* [PATCH 2/2] mkfs: pass a custom cowextsize into the created filesystem
@ 2017-09-01 16:40 Darrick J. Wong
2017-09-03 8:38 ` Christoph Hellwig
2017-09-04 17:45 ` [PATCH v2 " Darrick J. Wong
0 siblings, 2 replies; 7+ messages in thread
From: Darrick J. Wong @ 2017-09-01 16:40 UTC (permalink / raw)
To: Eric Sandeen; +Cc: xfs
Create a -d option to mkfs.xfs that enables administrators to set
the CoW extent size hint on the created files.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
libxfs/util.c | 26 ++++++++++++++++++++++++--
man/man8/mkfs.xfs.8 | 7 +++++++
mkfs/xfs_mkfs.c | 20 ++++++++++++++++++++
3 files changed, 51 insertions(+), 2 deletions(-)
diff --git a/libxfs/util.c b/libxfs/util.c
index 0e2f29e..4f82d7f 100644
--- a/libxfs/util.c
+++ b/libxfs/util.c
@@ -175,6 +175,26 @@ libxfs_trans_ichgtime(
}
}
+static inline uint16_t
+xflags_to_diflags(
+ __u32 xflags)
+{
+ /* bottom 15 xflag bits correspond to diflag bits */
+ return xflags & 0x7FFF;
+}
+
+static inline uint64_t
+xflags_to_diflags2(
+ __u32 xflags)
+{
+ uint64_t ret = 0;
+
+ if (xflags & FS_XFLAG_COWEXTSIZE)
+ ret |= XFS_DIFLAG2_COWEXTSIZE;
+
+ return ret;
+}
+
/*
* Allocate an inode on disk and return a copy of its in-core version.
* Set mode, nlink, and rdev appropriately within the inode.
@@ -254,15 +274,17 @@ libxfs_ialloc(
ip->i_d.di_extsize = pip ? 0 : fsx->fsx_extsize;
ip->i_d.di_dmevmask = 0;
ip->i_d.di_dmstate = 0;
- ip->i_d.di_flags = pip ? 0 : fsx->fsx_xflags;
+ ip->i_d.di_flags = pip ? 0 : xflags_to_diflags(fsx->fsx_xflags);
if (ip->i_d.di_version == 3) {
ASSERT(ip->i_d.di_ino == ino);
ASSERT(uuid_equal(&ip->i_d.di_uuid, &mp->m_sb.sb_meta_uuid));
VFS_I(ip)->i_version = 1;
- ip->i_d.di_flags2 = 0;
+ ip->i_d.di_flags2 = pip ? 0 :
+ xflags_to_diflags2(fsx->fsx_xflags);
ip->i_d.di_crtime.t_sec = (int32_t)VFS_I(ip)->i_mtime.tv_sec;
ip->i_d.di_crtime.t_nsec = (int32_t)VFS_I(ip)->i_mtime.tv_nsec;
+ ip->i_d.di_cowextsize = pip ? 0 : fsx->fsx_cowextsize;
}
flags = XFS_ILOG_CORE;
diff --git a/man/man8/mkfs.xfs.8 b/man/man8/mkfs.xfs.8
index 33831c9..bbbe1c5 100644
--- a/man/man8/mkfs.xfs.8
+++ b/man/man8/mkfs.xfs.8
@@ -279,6 +279,13 @@ and
.B agsize
suboptions are mutually exclusive.
.TP
+.BI cowextsize= value
+Set the copy-on-write extent size hint on all inodes created by
+.BR mkfs.xfs "."
+The value must be provided in units of filesystem blocks.
+If the value is zero, the default value (currently 32 blocks) will be used.
+Directories will pass on this hint to newly created children.
+.TP
.BI name= value
This can be used to specify the name of the special file containing
the filesystem. In this case, the log section must be specified as
diff --git a/mkfs/xfs_mkfs.c b/mkfs/xfs_mkfs.c
index 024cb02..ea0a317 100644
--- a/mkfs/xfs_mkfs.c
+++ b/mkfs/xfs_mkfs.c
@@ -187,6 +187,8 @@ struct opt_params dopts = {
"projinherit",
#define D_EXTSZINHERIT 14
"extszinherit",
+#define D_COWEXTSIZE 15
+ "cowextsize",
NULL
},
.subopt_params = {
@@ -303,6 +305,12 @@ struct opt_params dopts = {
.maxval = UINT_MAX,
.defaultval = SUBOPT_NEEDS_VAL,
},
+ { .index = D_COWEXTSIZE,
+ .conflicts = { LAST_CONFLICT },
+ .minval = 0,
+ .maxval = UINT_MAX,
+ .defaultval = SUBOPT_NEEDS_VAL,
+ },
},
};
@@ -1638,6 +1646,13 @@ main(
fsx.fsx_xflags |=
XFS_DIFLAG_EXTSZINHERIT;
break;
+ case D_COWEXTSIZE:
+ fsx.fsx_cowextsize = getnum(value,
+ &dopts,
+ D_COWEXTSIZE);
+ fsx.fsx_xflags |=
+ FS_XFLAG_COWEXTSIZE;
+ break;
default:
unknown('d', value);
}
@@ -2142,6 +2157,11 @@ _("reflink not supported without CRC support\n"));
sb_feat.reflink = false;
}
+ if ((fsx.fsx_xflags & FS_XFLAG_COWEXTSIZE) && !sb_feat.reflink) {
+ fprintf(stderr,
+_("cowextsize not supported without reflink support\n"));
+ usage();
+ }
if (nsflag || nlflag) {
if (dirblocksize < blocksize ||
^ permalink raw reply related [flat|nested] 7+ messages in thread
* Re: [PATCH 2/2] mkfs: pass a custom cowextsize into the created filesystem
2017-09-01 16:40 [PATCH 2/2] mkfs: pass a custom cowextsize into the created filesystem Darrick J. Wong
@ 2017-09-03 8:38 ` Christoph Hellwig
2017-09-03 15:35 ` Darrick J. Wong
2017-09-04 17:45 ` [PATCH v2 " Darrick J. Wong
1 sibling, 1 reply; 7+ messages in thread
From: Christoph Hellwig @ 2017-09-03 8:38 UTC (permalink / raw)
To: Darrick J. Wong; +Cc: Eric Sandeen, xfs
On Fri, Sep 01, 2017 at 09:40:35AM -0700, Darrick J. Wong wrote:
> Create a -d option to mkfs.xfs that enables administrators to set
> the CoW extent size hint on the created files.
>
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> ---
> libxfs/util.c | 26 ++++++++++++++++++++++++--
> man/man8/mkfs.xfs.8 | 7 +++++++
> mkfs/xfs_mkfs.c | 20 ++++++++++++++++++++
> 3 files changed, 51 insertions(+), 2 deletions(-)
>
> diff --git a/libxfs/util.c b/libxfs/util.c
> index 0e2f29e..4f82d7f 100644
> --- a/libxfs/util.c
> +++ b/libxfs/util.c
> @@ -175,6 +175,26 @@ libxfs_trans_ichgtime(
> }
> }
>
> +static inline uint16_t
> +xflags_to_diflags(
> + __u32 xflags)
> +{
> + /* bottom 15 xflag bits correspond to diflag bits */
> + return xflags & 0x7FFF;
Do we really want to rely on that fact?
> +}
> +
> +static inline uint64_t
> +xflags_to_diflags2(
> + __u32 xflags)
> +{
> + uint64_t ret = 0;
> +
> + if (xflags & FS_XFLAG_COWEXTSIZE)
> + ret |= XFS_DIFLAG2_COWEXTSIZE;
> +
> + return ret;
> +}
It seems like we should just lift the kernels xfs_flags2diflags and
xfs_flags2diflags2 to libxfs and use them here?
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH 2/2] mkfs: pass a custom cowextsize into the created filesystem
2017-09-03 8:38 ` Christoph Hellwig
@ 2017-09-03 15:35 ` Darrick J. Wong
0 siblings, 0 replies; 7+ messages in thread
From: Darrick J. Wong @ 2017-09-03 15:35 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: Eric Sandeen, xfs
On Sun, Sep 03, 2017 at 01:38:57AM -0700, Christoph Hellwig wrote:
> On Fri, Sep 01, 2017 at 09:40:35AM -0700, Darrick J. Wong wrote:
> > Create a -d option to mkfs.xfs that enables administrators to set
> > the CoW extent size hint on the created files.
> >
> > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> > ---
> > libxfs/util.c | 26 ++++++++++++++++++++++++--
> > man/man8/mkfs.xfs.8 | 7 +++++++
> > mkfs/xfs_mkfs.c | 20 ++++++++++++++++++++
> > 3 files changed, 51 insertions(+), 2 deletions(-)
> >
> > diff --git a/libxfs/util.c b/libxfs/util.c
> > index 0e2f29e..4f82d7f 100644
> > --- a/libxfs/util.c
> > +++ b/libxfs/util.c
> > @@ -175,6 +175,26 @@ libxfs_trans_ichgtime(
> > }
> > }
> >
> > +static inline uint16_t
> > +xflags_to_diflags(
> > + __u32 xflags)
> > +{
> > + /* bottom 15 xflag bits correspond to diflag bits */
> > + return xflags & 0x7FFF;
>
> Do we really want to rely on that fact?
>
> > +}
> > +
> > +static inline uint64_t
> > +xflags_to_diflags2(
> > + __u32 xflags)
> > +{
> > + uint64_t ret = 0;
> > +
> > + if (xflags & FS_XFLAG_COWEXTSIZE)
> > + ret |= XFS_DIFLAG2_COWEXTSIZE;
> > +
> > + return ret;
> > +}
>
> It seems like we should just lift the kernels xfs_flags2diflags and
> xfs_flags2diflags2 to libxfs and use them here?
Once that they exist upstream, sure! :)
--D
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 7+ messages in thread
* [PATCH v2 2/2] mkfs: pass a custom cowextsize into the created filesystem
2017-09-01 16:40 [PATCH 2/2] mkfs: pass a custom cowextsize into the created filesystem Darrick J. Wong
2017-09-03 8:38 ` Christoph Hellwig
@ 2017-09-04 17:45 ` Darrick J. Wong
2017-09-08 18:02 ` Eric Sandeen
1 sibling, 1 reply; 7+ messages in thread
From: Darrick J. Wong @ 2017-09-04 17:45 UTC (permalink / raw)
To: Eric Sandeen; +Cc: xfs
Create a -d option to mkfs.xfs that enables administrators to set
the CoW extent size hint on the created files.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
v2: port flags2diflags from the kernel
---
libxfs/util.c | 64 +++++++++++++++++++++++++++++++++++++++++++++++++--
man/man8/mkfs.xfs.8 | 7 ++++++
mkfs/xfs_mkfs.c | 20 ++++++++++++++++
3 files changed, 89 insertions(+), 2 deletions(-)
diff --git a/libxfs/util.c b/libxfs/util.c
index 0e2f29e..b600594 100644
--- a/libxfs/util.c
+++ b/libxfs/util.c
@@ -175,6 +175,64 @@ libxfs_trans_ichgtime(
}
}
+STATIC uint16_t
+xfs_flags2diflags(
+ struct xfs_inode *ip,
+ unsigned int xflags)
+{
+ /* can't set PREALLOC this way, just preserve it */
+ uint16_t di_flags =
+ (ip->i_d.di_flags & XFS_DIFLAG_PREALLOC);
+
+ if (xflags & FS_XFLAG_IMMUTABLE)
+ di_flags |= XFS_DIFLAG_IMMUTABLE;
+ if (xflags & FS_XFLAG_APPEND)
+ di_flags |= XFS_DIFLAG_APPEND;
+ if (xflags & FS_XFLAG_SYNC)
+ di_flags |= XFS_DIFLAG_SYNC;
+ if (xflags & FS_XFLAG_NOATIME)
+ di_flags |= XFS_DIFLAG_NOATIME;
+ if (xflags & FS_XFLAG_NODUMP)
+ di_flags |= XFS_DIFLAG_NODUMP;
+ if (xflags & FS_XFLAG_NODEFRAG)
+ di_flags |= XFS_DIFLAG_NODEFRAG;
+ if (xflags & FS_XFLAG_FILESTREAM)
+ di_flags |= XFS_DIFLAG_FILESTREAM;
+ if (S_ISDIR(VFS_I(ip)->i_mode)) {
+ if (xflags & FS_XFLAG_RTINHERIT)
+ di_flags |= XFS_DIFLAG_RTINHERIT;
+ if (xflags & FS_XFLAG_NOSYMLINKS)
+ di_flags |= XFS_DIFLAG_NOSYMLINKS;
+ if (xflags & FS_XFLAG_EXTSZINHERIT)
+ di_flags |= XFS_DIFLAG_EXTSZINHERIT;
+ if (xflags & FS_XFLAG_PROJINHERIT)
+ di_flags |= XFS_DIFLAG_PROJINHERIT;
+ } else if (S_ISREG(VFS_I(ip)->i_mode)) {
+ if (xflags & FS_XFLAG_REALTIME)
+ di_flags |= XFS_DIFLAG_REALTIME;
+ if (xflags & FS_XFLAG_EXTSIZE)
+ di_flags |= XFS_DIFLAG_EXTSIZE;
+ }
+
+ return di_flags;
+}
+
+STATIC uint64_t
+xfs_flags2diflags2(
+ struct xfs_inode *ip,
+ unsigned int xflags)
+{
+ uint64_t di_flags2 =
+ (ip->i_d.di_flags2 & XFS_DIFLAG2_REFLINK);
+
+ if (xflags & FS_XFLAG_DAX)
+ di_flags2 |= XFS_DIFLAG2_DAX;
+ if (xflags & FS_XFLAG_COWEXTSIZE)
+ di_flags2 |= XFS_DIFLAG2_COWEXTSIZE;
+
+ return di_flags2;
+}
+
/*
* Allocate an inode on disk and return a copy of its in-core version.
* Set mode, nlink, and rdev appropriately within the inode.
@@ -254,15 +312,17 @@ libxfs_ialloc(
ip->i_d.di_extsize = pip ? 0 : fsx->fsx_extsize;
ip->i_d.di_dmevmask = 0;
ip->i_d.di_dmstate = 0;
- ip->i_d.di_flags = pip ? 0 : fsx->fsx_xflags;
+ ip->i_d.di_flags = pip ? 0 : xfs_flags2diflags(ip, fsx->fsx_xflags);
if (ip->i_d.di_version == 3) {
ASSERT(ip->i_d.di_ino == ino);
ASSERT(uuid_equal(&ip->i_d.di_uuid, &mp->m_sb.sb_meta_uuid));
VFS_I(ip)->i_version = 1;
- ip->i_d.di_flags2 = 0;
+ ip->i_d.di_flags2 = pip ? 0 : xfs_flags2diflags2(ip,
+ fsx->fsx_xflags);
ip->i_d.di_crtime.t_sec = (int32_t)VFS_I(ip)->i_mtime.tv_sec;
ip->i_d.di_crtime.t_nsec = (int32_t)VFS_I(ip)->i_mtime.tv_nsec;
+ ip->i_d.di_cowextsize = pip ? 0 : fsx->fsx_cowextsize;
}
flags = XFS_ILOG_CORE;
diff --git a/man/man8/mkfs.xfs.8 b/man/man8/mkfs.xfs.8
index 33831c9..bbbe1c5 100644
--- a/man/man8/mkfs.xfs.8
+++ b/man/man8/mkfs.xfs.8
@@ -279,6 +279,13 @@ and
.B agsize
suboptions are mutually exclusive.
.TP
+.BI cowextsize= value
+Set the copy-on-write extent size hint on all inodes created by
+.BR mkfs.xfs "."
+The value must be provided in units of filesystem blocks.
+If the value is zero, the default value (currently 32 blocks) will be used.
+Directories will pass on this hint to newly created children.
+.TP
.BI name= value
This can be used to specify the name of the special file containing
the filesystem. In this case, the log section must be specified as
diff --git a/mkfs/xfs_mkfs.c b/mkfs/xfs_mkfs.c
index 024cb02..ea0a317 100644
--- a/mkfs/xfs_mkfs.c
+++ b/mkfs/xfs_mkfs.c
@@ -187,6 +187,8 @@ struct opt_params dopts = {
"projinherit",
#define D_EXTSZINHERIT 14
"extszinherit",
+#define D_COWEXTSIZE 15
+ "cowextsize",
NULL
},
.subopt_params = {
@@ -303,6 +305,12 @@ struct opt_params dopts = {
.maxval = UINT_MAX,
.defaultval = SUBOPT_NEEDS_VAL,
},
+ { .index = D_COWEXTSIZE,
+ .conflicts = { LAST_CONFLICT },
+ .minval = 0,
+ .maxval = UINT_MAX,
+ .defaultval = SUBOPT_NEEDS_VAL,
+ },
},
};
@@ -1638,6 +1646,13 @@ main(
fsx.fsx_xflags |=
XFS_DIFLAG_EXTSZINHERIT;
break;
+ case D_COWEXTSIZE:
+ fsx.fsx_cowextsize = getnum(value,
+ &dopts,
+ D_COWEXTSIZE);
+ fsx.fsx_xflags |=
+ FS_XFLAG_COWEXTSIZE;
+ break;
default:
unknown('d', value);
}
@@ -2142,6 +2157,11 @@ _("reflink not supported without CRC support\n"));
sb_feat.reflink = false;
}
+ if ((fsx.fsx_xflags & FS_XFLAG_COWEXTSIZE) && !sb_feat.reflink) {
+ fprintf(stderr,
+_("cowextsize not supported without reflink support\n"));
+ usage();
+ }
if (nsflag || nlflag) {
if (dirblocksize < blocksize ||
^ permalink raw reply related [flat|nested] 7+ messages in thread
* Re: [PATCH v2 2/2] mkfs: pass a custom cowextsize into the created filesystem
2017-09-04 17:45 ` [PATCH v2 " Darrick J. Wong
@ 2017-09-08 18:02 ` Eric Sandeen
2017-09-18 17:22 ` Darrick J. Wong
0 siblings, 1 reply; 7+ messages in thread
From: Eric Sandeen @ 2017-09-08 18:02 UTC (permalink / raw)
To: Darrick J. Wong, Eric Sandeen; +Cc: xfs
On 9/4/17 12:45 PM, Darrick J. Wong wrote:
> Create a -d option to mkfs.xfs that enables administrators to set
> the CoW extent size hint on the created files.
>
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> ---
> v2: port flags2diflags from the kernel
hm, I wonder if we need a libxfs/util.c in the kernel for stuff like
this so libxfs diff will keep such things in sync in the future.
a project for another day I guess. One question below.
> ---
> libxfs/util.c | 64 +++++++++++++++++++++++++++++++++++++++++++++++++--
> man/man8/mkfs.xfs.8 | 7 ++++++
> mkfs/xfs_mkfs.c | 20 ++++++++++++++++
> 3 files changed, 89 insertions(+), 2 deletions(-)
>
> diff --git a/libxfs/util.c b/libxfs/util.c
> index 0e2f29e..b600594 100644
> --- a/libxfs/util.c
> +++ b/libxfs/util.c
> @@ -175,6 +175,64 @@ libxfs_trans_ichgtime(
> }
> }
>
> +STATIC uint16_t
> +xfs_flags2diflags(
> + struct xfs_inode *ip,
> + unsigned int xflags)
> +{
> + /* can't set PREALLOC this way, just preserve it */
> + uint16_t di_flags =
> + (ip->i_d.di_flags & XFS_DIFLAG_PREALLOC);
> +
> + if (xflags & FS_XFLAG_IMMUTABLE)
> + di_flags |= XFS_DIFLAG_IMMUTABLE;
> + if (xflags & FS_XFLAG_APPEND)
> + di_flags |= XFS_DIFLAG_APPEND;
> + if (xflags & FS_XFLAG_SYNC)
> + di_flags |= XFS_DIFLAG_SYNC;
> + if (xflags & FS_XFLAG_NOATIME)
> + di_flags |= XFS_DIFLAG_NOATIME;
> + if (xflags & FS_XFLAG_NODUMP)
> + di_flags |= XFS_DIFLAG_NODUMP;
> + if (xflags & FS_XFLAG_NODEFRAG)
> + di_flags |= XFS_DIFLAG_NODEFRAG;
> + if (xflags & FS_XFLAG_FILESTREAM)
> + di_flags |= XFS_DIFLAG_FILESTREAM;
> + if (S_ISDIR(VFS_I(ip)->i_mode)) {
> + if (xflags & FS_XFLAG_RTINHERIT)
> + di_flags |= XFS_DIFLAG_RTINHERIT;
> + if (xflags & FS_XFLAG_NOSYMLINKS)
> + di_flags |= XFS_DIFLAG_NOSYMLINKS;
> + if (xflags & FS_XFLAG_EXTSZINHERIT)
> + di_flags |= XFS_DIFLAG_EXTSZINHERIT;
> + if (xflags & FS_XFLAG_PROJINHERIT)
> + di_flags |= XFS_DIFLAG_PROJINHERIT;
> + } else if (S_ISREG(VFS_I(ip)->i_mode)) {
> + if (xflags & FS_XFLAG_REALTIME)
> + di_flags |= XFS_DIFLAG_REALTIME;
> + if (xflags & FS_XFLAG_EXTSIZE)
> + di_flags |= XFS_DIFLAG_EXTSIZE;
> + }
> +
> + return di_flags;
> +}
> +
> +STATIC uint64_t
> +xfs_flags2diflags2(
> + struct xfs_inode *ip,
> + unsigned int xflags)
> +{
> + uint64_t di_flags2 =
> + (ip->i_d.di_flags2 & XFS_DIFLAG2_REFLINK);
> +
> + if (xflags & FS_XFLAG_DAX)
> + di_flags2 |= XFS_DIFLAG2_DAX;
> + if (xflags & FS_XFLAG_COWEXTSIZE)
> + di_flags2 |= XFS_DIFLAG2_COWEXTSIZE;
> +
> + return di_flags2;
> +}
> +
> /*
> * Allocate an inode on disk and return a copy of its in-core version.
> * Set mode, nlink, and rdev appropriately within the inode.
> @@ -254,15 +312,17 @@ libxfs_ialloc(
> ip->i_d.di_extsize = pip ? 0 : fsx->fsx_extsize;
> ip->i_d.di_dmevmask = 0;
> ip->i_d.di_dmstate = 0;
> - ip->i_d.di_flags = pip ? 0 : fsx->fsx_xflags;
> + ip->i_d.di_flags = pip ? 0 : xfs_flags2diflags(ip, fsx->fsx_xflags);
is this a bugfix?
The rest seems ok, just wondering if the above should be hidden in here -
at least a mention of the change in the changelog seems needed, but most
likely thjis is a separate change, no? Or am I missing something?
-Eric
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH v2 2/2] mkfs: pass a custom cowextsize into the created filesystem
2017-09-08 18:02 ` Eric Sandeen
@ 2017-09-18 17:22 ` Darrick J. Wong
2017-09-18 17:37 ` Eric Sandeen
0 siblings, 1 reply; 7+ messages in thread
From: Darrick J. Wong @ 2017-09-18 17:22 UTC (permalink / raw)
To: Eric Sandeen; +Cc: Eric Sandeen, xfs
On Fri, Sep 08, 2017 at 01:02:15PM -0500, Eric Sandeen wrote:
> On 9/4/17 12:45 PM, Darrick J. Wong wrote:
> > Create a -d option to mkfs.xfs that enables administrators to set
> > the CoW extent size hint on the created files.
> >
> > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> > ---
> > v2: port flags2diflags from the kernel
>
> hm, I wonder if we need a libxfs/util.c in the kernel for stuff like
> this so libxfs diff will keep such things in sync in the future.
>
> a project for another day I guess. One question below.
>
> > ---
> > libxfs/util.c | 64 +++++++++++++++++++++++++++++++++++++++++++++++++--
> > man/man8/mkfs.xfs.8 | 7 ++++++
> > mkfs/xfs_mkfs.c | 20 ++++++++++++++++
> > 3 files changed, 89 insertions(+), 2 deletions(-)
> >
> > diff --git a/libxfs/util.c b/libxfs/util.c
> > index 0e2f29e..b600594 100644
> > --- a/libxfs/util.c
> > +++ b/libxfs/util.c
> > @@ -175,6 +175,64 @@ libxfs_trans_ichgtime(
> > }
> > }
> >
> > +STATIC uint16_t
> > +xfs_flags2diflags(
> > + struct xfs_inode *ip,
> > + unsigned int xflags)
> > +{
> > + /* can't set PREALLOC this way, just preserve it */
> > + uint16_t di_flags =
> > + (ip->i_d.di_flags & XFS_DIFLAG_PREALLOC);
> > +
> > + if (xflags & FS_XFLAG_IMMUTABLE)
> > + di_flags |= XFS_DIFLAG_IMMUTABLE;
> > + if (xflags & FS_XFLAG_APPEND)
> > + di_flags |= XFS_DIFLAG_APPEND;
> > + if (xflags & FS_XFLAG_SYNC)
> > + di_flags |= XFS_DIFLAG_SYNC;
> > + if (xflags & FS_XFLAG_NOATIME)
> > + di_flags |= XFS_DIFLAG_NOATIME;
> > + if (xflags & FS_XFLAG_NODUMP)
> > + di_flags |= XFS_DIFLAG_NODUMP;
> > + if (xflags & FS_XFLAG_NODEFRAG)
> > + di_flags |= XFS_DIFLAG_NODEFRAG;
> > + if (xflags & FS_XFLAG_FILESTREAM)
> > + di_flags |= XFS_DIFLAG_FILESTREAM;
> > + if (S_ISDIR(VFS_I(ip)->i_mode)) {
> > + if (xflags & FS_XFLAG_RTINHERIT)
> > + di_flags |= XFS_DIFLAG_RTINHERIT;
> > + if (xflags & FS_XFLAG_NOSYMLINKS)
> > + di_flags |= XFS_DIFLAG_NOSYMLINKS;
> > + if (xflags & FS_XFLAG_EXTSZINHERIT)
> > + di_flags |= XFS_DIFLAG_EXTSZINHERIT;
> > + if (xflags & FS_XFLAG_PROJINHERIT)
> > + di_flags |= XFS_DIFLAG_PROJINHERIT;
> > + } else if (S_ISREG(VFS_I(ip)->i_mode)) {
> > + if (xflags & FS_XFLAG_REALTIME)
> > + di_flags |= XFS_DIFLAG_REALTIME;
> > + if (xflags & FS_XFLAG_EXTSIZE)
> > + di_flags |= XFS_DIFLAG_EXTSIZE;
> > + }
> > +
> > + return di_flags;
> > +}
> > +
> > +STATIC uint64_t
> > +xfs_flags2diflags2(
> > + struct xfs_inode *ip,
> > + unsigned int xflags)
> > +{
> > + uint64_t di_flags2 =
> > + (ip->i_d.di_flags2 & XFS_DIFLAG2_REFLINK);
> > +
> > + if (xflags & FS_XFLAG_DAX)
> > + di_flags2 |= XFS_DIFLAG2_DAX;
> > + if (xflags & FS_XFLAG_COWEXTSIZE)
> > + di_flags2 |= XFS_DIFLAG2_COWEXTSIZE;
> > +
> > + return di_flags2;
> > +}
> > +
> > /*
> > * Allocate an inode on disk and return a copy of its in-core version.
> > * Set mode, nlink, and rdev appropriately within the inode.
> > @@ -254,15 +312,17 @@ libxfs_ialloc(
> > ip->i_d.di_extsize = pip ? 0 : fsx->fsx_extsize;
> > ip->i_d.di_dmevmask = 0;
> > ip->i_d.di_dmstate = 0;
> > - ip->i_d.di_flags = pip ? 0 : fsx->fsx_xflags;
> > + ip->i_d.di_flags = pip ? 0 : xfs_flags2diflags(ip, fsx->fsx_xflags);
>
> is this a bugfix?
No.
Prior to this patch, the only fsx_xflags bits that mkfs could set are
the ones that correspond exactly to di_flags bits, so it was fine to set
them directly. Subtle and annoying, but it worked.
However, the xfs_mkfs.c changes enable us to set FS_XFLAG_COWEXTSIZE,
which doesn't correspond to a di_flags bit, so now we need translation
functions to return the correct di_flags/di_flags2 values for the given
fsx_xflags.
(Granted you could argue that COWEXTSIZE is bit 17 so it'll just get
truncated when we set di_flags and therefore it's not absolutely
necessary to have a translation function for di_flags but that's
sloppy.)
--D
> The rest seems ok, just wondering if the above should be hidden in here -
> at least a mention of the change in the changelog seems needed, but most
> likely thjis is a separate change, no? Or am I missing something?
>
> -Eric
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH v2 2/2] mkfs: pass a custom cowextsize into the created filesystem
2017-09-18 17:22 ` Darrick J. Wong
@ 2017-09-18 17:37 ` Eric Sandeen
0 siblings, 0 replies; 7+ messages in thread
From: Eric Sandeen @ 2017-09-18 17:37 UTC (permalink / raw)
To: Darrick J. Wong; +Cc: Eric Sandeen, xfs
On 9/18/17 12:22 PM, Darrick J. Wong wrote:
>>> @@ -254,15 +312,17 @@ libxfs_ialloc(
>>> ip->i_d.di_extsize = pip ? 0 : fsx->fsx_extsize;
>>> ip->i_d.di_dmevmask = 0;
>>> ip->i_d.di_dmstate = 0;
>>> - ip->i_d.di_flags = pip ? 0 : fsx->fsx_xflags;
>>> + ip->i_d.di_flags = pip ? 0 : xfs_flags2diflags(ip, fsx->fsx_xflags);
>> is this a bugfix?
> No.
>
> Prior to this patch, the only fsx_xflags bits that mkfs could set are
> the ones that correspond exactly to di_flags bits, so it was fine to set
> them directly. Subtle and annoying, but it worked.
>
> However, the xfs_mkfs.c changes enable us to set FS_XFLAG_COWEXTSIZE,
> which doesn't correspond to a di_flags bit, so now we need translation
> functions to return the correct di_flags/di_flags2 values for the given
> fsx_xflags.
Oh, ok. Before I guess it was only XFS_DIFLAG_RTINHERIT,
XFS_DIFLAG_PROJINHERIT and XFS_DIFLAG_EXTSZINHERIT.
Fair enough, maybe I should have seen that. Thanks.
(I might note it in the changelog on the way in, though, since
it's subtle and annoying) ;)
-Eric
^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2017-09-18 17:37 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-09-01 16:40 [PATCH 2/2] mkfs: pass a custom cowextsize into the created filesystem Darrick J. Wong
2017-09-03 8:38 ` Christoph Hellwig
2017-09-03 15:35 ` Darrick J. Wong
2017-09-04 17:45 ` [PATCH v2 " Darrick J. Wong
2017-09-08 18:02 ` Eric Sandeen
2017-09-18 17:22 ` Darrick J. Wong
2017-09-18 17:37 ` Eric Sandeen
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.