Re: [PATCH 7/7] fsstress: update for FIEXCHANGE_RANGE

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Tue, Feb 28, 2023 at 06:59:43PM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <djwong@xxxxxxxxxx>
> 
> Teach this stress tool to be able to use the file content exchange
> ioctl.
> 
> Signed-off-by: Darrick J. Wong <djwong@xxxxxxxxxx>
> ---
>  ltp/fsstress.c |  168 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 168 insertions(+)
> 
> 
> diff --git a/ltp/fsstress.c b/ltp/fsstress.c
> index 10608fb554..0fba3d92a0 100644
> --- a/ltp/fsstress.c
> +++ b/ltp/fsstress.c
> @@ -143,6 +143,7 @@ typedef enum {
>  	OP_URING_WRITE,
>  	OP_WRITE,
>  	OP_WRITEV,
> +	OP_XCHGRANGE,
>  	OP_LAST
>  } opty_t;
>  
> @@ -272,6 +273,8 @@ void	uring_read_f(opnum_t, long);
>  void	uring_write_f(opnum_t, long);
>  void	write_f(opnum_t, long);
>  void	writev_f(opnum_t, long);
> +void	xchgrange_f(opnum_t, long);
> +
>  char	*xattr_flag_to_string(int);
>  
>  struct opdesc	ops[OP_LAST]	= {
> @@ -340,6 +343,7 @@ struct opdesc	ops[OP_LAST]	= {
>  	[OP_URING_WRITE]   = {"uring_write",   uring_write_f,	1, 1 },
>  	[OP_WRITE]	   = {"write",	       write_f,		4, 1 },
>  	[OP_WRITEV]	   = {"writev",	       writev_f,	4, 1 },
> +	[OP_XCHGRANGE]	   = {"xchgrange",     xchgrange_f,	4, 1 },

Do you think this is a common operation which should use same frequency (4)
with read/write operations? I'd like to reduce the default freq=4 to 2 or 1
when I merge it. what do you think?

Thanks,
Zorro

>  }, *ops_end;
>  
>  flist_t	flist[FT_nft] = {
> @@ -2494,6 +2498,170 @@ chown_f(opnum_t opno, long r)
>  	free_pathname(&f);
>  }
>  
> +/* exchange some arbitrary range of f1 to f2...fn. */
> +void
> +xchgrange_f(
> +	opnum_t			opno,
> +	long			r)
> +{
> +#ifdef FIEXCHANGE_RANGE
> +	struct file_xchg_range	fxr = { 0 };
> +	static __u64		swap_flags = 0;
> +	struct pathname		fpath1;
> +	struct pathname		fpath2;
> +	struct stat64		stat1;
> +	struct stat64		stat2;
> +	char			inoinfo1[1024];
> +	char			inoinfo2[1024];
> +	off64_t			lr;
> +	off64_t			off1;
> +	off64_t			off2;
> +	off64_t			max_off2;
> +	size_t			len;
> +	int			v1;
> +	int			v2;
> +	int			fd1;
> +	int			fd2;
> +	int			ret;
> +	int			tries = 0;
> +	int			e;
> +
> +	/* Load paths */
> +	init_pathname(&fpath1);
> +	if (!get_fname(FT_REGm, r, &fpath1, NULL, NULL, &v1)) {
> +		if (v1)
> +			printf("%d/%lld: xchgrange read - no filename\n",
> +				procid, opno);
> +		goto out_fpath1;
> +	}
> +
> +	init_pathname(&fpath2);
> +	if (!get_fname(FT_REGm, random(), &fpath2, NULL, NULL, &v2)) {
> +		if (v2)
> +			printf("%d/%lld: xchgrange write - no filename\n",
> +				procid, opno);
> +		goto out_fpath2;
> +	}
> +
> +	/* Open files */
> +	fd1 = open_path(&fpath1, O_RDONLY);
> +	e = fd1 < 0 ? errno : 0;
> +	check_cwd();
> +	if (fd1 < 0) {
> +		if (v1)
> +			printf("%d/%lld: xchgrange read - open %s failed %d\n",
> +				procid, opno, fpath1.path, e);
> +		goto out_fpath2;
> +	}
> +
> +	fd2 = open_path(&fpath2, O_WRONLY);
> +	e = fd2 < 0 ? errno : 0;
> +	check_cwd();
> +	if (fd2 < 0) {
> +		if (v2)
> +			printf("%d/%lld: xchgrange write - open %s failed %d\n",
> +				procid, opno, fpath2.path, e);
> +		goto out_fd1;
> +	}
> +
> +	/* Get file stats */
> +	if (fstat64(fd1, &stat1) < 0) {
> +		if (v1)
> +			printf("%d/%lld: xchgrange read - fstat64 %s failed %d\n",
> +				procid, opno, fpath1.path, errno);
> +		goto out_fd2;
> +	}
> +	inode_info(inoinfo1, sizeof(inoinfo1), &stat1, v1);
> +
> +	if (fstat64(fd2, &stat2) < 0) {
> +		if (v2)
> +			printf("%d/%lld: xchgrange write - fstat64 %s failed %d\n",
> +				procid, opno, fpath2.path, errno);
> +		goto out_fd2;
> +	}
> +	inode_info(inoinfo2, sizeof(inoinfo2), &stat2, v2);
> +
> +	if (stat1.st_size < (stat1.st_blksize * 2) ||
> +	    stat2.st_size < (stat2.st_blksize * 2)) {
> +		if (v2)
> +			printf("%d/%lld: xchgrange - files are too small\n",
> +				procid, opno);
> +		goto out_fd2;
> +	}
> +
> +	/* Never let us swap more than 1/4 of the files. */
> +	len = (random() % FILELEN_MAX) + 1;
> +	if (len > stat1.st_size / 4)
> +		len = stat1.st_size / 4;
> +	if (len > stat2.st_size / 4)
> +		len = stat2.st_size / 4;
> +	len = rounddown_64(len, stat1.st_blksize);
> +	if (len == 0)
> +		len = stat1.st_blksize;
> +
> +	/* Calculate offsets */
> +	lr = ((int64_t)random() << 32) + random();
> +	if (stat1.st_size == len)
> +		off1 = 0;
> +	else
> +		off1 = (off64_t)(lr % MIN(stat1.st_size - len, MAXFSIZE));
> +	off1 %= maxfsize;
> +	off1 = rounddown_64(off1, stat1.st_blksize);
> +
> +	/*
> +	 * If srcfile == destfile, randomly generate destination ranges
> +	 * until we find one that doesn't overlap the source range.
> +	 */
> +	max_off2 = MIN(stat2.st_size  - len, MAXFSIZE);
> +	do {
> +		lr = ((int64_t)random() << 32) + random();
> +		if (stat2.st_size == len)
> +			off2 = 0;
> +		else
> +			off2 = (off64_t)(lr % max_off2);
> +		off2 %= maxfsize;
> +		off2 = rounddown_64(off2, stat2.st_blksize);
> +	} while (stat1.st_ino == stat2.st_ino &&
> +		 llabs(off2 - off1) < len &&
> +		 tries++ < 10);
> +
> +	/* Swap data blocks */
> +	fxr.file1_fd = fd1;
> +	fxr.file1_offset = off1;
> +	fxr.length = len;
> +	fxr.file2_offset = off2;
> +	fxr.flags = swap_flags;
> +
> +retry:
> +	ret = ioctl(fd2, FIEXCHANGE_RANGE, &fxr);
> +	e = ret < 0 ? errno : 0;
> +	if (e == EOPNOTSUPP && !(swap_flags & FILE_XCHG_RANGE_NONATOMIC)) {
> +		swap_flags = FILE_XCHG_RANGE_NONATOMIC;
> +		fxr.flags |= swap_flags;
> +		goto retry;
> +	}
> +	if (v1 || v2) {
> +		printf("%d/%lld: xchgrange %s%s [%lld,%lld] -> %s%s [%lld,%lld]",
> +			procid, opno,
> +			fpath1.path, inoinfo1, (long long)off1, (long long)len,
> +			fpath2.path, inoinfo2, (long long)off2, (long long)len);
> +
> +		if (ret < 0)
> +			printf(" error %d", e);
> +		printf("\n");
> +	}
> +
> +out_fd2:
> +	close(fd2);
> +out_fd1:
> +	close(fd1);
> +out_fpath2:
> +	free_pathname(&fpath2);
> +out_fpath1:
> +	free_pathname(&fpath1);
> +#endif
> +}
> +
>  /* reflink some arbitrary range of f1 to f2. */
>  void
>  clonerange_f(
> 




[Index of Archives]     [XFS Filesystem Development (older mail)]     [Linux Filesystem Development]     [Linux Audio Users]     [Yosemite Trails]     [Linux Kernel]     [Linux RAID]     [Linux SCSI]


  Powered by Linux