Document the new XFS_IOC_GETFSMAP ioctl that returns the physical layout of a (disk-based) filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@xxxxxxxxxx> --- man2/ioctl_xfs_ioc_getfsmap.2 | 294 +++++++++++++++++++++++++++++++++++++++++ 1 file changed, 294 insertions(+) create mode 100644 man2/ioctl_xfs_ioc_getfsmap.2 diff --git a/man2/ioctl_xfs_ioc_getfsmap.2 b/man2/ioctl_xfs_ioc_getfsmap.2 new file mode 100644 index 0000000..0d9ed47 --- /dev/null +++ b/man2/ioctl_xfs_ioc_getfsmap.2 @@ -0,0 +1,294 @@ +.\" Copyright (c) 2016, Oracle. All rights reserved. +.\" +.\" %%%LICENSE_START(GPLv2+_DOC_FULL) +.\" This is free documentation; you can redistribute it and/or +.\" modify it under the terms of the GNU General Public License as +.\" published by the Free Software Foundation; either version 2 of +.\" the License, or (at your option) any later version. +.\" +.\" The GNU General Public License's references to "object code" +.\" and "executables" are to be interpreted as the output of any +.\" document formatting or typesetting system, including +.\" intermediate and printed output. +.\" +.\" This manual is distributed in the hope that it will be useful, +.\" but WITHOUT ANY WARRANTY; without even the implied warranty of +.\" MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +.\" GNU General Public License for more details. +.\" +.\" You should have received a copy of the GNU General Public +.\" License along with this manual; if not, see +.\" <http://www.gnu.org/licenses/>. +.\" %%%LICENSE_END +.TH IOCTL-XFS_IOC_GETFSMAP 2 2016-07-20 "Linux" "Linux Programmer's Manual" +.SH NAME +ioctl_xfs_ioc_getfsmap \- retrieve the physical layout of the filesystem +.SH SYNOPSIS +.br +.B #include <sys/ioctl.h> +.br +.B #include <linux/fs.h> +.sp +.BI "int ioctl(int " fd ", XFS_IOC_GETFSMAP, struct getfsmap * " arg ); +.SH DESCRIPTION +This +.BR ioctl (2) +retrieves physical extent mappings for a filesystem. +This information can be used to discover which files are mapped to a physical +block, examine free space, or find known bad blocks, among other things. + +The sole argument to this ioctl should be an array of the following +structure: +.in +4n +.nf + +struct getfsmap { + __u32 fmv_device; /* device id */ + __u32 fmv_unused1; /* future use, must be zero */ + __u64 fmv_block; /* starting block */ + __u64 fmv_owner; /* owner id */ + __u64 fmv_offset; /* file offset of segment */ + __u64 fmv_length; /* length of segment, blocks */ + __u32 fmv_oflags; /* mapping flags */ + __u32 fmv_iflags; /* control flags (1st structure) */ + __u32 fmv_count; /* # of entries in array incl. input */ + __u32 fmv_entries; /* # of entries filled in (output). */ + __u64 fmv_unused2; /* future use, must be zero */ +}; + +.fi +.in +The array must contain at least two elements. +The first two array elements specify the lowest and highest reverse-mapping +keys, respectively, for which userspace would like physical mapping +information. +A reverse mapping key consists of the tuple (device, block, owner, offset). +The owner and offset fields are part of the key because some filesystems +support sharing physical blocks between multiple files and +therefore may return multiple mappings for a given physical block. + +.SS Fields of struct getfsmap +.PP +The +.I fmv_device +field contains a 32-bit cookie to uniquely identify the underlying storage +device. +If the +.B FMV_HOF_DEV_T +flag is set in the header's +.I fmv_oflags +field, this field contains a dev_t from which major and minor numbers can +be extracted. +If the flag is not set, this field contains a value that must be unique +for each storage device. + +.PP +The +.I fmv_unused1 +field must be zero in the first two array elements. + +.PP +The +.I fmv_block +field contains the 512-byte sector address of the extent. + +.PP +The +.I fmv_owner +field contains the owner of the extent. +This is generally an inode number, though if +.B FMV_OF_SPECIAL_OWNER +is set in the +.I fmv_oflags +field, then the owner value is one of the following special values: +.TP +.B FMV_OWN_FREE +Free space. +.TP +.B FMV_OWN_UNKNOWN +This extent has an unknown owner. +.TP +.B FMV_OWN_FS +Static filesystem metadata. +.TP +.B FMV_OWN_LOG +The filesystem journal. +.TP +.B FMV_OWN_AG +Allocation group metadata. +.TP +.B FMV_OWN_INOBT +The inode index, if one is provided. +.TP +.B FMV_OWN_INODES +Inodes. +.TP +.B FMV_OWN_REFC +Reference counting indexes. +.TP +.B FMV_OWN_COW +This extent is being used to stage a copy-on-write. +.TP +.B FMV_OWN_DEFECTIVE: +This extent has been marked defective either by the filesystem or the +underlying device. + +.PP +The +.I fmv_offset +field contains the logical address of the reverse mapping record, in units +of 512-byte blocks. +This field has no meaning if the +.BR FMV_OF_SPECIAL_OWNER " or " FMV_OF_EXTENT_MAP +flags are set in +.IR fmv_oflags "." + +.PP +The +.I fmv_length +field contains the length of the extent, in units of 512-byte blocks. +This field must be zero in the second array element. + +.PP +The +.I fmv_oflags +field is a bitmask of extent state flags. +In the header, the bits are: +.TP +.B FMV_HOF_DEV_T +All +.I fmv_device +values will be in dev_t format. +If this flag is not set, the value is merely a 32-bit cookie that will be +unique for each physical device. +.TP +In a non-header, the bits are: +.TP +.B FMV_OF_PREALLOC +The extent is allocated but not yet written. +.TP +.B FMV_OF_ATTR_FORK +This extent contains extended attribute data. +.TP +.B FMV_OF_EXTENT_MAP +This extent contains extent map information for the owner. +.TP +.B FMV_OF_SHARED +Parts of this extent may be shared. +.TP +.B FMV_OF_SPECIAL_OWNER +The +.I fmv_owner +field contains a special value instead of an inode number. +.TP +.B FMV_OF_LAST +This is the last record in the filesystem. + +.PP +The +.I fmv_iflags +field is a bitmask passed to the kernel to alter the output. +There are no flags defined, so this value must be zero in the first +two array elements. + +.PP +The +.I fmv_count +field contains the number of elements in the array being passed to the +kernel. +This count must include the two control elements at the start of the +array. +The value must be specified in the first array element; in the second +element this field must be zero. + +If this value is 2, +.I fmv_entries +will be set to the number of records that would have been returned had +the array been large enough; +no extent information will be returned. + +.PP +The +.I fmv_entries +field contains the number of elements in the array that contain useful +information if the ioctl returns a non-error value. +This value does not include the two control elements at the start of the array. +This value is only set in the first array element; +in the second element, this field must be zero. + +.PP +The +.I fmv_unused2 +field must be zero in the first two array elements. + +.SS Array Elements +.PP +The key fields (fmv_device, fmv_block, fmv_owner, fmv_offset) of the first +element of the array specify the lowest extent record in the keyspace that +the caller wants returned. +For example, if the key is set to (0, 36, 0, 0), the filesystem will +only return records for extents starting at or above sector 36 on +disk. +For convenience, the +.I fmv_length +field will be added to the +.IR fmv_block " and " fmv_offset +fields as appropriate so that the (fmv_device, fmv_block, fmv_owner, +fmv_offset, fmv_length) fields in the last array element can be copied +into the first element to seed the next ioctl call. + +The key fields of the second element of the array specify the highest +extent record in the keyspace that the caller wants returned. +Returning to our example above, if that example key were instead +passed in via the second array element, the filesystem will not return +records for extents going past sector 36 on disk. +For convenience, the four key fields can be set to ~0 (all ones) to +signify "end of filesystem". + +If +.I fmv_count +in the first element of the array is 2, then +.I fmv_entries +in the first element of the array will be set to the number of extent +records found in the filesystem. +Otherwise, +.I fmv_entries +will be set to the number of extents actually returned, and the subsequent +array elements will be filled out with extent information. +In these +subsequent array elements, the fields +.IR fmv_iflags ", " fmv_count ", " fmv_entries ", and " fmv_unused1 +will be set to zero by the filesystem. + +.SH RETURN VALUE +On error, \-1 is returned, and +.I errno +is set to indicate the error. +.PP +.SH ERRORS +Error codes can be one of, but are not limited to, the following: +.TP +.B EINVAL +The array is not long enough, or a non-zero value was passed in one of the +fields that must be zero. +.TP +.B EFAULT +The pointer passed in was not mapped to a valid memory address. +.TP +.B EBADF +.IR fd +is not open for reading. +.TP +.B EPERM +This query is not allowed. +.TP +.B EOPNOTSUPP +The filesystem does not support this command. + +.SH CONFORMING TO +This API is Linux-specific. +Not all filesystems support it. +.fi +.in +.SH SEE ALSO +.BR ioctl (2) -- To unsubscribe from this list: send the line "unsubscribe linux-api" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html