iv/linux
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
e7720afad0
linux/fs/xfs/xfs_fsmap.c

976 lines
26 KiB
C
Raw Normal View History

xfs: convert to SPDX license tags Remove the verbose license text from XFS files and replace them with SPDX tags. This does not change the license of any of the code, merely refers to the common, up-to-date license files in LICENSES/ This change was mostly scripted. fs/xfs/Makefile and fs/xfs/libxfs/xfs_fs.h were modified by hand, the rest were detected and modified by the following command: for f in `git grep -l "GNU General" fs/xfs/` ; do echo $f cat $f | awk -f hdr.awk > $f.new mv -f $f.new $f done And the hdr.awk script that did the modification (including detecting the difference between GPL-2.0 and GPL-2.0+ licenses) is as follows: $ cat hdr.awk BEGIN { hdr = 1.0 tag = "GPL-2.0" str = "" } /^ \* This program is free software/ { hdr = 2.0; next } /any later version./ { tag = "GPL-2.0+" next } /^ \*\// { if (hdr > 0.0) { print "// SPDX-License-Identifier: " tag print str print $0 str="" hdr = 0.0 next } print $0 next } /^ \* / { if (hdr > 1.0) next if (hdr > 0.0) { if (str != "") str = str "\n" str = str $0 next } print $0 next } /^ \*/ { if (hdr > 0.0) next print $0 next } // { if (hdr > 0.0) { if (str != "") str = str "\n" str = str $0 next } print $0 } END { } $ Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2018-06-06 05:42:14 +03:00
// SPDX-License-Identifier: GPL-2.0+
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
/*
* Copyright (C) 2017 Oracle. All Rights Reserved.
* Author: Darrick J. Wong <darrick.wong@oracle.com>
*/
#include "xfs.h"
#include "xfs_fs.h"
#include "xfs_shared.h"
#include "xfs_format.h"
#include "xfs_log_format.h"
#include "xfs_trans_resv.h"
#include "xfs_mount.h"
#include "xfs_inode.h"
#include "xfs_trans.h"
#include "xfs_btree.h"
#include "xfs_rmap_btree.h"
#include "xfs_trace.h"
#include "xfs_rmap.h"
#include "xfs_alloc.h"
#include "xfs_bit.h"
#include <linux/fsmap.h>
#include "xfs_fsmap.h"
#include "xfs_refcount.h"
#include "xfs_refcount_btree.h"
xfs: have getfsmap fall back to the freesp btrees when rmap is not present If the reverse-mapping btree isn't available, fall back to the free space btrees to provide partial reverse mapping information. The online scrub tool can make use of even partial information to speed up the data block scan. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
#include "xfs_alloc_btree.h"
xfs: report realtime space information via the rtbitmap Use the realtime bitmap to return free space information via getfsmap. Eventually this will be superseded by the realtime rmapbt code. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:38 +03:00
#include "xfs_rtalloc.h"
xfs: pass perags around in fsmap data dev functions Needs a [from, to] ranged AG walk, and the perag to be stuffed into the info structure for callouts to use. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2021-06-02 03:48:24 +03:00
#include "xfs_ag.h"
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
/* Convert an xfs_fsmap to an fsmap. */
xfs: fix deadlock and streamline xfs_getfsmap performance Refactor xfs_getfsmap to improve its performance: instead of indirectly calling a function that copies one record to userspace at a time, create a shadow buffer in the kernel and copy the whole array once at the end. On the author's computer, this reduces the runtime on his /home by ~20%. This also eliminates a deadlock when running GETFSMAP against the realtime device. The current code locks the rtbitmap to create fsmappings and copies them into userspace, having not released the rtbitmap lock. If the userspace buffer is an mmap of a sparse file that itself resides on the realtime device, the write page fault will recurse into the fs for allocation, which will deadlock on the rtbitmap lock. Fixes: 4c934c7dd60c ("xfs: report realtime space information via the rtbitmap") Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2020-10-01 20:56:07 +03:00
static void
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
xfs_fsmap_from_internal(
struct fsmap *dest,
struct xfs_fsmap *src)
{
dest->fmr_device = src->fmr_device;
dest->fmr_flags = src->fmr_flags;
dest->fmr_physical = BBTOB(src->fmr_physical);
dest->fmr_owner = src->fmr_owner;
dest->fmr_offset = BBTOB(src->fmr_offset);
dest->fmr_length = BBTOB(src->fmr_length);
dest->fmr_reserved[0] = 0;
dest->fmr_reserved[1] = 0;
dest->fmr_reserved[2] = 0;
}
/* Convert an fsmap to an xfs_fsmap. */
void
xfs_fsmap_to_internal(
struct xfs_fsmap *dest,
struct fsmap *src)
{
dest->fmr_device = src->fmr_device;
dest->fmr_flags = src->fmr_flags;
dest->fmr_physical = BTOBBT(src->fmr_physical);
dest->fmr_owner = src->fmr_owner;
dest->fmr_offset = BTOBBT(src->fmr_offset);
dest->fmr_length = BTOBBT(src->fmr_length);
}
/* Convert an fsmap owner into an rmapbt owner. */
static int
xfs_fsmap_owner_to_rmap(
struct xfs_rmap_irec *dest,
xfs: make fsmap backend function key parameters const There are several GETFSMAP backend functions for XFS to cover the three devices and various feature support. Each of these functions are passed pointers to the low and high keys for the dataset that userspace requested, and a pointer to scratchpad variables that are used to control the iteration and fill out records. The scratchpad data can be changed arbitrarily, but the keys are supposed to remain unchanged (and under the control of the outermost loop in xfs_getfsmap). Unfortunately, the data and rt backends modify the keys that are passed in from the main control loop, which causes subsequent calls to return incorrect query results. Specifically, each of those two functions set the block number in the high key to the size of their respective device. Since fsmap results are sorted in device number order, if the lower numbered device is smaller than the higher numbered device, the first function will set the high key to the small size, and the key remains unchanged as it is passed into the function for the higher numbered device. The second function will then fail to return all of the results for the dataset that userspace is asking for because the keyspace is incorrectly constrained. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2021-08-11 03:00:31 +03:00
const struct xfs_fsmap *src)
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
{
if (!(src->fmr_flags & FMR_OF_SPECIAL_OWNER)) {
dest->rm_owner = src->fmr_owner;
return 0;
}
switch (src->fmr_owner) {
case 0: /* "lowest owner id possible" */
case -1ULL: /* "highest owner id possible" */
dest->rm_owner = 0;
break;
case XFS_FMR_OWN_FREE:
dest->rm_owner = XFS_RMAP_OWN_NULL;
break;
case XFS_FMR_OWN_UNKNOWN:
dest->rm_owner = XFS_RMAP_OWN_UNKNOWN;
break;
case XFS_FMR_OWN_FS:
dest->rm_owner = XFS_RMAP_OWN_FS;
break;
case XFS_FMR_OWN_LOG:
dest->rm_owner = XFS_RMAP_OWN_LOG;
break;
case XFS_FMR_OWN_AG:
dest->rm_owner = XFS_RMAP_OWN_AG;
break;
case XFS_FMR_OWN_INOBT:
dest->rm_owner = XFS_RMAP_OWN_INOBT;
break;
case XFS_FMR_OWN_INODES:
dest->rm_owner = XFS_RMAP_OWN_INODES;
break;
case XFS_FMR_OWN_REFC:
dest->rm_owner = XFS_RMAP_OWN_REFC;
break;
case XFS_FMR_OWN_COW:
dest->rm_owner = XFS_RMAP_OWN_COW;
break;
case XFS_FMR_OWN_DEFECTIVE: /* not implemented */
/* fall through */
default:
return -EINVAL;
}
return 0;
}
/* Convert an rmapbt owner into an fsmap owner. */
static int
xfs_fsmap_owner_from_rmap(
xfs: make the record pointer passed to query_range functions const The query_range functions are supposed to call a caller-supplied function on each record found in the dataset. These functions don't own the memory storing the record, so don't let them change the record. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de>
2021-08-11 03:02:16 +03:00
struct xfs_fsmap *dest,
const struct xfs_rmap_irec *src)
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
{
dest->fmr_flags = 0;
if (!XFS_RMAP_NON_INODE_OWNER(src->rm_owner)) {
dest->fmr_owner = src->rm_owner;
return 0;
}
dest->fmr_flags |= FMR_OF_SPECIAL_OWNER;
switch (src->rm_owner) {
case XFS_RMAP_OWN_FS:
dest->fmr_owner = XFS_FMR_OWN_FS;
break;
case XFS_RMAP_OWN_LOG:
dest->fmr_owner = XFS_FMR_OWN_LOG;
break;
case XFS_RMAP_OWN_AG:
dest->fmr_owner = XFS_FMR_OWN_AG;
break;
case XFS_RMAP_OWN_INOBT:
dest->fmr_owner = XFS_FMR_OWN_INOBT;
break;
case XFS_RMAP_OWN_INODES:
dest->fmr_owner = XFS_FMR_OWN_INODES;
break;
case XFS_RMAP_OWN_REFC:
dest->fmr_owner = XFS_FMR_OWN_REFC;
break;
case XFS_RMAP_OWN_COW:
dest->fmr_owner = XFS_FMR_OWN_COW;
break;
xfs: have getfsmap fall back to the freesp btrees when rmap is not present If the reverse-mapping btree isn't available, fall back to the free space btrees to provide partial reverse mapping information. The online scrub tool can make use of even partial information to speed up the data block scan. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
case XFS_RMAP_OWN_NULL: /* "free" */
dest->fmr_owner = XFS_FMR_OWN_FREE;
break;
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
default:
xfs: add missing assert in xfs_fsmap_owner_from_rmap The fsmap handler shouldn't fail silently if the rmap code ever feeds it a special owner number that isn't known to the fsmap handler. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
2019-11-02 19:41:18 +03:00
ASSERT(0);
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
return -EFSCORRUPTED;
}
return 0;
}
/* getfsmap query state */
struct xfs_getfsmap_info {
struct xfs_fsmap_head *head;
xfs: fix deadlock and streamline xfs_getfsmap performance Refactor xfs_getfsmap to improve its performance: instead of indirectly calling a function that copies one record to userspace at a time, create a shadow buffer in the kernel and copy the whole array once at the end. On the author's computer, this reduces the runtime on his /home by ~20%. This also eliminates a deadlock when running GETFSMAP against the realtime device. The current code locks the rtbitmap to create fsmappings and copies them into userspace, having not released the rtbitmap lock. If the userspace buffer is an mmap of a sparse file that itself resides on the realtime device, the write page fault will recurse into the fs for allocation, which will deadlock on the rtbitmap lock. Fixes: 4c934c7dd60c ("xfs: report realtime space information via the rtbitmap") Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2020-10-01 20:56:07 +03:00
struct fsmap *fsmap_recs; /* mapping records */
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
struct xfs_buf *agf_bp; /* AGF, for refcount queries */
xfs: pass perags around in fsmap data dev functions Needs a [from, to] ranged AG walk, and the perag to be stuffed into the info structure for callouts to use. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2021-06-02 03:48:24 +03:00
struct xfs_perag *pag; /* AG info, if applicable */
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
xfs_daddr_t next_daddr; /* next daddr we expect */
u64 missing_owner; /* owner of holes */
u32 dev; /* device id */
struct xfs_rmap_irec low; /* low rmap key */
struct xfs_rmap_irec high; /* high rmap key */
bool last; /* last extent? */
};
/* Associate a device with a getfsmap handler. */
struct xfs_getfsmap_dev {
u32 dev;
int (*fn)(struct xfs_trans *tp,
xfs: make fsmap backend function key parameters const There are several GETFSMAP backend functions for XFS to cover the three devices and various feature support. Each of these functions are passed pointers to the low and high keys for the dataset that userspace requested, and a pointer to scratchpad variables that are used to control the iteration and fill out records. The scratchpad data can be changed arbitrarily, but the keys are supposed to remain unchanged (and under the control of the outermost loop in xfs_getfsmap). Unfortunately, the data and rt backends modify the keys that are passed in from the main control loop, which causes subsequent calls to return incorrect query results. Specifically, each of those two functions set the block number in the high key to the size of their respective device. Since fsmap results are sorted in device number order, if the lower numbered device is smaller than the higher numbered device, the first function will set the high key to the small size, and the key remains unchanged as it is passed into the function for the higher numbered device. The second function will then fail to return all of the results for the dataset that userspace is asking for because the keyspace is incorrectly constrained. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2021-08-11 03:00:31 +03:00
const struct xfs_fsmap *keys,
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
struct xfs_getfsmap_info *info);
};
/* Compare two getfsmap device handlers. */
static int
xfs_getfsmap_dev_compare(
const void *p1,
const void *p2)
{
const struct xfs_getfsmap_dev *d1 = p1;
const struct xfs_getfsmap_dev *d2 = p2;
return d1->dev - d2->dev;
}
/* Decide if this mapping is shared. */
STATIC int
xfs_getfsmap_is_shared(
struct xfs_trans *tp,
struct xfs_getfsmap_info *info,
xfs: make the record pointer passed to query_range functions const The query_range functions are supposed to call a caller-supplied function on each record found in the dataset. These functions don't own the memory storing the record, so don't let them change the record. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de>
2021-08-11 03:02:16 +03:00
const struct xfs_rmap_irec *rec,
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
bool *stat)
{
struct xfs_mount *mp = tp->t_mountp;
struct xfs_btree_cur *cur;
xfs_agblock_t fbno;
xfs_extlen_t flen;
int error;
*stat = false;
xfs: replace xfs_sb_version checks with feature flag checks Convert the xfs_sb_version_hasfoo() to checks against mp->m_features. Checks of the superblock itself during disk operations (e.g. in the read/write verifiers and the to/from disk formatters) are not converted - they operate purely on the superblock state. Everything else should use the mount features. Large parts of this conversion were done with sed with commands like this: for f in `git grep -l xfs_sb_version_has fs/xfs/*.c`; do sed -i -e 's/xfs_sb_version_has\(.*\)(&\(.*\)->m_sb)/xfs_has_\1(\2)/' $f done With manual cleanups for things like "xfs_has_extflgbit" and other little inconsistencies in naming. The result is ia lot less typing to check features and an XFS binary size reduced by a bit over 3kB: $ size -t fs/xfs/built-in.a text data bss dec hex filenam before 1130866 311352 484 1442702 16038e (TOTALS) after 1127727 311352 484 1439563 15f74b (TOTALS) Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Darrick J. Wong <djwong@kernel.org>
2021-08-19 04:46:37 +03:00
if (!xfs_has_reflink(mp))
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
return 0;
xfs: pass perags around in fsmap data dev functions Needs a [from, to] ranged AG walk, and the perag to be stuffed into the info structure for callouts to use. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2021-06-02 03:48:24 +03:00
/* rt files will have no perag structure */
if (!info->pag)
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
return 0;
/* Are there any shared blocks here? */
flen = 0;
xfs: convert refcount btree cursor to use perags Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2021-06-02 03:48:24 +03:00
cur = xfs_refcountbt_init_cursor(mp, tp, info->agf_bp, info->pag);
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
error = xfs_refcount_find_shared(cur, rec->rm_startblock,
rec->rm_blockcount, &fbno, &flen, false);
xfs: trivial xfs_btree_del_cursor cleanups The error argument to xfs_btree_del_cursor already understands the "nonzero for error" semantics, so remove pointless error testing in the callers and pass it directly. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
2018-07-19 22:26:31 +03:00
xfs_btree_del_cursor(cur, error);
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
if (error)
return error;
*stat = flen > 0;
return 0;
}
xfs: fix deadlock and streamline xfs_getfsmap performance Refactor xfs_getfsmap to improve its performance: instead of indirectly calling a function that copies one record to userspace at a time, create a shadow buffer in the kernel and copy the whole array once at the end. On the author's computer, this reduces the runtime on his /home by ~20%. This also eliminates a deadlock when running GETFSMAP against the realtime device. The current code locks the rtbitmap to create fsmappings and copies them into userspace, having not released the rtbitmap lock. If the userspace buffer is an mmap of a sparse file that itself resides on the realtime device, the write page fault will recurse into the fs for allocation, which will deadlock on the rtbitmap lock. Fixes: 4c934c7dd60c ("xfs: report realtime space information via the rtbitmap") Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2020-10-01 20:56:07 +03:00
static inline void
xfs_getfsmap_format(
struct xfs_mount *mp,
struct xfs_fsmap *xfm,
struct xfs_getfsmap_info *info)
{
struct fsmap *rec;
trace_xfs_getfsmap_mapping(mp, xfm);
rec = &info->fsmap_recs[info->head->fmh_entries++];
xfs_fsmap_from_internal(rec, xfm);
}
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
/*
* Format a reverse mapping for getfsmap, having translated rm_startblock
* into the appropriate daddr units.
*/
STATIC int
xfs_getfsmap_helper(
struct xfs_trans *tp,
struct xfs_getfsmap_info *info,
xfs: make the record pointer passed to query_range functions const The query_range functions are supposed to call a caller-supplied function on each record found in the dataset. These functions don't own the memory storing the record, so don't let them change the record. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de>
2021-08-11 03:02:16 +03:00
const struct xfs_rmap_irec *rec,
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
xfs_daddr_t rec_daddr)
{
struct xfs_fsmap fmr;
struct xfs_mount *mp = tp->t_mountp;
bool shared;
int error;
if (fatal_signal_pending(current))
return -EINTR;
/*
* Filter out records that start before our startpoint, if the
* caller requested that.
*/
if (xfs_rmap_compare(rec, &info->low) < 0) {
rec_daddr += XFS_FSB_TO_BB(mp, rec->rm_blockcount);
if (info->next_daddr < rec_daddr)
info->next_daddr = rec_daddr;
xfs: remove all *_ITER_CONTINUE values Iterator functions already use 0 to signal "continue iterating", so get rid of the #defines and just do it directly. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-08-29 00:39:46 +03:00
return 0;
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
}
/* Are we just counting mappings? */
if (info->head->fmh_count == 0) {
xfs: limit entries returned when counting fsmap records If userspace asked fsmap to count the number of entries, we cannot return more than UINT_MAX entries because fmh_entries is u32. Therefore, stop counting if we hit this limit or else we will waste time to return truncated results. Fixes: e89c041338ed ("xfs: implement the GETFSMAP ioctl") Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2020-10-01 20:56:07 +03:00
if (info->head->fmh_entries == UINT_MAX)
return -ECANCELED;
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
if (rec_daddr > info->next_daddr)
info->head->fmh_entries++;
if (info->last)
xfs: remove all *_ITER_CONTINUE values Iterator functions already use 0 to signal "continue iterating", so get rid of the #defines and just do it directly. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-08-29 00:39:46 +03:00
return 0;
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
info->head->fmh_entries++;
rec_daddr += XFS_FSB_TO_BB(mp, rec->rm_blockcount);
if (info->next_daddr < rec_daddr)
info->next_daddr = rec_daddr;
xfs: remove all *_ITER_CONTINUE values Iterator functions already use 0 to signal "continue iterating", so get rid of the #defines and just do it directly. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-08-29 00:39:46 +03:00
return 0;
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
}
/*
* If the record starts past the last physical block we saw,
* then we've found a gap. Report the gap as being owned by
* whatever the caller specified is the missing owner.
*/
if (rec_daddr > info->next_daddr) {
if (info->head->fmh_entries >= info->head->fmh_count)
xfs: remove all *_ITER_ABORT values Use -ECANCELED to signal "stop iterating" instead of these magical *_ITER_ABORT values, since it's duplicative. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-08-29 00:37:57 +03:00
return -ECANCELED;
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
fmr.fmr_device = info->dev;
fmr.fmr_physical = info->next_daddr;
fmr.fmr_owner = info->missing_owner;
fmr.fmr_offset = 0;
fmr.fmr_length = rec_daddr - info->next_daddr;
fmr.fmr_flags = FMR_OF_SPECIAL_OWNER;
xfs: fix deadlock and streamline xfs_getfsmap performance Refactor xfs_getfsmap to improve its performance: instead of indirectly calling a function that copies one record to userspace at a time, create a shadow buffer in the kernel and copy the whole array once at the end. On the author's computer, this reduces the runtime on his /home by ~20%. This also eliminates a deadlock when running GETFSMAP against the realtime device. The current code locks the rtbitmap to create fsmappings and copies them into userspace, having not released the rtbitmap lock. If the userspace buffer is an mmap of a sparse file that itself resides on the realtime device, the write page fault will recurse into the fs for allocation, which will deadlock on the rtbitmap lock. Fixes: 4c934c7dd60c ("xfs: report realtime space information via the rtbitmap") Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2020-10-01 20:56:07 +03:00
xfs_getfsmap_format(mp, &fmr, info);
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
}
if (info->last)
goto out;
/* Fill out the extent we found */
if (info->head->fmh_entries >= info->head->fmh_count)
xfs: remove all *_ITER_ABORT values Use -ECANCELED to signal "stop iterating" instead of these magical *_ITER_ABORT values, since it's duplicative. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-08-29 00:37:57 +03:00
return -ECANCELED;
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
xfs: pass perags around in fsmap data dev functions Needs a [from, to] ranged AG walk, and the perag to be stuffed into the info structure for callouts to use. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2021-06-02 03:48:24 +03:00
trace_xfs_fsmap_mapping(mp, info->dev,
info->pag ? info->pag->pag_agno : NULLAGNUMBER, rec);
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
fmr.fmr_device = info->dev;
fmr.fmr_physical = rec_daddr;
error = xfs_fsmap_owner_from_rmap(&fmr, rec);
if (error)
return error;
fmr.fmr_offset = XFS_FSB_TO_BB(mp, rec->rm_offset);
fmr.fmr_length = XFS_FSB_TO_BB(mp, rec->rm_blockcount);
if (rec->rm_flags & XFS_RMAP_UNWRITTEN)
fmr.fmr_flags |= FMR_OF_PREALLOC;
if (rec->rm_flags & XFS_RMAP_ATTR_FORK)
fmr.fmr_flags |= FMR_OF_ATTR_FORK;
if (rec->rm_flags & XFS_RMAP_BMBT_BLOCK)
fmr.fmr_flags |= FMR_OF_EXTENT_MAP;
if (fmr.fmr_flags == 0) {
error = xfs_getfsmap_is_shared(tp, info, rec, &shared);
if (error)
return error;
if (shared)
fmr.fmr_flags |= FMR_OF_SHARED;
}
xfs: fix deadlock and streamline xfs_getfsmap performance Refactor xfs_getfsmap to improve its performance: instead of indirectly calling a function that copies one record to userspace at a time, create a shadow buffer in the kernel and copy the whole array once at the end. On the author's computer, this reduces the runtime on his /home by ~20%. This also eliminates a deadlock when running GETFSMAP against the realtime device. The current code locks the rtbitmap to create fsmappings and copies them into userspace, having not released the rtbitmap lock. If the userspace buffer is an mmap of a sparse file that itself resides on the realtime device, the write page fault will recurse into the fs for allocation, which will deadlock on the rtbitmap lock. Fixes: 4c934c7dd60c ("xfs: report realtime space information via the rtbitmap") Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2020-10-01 20:56:07 +03:00
xfs_getfsmap_format(mp, &fmr, info);
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
out:
rec_daddr += XFS_FSB_TO_BB(mp, rec->rm_blockcount);
if (info->next_daddr < rec_daddr)
info->next_daddr = rec_daddr;
xfs: remove all *_ITER_CONTINUE values Iterator functions already use 0 to signal "continue iterating", so get rid of the #defines and just do it directly. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Dave Chinner <dchinner@redhat.com>
2019-08-29 00:39:46 +03:00
return 0;
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
}
/* Transform a rmapbt irec into a fsmap */
STATIC int
xfs_getfsmap_datadev_helper(
struct xfs_btree_cur *cur,
xfs: make the record pointer passed to query_range functions const The query_range functions are supposed to call a caller-supplied function on each record found in the dataset. These functions don't own the memory storing the record, so don't let them change the record. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de>
2021-08-11 03:02:16 +03:00
const struct xfs_rmap_irec *rec,
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
void *priv)
{
struct xfs_mount *mp = cur->bc_mp;
struct xfs_getfsmap_info *info = priv;
xfs_fsblock_t fsb;
xfs_daddr_t rec_daddr;
xfs: remove agno from btree cursor Now that everything passes a perag, the agno is not needed anymore. Convert all the users to use pag->pag_agno instead and remove the agno from the cursor. This was largely done as an automated search and replace. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2021-06-02 03:48:24 +03:00
fsb = XFS_AGB_TO_FSB(mp, cur->bc_ag.pag->pag_agno, rec->rm_startblock);
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
rec_daddr = XFS_FSB_TO_DADDR(mp, fsb);
return xfs_getfsmap_helper(cur->bc_tp, info, rec, rec_daddr);
}
xfs: have getfsmap fall back to the freesp btrees when rmap is not present If the reverse-mapping btree isn't available, fall back to the free space btrees to provide partial reverse mapping information. The online scrub tool can make use of even partial information to speed up the data block scan. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
/* Transform a bnobt irec into a fsmap */
STATIC int
xfs_getfsmap_datadev_bnobt_helper(
struct xfs_btree_cur *cur,
xfs: make the record pointer passed to query_range functions const The query_range functions are supposed to call a caller-supplied function on each record found in the dataset. These functions don't own the memory storing the record, so don't let them change the record. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de>
2021-08-11 03:02:16 +03:00
const struct xfs_alloc_rec_incore *rec,
xfs: have getfsmap fall back to the freesp btrees when rmap is not present If the reverse-mapping btree isn't available, fall back to the free space btrees to provide partial reverse mapping information. The online scrub tool can make use of even partial information to speed up the data block scan. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
void *priv)
{
struct xfs_mount *mp = cur->bc_mp;
struct xfs_getfsmap_info *info = priv;
struct xfs_rmap_irec irec;
xfs_daddr_t rec_daddr;
xfs: remove agno from btree cursor Now that everything passes a perag, the agno is not needed anymore. Convert all the users to use pag->pag_agno instead and remove the agno from the cursor. This was largely done as an automated search and replace. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2021-06-02 03:48:24 +03:00
rec_daddr = XFS_AGB_TO_DADDR(mp, cur->bc_ag.pag->pag_agno,
xfs: have getfsmap fall back to the freesp btrees when rmap is not present If the reverse-mapping btree isn't available, fall back to the free space btrees to provide partial reverse mapping information. The online scrub tool can make use of even partial information to speed up the data block scan. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
rec->ar_startblock);
irec.rm_startblock = rec->ar_startblock;
irec.rm_blockcount = rec->ar_blockcount;
irec.rm_owner = XFS_RMAP_OWN_NULL; /* "free" */
irec.rm_offset = 0;
irec.rm_flags = 0;
return xfs_getfsmap_helper(cur->bc_tp, info, &irec, rec_daddr);
}
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
/* Set rmap flags based on the getfsmap flags */
static void
xfs_getfsmap_set_irec_flags(
struct xfs_rmap_irec *irec,
xfs: make fsmap backend function key parameters const There are several GETFSMAP backend functions for XFS to cover the three devices and various feature support. Each of these functions are passed pointers to the low and high keys for the dataset that userspace requested, and a pointer to scratchpad variables that are used to control the iteration and fill out records. The scratchpad data can be changed arbitrarily, but the keys are supposed to remain unchanged (and under the control of the outermost loop in xfs_getfsmap). Unfortunately, the data and rt backends modify the keys that are passed in from the main control loop, which causes subsequent calls to return incorrect query results. Specifically, each of those two functions set the block number in the high key to the size of their respective device. Since fsmap results are sorted in device number order, if the lower numbered device is smaller than the higher numbered device, the first function will set the high key to the small size, and the key remains unchanged as it is passed into the function for the higher numbered device. The second function will then fail to return all of the results for the dataset that userspace is asking for because the keyspace is incorrectly constrained. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2021-08-11 03:00:31 +03:00
const struct xfs_fsmap *fmr)
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
{
irec->rm_flags = 0;
if (fmr->fmr_flags & FMR_OF_ATTR_FORK)
irec->rm_flags |= XFS_RMAP_ATTR_FORK;
if (fmr->fmr_flags & FMR_OF_EXTENT_MAP)
irec->rm_flags |= XFS_RMAP_BMBT_BLOCK;
if (fmr->fmr_flags & FMR_OF_PREALLOC)
irec->rm_flags |= XFS_RMAP_UNWRITTEN;
}
/* Execute a getfsmap query against the log device. */
STATIC int
xfs_getfsmap_logdev(
struct xfs_trans *tp,
xfs: make fsmap backend function key parameters const There are several GETFSMAP backend functions for XFS to cover the three devices and various feature support. Each of these functions are passed pointers to the low and high keys for the dataset that userspace requested, and a pointer to scratchpad variables that are used to control the iteration and fill out records. The scratchpad data can be changed arbitrarily, but the keys are supposed to remain unchanged (and under the control of the outermost loop in xfs_getfsmap). Unfortunately, the data and rt backends modify the keys that are passed in from the main control loop, which causes subsequent calls to return incorrect query results. Specifically, each of those two functions set the block number in the high key to the size of their respective device. Since fsmap results are sorted in device number order, if the lower numbered device is smaller than the higher numbered device, the first function will set the high key to the small size, and the key remains unchanged as it is passed into the function for the higher numbered device. The second function will then fail to return all of the results for the dataset that userspace is asking for because the keyspace is incorrectly constrained. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2021-08-11 03:00:31 +03:00
const struct xfs_fsmap *keys,
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
struct xfs_getfsmap_info *info)
{
struct xfs_mount *mp = tp->t_mountp;
struct xfs_rmap_irec rmap;
int error;
/* Set up search keys */
info->low.rm_startblock = XFS_BB_TO_FSBT(mp, keys[0].fmr_physical);
info->low.rm_offset = XFS_BB_TO_FSBT(mp, keys[0].fmr_offset);
error = xfs_fsmap_owner_to_rmap(&info->low, keys);
if (error)
return error;
info->low.rm_blockcount = 0;
xfs_getfsmap_set_irec_flags(&info->low, &keys[0]);
error = xfs_fsmap_owner_to_rmap(&info->high, keys + 1);
if (error)
return error;
info->high.rm_startblock = -1U;
info->high.rm_owner = ULLONG_MAX;
info->high.rm_offset = ULLONG_MAX;
info->high.rm_blockcount = 0;
info->high.rm_flags = XFS_RMAP_KEY_FLAGS | XFS_RMAP_REC_FLAGS;
info->missing_owner = XFS_FMR_OWN_FREE;
xfs: pass perags around in fsmap data dev functions Needs a [from, to] ranged AG walk, and the perag to be stuffed into the info structure for callouts to use. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2021-06-02 03:48:24 +03:00
trace_xfs_fsmap_low_key(mp, info->dev, NULLAGNUMBER, &info->low);
trace_xfs_fsmap_high_key(mp, info->dev, NULLAGNUMBER, &info->high);
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
if (keys[0].fmr_physical > 0)
return 0;
/* Fabricate an rmap entry for the external log device. */
rmap.rm_startblock = 0;
rmap.rm_blockcount = mp->m_sb.sb_logblocks;
rmap.rm_owner = XFS_RMAP_OWN_LOG;
rmap.rm_offset = 0;
rmap.rm_flags = 0;
return xfs_getfsmap_helper(tp, info, &rmap, 0);
}
xfs: move two more RT specific functions into CONFIG_XFS_RT The last cleanup introduced two harmless warnings: fs/xfs/xfs_fsmap.c:480:1: warning: '__xfs_getfsmap_rtdev' defined but not used fs/xfs/xfs_fsmap.c:372:1: warning: 'xfs_getfsmap_rtdev_rtbitmap_helper' defined but not used This moves those two functions as well. Fixes: bb9c2e543325 ("xfs: move more RT specific code under CONFIG_XFS_RT") Signed-off-by: Arnd Bergmann <arnd@arndb.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Acked-by: Geert Uytterhoeven <geert@linux-m68k.org> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-10-13 19:47:47 +03:00
#ifdef CONFIG_XFS_RT
/* Transform a rtbitmap "record" into a fsmap */
STATIC int
xfs_getfsmap_rtdev_rtbitmap_helper(
struct xfs_trans *tp,
xfs: make the record pointer passed to query_range functions const The query_range functions are supposed to call a caller-supplied function on each record found in the dataset. These functions don't own the memory storing the record, so don't let them change the record. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de>
2021-08-11 03:02:16 +03:00
const struct xfs_rtalloc_rec *rec,
xfs: move two more RT specific functions into CONFIG_XFS_RT The last cleanup introduced two harmless warnings: fs/xfs/xfs_fsmap.c:480:1: warning: '__xfs_getfsmap_rtdev' defined but not used fs/xfs/xfs_fsmap.c:372:1: warning: 'xfs_getfsmap_rtdev_rtbitmap_helper' defined but not used This moves those two functions as well. Fixes: bb9c2e543325 ("xfs: move more RT specific code under CONFIG_XFS_RT") Signed-off-by: Arnd Bergmann <arnd@arndb.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Acked-by: Geert Uytterhoeven <geert@linux-m68k.org> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-10-13 19:47:47 +03:00
void *priv)
{
struct xfs_mount *mp = tp->t_mountp;
struct xfs_getfsmap_info *info = priv;
struct xfs_rmap_irec irec;
xfs_daddr_t rec_daddr;
xfs: fix xfs_rtalloc_rec units All the realtime allocation functions deal with space on the rtdev in units of realtime extents. However, struct xfs_rtalloc_rec confusingly uses the word 'block' in the name, even though they're really extents. Fix the naming problem and fix all the unit handling problems in the two existing users. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Allison Henderson <allison.henderson@oracle.com> Reviewed-by: Bill O'Donnell <billodo@redhat.com>
2018-05-31 19:12:10 +03:00
irec.rm_startblock = rec->ar_startext * mp->m_sb.sb_rextsize;
rec_daddr = XFS_FSB_TO_BB(mp, irec.rm_startblock);
irec.rm_blockcount = rec->ar_extcount * mp->m_sb.sb_rextsize;
xfs: move two more RT specific functions into CONFIG_XFS_RT The last cleanup introduced two harmless warnings: fs/xfs/xfs_fsmap.c:480:1: warning: '__xfs_getfsmap_rtdev' defined but not used fs/xfs/xfs_fsmap.c:372:1: warning: 'xfs_getfsmap_rtdev_rtbitmap_helper' defined but not used This moves those two functions as well. Fixes: bb9c2e543325 ("xfs: move more RT specific code under CONFIG_XFS_RT") Signed-off-by: Arnd Bergmann <arnd@arndb.de> Reviewed-by: Brian Foster <bfoster@redhat.com> Acked-by: Geert Uytterhoeven <geert@linux-m68k.org> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-10-13 19:47:47 +03:00
irec.rm_owner = XFS_RMAP_OWN_NULL; /* "free" */
irec.rm_offset = 0;
irec.rm_flags = 0;
return xfs_getfsmap_helper(tp, info, &irec, rec_daddr);
}
xfs: report realtime space information via the rtbitmap Use the realtime bitmap to return free space information via getfsmap. Eventually this will be superseded by the realtime rmapbt code. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:38 +03:00
/* Execute a getfsmap query against the realtime device. */
STATIC int
__xfs_getfsmap_rtdev(
struct xfs_trans *tp,
xfs: make fsmap backend function key parameters const There are several GETFSMAP backend functions for XFS to cover the three devices and various feature support. Each of these functions are passed pointers to the low and high keys for the dataset that userspace requested, and a pointer to scratchpad variables that are used to control the iteration and fill out records. The scratchpad data can be changed arbitrarily, but the keys are supposed to remain unchanged (and under the control of the outermost loop in xfs_getfsmap). Unfortunately, the data and rt backends modify the keys that are passed in from the main control loop, which causes subsequent calls to return incorrect query results. Specifically, each of those two functions set the block number in the high key to the size of their respective device. Since fsmap results are sorted in device number order, if the lower numbered device is smaller than the higher numbered device, the first function will set the high key to the small size, and the key remains unchanged as it is passed into the function for the higher numbered device. The second function will then fail to return all of the results for the dataset that userspace is asking for because the keyspace is incorrectly constrained. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2021-08-11 03:00:31 +03:00
const struct xfs_fsmap *keys,
xfs: report realtime space information via the rtbitmap Use the realtime bitmap to return free space information via getfsmap. Eventually this will be superseded by the realtime rmapbt code. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:38 +03:00
int (*query_fn)(struct xfs_trans *,
struct xfs_getfsmap_info *),
struct xfs_getfsmap_info *info)
{
struct xfs_mount *mp = tp->t_mountp;
xfs_fsblock_t start_fsb;
xfs_fsblock_t end_fsb;
xfs: make fsmap backend function key parameters const There are several GETFSMAP backend functions for XFS to cover the three devices and various feature support. Each of these functions are passed pointers to the low and high keys for the dataset that userspace requested, and a pointer to scratchpad variables that are used to control the iteration and fill out records. The scratchpad data can be changed arbitrarily, but the keys are supposed to remain unchanged (and under the control of the outermost loop in xfs_getfsmap). Unfortunately, the data and rt backends modify the keys that are passed in from the main control loop, which causes subsequent calls to return incorrect query results. Specifically, each of those two functions set the block number in the high key to the size of their respective device. Since fsmap results are sorted in device number order, if the lower numbered device is smaller than the higher numbered device, the first function will set the high key to the small size, and the key remains unchanged as it is passed into the function for the higher numbered device. The second function will then fail to return all of the results for the dataset that userspace is asking for because the keyspace is incorrectly constrained. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2021-08-11 03:00:31 +03:00
uint64_t eofs;
xfs: report realtime space information via the rtbitmap Use the realtime bitmap to return free space information via getfsmap. Eventually this will be superseded by the realtime rmapbt code. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:38 +03:00
int error = 0;
eofs = XFS_FSB_TO_BB(mp, mp->m_sb.sb_rblocks);
if (keys[0].fmr_physical >= eofs)
return 0;
start_fsb = XFS_BB_TO_FSBT(mp, keys[0].fmr_physical);
xfs: make fsmap backend function key parameters const There are several GETFSMAP backend functions for XFS to cover the three devices and various feature support. Each of these functions are passed pointers to the low and high keys for the dataset that userspace requested, and a pointer to scratchpad variables that are used to control the iteration and fill out records. The scratchpad data can be changed arbitrarily, but the keys are supposed to remain unchanged (and under the control of the outermost loop in xfs_getfsmap). Unfortunately, the data and rt backends modify the keys that are passed in from the main control loop, which causes subsequent calls to return incorrect query results. Specifically, each of those two functions set the block number in the high key to the size of their respective device. Since fsmap results are sorted in device number order, if the lower numbered device is smaller than the higher numbered device, the first function will set the high key to the small size, and the key remains unchanged as it is passed into the function for the higher numbered device. The second function will then fail to return all of the results for the dataset that userspace is asking for because the keyspace is incorrectly constrained. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2021-08-11 03:00:31 +03:00
end_fsb = XFS_BB_TO_FSB(mp, min(eofs - 1, keys[1].fmr_physical));
xfs: report realtime space information via the rtbitmap Use the realtime bitmap to return free space information via getfsmap. Eventually this will be superseded by the realtime rmapbt code. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:38 +03:00
/* Set up search keys */
info->low.rm_startblock = start_fsb;
error = xfs_fsmap_owner_to_rmap(&info->low, &keys[0]);
if (error)
return error;
info->low.rm_offset = XFS_BB_TO_FSBT(mp, keys[0].fmr_offset);
info->low.rm_blockcount = 0;
xfs_getfsmap_set_irec_flags(&info->low, &keys[0]);
info->high.rm_startblock = end_fsb;
error = xfs_fsmap_owner_to_rmap(&info->high, &keys[1]);
if (error)
return error;
info->high.rm_offset = XFS_BB_TO_FSBT(mp, keys[1].fmr_offset);
info->high.rm_blockcount = 0;
xfs_getfsmap_set_irec_flags(&info->high, &keys[1]);
xfs: pass perags around in fsmap data dev functions Needs a [from, to] ranged AG walk, and the perag to be stuffed into the info structure for callouts to use. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2021-06-02 03:48:24 +03:00
trace_xfs_fsmap_low_key(mp, info->dev, NULLAGNUMBER, &info->low);
trace_xfs_fsmap_high_key(mp, info->dev, NULLAGNUMBER, &info->high);
xfs: report realtime space information via the rtbitmap Use the realtime bitmap to return free space information via getfsmap. Eventually this will be superseded by the realtime rmapbt code. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:38 +03:00
return query_fn(tp, info);
}
/* Actually query the realtime bitmap. */
STATIC int
xfs_getfsmap_rtdev_rtbitmap_query(
struct xfs_trans *tp,
struct xfs_getfsmap_info *info)
{
xfs: fix uninitialized field in rtbitmap fsmap backend Initialize the extent count field of the high key so that when we use the high key to synthesize an 'unknown owner' record (i.e. used space record) at the end of the queried range we have a field with which to compute rm_blockcount. This is not strictly necessary because the synthesizer never uses the rm_blockcount field, but we can shut up the static code analysis anyway. Coverity-id: 1437358 Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Allison Henderson <allison.henderson@oracle.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
2018-06-22 09:26:57 +03:00
struct xfs_rtalloc_rec alow = { 0 };
struct xfs_rtalloc_rec ahigh = { 0 };
xfs: fix off-by-one error when the last rt extent is in use The fsmap implementation for realtime devices uses the gap between info->next_daddr and a free rtextent reported by xfs_rtalloc_query_range to feed userspace fsmap records with an "unknown" owner. We use this trick to report to userspace when the last rtextent in the filesystem is in use by synthesizing a null rmap record starting at the next block after the query range. Unfortunately, there's a minor accounting bug in the way that we construct the null rmap record. Originally, ahigh.ar_startext contains the last rtextent for which the user wants records. It's entirely possible that number is beyond the end of the rt volume, so the location synthesized rmap record /must/ be constrained to the minimum of the high key and the number of extents in the rt volume. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2021-08-11 03:00:31 +03:00
struct xfs_mount *mp = tp->t_mountp;
xfs: report realtime space information via the rtbitmap Use the realtime bitmap to return free space information via getfsmap. Eventually this will be superseded by the realtime rmapbt code. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:38 +03:00
int error;
xfs: fix off-by-one error when the last rt extent is in use The fsmap implementation for realtime devices uses the gap between info->next_daddr and a free rtextent reported by xfs_rtalloc_query_range to feed userspace fsmap records with an "unknown" owner. We use this trick to report to userspace when the last rtextent in the filesystem is in use by synthesizing a null rmap record starting at the next block after the query range. Unfortunately, there's a minor accounting bug in the way that we construct the null rmap record. Originally, ahigh.ar_startext contains the last rtextent for which the user wants records. It's entirely possible that number is beyond the end of the rt volume, so the location synthesized rmap record /must/ be constrained to the minimum of the high key and the number of extents in the rt volume. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2021-08-11 03:00:31 +03:00
xfs_ilock(mp->m_rbmip, XFS_ILOCK_SHARED);
xfs: report realtime space information via the rtbitmap Use the realtime bitmap to return free space information via getfsmap. Eventually this will be superseded by the realtime rmapbt code. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:38 +03:00
xfs: fix off-by-one error when the last rt extent is in use The fsmap implementation for realtime devices uses the gap between info->next_daddr and a free rtextent reported by xfs_rtalloc_query_range to feed userspace fsmap records with an "unknown" owner. We use this trick to report to userspace when the last rtextent in the filesystem is in use by synthesizing a null rmap record starting at the next block after the query range. Unfortunately, there's a minor accounting bug in the way that we construct the null rmap record. Originally, ahigh.ar_startext contains the last rtextent for which the user wants records. It's entirely possible that number is beyond the end of the rt volume, so the location synthesized rmap record /must/ be constrained to the minimum of the high key and the number of extents in the rt volume. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2021-08-11 03:00:31 +03:00
/*
* Set up query parameters to return free rtextents covering the range
* we want.
*/
xfs: fix xfs_rtalloc_rec units All the realtime allocation functions deal with space on the rtdev in units of realtime extents. However, struct xfs_rtalloc_rec confusingly uses the word 'block' in the name, even though they're really extents. Fix the naming problem and fix all the unit handling problems in the two existing users. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Allison Henderson <allison.henderson@oracle.com> Reviewed-by: Bill O'Donnell <billodo@redhat.com>
2018-05-31 19:12:10 +03:00
alow.ar_startext = info->low.rm_startblock;
ahigh.ar_startext = info->high.rm_startblock;
xfs: fix off-by-one error when the last rt extent is in use The fsmap implementation for realtime devices uses the gap between info->next_daddr and a free rtextent reported by xfs_rtalloc_query_range to feed userspace fsmap records with an "unknown" owner. We use this trick to report to userspace when the last rtextent in the filesystem is in use by synthesizing a null rmap record starting at the next block after the query range. Unfortunately, there's a minor accounting bug in the way that we construct the null rmap record. Originally, ahigh.ar_startext contains the last rtextent for which the user wants records. It's entirely possible that number is beyond the end of the rt volume, so the location synthesized rmap record /must/ be constrained to the minimum of the high key and the number of extents in the rt volume. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2021-08-11 03:00:31 +03:00
do_div(alow.ar_startext, mp->m_sb.sb_rextsize);
if (do_div(ahigh.ar_startext, mp->m_sb.sb_rextsize))
xfs: fix xfs_rtalloc_rec units All the realtime allocation functions deal with space on the rtdev in units of realtime extents. However, struct xfs_rtalloc_rec confusingly uses the word 'block' in the name, even though they're really extents. Fix the naming problem and fix all the unit handling problems in the two existing users. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Allison Henderson <allison.henderson@oracle.com> Reviewed-by: Bill O'Donnell <billodo@redhat.com>
2018-05-31 19:12:10 +03:00
ahigh.ar_startext++;
xfs: report realtime space information via the rtbitmap Use the realtime bitmap to return free space information via getfsmap. Eventually this will be superseded by the realtime rmapbt code. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:38 +03:00
error = xfs_rtalloc_query_range(tp, &alow, &ahigh,
xfs_getfsmap_rtdev_rtbitmap_helper, info);
if (error)
goto err;
xfs: fix off-by-one error when the last rt extent is in use The fsmap implementation for realtime devices uses the gap between info->next_daddr and a free rtextent reported by xfs_rtalloc_query_range to feed userspace fsmap records with an "unknown" owner. We use this trick to report to userspace when the last rtextent in the filesystem is in use by synthesizing a null rmap record starting at the next block after the query range. Unfortunately, there's a minor accounting bug in the way that we construct the null rmap record. Originally, ahigh.ar_startext contains the last rtextent for which the user wants records. It's entirely possible that number is beyond the end of the rt volume, so the location synthesized rmap record /must/ be constrained to the minimum of the high key and the number of extents in the rt volume. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2021-08-11 03:00:31 +03:00
/*
* Report any gaps at the end of the rtbitmap by simulating a null
* rmap starting at the block after the end of the query range.
*/
xfs: report realtime space information via the rtbitmap Use the realtime bitmap to return free space information via getfsmap. Eventually this will be superseded by the realtime rmapbt code. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:38 +03:00
info->last = true;
xfs: fix off-by-one error when the last rt extent is in use The fsmap implementation for realtime devices uses the gap between info->next_daddr and a free rtextent reported by xfs_rtalloc_query_range to feed userspace fsmap records with an "unknown" owner. We use this trick to report to userspace when the last rtextent in the filesystem is in use by synthesizing a null rmap record starting at the next block after the query range. Unfortunately, there's a minor accounting bug in the way that we construct the null rmap record. Originally, ahigh.ar_startext contains the last rtextent for which the user wants records. It's entirely possible that number is beyond the end of the rt volume, so the location synthesized rmap record /must/ be constrained to the minimum of the high key and the number of extents in the rt volume. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2021-08-11 03:00:31 +03:00
ahigh.ar_startext = min(mp->m_sb.sb_rextents, ahigh.ar_startext);
xfs: report realtime space information via the rtbitmap Use the realtime bitmap to return free space information via getfsmap. Eventually this will be superseded by the realtime rmapbt code. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:38 +03:00
error = xfs_getfsmap_rtdev_rtbitmap_helper(tp, &ahigh, info);
if (error)
goto err;
err:
xfs: fix off-by-one error when the last rt extent is in use The fsmap implementation for realtime devices uses the gap between info->next_daddr and a free rtextent reported by xfs_rtalloc_query_range to feed userspace fsmap records with an "unknown" owner. We use this trick to report to userspace when the last rtextent in the filesystem is in use by synthesizing a null rmap record starting at the next block after the query range. Unfortunately, there's a minor accounting bug in the way that we construct the null rmap record. Originally, ahigh.ar_startext contains the last rtextent for which the user wants records. It's entirely possible that number is beyond the end of the rt volume, so the location synthesized rmap record /must/ be constrained to the minimum of the high key and the number of extents in the rt volume. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2021-08-11 03:00:31 +03:00
xfs_iunlock(mp->m_rbmip, XFS_ILOCK_SHARED);
xfs: report realtime space information via the rtbitmap Use the realtime bitmap to return free space information via getfsmap. Eventually this will be superseded by the realtime rmapbt code. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:38 +03:00
return error;
}
/* Execute a getfsmap query against the realtime device rtbitmap. */
STATIC int
xfs_getfsmap_rtdev_rtbitmap(
struct xfs_trans *tp,
xfs: make fsmap backend function key parameters const There are several GETFSMAP backend functions for XFS to cover the three devices and various feature support. Each of these functions are passed pointers to the low and high keys for the dataset that userspace requested, and a pointer to scratchpad variables that are used to control the iteration and fill out records. The scratchpad data can be changed arbitrarily, but the keys are supposed to remain unchanged (and under the control of the outermost loop in xfs_getfsmap). Unfortunately, the data and rt backends modify the keys that are passed in from the main control loop, which causes subsequent calls to return incorrect query results. Specifically, each of those two functions set the block number in the high key to the size of their respective device. Since fsmap results are sorted in device number order, if the lower numbered device is smaller than the higher numbered device, the first function will set the high key to the small size, and the key remains unchanged as it is passed into the function for the higher numbered device. The second function will then fail to return all of the results for the dataset that userspace is asking for because the keyspace is incorrectly constrained. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2021-08-11 03:00:31 +03:00
const struct xfs_fsmap *keys,
xfs: report realtime space information via the rtbitmap Use the realtime bitmap to return free space information via getfsmap. Eventually this will be superseded by the realtime rmapbt code. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:38 +03:00
struct xfs_getfsmap_info *info)
{
info->missing_owner = XFS_FMR_OWN_UNKNOWN;
return __xfs_getfsmap_rtdev(tp, keys, xfs_getfsmap_rtdev_rtbitmap_query,
info);
}
xfs: move more RT specific code under CONFIG_XFS_RT Various utility functions and interfaces that iterate internal devices try to reference the realtime device even when RT support is not compiled into the kernel. Make sure this code is excluded from the CONFIG_XFS_RT=n build, and where appropriate stub functions to return fatal errors if they ever get called when RT support is not present. Signed-Off-By: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-10-09 21:37:22 +03:00
#endif /* CONFIG_XFS_RT */
xfs: report realtime space information via the rtbitmap Use the realtime bitmap to return free space information via getfsmap. Eventually this will be superseded by the realtime rmapbt code. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:38 +03:00
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
/* Execute a getfsmap query against the regular data device. */
STATIC int
__xfs_getfsmap_datadev(
struct xfs_trans *tp,
xfs: make fsmap backend function key parameters const There are several GETFSMAP backend functions for XFS to cover the three devices and various feature support. Each of these functions are passed pointers to the low and high keys for the dataset that userspace requested, and a pointer to scratchpad variables that are used to control the iteration and fill out records. The scratchpad data can be changed arbitrarily, but the keys are supposed to remain unchanged (and under the control of the outermost loop in xfs_getfsmap). Unfortunately, the data and rt backends modify the keys that are passed in from the main control loop, which causes subsequent calls to return incorrect query results. Specifically, each of those two functions set the block number in the high key to the size of their respective device. Since fsmap results are sorted in device number order, if the lower numbered device is smaller than the higher numbered device, the first function will set the high key to the small size, and the key remains unchanged as it is passed into the function for the higher numbered device. The second function will then fail to return all of the results for the dataset that userspace is asking for because the keyspace is incorrectly constrained. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2021-08-11 03:00:31 +03:00
const struct xfs_fsmap *keys,
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
struct xfs_getfsmap_info *info,
int (*query_fn)(struct xfs_trans *,
struct xfs_getfsmap_info *,
struct xfs_btree_cur **,
void *),
void *priv)
{
struct xfs_mount *mp = tp->t_mountp;
xfs: pass perags around in fsmap data dev functions Needs a [from, to] ranged AG walk, and the perag to be stuffed into the info structure for callouts to use. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2021-06-02 03:48:24 +03:00
struct xfs_perag *pag;
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
struct xfs_btree_cur *bt_cur = NULL;
xfs_fsblock_t start_fsb;
xfs_fsblock_t end_fsb;
xfs_agnumber_t start_ag;
xfs_agnumber_t end_ag;
xfs: make fsmap backend function key parameters const There are several GETFSMAP backend functions for XFS to cover the three devices and various feature support. Each of these functions are passed pointers to the low and high keys for the dataset that userspace requested, and a pointer to scratchpad variables that are used to control the iteration and fill out records. The scratchpad data can be changed arbitrarily, but the keys are supposed to remain unchanged (and under the control of the outermost loop in xfs_getfsmap). Unfortunately, the data and rt backends modify the keys that are passed in from the main control loop, which causes subsequent calls to return incorrect query results. Specifically, each of those two functions set the block number in the high key to the size of their respective device. Since fsmap results are sorted in device number order, if the lower numbered device is smaller than the higher numbered device, the first function will set the high key to the small size, and the key remains unchanged as it is passed into the function for the higher numbered device. The second function will then fail to return all of the results for the dataset that userspace is asking for because the keyspace is incorrectly constrained. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2021-08-11 03:00:31 +03:00
uint64_t eofs;
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
int error = 0;
eofs = XFS_FSB_TO_BB(mp, mp->m_sb.sb_dblocks);
if (keys[0].fmr_physical >= eofs)
return 0;
start_fsb = XFS_DADDR_TO_FSB(mp, keys[0].fmr_physical);
xfs: make fsmap backend function key parameters const There are several GETFSMAP backend functions for XFS to cover the three devices and various feature support. Each of these functions are passed pointers to the low and high keys for the dataset that userspace requested, and a pointer to scratchpad variables that are used to control the iteration and fill out records. The scratchpad data can be changed arbitrarily, but the keys are supposed to remain unchanged (and under the control of the outermost loop in xfs_getfsmap). Unfortunately, the data and rt backends modify the keys that are passed in from the main control loop, which causes subsequent calls to return incorrect query results. Specifically, each of those two functions set the block number in the high key to the size of their respective device. Since fsmap results are sorted in device number order, if the lower numbered device is smaller than the higher numbered device, the first function will set the high key to the small size, and the key remains unchanged as it is passed into the function for the higher numbered device. The second function will then fail to return all of the results for the dataset that userspace is asking for because the keyspace is incorrectly constrained. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2021-08-11 03:00:31 +03:00
end_fsb = XFS_DADDR_TO_FSB(mp, min(eofs - 1, keys[1].fmr_physical));
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
/*
* Convert the fsmap low/high keys to AG based keys. Initialize
* low to the fsmap low key and max out the high key to the end
* of the AG.
*/
info->low.rm_startblock = XFS_FSB_TO_AGBNO(mp, start_fsb);
info->low.rm_offset = XFS_BB_TO_FSBT(mp, keys[0].fmr_offset);
error = xfs_fsmap_owner_to_rmap(&info->low, &keys[0]);
if (error)
return error;
info->low.rm_blockcount = 0;
xfs_getfsmap_set_irec_flags(&info->low, &keys[0]);
info->high.rm_startblock = -1U;
info->high.rm_owner = ULLONG_MAX;
info->high.rm_offset = ULLONG_MAX;
info->high.rm_blockcount = 0;
info->high.rm_flags = XFS_RMAP_KEY_FLAGS | XFS_RMAP_REC_FLAGS;
start_ag = XFS_FSB_TO_AGNO(mp, start_fsb);
end_ag = XFS_FSB_TO_AGNO(mp, end_fsb);
xfs: pass perags around in fsmap data dev functions Needs a [from, to] ranged AG walk, and the perag to be stuffed into the info structure for callouts to use. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2021-06-02 03:48:24 +03:00
for_each_perag_range(mp, start_ag, end_ag, pag) {
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
/*
* Set the AG high key from the fsmap high key if this
* is the last AG that we're querying.
*/
xfs: pass perags around in fsmap data dev functions Needs a [from, to] ranged AG walk, and the perag to be stuffed into the info structure for callouts to use. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2021-06-02 03:48:24 +03:00
info->pag = pag;
if (pag->pag_agno == end_ag) {
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
info->high.rm_startblock = XFS_FSB_TO_AGBNO(mp,
end_fsb);
info->high.rm_offset = XFS_BB_TO_FSBT(mp,
keys[1].fmr_offset);
error = xfs_fsmap_owner_to_rmap(&info->high, &keys[1]);
if (error)
xfs: pass perags around in fsmap data dev functions Needs a [from, to] ranged AG walk, and the perag to be stuffed into the info structure for callouts to use. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2021-06-02 03:48:24 +03:00
break;
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
xfs_getfsmap_set_irec_flags(&info->high, &keys[1]);
}
if (bt_cur) {
xfs_btree_del_cursor(bt_cur, XFS_BTREE_NOERROR);
bt_cur = NULL;
xfs_trans_brelse(tp, info->agf_bp);
info->agf_bp = NULL;
}
xfs: pass perags around in fsmap data dev functions Needs a [from, to] ranged AG walk, and the perag to be stuffed into the info structure for callouts to use. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2021-06-02 03:48:24 +03:00
error = xfs_alloc_read_agf(mp, tp, pag->pag_agno, 0,
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
&info->agf_bp);
if (error)
xfs: pass perags around in fsmap data dev functions Needs a [from, to] ranged AG walk, and the perag to be stuffed into the info structure for callouts to use. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2021-06-02 03:48:24 +03:00
break;
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
xfs: pass perags around in fsmap data dev functions Needs a [from, to] ranged AG walk, and the perag to be stuffed into the info structure for callouts to use. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2021-06-02 03:48:24 +03:00
trace_xfs_fsmap_low_key(mp, info->dev, pag->pag_agno,
&info->low);
trace_xfs_fsmap_high_key(mp, info->dev, pag->pag_agno,
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
&info->high);
error = query_fn(tp, info, &bt_cur, priv);
if (error)
xfs: pass perags around in fsmap data dev functions Needs a [from, to] ranged AG walk, and the perag to be stuffed into the info structure for callouts to use. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2021-06-02 03:48:24 +03:00
break;
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
/*
* Set the AG low key to the start of the AG prior to
* moving on to the next AG.
*/
xfs: pass perags around in fsmap data dev functions Needs a [from, to] ranged AG walk, and the perag to be stuffed into the info structure for callouts to use. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2021-06-02 03:48:24 +03:00
if (pag->pag_agno == start_ag) {
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
info->low.rm_startblock = 0;
info->low.rm_owner = 0;
info->low.rm_offset = 0;
info->low.rm_flags = 0;
}
xfs: pass perags around in fsmap data dev functions Needs a [from, to] ranged AG walk, and the perag to be stuffed into the info structure for callouts to use. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2021-06-02 03:48:24 +03:00
/*
* If this is the last AG, report any gap at the end of it
* before we drop the reference to the perag when the loop
* terminates.
*/
if (pag->pag_agno == end_ag) {
info->last = true;
error = query_fn(tp, info, &bt_cur, priv);
if (error)
break;
}
info->pag = NULL;
}
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
if (bt_cur)
xfs_btree_del_cursor(bt_cur, error < 0 ? XFS_BTREE_ERROR :
XFS_BTREE_NOERROR);
if (info->agf_bp) {
xfs_trans_brelse(tp, info->agf_bp);
info->agf_bp = NULL;
}
xfs: pass perags around in fsmap data dev functions Needs a [from, to] ranged AG walk, and the perag to be stuffed into the info structure for callouts to use. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2021-06-02 03:48:24 +03:00
if (info->pag) {
xfs_perag_put(info->pag);
info->pag = NULL;
} else if (pag) {
/* loop termination case */
xfs_perag_put(pag);
}
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
return error;
}
/* Actually query the rmap btree. */
STATIC int
xfs_getfsmap_datadev_rmapbt_query(
struct xfs_trans *tp,
struct xfs_getfsmap_info *info,
struct xfs_btree_cur **curpp,
void *priv)
{
/* Report any gap at the end of the last AG. */
if (info->last)
return xfs_getfsmap_datadev_helper(*curpp, &info->high, info);
/* Allocate cursor for this AG and query_range it. */
*curpp = xfs_rmapbt_init_cursor(tp->t_mountp, tp, info->agf_bp,
xfs: convert rmap btree cursor to using a perag Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2021-06-02 03:48:24 +03:00
info->pag);
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
return xfs_rmap_query_range(*curpp, &info->low, &info->high,
xfs_getfsmap_datadev_helper, info);
}
/* Execute a getfsmap query against the regular data device rmapbt. */
STATIC int
xfs_getfsmap_datadev_rmapbt(
struct xfs_trans *tp,
xfs: make fsmap backend function key parameters const There are several GETFSMAP backend functions for XFS to cover the three devices and various feature support. Each of these functions are passed pointers to the low and high keys for the dataset that userspace requested, and a pointer to scratchpad variables that are used to control the iteration and fill out records. The scratchpad data can be changed arbitrarily, but the keys are supposed to remain unchanged (and under the control of the outermost loop in xfs_getfsmap). Unfortunately, the data and rt backends modify the keys that are passed in from the main control loop, which causes subsequent calls to return incorrect query results. Specifically, each of those two functions set the block number in the high key to the size of their respective device. Since fsmap results are sorted in device number order, if the lower numbered device is smaller than the higher numbered device, the first function will set the high key to the small size, and the key remains unchanged as it is passed into the function for the higher numbered device. The second function will then fail to return all of the results for the dataset that userspace is asking for because the keyspace is incorrectly constrained. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2021-08-11 03:00:31 +03:00
const struct xfs_fsmap *keys,
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
struct xfs_getfsmap_info *info)
{
info->missing_owner = XFS_FMR_OWN_FREE;
return __xfs_getfsmap_datadev(tp, keys, info,
xfs_getfsmap_datadev_rmapbt_query, NULL);
}
xfs: have getfsmap fall back to the freesp btrees when rmap is not present If the reverse-mapping btree isn't available, fall back to the free space btrees to provide partial reverse mapping information. The online scrub tool can make use of even partial information to speed up the data block scan. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
/* Actually query the bno btree. */
STATIC int
xfs_getfsmap_datadev_bnobt_query(
struct xfs_trans *tp,
struct xfs_getfsmap_info *info,
struct xfs_btree_cur **curpp,
void *priv)
{
struct xfs_alloc_rec_incore *key = priv;
/* Report any gap at the end of the last AG. */
if (info->last)
return xfs_getfsmap_datadev_bnobt_helper(*curpp, &key[1], info);
/* Allocate cursor for this AG and query_range it. */
*curpp = xfs_allocbt_init_cursor(tp->t_mountp, tp, info->agf_bp,
xfs: convert allocbt cursors to use perags Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2021-06-02 03:48:24 +03:00
info->pag, XFS_BTNUM_BNO);
xfs: have getfsmap fall back to the freesp btrees when rmap is not present If the reverse-mapping btree isn't available, fall back to the free space btrees to provide partial reverse mapping information. The online scrub tool can make use of even partial information to speed up the data block scan. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
key->ar_startblock = info->low.rm_startblock;
key[1].ar_startblock = info->high.rm_startblock;
return xfs_alloc_query_range(*curpp, key, &key[1],
xfs_getfsmap_datadev_bnobt_helper, info);
}
/* Execute a getfsmap query against the regular data device's bnobt. */
STATIC int
xfs_getfsmap_datadev_bnobt(
struct xfs_trans *tp,
xfs: make fsmap backend function key parameters const There are several GETFSMAP backend functions for XFS to cover the three devices and various feature support. Each of these functions are passed pointers to the low and high keys for the dataset that userspace requested, and a pointer to scratchpad variables that are used to control the iteration and fill out records. The scratchpad data can be changed arbitrarily, but the keys are supposed to remain unchanged (and under the control of the outermost loop in xfs_getfsmap). Unfortunately, the data and rt backends modify the keys that are passed in from the main control loop, which causes subsequent calls to return incorrect query results. Specifically, each of those two functions set the block number in the high key to the size of their respective device. Since fsmap results are sorted in device number order, if the lower numbered device is smaller than the higher numbered device, the first function will set the high key to the small size, and the key remains unchanged as it is passed into the function for the higher numbered device. The second function will then fail to return all of the results for the dataset that userspace is asking for because the keyspace is incorrectly constrained. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2021-08-11 03:00:31 +03:00
const struct xfs_fsmap *keys,
xfs: have getfsmap fall back to the freesp btrees when rmap is not present If the reverse-mapping btree isn't available, fall back to the free space btrees to provide partial reverse mapping information. The online scrub tool can make use of even partial information to speed up the data block scan. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
struct xfs_getfsmap_info *info)
{
struct xfs_alloc_rec_incore akeys[2];
info->missing_owner = XFS_FMR_OWN_UNKNOWN;
return __xfs_getfsmap_datadev(tp, keys, info,
xfs_getfsmap_datadev_bnobt_query, &akeys[0]);
}
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
/* Do we recognize the device? */
STATIC bool
xfs_getfsmap_is_valid_device(
struct xfs_mount *mp,
struct xfs_fsmap *fm)
{
if (fm->fmr_device == 0 || fm->fmr_device == UINT_MAX ||
fm->fmr_device == new_encode_dev(mp->m_ddev_targp->bt_dev))
return true;
if (mp->m_logdev_targp &&
fm->fmr_device == new_encode_dev(mp->m_logdev_targp->bt_dev))
return true;
xfs: report realtime space information via the rtbitmap Use the realtime bitmap to return free space information via getfsmap. Eventually this will be superseded by the realtime rmapbt code. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:38 +03:00
if (mp->m_rtdev_targp &&
fm->fmr_device == new_encode_dev(mp->m_rtdev_targp->bt_dev))
return true;
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
return false;
}
/* Ensure that the low key is less than the high key. */
STATIC bool
xfs_getfsmap_check_keys(
struct xfs_fsmap *low_key,
struct xfs_fsmap *high_key)
{
if (low_key->fmr_device > high_key->fmr_device)
return false;
if (low_key->fmr_device < high_key->fmr_device)
return true;
if (low_key->fmr_physical > high_key->fmr_physical)
return false;
if (low_key->fmr_physical < high_key->fmr_physical)
return true;
if (low_key->fmr_owner > high_key->fmr_owner)
return false;
if (low_key->fmr_owner < high_key->fmr_owner)
return true;
if (low_key->fmr_offset > high_key->fmr_offset)
return false;
if (low_key->fmr_offset < high_key->fmr_offset)
return true;
return false;
}
xfs: move more RT specific code under CONFIG_XFS_RT Various utility functions and interfaces that iterate internal devices try to reference the realtime device even when RT support is not compiled into the kernel. Make sure this code is excluded from the CONFIG_XFS_RT=n build, and where appropriate stub functions to return fatal errors if they ever get called when RT support is not present. Signed-Off-By: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-10-09 21:37:22 +03:00
/*
* There are only two devices if we didn't configure RT devices at build time.
*/
#ifdef CONFIG_XFS_RT
xfs: report realtime space information via the rtbitmap Use the realtime bitmap to return free space information via getfsmap. Eventually this will be superseded by the realtime rmapbt code. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:38 +03:00
#define XFS_GETFSMAP_DEVS 3
xfs: move more RT specific code under CONFIG_XFS_RT Various utility functions and interfaces that iterate internal devices try to reference the realtime device even when RT support is not compiled into the kernel. Make sure this code is excluded from the CONFIG_XFS_RT=n build, and where appropriate stub functions to return fatal errors if they ever get called when RT support is not present. Signed-Off-By: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-10-09 21:37:22 +03:00
#else
#define XFS_GETFSMAP_DEVS 2
#endif /* CONFIG_XFS_RT */
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
/*
xfs: fix deadlock and streamline xfs_getfsmap performance Refactor xfs_getfsmap to improve its performance: instead of indirectly calling a function that copies one record to userspace at a time, create a shadow buffer in the kernel and copy the whole array once at the end. On the author's computer, this reduces the runtime on his /home by ~20%. This also eliminates a deadlock when running GETFSMAP against the realtime device. The current code locks the rtbitmap to create fsmappings and copies them into userspace, having not released the rtbitmap lock. If the userspace buffer is an mmap of a sparse file that itself resides on the realtime device, the write page fault will recurse into the fs for allocation, which will deadlock on the rtbitmap lock. Fixes: 4c934c7dd60c ("xfs: report realtime space information via the rtbitmap") Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2020-10-01 20:56:07 +03:00
* Get filesystem's extents as described in head, and format for output. Fills
* in the supplied records array until there are no more reverse mappings to
* return or head.fmh_entries == head.fmh_count. In the second case, this
* function returns -ECANCELED to indicate that more records would have been
* returned.
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
*
* Key to Confusion
* ----------------
* There are multiple levels of keys and counters at work here:
* xfs_fsmap_head.fmh_keys -- low and high fsmap keys passed in;
* these reflect fs-wide sector addrs.
* dkeys -- fmh_keys used to query each device;
* these are fmh_keys but w/ the low key
* bumped up by fmr_length.
* xfs_getfsmap_info.next_daddr -- next disk addr we expect to see; this
* is how we detect gaps in the fsmap
records and report them.
* xfs_getfsmap_info.low/high -- per-AG low/high keys computed from
* dkeys; used to query the metadata.
*/
int
xfs_getfsmap(
struct xfs_mount *mp,
struct xfs_fsmap_head *head,
xfs: fix deadlock and streamline xfs_getfsmap performance Refactor xfs_getfsmap to improve its performance: instead of indirectly calling a function that copies one record to userspace at a time, create a shadow buffer in the kernel and copy the whole array once at the end. On the author's computer, this reduces the runtime on his /home by ~20%. This also eliminates a deadlock when running GETFSMAP against the realtime device. The current code locks the rtbitmap to create fsmappings and copies them into userspace, having not released the rtbitmap lock. If the userspace buffer is an mmap of a sparse file that itself resides on the realtime device, the write page fault will recurse into the fs for allocation, which will deadlock on the rtbitmap lock. Fixes: 4c934c7dd60c ("xfs: report realtime space information via the rtbitmap") Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2020-10-01 20:56:07 +03:00
struct fsmap *fsmap_recs)
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
{
struct xfs_trans *tp = NULL;
struct xfs_fsmap dkeys[2]; /* per-dev keys */
struct xfs_getfsmap_dev handlers[XFS_GETFSMAP_DEVS];
xfs: use NULL instead of 0 to initialize a pointer in xfs_getfsmap Found by sparse. Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-04-21 21:24:39 +03:00
struct xfs_getfsmap_info info = { NULL };
xfs: only return detailed fsmap info if the caller has CAP_SYS_ADMIN There were a number of handwaving complaints that one could "possibly" use inode numbers and extent maps to fingerprint a filesystem hosting multiple containers and somehow use the information to guess at the contents of other containers and attack them. Despite the total lack of any demonstration that this is actually possible, it's easier to restrict access now and broaden it later, so use the rmapbt fsmap backends only if the caller has CAP_SYS_ADMIN. Unprivileged users will just have to make do with only getting the free space and static metadata placement information. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>
2017-05-12 20:44:10 +03:00
bool use_rmap;
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
int i;
int error = 0;
if (head->fmh_iflags & ~FMH_IF_VALID)
return -EINVAL;
if (!xfs_getfsmap_is_valid_device(mp, &head->fmh_keys[0]) ||
!xfs_getfsmap_is_valid_device(mp, &head->fmh_keys[1]))
return -EINVAL;
xfs: only return detailed fsmap info if the caller has CAP_SYS_ADMIN There were a number of handwaving complaints that one could "possibly" use inode numbers and extent maps to fingerprint a filesystem hosting multiple containers and somehow use the information to guess at the contents of other containers and attack them. Despite the total lack of any demonstration that this is actually possible, it's easier to restrict access now and broaden it later, so use the rmapbt fsmap backends only if the caller has CAP_SYS_ADMIN. Unprivileged users will just have to make do with only getting the free space and static metadata placement information. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>
2017-05-12 20:44:10 +03:00
use_rmap = capable(CAP_SYS_ADMIN) &&
xfs: replace xfs_sb_version checks with feature flag checks Convert the xfs_sb_version_hasfoo() to checks against mp->m_features. Checks of the superblock itself during disk operations (e.g. in the read/write verifiers and the to/from disk formatters) are not converted - they operate purely on the superblock state. Everything else should use the mount features. Large parts of this conversion were done with sed with commands like this: for f in `git grep -l xfs_sb_version_has fs/xfs/*.c`; do sed -i -e 's/xfs_sb_version_has\(.*\)(&\(.*\)->m_sb)/xfs_has_\1(\2)/' $f done With manual cleanups for things like "xfs_has_extflgbit" and other little inconsistencies in naming. The result is ia lot less typing to check features and an XFS binary size reduced by a bit over 3kB: $ size -t fs/xfs/built-in.a text data bss dec hex filenam before 1130866 311352 484 1442702 16038e (TOTALS) after 1127727 311352 484 1439563 15f74b (TOTALS) Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Darrick J. Wong <djwong@kernel.org> Signed-off-by: Darrick J. Wong <djwong@kernel.org>
2021-08-19 04:46:37 +03:00
xfs_has_rmapbt(mp);
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
head->fmh_entries = 0;
/* Set up our device handlers. */
memset(handlers, 0, sizeof(handlers));
handlers[0].dev = new_encode_dev(mp->m_ddev_targp->bt_dev);
xfs: only return detailed fsmap info if the caller has CAP_SYS_ADMIN There were a number of handwaving complaints that one could "possibly" use inode numbers and extent maps to fingerprint a filesystem hosting multiple containers and somehow use the information to guess at the contents of other containers and attack them. Despite the total lack of any demonstration that this is actually possible, it's easier to restrict access now and broaden it later, so use the rmapbt fsmap backends only if the caller has CAP_SYS_ADMIN. Unprivileged users will just have to make do with only getting the free space and static metadata placement information. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>
2017-05-12 20:44:10 +03:00
if (use_rmap)
xfs: have getfsmap fall back to the freesp btrees when rmap is not present If the reverse-mapping btree isn't available, fall back to the free space btrees to provide partial reverse mapping information. The online scrub tool can make use of even partial information to speed up the data block scan. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
handlers[0].fn = xfs_getfsmap_datadev_rmapbt;
else
handlers[0].fn = xfs_getfsmap_datadev_bnobt;
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
if (mp->m_logdev_targp != mp->m_ddev_targp) {
handlers[1].dev = new_encode_dev(mp->m_logdev_targp->bt_dev);
handlers[1].fn = xfs_getfsmap_logdev;
}
xfs: move more RT specific code under CONFIG_XFS_RT Various utility functions and interfaces that iterate internal devices try to reference the realtime device even when RT support is not compiled into the kernel. Make sure this code is excluded from the CONFIG_XFS_RT=n build, and where appropriate stub functions to return fatal errors if they ever get called when RT support is not present. Signed-Off-By: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-10-09 21:37:22 +03:00
#ifdef CONFIG_XFS_RT
xfs: report realtime space information via the rtbitmap Use the realtime bitmap to return free space information via getfsmap. Eventually this will be superseded by the realtime rmapbt code. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:38 +03:00
if (mp->m_rtdev_targp) {
handlers[2].dev = new_encode_dev(mp->m_rtdev_targp->bt_dev);
handlers[2].fn = xfs_getfsmap_rtdev_rtbitmap;
}
xfs: move more RT specific code under CONFIG_XFS_RT Various utility functions and interfaces that iterate internal devices try to reference the realtime device even when RT support is not compiled into the kernel. Make sure this code is excluded from the CONFIG_XFS_RT=n build, and where appropriate stub functions to return fatal errors if they ever get called when RT support is not present. Signed-Off-By: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-10-09 21:37:22 +03:00
#endif /* CONFIG_XFS_RT */
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
xfs_sort(handlers, XFS_GETFSMAP_DEVS, sizeof(struct xfs_getfsmap_dev),
xfs_getfsmap_dev_compare);
/*
* To continue where we left off, we allow userspace to use the
* last mapping from a previous call as the low key of the next.
* This is identified by a non-zero length in the low key. We
* have to increment the low key in this scenario to ensure we
* don't return the same mapping again, and instead return the
* very next mapping.
*
* If the low key mapping refers to file data, the same physical
* blocks could be mapped to several other files/offsets.
* According to rmapbt record ordering, the minimal next
* possible record for the block range is the next starting
* offset in the same inode. Therefore, bump the file offset to
* continue the search appropriately. For all other low key
* mapping types (attr blocks, metadata), bump the physical
* offset as there can be no other mapping for the same physical
* block range.
*/
dkeys[0] = head->fmh_keys[0];
if (dkeys[0].fmr_flags & (FMR_OF_SPECIAL_OWNER | FMR_OF_EXTENT_MAP)) {
dkeys[0].fmr_physical += dkeys[0].fmr_length;
dkeys[0].fmr_owner = 0;
if (dkeys[0].fmr_offset)
return -EINVAL;
} else
dkeys[0].fmr_offset += dkeys[0].fmr_length;
dkeys[0].fmr_length = 0;
memset(&dkeys[1], 0xFF, sizeof(struct xfs_fsmap));
if (!xfs_getfsmap_check_keys(dkeys, &head->fmh_keys[1]))
return -EINVAL;
info.next_daddr = head->fmh_keys[0].fmr_physical +
head->fmh_keys[0].fmr_length;
xfs: fix deadlock and streamline xfs_getfsmap performance Refactor xfs_getfsmap to improve its performance: instead of indirectly calling a function that copies one record to userspace at a time, create a shadow buffer in the kernel and copy the whole array once at the end. On the author's computer, this reduces the runtime on his /home by ~20%. This also eliminates a deadlock when running GETFSMAP against the realtime device. The current code locks the rtbitmap to create fsmappings and copies them into userspace, having not released the rtbitmap lock. If the userspace buffer is an mmap of a sparse file that itself resides on the realtime device, the write page fault will recurse into the fs for allocation, which will deadlock on the rtbitmap lock. Fixes: 4c934c7dd60c ("xfs: report realtime space information via the rtbitmap") Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Chandan Babu R <chandanrlinux@gmail.com>
2020-10-01 20:56:07 +03:00
info.fsmap_recs = fsmap_recs;
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
info.head = head;
/* For each device we support... */
for (i = 0; i < XFS_GETFSMAP_DEVS; i++) {
/* Is this device within the range the user asked for? */
if (!handlers[i].fn)
continue;
if (head->fmh_keys[0].fmr_device > handlers[i].dev)
continue;
if (head->fmh_keys[1].fmr_device < handlers[i].dev)
break;
/*
* If this device number matches the high key, we have
* to pass the high key to the handler to limit the
* query results. If the device number exceeds the
* low key, zero out the low key so that we get
* everything from the beginning.
*/
if (handlers[i].dev == head->fmh_keys[1].fmr_device)
dkeys[1] = head->fmh_keys[1];
if (handlers[i].dev > head->fmh_keys[0].fmr_device)
memset(&dkeys[0], 0, sizeof(struct xfs_fsmap));
xfs: drop freeze protection when running GETFSMAP A recent log refactoring patchset from Brian Foster relaxed fsfreeze behavior with regards to the buffer cache -- now freeze only waits for pending buffer IO to finish, and does not try to drain the buffer cache LRU. As a result, fsfreeze should no longer stall indefinitely while fsmap runs. Drop the sb_start_write calls around fsmap invocations. While we're cleaning things, add a comment to the xfs_trans_alloc_empty call explaining why we're running around with empty transactions. Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Christoph Hellwig <hch@lst.de>
2021-03-22 19:51:50 +03:00
/*
* Grab an empty transaction so that we can use its recursive
* buffer locking abilities to detect cycles in the rmapbt
* without deadlocking.
*/
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
error = xfs_trans_alloc_empty(mp, &tp);
if (error)
break;
info.dev = handlers[i].dev;
info.last = false;
xfs: pass perags around in fsmap data dev functions Needs a [from, to] ranged AG walk, and the perag to be stuffed into the info structure for callouts to use. Signed-off-by: Dave Chinner <dchinner@redhat.com> Reviewed-by: Brian Foster <bfoster@redhat.com> Reviewed-by: Darrick J. Wong <djwong@kernel.org>
2021-06-02 03:48:24 +03:00
info.pag = NULL;
xfs: implement the GETFSMAP ioctl Introduce a new ioctl that uses the reverse mapping btree to return information about the physical layout of the filesystem. Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com> Reviewed-by: Brian Foster <bfoster@redhat.com>
2017-03-29 00:56:37 +03:00
error = handlers[i].fn(tp, dkeys, &info);
if (error)
break;
xfs_trans_cancel(tp);
tp = NULL;
info.next_daddr = 0;
}
if (tp)
xfs_trans_cancel(tp);
head->fmh_oflags = FMH_OF_DEV_T;
return error;
}
Reference in New Issue Copy Permalink