linux-xfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Eryu Guan <guan@eryu.me>
To: "Darrick J. Wong" <djwong@kernel.org>
Cc: guaneryu@gmail.com, linux-xfs@vger.kernel.org, fstests@vger.kernel.org
Subject: Re: [PATCH 7/8] iogen: upgrade to fallocate
Date: Sun, 16 Jan 2022 15:01:35 +0800	[thread overview]
Message-ID: <YePCz96eXSb66ppd@desktop> (raw)
In-Reply-To: <164193784690.3008286.8689130816813600863.stgit@magnolia>

On Tue, Jan 11, 2022 at 01:50:46PM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <djwong@kernel.org>
> 
> Update this utility to use fallocate to preallocate/reserve space to a
> file so that we're not so dependent on legacy XFS ioctls.
> 
> Signed-off-by: Darrick J. Wong <djwong@kernel.org>
> ---
>  ltp/iogen.c |   32 ++++++++++++++++++++++----------
>  1 file changed, 22 insertions(+), 10 deletions(-)
> 
> 
> diff --git a/ltp/iogen.c b/ltp/iogen.c
> index 2b6644d5..6c2c1534 100644
> --- a/ltp/iogen.c
> +++ b/ltp/iogen.c
> @@ -928,7 +928,15 @@ bozo!
>  		   fd, f.l_whence, (long long)f.l_start, (long long)f.l_len);*/
>  
>  	    /* non-zeroing reservation */
> -#ifdef XFS_IOC_RESVSP
> +#if defined(HAVE_FALLOCATE)
> +	    if (fallocate(fd, 0, 0, nbytes) == -1) {

Seems this is not a identical replacement for XFS_IOC_RESVSP (is that
what we want here?), as fallocate(2) here zeros space and change i_size
as well.

And from xfsctl(3), after XFS_IOC_RESVSP "The blocks are allocated, but
not zeroed, and the file size does not change." And the comments above
indicates "non-zeroing reservation" as well.

If identical replacement is not required, we could drop "non-zeroing"
part, but add FALLOC_FL_KEEP_SIZE?

Thanks,
Eryu

> +		fprintf(stderr,
> +			"iogen%s:  Could not fallocate %d bytes in file %s: %s (%d)\n",
> +			TagName, nbytes, path, SYSERR, errno);
> +		close(fd);
> +		return -1;
> +	    }
> +#elif defined(XFS_IOC_RESVSP)
>  	    if( xfsctl( path, fd, XFS_IOC_RESVSP, &f ) == -1) {
>  		fprintf(stderr,
>  			"iogen%s:  Could not xfsctl(XFS_IOC_RESVSP) %d bytes in file %s: %s (%d)\n",
> @@ -936,8 +944,7 @@ bozo!
>  		close(fd);
>  		return -1;
>  	    }
> -#else
> -#ifdef F_RESVSP
> +#elif defined(F_RESVSP)
>  	    if( fcntl( fd, F_RESVSP, &f ) == -1) {
>  		fprintf(stderr,
>  			"iogen%s:  Could not fcntl(F_RESVSP) %d bytes in file %s: %s (%d)\n",
> @@ -946,8 +953,7 @@ bozo!
>  		return -1;
>  	    }
>  #else
> -bozo!
> -#endif
> +# error Dont know how to reserve space!
>  #endif
>  	}
>  
> @@ -962,7 +968,15 @@ bozo!
>  		    (long long)f.l_len);*/
>  
>  	    /* zeroing reservation */
> -#ifdef XFS_IOC_ALLOCSP
> +#if defined(HAVE_FALLOCATE)
> +	    if (fallocate(fd, 0, sbuf.st_size, nbytes - sbuf.st_size) == -1) {
> +		fprintf(stderr,
> +			"iogen%s:  Could not fallocate %d bytes in file %s: %s (%d)\n",
> +			TagName, nbytes, path, SYSERR, errno);
> +		close(fd);
> +		return -1;
> +	    }
> +#elif defined(XFS_IOC_ALLOCSP)
>  	    if( xfsctl( path, fd, XFS_IOC_ALLOCSP, &f ) == -1) {
>  		fprintf(stderr,
>  			"iogen%s:  Could not xfsctl(XFS_IOC_ALLOCSP) %d bytes in file %s: %s (%d)\n",
> @@ -970,8 +984,7 @@ bozo!
>  		close(fd);
>  		return -1;
>  	    }
> -#else
> -#ifdef F_ALLOCSP
> +#elif defined(F_ALLOCSP)
>  	    if ( fcntl(fd, F_ALLOCSP, &f) < 0) {
>  		fprintf(stderr,
>  			"iogen%s:  Could not fcntl(F_ALLOCSP) %d bytes in file %s: %s (%d)\n",
> @@ -980,8 +993,7 @@ bozo!
>  		return -1;
>  	    }
>  #else
> -bozo!
> -#endif
> +# error Dont know how to (pre)allocate space!
>  #endif
>  	}
>  #endif

  reply	other threads:[~2022-01-16  7:01 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-01-11 21:50 [PATCHSET 0/8] fstests: random fixes Darrick J. Wong
2022-01-11 21:50 ` [PATCH 1/8] common/rc: fix _check_xfs_scrub_does_unicode on newer versions of libc-bin Darrick J. Wong
2022-01-11 21:50 ` [PATCH 2/8] common/rc: fix unicode checker detection in xfs_scrub Darrick J. Wong
2022-01-11 21:50 ` [PATCH 3/8] xfs/220: fix quotarm syscall test Darrick J. Wong
2022-01-12  1:14   ` xuyang2018.jy
2022-01-14  3:23     ` xuyang2018.jy
2022-01-14 17:08       ` Darrick J. Wong
2022-01-17  1:06         ` xuyang2018.jy
2022-01-11 21:50 ` [PATCH 4/8] xfs: test fixes for new 5.17 behaviors Darrick J. Wong
2022-01-11 21:50 ` [PATCH 5/8] xfs: regression test for allocsp handing out stale disk contents Darrick J. Wong
2022-01-16  7:12   ` Eryu Guan
2022-01-18  3:37   ` Zorro Lang
2022-01-18 18:09     ` Darrick J. Wong
2022-01-19  4:07       ` Zorro Lang
2022-01-19  4:52         ` Darrick J. Wong
2022-01-11 21:50 ` [PATCH 6/8] fsx: add support for XFS_IOC_ALLOCSP Darrick J. Wong
2022-01-11 21:50 ` [PATCH 7/8] iogen: upgrade to fallocate Darrick J. Wong
2022-01-16  7:01   ` Eryu Guan [this message]
2022-01-17 17:15     ` Darrick J. Wong
2022-01-11 21:50 ` [PATCH 8/8] alloc: " Darrick J. Wong
2022-01-16  7:16 ` [PATCHSET 0/8] fstests: random fixes Eryu Guan

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=YePCz96eXSb66ppd@desktop \
    --to=guan@eryu.me \
    --cc=djwong@kernel.org \
    --cc=fstests@vger.kernel.org \
    --cc=guaneryu@gmail.com \
    --cc=linux-xfs@vger.kernel.org \
    /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;
as well as URLs for NNTP newsgroup(s).