| Message ID | 1427194266-2885-9-git-send-email-david@fromorbit.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On 03/24/2015 12:51 PM, Dave Chinner wrote: > From: Dave Chinner <dchinner@redhat.com> > > Add initial DAX support to XFS. To do this we need a new mount > option to turn DAX on filesystem, and we need to propagate thi into > the inode flags whenever an inode is instantiated so that the > per-inode checks throughout the code Do The Right Thing. > > There are still some things remaining to be done: > > - needs per-inode flags to mark inodes as DAX enabled, and > an inheritance flag to enable automatic filesystem > propagation of the property > - fails occasionally with zero length writes instead of > ENOSPC errors, so error propagation inside/from the DAX > code need work > - occasionally creates two extents rather than a single > larger extent like non-dax filesystems. > - much more testing > > Signed-off-by: Dave Chinner <dchinner@redhat.com> > --- > fs/xfs/xfs_iops.c | 24 ++++++++++++------------ > fs/xfs/xfs_mount.h | 2 ++ > fs/xfs/xfs_super.c | 25 +++++++++++++++++++++++-- > 3 files changed, 37 insertions(+), 14 deletions(-) > > diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c > index 9ca5352..695d857 100644 > --- a/fs/xfs/xfs_iops.c > +++ b/fs/xfs/xfs_iops.c > @@ -1195,22 +1195,22 @@ xfs_diflags_to_iflags( > struct inode *inode, > struct xfs_inode *ip) > { > - if (ip->i_d.di_flags & XFS_DIFLAG_IMMUTABLE) > + uint16_t flags = ip->i_d.di_flags; > + > + inode->i_flags &= ~(S_IMMUTABLE | S_APPEND | S_SYNC | > + S_NOATIME | S_DAX); > + > + if (flags & XFS_DIFLAG_IMMUTABLE) > inode->i_flags |= S_IMMUTABLE; > - else > - inode->i_flags &= ~S_IMMUTABLE; > - if (ip->i_d.di_flags & XFS_DIFLAG_APPEND) > + if (flags & XFS_DIFLAG_APPEND) > inode->i_flags |= S_APPEND; > - else > - inode->i_flags &= ~S_APPEND; > - if (ip->i_d.di_flags & XFS_DIFLAG_SYNC) > + if (flags & XFS_DIFLAG_SYNC) > inode->i_flags |= S_SYNC; > - else > - inode->i_flags &= ~S_SYNC; > - if (ip->i_d.di_flags & XFS_DIFLAG_NOATIME) > + if (flags & XFS_DIFLAG_NOATIME) > inode->i_flags |= S_NOATIME; > - else > - inode->i_flags &= ~S_NOATIME; > + /* XXX: Also needs an on-disk per inode flag! */ > + if (ip->i_mount->m_flags & XFS_MOUNT_DAX) > + inode->i_flags |= S_DAX; > } > > /* > diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h > index 8c995a2..cd44e88 100644 > --- a/fs/xfs/xfs_mount.h > +++ b/fs/xfs/xfs_mount.h > @@ -179,6 +179,8 @@ typedef struct xfs_mount { > allocator */ > #define XFS_MOUNT_NOATTR2 (1ULL << 25) /* disable use of attr2 format */ > > +#define XFS_MOUNT_DAX (1ULL << 62) /* TEST ONLY! */ > + > > /* > * Default minimum read and write sizes. > diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c > index 3ad0b17..0f26d7a 100644 > --- a/fs/xfs/xfs_super.c > +++ b/fs/xfs/xfs_super.c > @@ -112,6 +112,8 @@ static struct xfs_kobj xfs_dbg_kobj; /* global debug sysfs attrs */ > #define MNTOPT_DISCARD "discard" /* Discard unused blocks */ > #define MNTOPT_NODISCARD "nodiscard" /* Do not discard unused blocks */ > > +#define MNTOPT_DAX "dax" /* Enable direct access to bdev pages */ > + > /* > * Table driven mount option parser. > * > @@ -363,6 +365,10 @@ xfs_parseargs( > mp->m_flags |= XFS_MOUNT_DISCARD; > } else if (!strcmp(this_char, MNTOPT_NODISCARD)) { > mp->m_flags &= ~XFS_MOUNT_DISCARD; > +#ifdef CONFIG_FS_DAX > + } else if (!strcmp(this_char, MNTOPT_DAX)) { > + mp->m_flags |= XFS_MOUNT_DAX; Hi So what I see, (I might be wrong), is that once this flag is set here the fs (At above xfs_diflags_to_iflags() ) will start serving DAX inodes. This is a problem because the bdev passed in might not support direct_access at all. I think we might want a dax_supported(sb) and call somewhere at mount time. > +#endif > } else { > xfs_warn(mp, "unknown mount option [%s].", this_char); > return -EINVAL; > @@ -452,8 +458,8 @@ done: > } > > struct proc_xfs_info { > - int flag; > - char *str; > + uint64_t flag; > + char *str; > }; > > STATIC int > @@ -474,6 +480,7 @@ xfs_showargs( > { XFS_MOUNT_GRPID, "," MNTOPT_GRPID }, > { XFS_MOUNT_DISCARD, "," MNTOPT_DISCARD }, > { XFS_MOUNT_SMALL_INUMS, "," MNTOPT_32BITINODE }, > + { XFS_MOUNT_DAX, "," MNTOPT_DAX }, > { 0, NULL } > }; > static struct proc_xfs_info xfs_info_unset[] = { > @@ -1501,6 +1508,20 @@ xfs_fs_fill_super( > if (XFS_SB_VERSION_NUM(&mp->m_sb) == XFS_SB_VERSION_5) > sb->s_flags |= MS_I_VERSION; > > + if (mp->m_flags & XFS_MOUNT_DAX) { > + xfs_warn(mp, > + "DAX enabled. Warning: EXPERIMENTAL, use at your own risk"); > + if (sb->s_blocksize != PAGE_SIZE) { > + xfs_alert(mp, > + "Filesystem block size invalid for DAX Turning DAX off."); > + mp->m_flags &= ~XFS_MOUNT_DAX; > + } else if (!sb->s_bdev->bd_disk->fops->direct_access) { > + xfs_alert(mp, > + "Block device does not support DAX Turning DAX off."); > + mp->m_flags &= ~XFS_MOUNT_DAX; > + } > + } > If we agree about the s_flags MS_MOUNT_DAX then we can define a if (MNTOPT_DAX) dax_enable_if_supported(sb); This will try a call to bdev_direct_access(sb->s_bdev, ...) and set the flag if everything is OK, else will leave it off. (I can do this patch if you want) > error = xfs_mountfs(mp); > if (error) > goto out_filestream_unmount; > Thanks Boaz -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Tue, Mar 24, 2015 at 02:52:48PM +0200, Boaz Harrosh wrote: > On 03/24/2015 12:51 PM, Dave Chinner wrote: > > From: Dave Chinner <dchinner@redhat.com> > > @@ -363,6 +365,10 @@ xfs_parseargs( > > mp->m_flags |= XFS_MOUNT_DISCARD; > > } else if (!strcmp(this_char, MNTOPT_NODISCARD)) { > > mp->m_flags &= ~XFS_MOUNT_DISCARD; > > +#ifdef CONFIG_FS_DAX > > + } else if (!strcmp(this_char, MNTOPT_DAX)) { > > + mp->m_flags |= XFS_MOUNT_DAX; > > Hi > > So what I see, (I might be wrong), is that once this flag is set here the > fs (At above xfs_diflags_to_iflags() ) will start serving DAX inodes. No, it won't, because... > > @@ -1501,6 +1508,20 @@ xfs_fs_fill_super( > > if (XFS_SB_VERSION_NUM(&mp->m_sb) == XFS_SB_VERSION_5) > > sb->s_flags |= MS_I_VERSION; > > > > + if (mp->m_flags & XFS_MOUNT_DAX) { > > + xfs_warn(mp, > > + "DAX enabled. Warning: EXPERIMENTAL, use at your own risk"); > > + if (sb->s_blocksize != PAGE_SIZE) { > > + xfs_alert(mp, > > + "Filesystem block size invalid for DAX Turning DAX off."); > > + mp->m_flags &= ~XFS_MOUNT_DAX; > > + } else if (!sb->s_bdev->bd_disk->fops->direct_access) { > > + xfs_alert(mp, > > + "Block device does not support DAX Turning DAX off."); > > + mp->m_flags &= ~XFS_MOUNT_DAX; > > + } > > + } We run these tests and then clear the XFS_MOUNT_DAX flag later. > If we agree about the s_flags MS_MOUNT_DAX then we can define a > if (MNTOPT_DAX) > dax_enable_if_supported(sb); No, I don't see any reason for a mount flag for this, because we do not want to be stuck relying on a mount option forever. Once This code has been shaken out, I fully intend DAX to be turned on automatically for any device that supports, and for it to be turned on and off on a per-inode basis through on-disk inode flags. i.e. the mount option is really a temporary solution and I don't want to code in any dependencies on DAX being mount wide existing. Cheers, Dave.
On 03/24/2015 11:25 PM, Dave Chinner wrote: <> > > No, I don't see any reason for a mount flag for this, because we > do not want to be stuck relying on a mount option forever. Once > This code has been shaken out, I fully intend DAX to be turned on > automatically for any device that supports, and for it to be turned > on and off on a per-inode basis through on-disk inode flags. i.e. > the mount option is really a temporary solution and I don't want to > code in any dependencies on DAX being mount wide existing. > OK I dropped this flag. I will just check for the direct_access vector. I'll need a wrapper that defines to nothing in the !CONFIG_FS_DAX case > Cheers, > Dave. > Thanks for your help Boaz -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Tue, Mar 24, 2015 at 09:51:06PM +1100, Dave Chinner wrote: > From: Dave Chinner <dchinner@redhat.com> > > Add initial DAX support to XFS. To do this we need a new mount > option to turn DAX on filesystem, and we need to propagate thi into > the inode flags whenever an inode is instantiated so that the > per-inode checks throughout the code Do The Right Thing. > > There are still some things remaining to be done: > > - needs per-inode flags to mark inodes as DAX enabled, and > an inheritance flag to enable automatic filesystem > propagation of the property > - fails occasionally with zero length writes instead of > ENOSPC errors, so error propagation inside/from the DAX > code need work > - occasionally creates two extents rather than a single > larger extent like non-dax filesystems. > - much more testing > > Signed-off-by: Dave Chinner <dchinner@redhat.com> > --- > fs/xfs/xfs_iops.c | 24 ++++++++++++------------ > fs/xfs/xfs_mount.h | 2 ++ > fs/xfs/xfs_super.c | 25 +++++++++++++++++++++++-- > 3 files changed, 37 insertions(+), 14 deletions(-) > > diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c > index 9ca5352..695d857 100644 > --- a/fs/xfs/xfs_iops.c > +++ b/fs/xfs/xfs_iops.c > @@ -1195,22 +1195,22 @@ xfs_diflags_to_iflags( > struct inode *inode, > struct xfs_inode *ip) > { > - if (ip->i_d.di_flags & XFS_DIFLAG_IMMUTABLE) > + uint16_t flags = ip->i_d.di_flags; > + > + inode->i_flags &= ~(S_IMMUTABLE | S_APPEND | S_SYNC | > + S_NOATIME | S_DAX); > + > + if (flags & XFS_DIFLAG_IMMUTABLE) > inode->i_flags |= S_IMMUTABLE; > - else > - inode->i_flags &= ~S_IMMUTABLE; > - if (ip->i_d.di_flags & XFS_DIFLAG_APPEND) > + if (flags & XFS_DIFLAG_APPEND) > inode->i_flags |= S_APPEND; > - else > - inode->i_flags &= ~S_APPEND; > - if (ip->i_d.di_flags & XFS_DIFLAG_SYNC) > + if (flags & XFS_DIFLAG_SYNC) > inode->i_flags |= S_SYNC; > - else > - inode->i_flags &= ~S_SYNC; > - if (ip->i_d.di_flags & XFS_DIFLAG_NOATIME) > + if (flags & XFS_DIFLAG_NOATIME) > inode->i_flags |= S_NOATIME; > - else > - inode->i_flags &= ~S_NOATIME; > + /* XXX: Also needs an on-disk per inode flag! */ > + if (ip->i_mount->m_flags & XFS_MOUNT_DAX) > + inode->i_flags |= S_DAX; This is a temporary hack until we have some kind of inode flag inheritance mechanism as mentioned in the commit log, correct? > } > > /* > diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h > index 8c995a2..cd44e88 100644 > --- a/fs/xfs/xfs_mount.h > +++ b/fs/xfs/xfs_mount.h > @@ -179,6 +179,8 @@ typedef struct xfs_mount { > allocator */ > #define XFS_MOUNT_NOATTR2 (1ULL << 25) /* disable use of attr2 format */ > > +#define XFS_MOUNT_DAX (1ULL << 62) /* TEST ONLY! */ > + > > /* > * Default minimum read and write sizes. > diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c > index 3ad0b17..0f26d7a 100644 > --- a/fs/xfs/xfs_super.c > +++ b/fs/xfs/xfs_super.c > @@ -112,6 +112,8 @@ static struct xfs_kobj xfs_dbg_kobj; /* global debug sysfs attrs */ > #define MNTOPT_DISCARD "discard" /* Discard unused blocks */ > #define MNTOPT_NODISCARD "nodiscard" /* Do not discard unused blocks */ > > +#define MNTOPT_DAX "dax" /* Enable direct access to bdev pages */ > + > /* > * Table driven mount option parser. > * > @@ -363,6 +365,10 @@ xfs_parseargs( > mp->m_flags |= XFS_MOUNT_DISCARD; > } else if (!strcmp(this_char, MNTOPT_NODISCARD)) { > mp->m_flags &= ~XFS_MOUNT_DISCARD; > +#ifdef CONFIG_FS_DAX > + } else if (!strcmp(this_char, MNTOPT_DAX)) { > + mp->m_flags |= XFS_MOUNT_DAX; > +#endif Something like what we do for !CONFIG_XFS_QUOTA just a few lines below might be a bit nicer. Then we can have a slightly more useful error message. > } else { > xfs_warn(mp, "unknown mount option [%s].", this_char); > return -EINVAL; > @@ -452,8 +458,8 @@ done: > } > > struct proc_xfs_info { > - int flag; > - char *str; > + uint64_t flag; > + char *str; > }; > > STATIC int > @@ -474,6 +480,7 @@ xfs_showargs( > { XFS_MOUNT_GRPID, "," MNTOPT_GRPID }, > { XFS_MOUNT_DISCARD, "," MNTOPT_DISCARD }, > { XFS_MOUNT_SMALL_INUMS, "," MNTOPT_32BITINODE }, > + { XFS_MOUNT_DAX, "," MNTOPT_DAX }, > { 0, NULL } > }; > static struct proc_xfs_info xfs_info_unset[] = { > @@ -1501,6 +1508,20 @@ xfs_fs_fill_super( > if (XFS_SB_VERSION_NUM(&mp->m_sb) == XFS_SB_VERSION_5) > sb->s_flags |= MS_I_VERSION; > > + if (mp->m_flags & XFS_MOUNT_DAX) { > + xfs_warn(mp, > + "DAX enabled. Warning: EXPERIMENTAL, use at your own risk"); > + if (sb->s_blocksize != PAGE_SIZE) { > + xfs_alert(mp, > + "Filesystem block size invalid for DAX Turning DAX off."); > + mp->m_flags &= ~XFS_MOUNT_DAX; > + } else if (!sb->s_bdev->bd_disk->fops->direct_access) { > + xfs_alert(mp, > + "Block device does not support DAX Turning DAX off."); > + mp->m_flags &= ~XFS_MOUNT_DAX; > + } Missing period in both error messages above. ;) Brian > + } > + > error = xfs_mountfs(mp); > if (error) > goto out_filestream_unmount; > -- > 2.0.0 > > _______________________________________________ > xfs mailing list > xfs@oss.sgi.com > http://oss.sgi.com/mailman/listinfo/xfs -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c index 9ca5352..695d857 100644 --- a/fs/xfs/xfs_iops.c +++ b/fs/xfs/xfs_iops.c @@ -1195,22 +1195,22 @@ xfs_diflags_to_iflags( struct inode *inode, struct xfs_inode *ip) { - if (ip->i_d.di_flags & XFS_DIFLAG_IMMUTABLE) + uint16_t flags = ip->i_d.di_flags; + + inode->i_flags &= ~(S_IMMUTABLE | S_APPEND | S_SYNC | + S_NOATIME | S_DAX); + + if (flags & XFS_DIFLAG_IMMUTABLE) inode->i_flags |= S_IMMUTABLE; - else - inode->i_flags &= ~S_IMMUTABLE; - if (ip->i_d.di_flags & XFS_DIFLAG_APPEND) + if (flags & XFS_DIFLAG_APPEND) inode->i_flags |= S_APPEND; - else - inode->i_flags &= ~S_APPEND; - if (ip->i_d.di_flags & XFS_DIFLAG_SYNC) + if (flags & XFS_DIFLAG_SYNC) inode->i_flags |= S_SYNC; - else - inode->i_flags &= ~S_SYNC; - if (ip->i_d.di_flags & XFS_DIFLAG_NOATIME) + if (flags & XFS_DIFLAG_NOATIME) inode->i_flags |= S_NOATIME; - else - inode->i_flags &= ~S_NOATIME; + /* XXX: Also needs an on-disk per inode flag! */ + if (ip->i_mount->m_flags & XFS_MOUNT_DAX) + inode->i_flags |= S_DAX; } /* diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index 8c995a2..cd44e88 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -179,6 +179,8 @@ typedef struct xfs_mount { allocator */ #define XFS_MOUNT_NOATTR2 (1ULL << 25) /* disable use of attr2 format */ +#define XFS_MOUNT_DAX (1ULL << 62) /* TEST ONLY! */ + /* * Default minimum read and write sizes. diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index 3ad0b17..0f26d7a 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -112,6 +112,8 @@ static struct xfs_kobj xfs_dbg_kobj; /* global debug sysfs attrs */ #define MNTOPT_DISCARD "discard" /* Discard unused blocks */ #define MNTOPT_NODISCARD "nodiscard" /* Do not discard unused blocks */ +#define MNTOPT_DAX "dax" /* Enable direct access to bdev pages */ + /* * Table driven mount option parser. * @@ -363,6 +365,10 @@ xfs_parseargs( mp->m_flags |= XFS_MOUNT_DISCARD; } else if (!strcmp(this_char, MNTOPT_NODISCARD)) { mp->m_flags &= ~XFS_MOUNT_DISCARD; +#ifdef CONFIG_FS_DAX + } else if (!strcmp(this_char, MNTOPT_DAX)) { + mp->m_flags |= XFS_MOUNT_DAX; +#endif } else { xfs_warn(mp, "unknown mount option [%s].", this_char); return -EINVAL; @@ -452,8 +458,8 @@ done: } struct proc_xfs_info { - int flag; - char *str; + uint64_t flag; + char *str; }; STATIC int @@ -474,6 +480,7 @@ xfs_showargs( { XFS_MOUNT_GRPID, "," MNTOPT_GRPID }, { XFS_MOUNT_DISCARD, "," MNTOPT_DISCARD }, { XFS_MOUNT_SMALL_INUMS, "," MNTOPT_32BITINODE }, + { XFS_MOUNT_DAX, "," MNTOPT_DAX }, { 0, NULL } }; static struct proc_xfs_info xfs_info_unset[] = { @@ -1501,6 +1508,20 @@ xfs_fs_fill_super( if (XFS_SB_VERSION_NUM(&mp->m_sb) == XFS_SB_VERSION_5) sb->s_flags |= MS_I_VERSION; + if (mp->m_flags & XFS_MOUNT_DAX) { + xfs_warn(mp, + "DAX enabled. Warning: EXPERIMENTAL, use at your own risk"); + if (sb->s_blocksize != PAGE_SIZE) { + xfs_alert(mp, + "Filesystem block size invalid for DAX Turning DAX off."); + mp->m_flags &= ~XFS_MOUNT_DAX; + } else if (!sb->s_bdev->bd_disk->fops->direct_access) { + xfs_alert(mp, + "Block device does not support DAX Turning DAX off."); + mp->m_flags &= ~XFS_MOUNT_DAX; + } + } + error = xfs_mountfs(mp); if (error) goto out_filestream_unmount;