Files
thirdparty-littlefs/scripts/readtree.py
Christopher Haster f4b6a6b328 Fixed issues with neighbor updates during moves
The root of the problem was some assumptions about what tags could be
sent to lfs_dir_commit.

- The first assumption is that there could be only one splice (create/delete)
  tag at a time, which is trivially broken by the core commit in lfs_rename.

- The second assumption is that there is at most one create and one delete in
  a single commit. This is less obvious but turns out to not be true in
  the case that we rename a file such that it overwrites another file in
  the same directory (1 delete for source file, 1 delete for destination).

- The third assumption was that there was an ordering to the
  delete/creates passed to lfs_dir_commit. It may be possible to force all
  deletes to follow creates by rearranging the tags in lfs_rename, but
  this risks overflowing tag ids.

The way the lfs_dir_commit first collected the "deletetag" and "createtag"
broke all three of these assumptions. And because we lose the ordering
information we can no longer apply the directory changes to open files
correctly. The file ids may be shifted in a way that doesn't reflect the
actual operations on disk.

These problems were made worst by lfs_dir_commit cleaning up moves
implicitly, which also creates deletes implicitly. While cleaning up moves
in lfs_dir_commit may save some code size, it makes the commit logic much more
difficult to implement correctly.

This bug turned into pulling out a dead tree stump, roots and all.

I ended up reworking how lfs_dir_commit updates open files so that it
has less assumptions, now it just traverses the commit tags multiple
times in order to update file ids after a successful commit in the
correct order.

This also got rid of the dir copy by carefully updating split dirs
after all files have an up-to-date copy of the original dir.

I also just removed the implicit move cleanup. It turns out the only
commits that can occur before we have cleaned up the move is in
lfs_fs_relocate, so it was simple enough to explicitly handle this case
when we update our parent and pred during a relocate.

Cases where we may need to fix moves:
- In lfs_rename when we move a file/dir
- In lfs_demove if we lose power
- In lfs_fs_relocate if we have to relocate our parent and we find it
  had a pending move (or else the move will be outdated)
- In lfs_fs_relocate if we have to relocate our predecessor and we find it
  had a pending move (or else the move will be outdated)

Note the two cases in lfs_fs_relocate may be recursive. But
lfs_fs_relocate can only trigger other lfs_fs_relocates so it's not
possible for pending moves to spill out into other filesystem commits

And of couse, I added several tests to cover these situations. Hopefully
the rename-with-open-files logic should be fairly locked down now.

found with initial fix by eastmoutain
2020-01-20 19:27:27 -06:00

279 lines
10 KiB
Python
Executable File

#!/usr/bin/env python3
import struct
import sys
import json
import io
import itertools as it
from readmdir import Tag, MetadataPair
def popc(x):
return bin(x).count('1')
def ctz(x):
return len(bin(x)) - len(bin(x).rstrip('0'))
def dumptags(args, mdir, f):
if args.all:
tags = mdir.all_
elif args.log:
tags = mdir.log
else:
tags = mdir.tags
for k, tag in enumerate(tags):
f.write("tag %08x %s" % (tag, tag.typerepr()))
if tag.id != 0x3ff:
f.write(" id %d" % tag.id)
if tag.size != 0x3ff:
f.write(" size %d" % tag.size)
if tag.is_('name'):
f.write(" name %s" %
json.dumps(tag.data.decode('utf8')))
if tag.is_('dirstruct'):
f.write(" dir {%#x, %#x}" % struct.unpack(
'<II', tag.data[:8].ljust(8, b'\xff')))
if tag.is_('ctzstruct'):
f.write(" ctz {%#x} size %d" % struct.unpack(
'<II', tag.data[:8].ljust(8, b'\xff')))
if tag.is_('inlinestruct'):
f.write(" inline size %d" % tag.size)
if tag.is_('gstate'):
f.write(" 0x%s" % ''.join('%02x' % c for c in tag.data))
if tag.is_('tail'):
f.write(" tail {%#x, %#x}" % struct.unpack(
'<II', tag.data[:8].ljust(8, b'\xff')))
f.write("\n")
if args.data:
for i in range(0, len(tag.data), 16):
f.write(" %-47s %-16s\n" % (
' '.join('%02x' % c for c in tag.data[i:i+16]),
''.join(c if c >= ' ' and c <= '~' else '.'
for c in map(chr, tag.data[i:i+16]))))
def dumpentries(args, mdir, f):
for k, id_ in enumerate(mdir.ids):
name = mdir[Tag('name', id_, 0)]
struct_ = mdir[Tag('struct', id_, 0)]
f.write("id %d %s %s" % (
id_, name.typerepr(),
json.dumps(name.data.decode('utf8'))))
if struct_.is_('dirstruct'):
f.write(" dir {%#x, %#x}" % struct.unpack(
'<II', struct_.data[:8].ljust(8, b'\xff')))
if struct_.is_('ctzstruct'):
f.write(" ctz {%#x} size %d" % struct.unpack(
'<II', struct_.data[:8].ljust(8, b'\xff')))
if struct_.is_('inlinestruct'):
f.write(" inline size %d" % struct_.size)
f.write("\n")
if args.data and struct_.is_('inlinestruct'):
for i in range(0, len(struct_.data), 16):
f.write(" %-47s %-16s\n" % (
' '.join('%02x' % c for c in struct_.data[i:i+16]),
''.join(c if c >= ' ' and c <= '~' else '.'
for c in map(chr, struct_.data[i:i+16]))))
elif args.data and struct_.is_('ctzstruct'):
block, size = struct.unpack(
'<II', struct_.data[:8].ljust(8, b'\xff'))
data = []
i = 0 if size == 0 else (size-1) // (args.block_size - 8)
if i != 0:
i = ((size-1) - 4*popc(i-1)+2) // (args.block_size - 8)
with open(args.disk, 'rb') as f2:
while i >= 0:
f2.seek(block * args.block_size)
dat = f2.read(args.block_size)
data.append(dat[4*(ctz(i)+1) if i != 0 else 0:])
block, = struct.unpack('<I', dat[:4].ljust(4, b'\xff'))
i -= 1
data = bytes(it.islice(
it.chain.from_iterable(reversed(data)), size))
for i in range(0, min(len(data), 256)
if not args.no_truncate else len(data), 16):
f.write(" %-47s %-16s\n" % (
' '.join('%02x' % c for c in data[i:i+16]),
''.join(c if c >= ' ' and c <= '~' else '.'
for c in map(chr, data[i:i+16]))))
for tag in mdir.tags:
if tag.id==id_ and tag.is_('userattr'):
f.write("id %d %s size %d\n" % (
id_, tag.typerepr(), tag.size))
if args.data:
for i in range(0, len(tag.data), 16):
f.write(" %-47s %-16s\n" % (
' '.join('%02x' % c for c in tag.data[i:i+16]),
''.join(c if c >= ' ' and c <= '~' else '.'
for c in map(chr, tag.data[i:i+16]))))
def main(args):
with open(args.disk, 'rb') as f:
dirs = []
superblock = None
gstate = b''
mdirs = []
tail = (args.block1, args.block2)
hard = False
while True:
# load mdir
data = []
blocks = {}
for block in tail:
f.seek(block * args.block_size)
data.append(f.read(args.block_size)
.ljust(args.block_size, b'\xff'))
blocks[id(data[-1])] = block
mdir = MetadataPair(data)
mdir.blocks = tuple(blocks[id(p.data)] for p in mdir.pair)
# fetch some key metadata as a we scan
try:
mdir.tail = mdir[Tag('tail', 0, 0)]
if mdir.tail.size != 8 or mdir.tail.data == 8*b'\xff':
mdir.tail = None
except KeyError:
mdir.tail = None
# have superblock?
try:
nsuperblock = mdir[
Tag(0x7ff, 0x3ff, 0), Tag('superblock', 0, 0)]
superblock = nsuperblock, mdir[Tag('inlinestruct', 0, 0)]
except KeyError:
pass
# have gstate?
try:
ngstate = mdir[Tag('movestate', 0, 0)]
gstate = bytes((a or 0) ^ (b or 0)
for a,b in it.zip_longest(gstate, ngstate.data))
except KeyError:
pass
# add to directories
mdirs.append(mdir)
if mdir.tail is None or not mdir.tail.is_('hardtail'):
dirs.append(mdirs)
mdirs = []
if mdir.tail is None:
break
tail = struct.unpack('<II', mdir.tail.data)
hard = mdir.tail.is_('hardtail')
# find paths
dirtable = {}
for dir in dirs:
dirtable[tuple(sorted(dir[0].blocks))] = dir
pending = [("/", dirs[0])]
while pending:
path, dir = pending.pop(0)
for mdir in dir:
for tag in mdir.tags:
if tag.is_('dir'):
try:
npath = tag.data.decode('utf8')
dirstruct = mdir[Tag('dirstruct', tag.id, 0)]
nblocks = struct.unpack('<II', dirstruct.data)
nmdir = dirtable[tuple(sorted(nblocks))]
pending.append(((path + '/' + npath), nmdir))
except KeyError:
pass
dir[0].path = path.replace('//', '/')
# dump tree
if not args.superblock and not args.gstate and not args.mdirs:
args.superblock = True
args.gstate = True
args.mdirs = True
if args.superblock and superblock:
print("superblock %s" % json.dumps(superblock[0].data.decode('utf8')))
print(
" version v{1}.{0}\n"
" block_size {2}\n"
" block_count {3}\n"
" name_max {4}\n"
" file_max {5}\n"
" attr_max {6}"
.format(*struct.unpack(
'<HHIIIII', superblock[1].data[:24].ljust(24, b'\xff'))))
if args.gstate and gstate:
print("gstate 0x%s" % ''.join('%02x' % c for c in gstate))
tag = Tag(struct.unpack('<I', gstate[0:4].ljust(4, b'\xff'))[0])
blocks = struct.unpack('<II', gstate[4:4+8].ljust(8, b'\xff'))
if tag.size:
print(" orphans %d" % tag.size)
if tag.type:
print(" move dir {%#x, %#x} id %d" % (
blocks[0], blocks[1], tag.id))
if args.mdirs:
for i, dir in enumerate(dirs):
print("dir %s" % (json.dumps(dir[0].path)
if hasattr(dir[0], 'path') else '(orphan)'))
for j, mdir in enumerate(dir):
print("mdir {%#x, %#x} rev %d%s" % (
mdir.blocks[0], mdir.blocks[1], mdir.rev,
' (corrupted)' if not mdir else ''))
f = io.StringIO()
if args.tags or args.all or args.log:
dumptags(args, mdir, f)
else:
dumpentries(args, mdir, f)
lines = list(filter(None, f.getvalue().split('\n')))
for k, line in enumerate(lines):
print("%s %s" % (
' ' if j == len(dir)-1 else
'v' if k == len(lines)-1 else
'|',
line))
return 0 if all(mdir for dir in dirs for mdir in dir) else 1
if __name__ == "__main__":
import argparse
import sys
parser = argparse.ArgumentParser(
description="Dump semantic info about the metadata tree in littlefs")
parser.add_argument('disk',
help="File representing the block device.")
parser.add_argument('block_size', type=lambda x: int(x, 0),
help="Size of a block in bytes.")
parser.add_argument('block1', nargs='?', default=0,
type=lambda x: int(x, 0),
help="Optional first block address for finding the root.")
parser.add_argument('block2', nargs='?', default=1,
type=lambda x: int(x, 0),
help="Optional second block address for finding the root.")
parser.add_argument('-s', '--superblock', action='store_true',
help="Show contents of the superblock.")
parser.add_argument('-g', '--gstate', action='store_true',
help="Show contents of global-state.")
parser.add_argument('-m', '--mdirs', action='store_true',
help="Show contents of metadata-pairs/directories.")
parser.add_argument('-t', '--tags', action='store_true',
help="Show metadata tags instead of reconstructing entries.")
parser.add_argument('-a', '--all', action='store_true',
help="Show all tags in log, included tags in corrupted commits.")
parser.add_argument('-l', '--log', action='store_true',
help="Show tags in log.")
parser.add_argument('-d', '--data', action='store_true',
help="Also show the raw contents of files/attrs/tags.")
parser.add_argument('-T', '--no-truncate', action='store_true',
help="Don't truncate large amounts of data in files.")
sys.exit(main(parser.parse_args()))