* [PATCH v2] xfs_io: copy_range don't truncate dst_file, and add smart length
@ 2019-09-06 17:02 Jianhong.Yin
2019-09-06 21:04 ` Eric Sandeen
0 siblings, 1 reply; 3+ messages in thread
From: Jianhong.Yin @ 2019-09-06 17:02 UTC (permalink / raw)
To: linux-xfs; +Cc: jiyin, darrick.wong, Jianhong.Yin
1. copy_range should be a simple wrapper for copy_file_range(2)
and nothing else. and there's already -t option for truncate.
so here we remove the truncate action in copy_range.
see: https://patchwork.kernel.org/comment/22863587/#1
2. improve the default length value generation:
if -l option is omitted use the length that from src_offset to end
(src_file's size - src_offset) instead.
if src_offset is greater than file size, length is 0.
3. update manpage
and have confirmed that this change will not affect xfstests.
Signed-off-by: Jianhong Yin <yin-jianhong@163.com>
---
io/copy_file_range.c | 22 +++++-----------------
man/man8/xfs_io.8 | 12 ++++++------
2 files changed, 11 insertions(+), 23 deletions(-)
diff --git a/io/copy_file_range.c b/io/copy_file_range.c
index b7b9fd88..2bc8494e 100644
--- a/io/copy_file_range.c
+++ b/io/copy_file_range.c
@@ -66,21 +66,13 @@ copy_src_filesize(int fd)
return st.st_size;
}
-static int
-copy_dst_truncate(void)
-{
- int ret = ftruncate(file->fd, 0);
- if (ret < 0)
- perror("ftruncate");
- return ret;
-}
-
static int
copy_range_f(int argc, char **argv)
{
long long src = 0;
long long dst = 0;
size_t len = 0;
+ bool len_specified = false;
int opt;
int ret;
int fd;
@@ -112,6 +104,7 @@ copy_range_f(int argc, char **argv)
printf(_("invalid length -- %s\n"), optarg);
return 0;
}
+ len_specified = true;
break;
case 'f':
src_file_nr = atoi(argv[1]);
@@ -137,7 +130,7 @@ copy_range_f(int argc, char **argv)
fd = filetable[src_file_nr].fd;
}
- if (src == 0 && dst == 0 && len == 0) {
+ if (! len_specified) {
off64_t sz;
sz = copy_src_filesize(fd);
@@ -145,13 +138,8 @@ copy_range_f(int argc, char **argv)
ret = 1;
goto out;
}
- len = sz;
-
- ret = copy_dst_truncate();
- if (ret < 0) {
- ret = 1;
- goto out;
- }
+ if (sz > src)
+ len = sz - src;
}
ret = copy_file_range_cmd(fd, &src, &dst, len);
diff --git a/man/man8/xfs_io.8 b/man/man8/xfs_io.8
index 6e064bdd..61c35c8e 100644
--- a/man/man8/xfs_io.8
+++ b/man/man8/xfs_io.8
@@ -669,13 +669,13 @@ The source must be specified either by path
or as another open file
.RB ( \-f ).
If
-.I src_file
-.IR src_offset ,
-.IR dst_offset ,
-and
.I length
-are omitted the contents of src_file will be copied to the beginning of the
-open file, overwriting any data already there.
+is not specified, this command copies data from
+.I src_offset
+to the end of
+.BI src_file
+into the dst_file at
+.IR dst_offset .
.RS 1.0i
.PD 0
.TP 0.4i
--
2.21.0
^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [PATCH v2] xfs_io: copy_range don't truncate dst_file, and add smart length
2019-09-06 17:02 [PATCH v2] xfs_io: copy_range don't truncate dst_file, and add smart length Jianhong.Yin
@ 2019-09-06 21:04 ` Eric Sandeen
2019-09-09 6:11 ` Jianhong Yin
0 siblings, 1 reply; 3+ messages in thread
From: Eric Sandeen @ 2019-09-06 21:04 UTC (permalink / raw)
To: Jianhong.Yin, linux-xfs; +Cc: jiyin, darrick.wong
On 9/6/19 12:02 PM, Jianhong.Yin wrote:
> 1. copy_range should be a simple wrapper for copy_file_range(2)
> and nothing else. and there's already -t option for truncate.
> so here we remove the truncate action in copy_range.
> see: https://patchwork.kernel.org/comment/22863587/#1
>
> 2. improve the default length value generation:
> if -l option is omitted use the length that from src_offset to end
> (src_file's size - src_offset) instead.
> if src_offset is greater than file size, length is 0.
>
> 3. update manpage
>
> and have confirmed that this change will not affect xfstests.
>
> Signed-off-by: Jianhong Yin <yin-jianhong@163.com>
> ---
> io/copy_file_range.c | 22 +++++-----------------
> man/man8/xfs_io.8 | 12 ++++++------
> 2 files changed, 11 insertions(+), 23 deletions(-)
>
> diff --git a/io/copy_file_range.c b/io/copy_file_range.c
> index b7b9fd88..2bc8494e 100644
> --- a/io/copy_file_range.c
> +++ b/io/copy_file_range.c
> @@ -66,21 +66,13 @@ copy_src_filesize(int fd)
> return st.st_size;
> }
>
> -static int
> -copy_dst_truncate(void)
> -{
> - int ret = ftruncate(file->fd, 0);
> - if (ret < 0)
> - perror("ftruncate");
> - return ret;
> -}
> -
> static int
> copy_range_f(int argc, char **argv)
> {
> long long src = 0;
> long long dst = 0;
I'd like to change these to src_off & dst_off, just to help keep track
of things.
> size_t len = 0;
> + bool len_specified = false;
> int opt;
> int ret;
> int fd;
> @@ -112,6 +104,7 @@ copy_range_f(int argc, char **argv)
> printf(_("invalid length -- %s\n"), optarg);
> return 0;
> }
> + len_specified = true;
> break;
> case 'f':
> src_file_nr = atoi(argv[1]);
> @@ -137,7 +130,7 @@ copy_range_f(int argc, char **argv)
> fd = filetable[src_file_nr].fd;
> }
>
> - if (src == 0 && dst == 0 && len == 0) {
> + if (! len_specified) {
normal xfsprogs style is no space after the !
> off64_t sz;
>
> sz = copy_src_filesize(fd);
> @@ -145,13 +138,8 @@ copy_range_f(int argc, char **argv)
> ret = 1;
> goto out;
> }
> - len = sz;
> -
> - ret = copy_dst_truncate();
> - if (ret < 0) {
> - ret = 1;
> - goto out;
> - }
> + if (sz > src)
> + len = sz - src;
Ok, so if source offset is past EOF, we keep len = 0.
Looks like the kernel does that internally too, so no problem there.
I'll make the cosmetic changes I mentioned above when I commit it, unless
you have any concerns about that.
/* Shorten the copy to EOF */
size_in = i_size_read(inode_in);
if (pos_in >= size_in)
count = 0;
Reviewed-by: Eric Sandeen <sandeen@redhat.com>
> }
>
> ret = copy_file_range_cmd(fd, &src, &dst, len);
> diff --git a/man/man8/xfs_io.8 b/man/man8/xfs_io.8
> index 6e064bdd..61c35c8e 100644
> --- a/man/man8/xfs_io.8
> +++ b/man/man8/xfs_io.8
> @@ -669,13 +669,13 @@ The source must be specified either by path
> or as another open file
> .RB ( \-f ).
> If
> -.I src_file
> -.IR src_offset ,
> -.IR dst_offset ,
> -and
> .I length
> -are omitted the contents of src_file will be copied to the beginning of the
> -open file, overwriting any data already there.
> +is not specified, this command copies data from
> +.I src_offset
> +to the end of
> +.BI src_file
> +into the dst_file at
> +.IR dst_offset .
> .RS 1.0i
> .PD 0
> .TP 0.4i
>
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH v2] xfs_io: copy_range don't truncate dst_file, and add smart length
2019-09-06 21:04 ` Eric Sandeen
@ 2019-09-09 6:11 ` Jianhong Yin
0 siblings, 0 replies; 3+ messages in thread
From: Jianhong Yin @ 2019-09-09 6:11 UTC (permalink / raw)
To: Eric Sandeen; +Cc: Jianhong.Yin, linux-xfs, darrick wong
----- 原始邮件 -----
> 发件人: "Eric Sandeen" <sandeen@sandeen.net>
> 收件人: "Jianhong.Yin" <yin-jianhong@163.com>, linux-xfs@vger.kernel.org
> 抄送: jiyin@redhat.com, "darrick wong" <darrick.wong@oracle.com>
> 发送时间: 星期六, 2019年 9 月 07日 上午 5:04:13
> 主题: Re: [PATCH v2] xfs_io: copy_range don't truncate dst_file, and add smart length
>
> On 9/6/19 12:02 PM, Jianhong.Yin wrote:
> > 1. copy_range should be a simple wrapper for copy_file_range(2)
> > and nothing else. and there's already -t option for truncate.
> > so here we remove the truncate action in copy_range.
> > see: https://patchwork.kernel.org/comment/22863587/#1
> >
> > 2. improve the default length value generation:
> > if -l option is omitted use the length that from src_offset to end
> > (src_file's size - src_offset) instead.
> > if src_offset is greater than file size, length is 0.
> >
> > 3. update manpage
> >
> > and have confirmed that this change will not affect xfstests.
> >
> > Signed-off-by: Jianhong Yin <yin-jianhong@163.com>
> > ---
> > io/copy_file_range.c | 22 +++++-----------------
> > man/man8/xfs_io.8 | 12 ++++++------
> > 2 files changed, 11 insertions(+), 23 deletions(-)
> >
> > diff --git a/io/copy_file_range.c b/io/copy_file_range.c
> > index b7b9fd88..2bc8494e 100644
> > --- a/io/copy_file_range.c
> > +++ b/io/copy_file_range.c
> > @@ -66,21 +66,13 @@ copy_src_filesize(int fd)
> > return st.st_size;
> > }
> >
> > -static int
> > -copy_dst_truncate(void)
> > -{
> > - int ret = ftruncate(file->fd, 0);
> > - if (ret < 0)
> > - perror("ftruncate");
> > - return ret;
> > -}
> > -
> > static int
> > copy_range_f(int argc, char **argv)
> > {
> > long long src = 0;
> > long long dst = 0;
>
> I'd like to change these to src_off & dst_off, just to help keep track
> of things.
that's better, suggest apply same change to copy_file_range_cmd():
-copy_file_range_cmd(int fd, long long *src, long long *dst, size_t len)
+copy_file_range_cmd(int fd, long long *src_off, long long *dst_off, size_t len)
{
loff_t ret;
do {
- ret = syscall(__NR_copy_file_range, fd, src, file->fd, dst,
+ ret = syscall(__NR_copy_file_range, fd, src_off, file->fd, dst_off,
len, 0);
if (ret == -1) {
perror("copy_range");
@@ -66,21 +66,13 @@ copy_src_filesize(int fd)
return st.st_size;
}
>
> > size_t len = 0;
> > + bool len_specified = false;
> > int opt;
> > int ret;
> > int fd;
> > @@ -112,6 +104,7 @@ copy_range_f(int argc, char **argv)
> > printf(_("invalid length -- %s\n"), optarg);
> > return 0;
> > }
> > + len_specified = true;
> > break;
> > case 'f':
> > src_file_nr = atoi(argv[1]);
> > @@ -137,7 +130,7 @@ copy_range_f(int argc, char **argv)
> > fd = filetable[src_file_nr].fd;
> > }
> >
> > - if (src == 0 && dst == 0 && len == 0) {
> > + if (! len_specified) {
>
> normal xfsprogs style is no space after the !
Good to know.
>
> > off64_t sz;
> >
> > sz = copy_src_filesize(fd);
> > @@ -145,13 +138,8 @@ copy_range_f(int argc, char **argv)
> > ret = 1;
> > goto out;
> > }
> > - len = sz;
> > -
> > - ret = copy_dst_truncate();
> > - if (ret < 0) {
> > - ret = 1;
> > - goto out;
> > - }
> > + if (sz > src)
> > + len = sz - src;
>
> Ok, so if source offset is past EOF, we keep len = 0.
>
> Looks like the kernel does that internally too, so no problem there.
>
> I'll make the cosmetic changes I mentioned above when I commit it, unless
> you have any concerns about that.
looks good to me. thanks Eric Darrick and Zorro.
>
> /* Shorten the copy to EOF */
> size_in = i_size_read(inode_in);
> if (pos_in >= size_in)
> count = 0;
>
> Reviewed-by: Eric Sandeen <sandeen@redhat.com>
>
> > }
> >
> > ret = copy_file_range_cmd(fd, &src, &dst, len);
> > diff --git a/man/man8/xfs_io.8 b/man/man8/xfs_io.8
> > index 6e064bdd..61c35c8e 100644
> > --- a/man/man8/xfs_io.8
> > +++ b/man/man8/xfs_io.8
> > @@ -669,13 +669,13 @@ The source must be specified either by path
> > or as another open file
> > .RB ( \-f ).
> > If
> > -.I src_file
> > -.IR src_offset ,
> > -.IR dst_offset ,
> > -and
> > .I length
> > -are omitted the contents of src_file will be copied to the beginning of
> > the
> > -open file, overwriting any data already there.
> > +is not specified, this command copies data from
> > +.I src_offset
> > +to the end of
> > +.BI src_file
> > +into the dst_file at
> > +.IR dst_offset .
> > .RS 1.0i
> > .PD 0
> > .TP 0.4i
> >
>
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2019-09-09 6:11 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-09-06 17:02 [PATCH v2] xfs_io: copy_range don't truncate dst_file, and add smart length Jianhong.Yin
2019-09-06 21:04 ` Eric Sandeen
2019-09-09 6:11 ` Jianhong Yin
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.