7DD96.6030001@sandeen.net> (raw)
Defining XFS_DM_EVENT* macros to 0 in the absence of
CONFIG_XFS_DMAPI allows gcc to optimize away tests that
should never be true. Also wrap one hunk of xfs_unmount
in #ifdef CONFIG_XFS_DMAPI
Stack deltas on x86, gcc 4.1:
xfs_create -16
xfs_free_file_space -4
xfs_getbmap +4
xfs_link -8
xfs_mkidr -20
xfs_read -24
xfs_remove -12
xfs_rename -8
xfs_rmdir -16
xfs_setattr -20
xfs_splice_read -20
xfs_splice_write -20
xfs_unmount -48
xfs_write -20
Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
Index: linux-2.6.22.i386/fs/xfs/xfs_dmapi.h
===================================================================
--- linux-2.6.22.i386.orig/fs/xfs/xfs_dmapi.h
+++ linux-2.6.22.i386/fs/xfs/xfs_dmapi.h
@@ -66,6 +66,7 @@ typedef enum {
} dm_right_t;
#define HAVE_DM_RIGHT_T
+#ifdef CONFIG_XFS_DMAPI
/* Defines for determining if an event message should be sent. */
#define DM_EVENT_ENABLED(vfsp, ip, event) ( \
unlikely ((vfsp)->vfs_flag & VFS_DMI) && \
@@ -78,6 +79,11 @@ typedef enum {
( ((io)->io_dmevmask & (1 << event)) || \
((io)->io_mount->m_dmevmask & (1 << event)) ) \
)
+#else
+#define DM_EVENT_ENABLED(vfsp, ip, event) (0)
+#define DM_EVENT_ENABLED_IO(vfsp, io, event) (0)
+#endif
+
#define DM_XFS_VALID_FS_EVENTS ( \
(1 << DM_EVENT_PREUNMOUNT) | \
Index: linux-2.6.22.i386/fs/xfs/xfs_vfsops.c
===================================================================
--- linux-2.6.22.i386.orig/fs/xfs/xfs_vfsops.c
+++ linux-2.6.22.i386/fs/xfs/xfs_vfsops.c
@@ -568,6 +568,7 @@ xfs_unmount(
rip = mp->m_rootip;
rvp = XFS_ITOV(rip);
+#ifdef CONFIG_XFS_DMAPI
if (vfsp->vfs_flag & VFS_DMI) {
error = XFS_SEND_PREUNMOUNT(mp, vfsp,
rvp, DM_RIGHT_NULL, rvp, DM_RIGHT_NULL,
@@ -580,7 +581,7 @@ xfs_unmount(
unmount_event_flags = (mp->m_dmevmask & (1<<DM_EVENT_UNMOUNT))?
0 : DM_FLAGS_UNWANTED;
}
-
+#endif
/*
* First blow any referenced inode from this file system
* out of the reference cache, and delete the timer.
next reply other threads:[~2007-08-19 6:05 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-08-19 6:05 Eric Sandeen [this message]
2007-08-19 19:07 ` [PATCH] optimize dmapi event tests w/o dmapi config Christoph Hellwig
2007-08-19 23:48 ` Vlad Apostolov
2007-08-20 3:37 ` Eric Sandeen
2007-08-20 5:22 ` Vlad Apostolov
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=46C7DD96.6030001@sandeen.net \
--to=sandeen@sandeen.net \
--cc=xfs@oss.sgi.com \
/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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox