mirror of
https://github.com/kdave/btrfs-progs
synced 2024-12-15 10:55:51 +00:00
b3477244f9
The in-kernel version of read_tree_block adds some extra sanity checks to make sure we don't return blocks that don't match what we expect. This includes the owning root, the level, and the expected first key. We don't actually do these checks in btrfs-progs, however kernel code we're going to sync will expect this calling convention, so update it to match the in-kernel code and then update all the callers. Signed-off-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: David Sterba <dsterba@suse.com>
514 lines
14 KiB
C
514 lines
14 KiB
C
/*
|
|
* Copyright (C) 2011 Red Hat. All rights reserved.
|
|
*
|
|
* This program is free software; you can redistribute it and/or
|
|
* modify it under the terms of the GNU General Public
|
|
* License v2 as published by the Free Software Foundation.
|
|
*
|
|
* This program is distributed in the hope that it will be useful,
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
* General Public License for more details.
|
|
*
|
|
* You should have received a copy of the GNU General Public
|
|
* License along with this program; if not, write to the
|
|
* Free Software Foundation, Inc., 59 Temple Place - Suite 330,
|
|
* Boston, MA 021110-1307, USA.
|
|
*/
|
|
|
|
#include "kerncompat.h"
|
|
#include <sys/time.h>
|
|
#include <stdio.h>
|
|
#include <stdlib.h>
|
|
#include <errno.h>
|
|
#include <getopt.h>
|
|
#include <string.h>
|
|
#include "kernel-lib/rbtree.h"
|
|
#include "kernel-lib/rbtree_types.h"
|
|
#include "kernel-shared/ctree.h"
|
|
#include "kernel-shared/disk-io.h"
|
|
#include "kernel-shared/extent_io.h"
|
|
#include "kernel-shared/file-item.h"
|
|
#include "common/utils.h"
|
|
#include "common/help.h"
|
|
#include "common/messages.h"
|
|
#include "common/open-utils.h"
|
|
#include "common/units.h"
|
|
#include "cmds/commands.h"
|
|
|
|
static int verbose = 0;
|
|
static bool no_pretty = false;
|
|
|
|
struct seek {
|
|
u64 distance;
|
|
u64 count;
|
|
struct rb_node n;
|
|
};
|
|
|
|
struct root_stats {
|
|
u64 total_nodes;
|
|
u64 total_bytes;
|
|
u64 total_inline;
|
|
u64 total_seeks;
|
|
u64 forward_seeks;
|
|
u64 backward_seeks;
|
|
u64 total_seek_len;
|
|
u64 max_seek_len;
|
|
u64 total_clusters;
|
|
u64 total_cluster_size;
|
|
u64 min_cluster_size;
|
|
u64 max_cluster_size;
|
|
u64 lowest_bytenr;
|
|
u64 highest_bytenr;
|
|
u64 node_counts[BTRFS_MAX_LEVEL];
|
|
struct rb_root seek_root;
|
|
int total_levels;
|
|
};
|
|
|
|
static int add_seek(struct rb_root *root, u64 dist)
|
|
{
|
|
struct rb_node **p = &root->rb_node;
|
|
struct rb_node *parent = NULL;
|
|
struct seek *seek = NULL;
|
|
|
|
while (*p) {
|
|
parent = *p;
|
|
seek = rb_entry(parent, struct seek, n);
|
|
|
|
if (dist < seek->distance) {
|
|
p = &(*p)->rb_left;
|
|
} else if (dist > seek->distance) {
|
|
p = &(*p)->rb_right;
|
|
} else {
|
|
seek->count++;
|
|
return 0;
|
|
}
|
|
}
|
|
|
|
seek = malloc(sizeof(struct seek));
|
|
if (!seek)
|
|
return -ENOMEM;
|
|
seek->distance = dist;
|
|
seek->count = 1;
|
|
rb_link_node(&seek->n, parent, p);
|
|
rb_insert_color(&seek->n, root);
|
|
return 0;
|
|
}
|
|
|
|
static int walk_leaf(struct btrfs_root *root, struct btrfs_path *path,
|
|
struct root_stats *stat, int find_inline)
|
|
{
|
|
struct extent_buffer *b = path->nodes[0];
|
|
struct btrfs_file_extent_item *fi;
|
|
struct btrfs_key found_key;
|
|
int i;
|
|
|
|
stat->total_bytes += root->fs_info->nodesize;
|
|
stat->total_nodes++;
|
|
stat->node_counts[0]++;
|
|
|
|
if (!find_inline)
|
|
return 0;
|
|
|
|
for (i = 0; i < btrfs_header_nritems(b); i++) {
|
|
btrfs_item_key_to_cpu(b, &found_key, i);
|
|
if (found_key.type != BTRFS_EXTENT_DATA_KEY)
|
|
continue;
|
|
|
|
fi = btrfs_item_ptr(b, i, struct btrfs_file_extent_item);
|
|
if (btrfs_file_extent_type(b, fi) == BTRFS_FILE_EXTENT_INLINE)
|
|
stat->total_inline +=
|
|
btrfs_file_extent_inline_item_len(b, i);
|
|
}
|
|
|
|
return 0;
|
|
}
|
|
|
|
static u64 calc_distance(u64 block1, u64 block2)
|
|
{
|
|
if (block1 < block2)
|
|
return block2 - block1;
|
|
return block1 - block2;
|
|
}
|
|
|
|
static int walk_nodes(struct btrfs_root *root, struct btrfs_path *path,
|
|
struct root_stats *stat, int level, int find_inline)
|
|
{
|
|
struct extent_buffer *b = path->nodes[level];
|
|
u32 nodesize = root->fs_info->nodesize;
|
|
u64 last_block;
|
|
u64 cluster_size = nodesize;
|
|
int i;
|
|
int ret = 0;
|
|
|
|
stat->total_bytes += nodesize;
|
|
stat->total_nodes++;
|
|
stat->node_counts[level]++;
|
|
|
|
last_block = btrfs_header_bytenr(b);
|
|
for (i = 0; i < btrfs_header_nritems(b); i++) {
|
|
struct extent_buffer *tmp = NULL;
|
|
u64 cur_blocknr = btrfs_node_blockptr(b, i);
|
|
|
|
path->slots[level] = i;
|
|
if ((level - 1) > 0 || find_inline) {
|
|
tmp = read_tree_block(root->fs_info, cur_blocknr,
|
|
btrfs_header_owner(b),
|
|
btrfs_node_ptr_generation(b, i),
|
|
level - 1, NULL);
|
|
if (!extent_buffer_uptodate(tmp)) {
|
|
error("failed to read blocknr %llu",
|
|
btrfs_node_blockptr(b, i));
|
|
continue;
|
|
}
|
|
path->nodes[level - 1] = tmp;
|
|
}
|
|
if (level - 1)
|
|
ret = walk_nodes(root, path, stat, level - 1,
|
|
find_inline);
|
|
else
|
|
ret = walk_leaf(root, path, stat, find_inline);
|
|
if (last_block + nodesize != cur_blocknr) {
|
|
u64 distance = calc_distance(last_block +
|
|
nodesize,
|
|
cur_blocknr);
|
|
stat->total_seeks++;
|
|
stat->total_seek_len += distance;
|
|
if (stat->max_seek_len < distance)
|
|
stat->max_seek_len = distance;
|
|
if (add_seek(&stat->seek_root, distance)) {
|
|
error("cannot add new seek at distance %llu", distance);
|
|
ret = -ENOMEM;
|
|
break;
|
|
}
|
|
|
|
if (last_block < cur_blocknr)
|
|
stat->forward_seeks++;
|
|
else
|
|
stat->backward_seeks++;
|
|
if (cluster_size != nodesize) {
|
|
stat->total_cluster_size += cluster_size;
|
|
stat->total_clusters++;
|
|
if (cluster_size < stat->min_cluster_size)
|
|
stat->min_cluster_size = cluster_size;
|
|
if (cluster_size > stat->max_cluster_size)
|
|
stat->max_cluster_size = cluster_size;
|
|
}
|
|
cluster_size = nodesize;
|
|
} else {
|
|
cluster_size += nodesize;
|
|
}
|
|
last_block = cur_blocknr;
|
|
if (cur_blocknr < stat->lowest_bytenr)
|
|
stat->lowest_bytenr = cur_blocknr;
|
|
if (cur_blocknr > stat->highest_bytenr)
|
|
stat->highest_bytenr = cur_blocknr;
|
|
free_extent_buffer(tmp);
|
|
if (ret) {
|
|
error("walking down path failed: %d", ret);
|
|
break;
|
|
}
|
|
}
|
|
|
|
return ret;
|
|
}
|
|
|
|
static void print_seek_histogram(struct root_stats *stat)
|
|
{
|
|
struct rb_node *n = rb_first(&stat->seek_root);
|
|
struct seek *seek;
|
|
u64 tick_interval;
|
|
u64 group_start = 0;
|
|
u64 group_count = 0;
|
|
u64 group_end = 0;
|
|
u64 i;
|
|
u64 max_seek = stat->max_seek_len;
|
|
int digits = 1;
|
|
|
|
if (stat->total_seeks < 20)
|
|
return;
|
|
|
|
while ((max_seek /= 10))
|
|
digits++;
|
|
|
|
/* Make a tick count as 5% of the total seeks */
|
|
tick_interval = stat->total_seeks / 20;
|
|
pr_verbose(LOG_DEFAULT, "\tSeek histogram\n");
|
|
for (; n; n = rb_next(n)) {
|
|
u64 ticks, gticks = 0;
|
|
|
|
seek = rb_entry(n, struct seek, n);
|
|
ticks = seek->count / tick_interval;
|
|
if (group_count)
|
|
gticks = group_count / tick_interval;
|
|
|
|
if (ticks <= 2 && gticks <= 2) {
|
|
if (group_count == 0)
|
|
group_start = seek->distance;
|
|
group_end = seek->distance;
|
|
group_count += seek->count;
|
|
continue;
|
|
}
|
|
|
|
if (group_count) {
|
|
|
|
gticks = group_count / tick_interval;
|
|
pr_verbose(LOG_DEFAULT, "\t\t%*llu - %*llu: %*llu ", digits, group_start,
|
|
digits, group_end, digits, group_count);
|
|
if (gticks) {
|
|
for (i = 0; i < gticks; i++)
|
|
pr_verbose(LOG_DEFAULT, "#");
|
|
pr_verbose(LOG_DEFAULT, "\n");
|
|
} else {
|
|
pr_verbose(LOG_DEFAULT, "|\n");
|
|
}
|
|
group_count = 0;
|
|
}
|
|
|
|
if (ticks <= 2)
|
|
continue;
|
|
|
|
pr_verbose(LOG_DEFAULT, "\t\t%*llu - %*llu: %*llu ", digits, seek->distance,
|
|
digits, seek->distance, digits, seek->count);
|
|
for (i = 0; i < ticks; i++)
|
|
pr_verbose(LOG_DEFAULT, "#");
|
|
pr_verbose(LOG_DEFAULT, "\n");
|
|
}
|
|
if (group_count) {
|
|
u64 gticks;
|
|
|
|
gticks = group_count / tick_interval;
|
|
pr_verbose(LOG_DEFAULT, "\t\t%*llu - %*llu: %*llu ", digits, group_start,
|
|
digits, group_end, digits, group_count);
|
|
if (gticks) {
|
|
for (i = 0; i < gticks; i++)
|
|
pr_verbose(LOG_DEFAULT, "#");
|
|
pr_verbose(LOG_DEFAULT, "\n");
|
|
} else {
|
|
pr_verbose(LOG_DEFAULT, "|\n");
|
|
}
|
|
group_count = 0;
|
|
}
|
|
}
|
|
|
|
static void timeval_subtract(struct timeval *result, struct timeval *x,
|
|
struct timeval *y)
|
|
{
|
|
if (x->tv_usec < y->tv_usec) {
|
|
int nsec = (y->tv_usec - x->tv_usec) / 1000000 + 1;
|
|
y->tv_usec -= 1000000 * nsec;
|
|
y->tv_sec += nsec;
|
|
}
|
|
|
|
if (x->tv_usec - y->tv_usec > 1000000) {
|
|
int nsec = (x->tv_usec - y->tv_usec) / 1000000;
|
|
y->tv_usec += 1000000 * nsec;
|
|
y->tv_sec -= nsec;
|
|
}
|
|
|
|
result->tv_sec = x->tv_sec - y->tv_sec;
|
|
result->tv_usec = x->tv_usec - y->tv_usec;
|
|
}
|
|
|
|
static int calc_root_size(struct btrfs_root *tree_root, struct btrfs_key *key,
|
|
int find_inline)
|
|
{
|
|
struct btrfs_root *root;
|
|
struct btrfs_path path;
|
|
struct rb_node *n;
|
|
struct timeval start, end, diff = {0};
|
|
struct root_stats stat;
|
|
int level;
|
|
int ret = 0;
|
|
int size_fail = 0;
|
|
int i;
|
|
|
|
root = btrfs_read_fs_root(tree_root->fs_info, key);
|
|
if (IS_ERR(root)) {
|
|
error("failed to read root %llu", key->objectid);
|
|
return 1;
|
|
}
|
|
|
|
btrfs_init_path(&path);
|
|
memset(&stat, 0, sizeof(stat));
|
|
level = btrfs_header_level(root->node);
|
|
stat.lowest_bytenr = btrfs_header_bytenr(root->node);
|
|
stat.highest_bytenr = stat.lowest_bytenr;
|
|
stat.min_cluster_size = (u64)-1;
|
|
stat.max_cluster_size = root->fs_info->nodesize;
|
|
path.nodes[level] = root->node;
|
|
if (gettimeofday(&start, NULL)) {
|
|
error("cannot get time: %m");
|
|
goto out;
|
|
}
|
|
if (!level) {
|
|
ret = walk_leaf(root, &path, &stat, find_inline);
|
|
if (ret)
|
|
goto out;
|
|
goto out_print;
|
|
}
|
|
|
|
ret = walk_nodes(root, &path, &stat, level, find_inline);
|
|
if (ret)
|
|
goto out;
|
|
if (gettimeofday(&end, NULL)) {
|
|
error("cannot get time: %m");
|
|
goto out;
|
|
}
|
|
timeval_subtract(&diff, &end, &start);
|
|
out_print:
|
|
if (stat.min_cluster_size == (u64)-1) {
|
|
stat.min_cluster_size = 0;
|
|
stat.total_clusters = 1;
|
|
}
|
|
|
|
if (no_pretty || size_fail) {
|
|
pr_verbose(LOG_DEFAULT, "\tTotal size: %llu\n", stat.total_bytes);
|
|
pr_verbose(LOG_DEFAULT, "\t\tInline data: %llu\n", stat.total_inline);
|
|
pr_verbose(LOG_DEFAULT, "\tTotal seeks: %llu\n", stat.total_seeks);
|
|
pr_verbose(LOG_DEFAULT, "\t\tForward seeks: %llu\n", stat.forward_seeks);
|
|
pr_verbose(LOG_DEFAULT, "\t\tBackward seeks: %llu\n", stat.backward_seeks);
|
|
pr_verbose(LOG_DEFAULT, "\t\tAvg seek len: %llu\n", stat.total_seeks ?
|
|
stat.total_seek_len / stat.total_seeks : 0);
|
|
print_seek_histogram(&stat);
|
|
pr_verbose(LOG_DEFAULT, "\tTotal clusters: %llu\n", stat.total_clusters);
|
|
pr_verbose(LOG_DEFAULT, "\t\tAvg cluster size: %llu\n", stat.total_cluster_size /
|
|
stat.total_clusters);
|
|
pr_verbose(LOG_DEFAULT, "\t\tMin cluster size: %llu\n", stat.min_cluster_size);
|
|
pr_verbose(LOG_DEFAULT, "\t\tMax cluster size: %llu\n", stat.max_cluster_size);
|
|
pr_verbose(LOG_DEFAULT, "\tTotal disk spread: %llu\n", stat.highest_bytenr -
|
|
stat.lowest_bytenr);
|
|
pr_verbose(LOG_DEFAULT, "\tTotal read time: %d s %d us\n", (int)diff.tv_sec,
|
|
(int)diff.tv_usec);
|
|
} else {
|
|
pr_verbose(LOG_DEFAULT, "\tTotal size: %s\n", pretty_size(stat.total_bytes));
|
|
pr_verbose(LOG_DEFAULT, "\t\tInline data: %s\n", pretty_size(stat.total_inline));
|
|
pr_verbose(LOG_DEFAULT, "\tTotal seeks: %llu\n", stat.total_seeks);
|
|
pr_verbose(LOG_DEFAULT, "\t\tForward seeks: %llu\n", stat.forward_seeks);
|
|
pr_verbose(LOG_DEFAULT, "\t\tBackward seeks: %llu\n", stat.backward_seeks);
|
|
pr_verbose(LOG_DEFAULT, "\t\tAvg seek len: %s\n", stat.total_seeks ?
|
|
pretty_size(stat.total_seek_len / stat.total_seeks) :
|
|
pretty_size(0));
|
|
print_seek_histogram(&stat);
|
|
pr_verbose(LOG_DEFAULT, "\tTotal clusters: %llu\n", stat.total_clusters);
|
|
pr_verbose(LOG_DEFAULT, "\t\tAvg cluster size: %s\n",
|
|
pretty_size((stat.total_cluster_size /
|
|
stat.total_clusters)));
|
|
pr_verbose(LOG_DEFAULT, "\t\tMin cluster size: %s\n",
|
|
pretty_size(stat.min_cluster_size));
|
|
pr_verbose(LOG_DEFAULT, "\t\tMax cluster size: %s\n",
|
|
pretty_size(stat.max_cluster_size));
|
|
pr_verbose(LOG_DEFAULT, "\tTotal disk spread: %s\n",
|
|
pretty_size(stat.highest_bytenr -
|
|
stat.lowest_bytenr));
|
|
pr_verbose(LOG_DEFAULT, "\tTotal read time: %d s %d us\n", (int)diff.tv_sec,
|
|
(int)diff.tv_usec);
|
|
}
|
|
pr_verbose(LOG_DEFAULT, "\tLevels: %d\n", level + 1);
|
|
pr_verbose(LOG_DEFAULT, "\tTotal nodes: %llu\n", stat.total_nodes);
|
|
for (i = 0; i < level + 1; i++) {
|
|
pr_verbose(LOG_DEFAULT, "\t\tOn level %d: %8llu", i, stat.node_counts[i]);
|
|
if (i > 0) {
|
|
u64 fanout;
|
|
|
|
fanout = stat.node_counts[i - 1];
|
|
fanout /= stat.node_counts[i];
|
|
pr_verbose(LOG_DEFAULT, " (avg fanout %llu)", fanout);
|
|
}
|
|
pr_verbose(LOG_DEFAULT, "\n");
|
|
}
|
|
out:
|
|
while ((n = rb_first(&stat.seek_root)) != NULL) {
|
|
struct seek *seek = rb_entry(n, struct seek, n);
|
|
rb_erase(n, &stat.seek_root);
|
|
free(seek);
|
|
}
|
|
|
|
/*
|
|
* We only use path to save node data in iterating, without holding
|
|
* eb's ref_cnt in path. Don't use btrfs_release_path() here, it will
|
|
* free these eb again, and cause many problems, as negative ref_cnt or
|
|
* invalid memory access.
|
|
*/
|
|
return ret;
|
|
}
|
|
|
|
static const char * const cmd_inspect_tree_stats_usage[] = {
|
|
"btrfs inspect-internal tree-stats [options] <device>",
|
|
"Print various stats for trees",
|
|
"",
|
|
OPTLINE("-b", "raw numbers in bytes"),
|
|
NULL
|
|
};
|
|
|
|
static int cmd_inspect_tree_stats(const struct cmd_struct *cmd,
|
|
int argc, char **argv)
|
|
{
|
|
struct btrfs_key key = { .type = BTRFS_ROOT_ITEM_KEY };
|
|
struct btrfs_root *root;
|
|
int opt;
|
|
int ret = 0;
|
|
|
|
optind = 0;
|
|
while ((opt = getopt(argc, argv, "vb")) != -1) {
|
|
switch (opt) {
|
|
case 'v':
|
|
verbose++;
|
|
break;
|
|
case 'b':
|
|
no_pretty = true;
|
|
break;
|
|
default:
|
|
usage_unknown_option(cmd, argv);
|
|
}
|
|
}
|
|
|
|
if (check_argc_exact(argc - optind, 1))
|
|
return 1;
|
|
|
|
ret = check_mounted(argv[optind]);
|
|
if (ret < 0) {
|
|
errno = -ret;
|
|
warning("unable to check mount status of: %m");
|
|
} else if (ret) {
|
|
warning("%s already mounted, tree-stats accesses the block devices directly, this may\n"
|
|
"\tresult in inaccurate numbers, various errors or it may crash if the filesystem\n"
|
|
"\tchanges unexpectedly, restart if needed or remount read-only", argv[optind]);
|
|
}
|
|
|
|
root = open_ctree(argv[optind], 0, 0);
|
|
if (!root) {
|
|
error("cannot open ctree");
|
|
exit(1);
|
|
}
|
|
|
|
pr_verbose(LOG_DEFAULT, "Calculating size of root tree\n");
|
|
key.objectid = BTRFS_ROOT_TREE_OBJECTID;
|
|
ret = calc_root_size(root, &key, 0);
|
|
if (ret)
|
|
goto out;
|
|
|
|
pr_verbose(LOG_DEFAULT, "Calculating size of extent tree\n");
|
|
key.objectid = BTRFS_EXTENT_TREE_OBJECTID;
|
|
ret = calc_root_size(root, &key, 0);
|
|
if (ret)
|
|
goto out;
|
|
|
|
pr_verbose(LOG_DEFAULT, "Calculating size of csum tree\n");
|
|
key.objectid = BTRFS_CSUM_TREE_OBJECTID;
|
|
ret = calc_root_size(root, &key, 0);
|
|
if (ret)
|
|
goto out;
|
|
|
|
key.objectid = BTRFS_FS_TREE_OBJECTID;
|
|
key.offset = (u64)-1;
|
|
pr_verbose(LOG_DEFAULT, "Calculating size of fs tree\n");
|
|
ret = calc_root_size(root, &key, 1);
|
|
if (ret)
|
|
goto out;
|
|
out:
|
|
close_ctree(root);
|
|
return ret;
|
|
}
|
|
DEFINE_SIMPLE_COMMAND(inspect_tree_stats, "tree-stats");
|