mirror/borg
1
0
Fork 0
mirror of https://github.com/borgbackup/borg.git synced 2024-12-27 02:08:54 +00:00
Code Issues Releases Wiki Activity

Outsource recursive directory walking (#5492)

Browse source 
Split recursive directory walking/processing into walking and item processing.
This commit is contained in:
Lapinot 2020-11-15 15:31:01 +01:00 committed by GitHub
parent b8fbe198dd
commit 34f6cfcd81
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
2 changed files with 105 additions and 88 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

15
src/borg/archive.py
View file

@ -38,7 +38,7 @@
from .helpers import bin_to_hex
from .helpers import safe_ns
from .helpers import ellipsis_truncate, ProgressIndicatorPercent, log_multi
from .helpers import os_open, flags_normal
from .helpers import os_open, flags_normal, flags_dir
from .helpers import msgpack
from .helpers import sig_int
from .patterns import PathPrefixPattern, FnmatchPattern, IECommand
@ -1200,11 +1200,22 @@ def create_helper(self, path, st, status=None, hardlinkable=True):
if hardlink_master:
self.hard_links[(st.st_ino, st.st_dev)] = safe_path
def process_dir(self, *, path, fd, st):
def process_dir_with_fd(self, *, path, fd, st):
with self.create_helper(path, st, 'd', hardlinkable=False) as (item, status, hardlinked, hardlink_master):
item.update(self.metadata_collector.stat_attrs(st, path, fd=fd))
return status
def process_dir(self, *, path, parent_fd, name, st):
with self.create_helper(path, st, 'd', hardlinkable=False) as (item, status, hardlinked, hardlink_master):
with OsOpen(path=path, parent_fd=parent_fd, name=name, flags=flags_dir,
noatime=True, op='dir_open') as fd:
# fd is None for directories on windows, in that case a race condition check is not possible.
if fd is not None:
with backup_io('fstat'):
st = stat_update_check(st, os.fstat(fd))
item.update(self.metadata_collector.stat_attrs(st, path, fd=fd))
return status
def process_fifo(self, *, path, parent_fd, name, st):
with self.create_helper(path, st, 'f') as (item, status, hardlinked, hardlink_master): # fifo
with OsOpen(path=path, parent_fd=parent_fd, name=name, flags=flags_normal, noatime=True) as fd:

178
src/borg/archiver.py
View file

@ -565,11 +565,11 @@ def create_inner(archive, cache, fso):
restrict_dev = st.st_dev
else:
restrict_dev = None
self._process(path=path, parent_fd=parent_fd, name=name,
fso=fso, cache=cache, matcher=matcher,
exclude_caches=args.exclude_caches, exclude_if_present=args.exclude_if_present,
keep_exclude_tags=args.keep_exclude_tags, skip_inodes=skip_inodes,
restrict_dev=restrict_dev, read_special=args.read_special, dry_run=dry_run)
self._rec_walk(path=path, parent_fd=parent_fd, name=name,
fso=fso, cache=cache, matcher=matcher,
exclude_caches=args.exclude_caches, exclude_if_present=args.exclude_if_present,
keep_exclude_tags=args.keep_exclude_tags, skip_inodes=skip_inodes,
restrict_dev=restrict_dev, read_special=args.read_special, dry_run=dry_run)
if not dry_run:
if args.progress:
archive.stats.show_progress(final=True)
@ -626,10 +626,66 @@ def create_inner(archive, cache, fso):
create_inner(None, None, None)
return self.exit_code
def _process(self, *, path, parent_fd=None, name=None,
fso, cache, matcher,
exclude_caches, exclude_if_present, keep_exclude_tags, skip_inodes,
restrict_dev, read_special=False, dry_run=False):
def _process_any(self, *, path, parent_fd, name, st, fso, cache, read_special, dry_run):
"""
Call the right method on the given FilesystemObjectProcessor.
"""
if dry_run:
return '-'
elif stat.S_ISREG(st.st_mode):
return fso.process_file(path=path, parent_fd=parent_fd, name=name, st=st, cache=cache)
elif stat.S_ISDIR(st.st_mode):
return fso.process_dir(path=path, parent_fd=parent_fd, name=name, st=st)
elif stat.S_ISLNK(st.st_mode):
if not read_special:
return fso.process_symlink(path=path, parent_fd=parent_fd, name=name, st=st)
else:
try:
st_target = os.stat(name, dir_fd=parent_fd, follow_symlinks=True)
except OSError:
special = False
else:
special = is_special(st_target.st_mode)
if special:
return fso.process_file(path=path, parent_fd=parent_fd, name=name, st=st_target,
cache=cache, flags=flags_special_follow)
else:
return fso.process_symlink(path=path, parent_fd=parent_fd, name=name, st=st)
elif stat.S_ISFIFO(st.st_mode):
if not read_special:
return fso.process_fifo(path=path, parent_fd=parent_fd, name=name, st=st)
else:
return fso.process_file(path=path, parent_fd=parent_fd, name=name, st=st,
cache=cache, flags=flags_special)
elif stat.S_ISCHR(st.st_mode):
if not read_special:
return fso.process_dev(path=path, parent_fd=parent_fd, name=name, st=st, dev_type='c')
else:
return fso.process_file(path=path, parent_fd=parent_fd, name=name, st=st,
cache=cache, flags=flags_special)
elif stat.S_ISBLK(st.st_mode):
if not read_special:
return fso.process_dev(path=path, parent_fd=parent_fd, name=name, st=st, dev_type='b')
else:
return fso.process_file(path=path, parent_fd=parent_fd, name=name, st=st,
cache=cache, flags=flags_special)
elif stat.S_ISSOCK(st.st_mode):
# Ignore unix sockets
return
elif stat.S_ISDOOR(st.st_mode):
# Ignore Solaris doors
return
elif stat.S_ISPORT(st.st_mode):
# Ignore Solaris event ports
return
else:
self.print_warning('Unknown file type: %s', path)
return
def _rec_walk(self, *, path, parent_fd, name, fso, cache, matcher,
exclude_caches, exclude_if_present, keep_exclude_tags,
skip_inodes, restrict_dev, read_special, dry_run):
"""
Process *path* (or, preferably, parent_fd/name) recursively according to the various parameters.
@ -639,6 +695,7 @@ def _process(self, *, path, parent_fd=None, name=None,
# the user says "get out of here!" and we have already completed the desired action.
return
status = None
try:
recurse_excluded_dir = False
if matcher.match(path):
@ -652,10 +709,10 @@ def _process(self, *, path, parent_fd=None, name=None,
# could trigger an error, e.g. if access is forbidden, see #3209.
if not matcher.recurse_dir:
return
recurse_excluded_dir = True
with backup_io('stat'):
st = os_stat(path=path, parent_fd=parent_fd, name=name, follow_symlinks=False)
recurse_excluded_dir = stat.S_ISDIR(st.st_mode)
if not recurse_excluded_dir:
if not stat.S_ISDIR(st.st_mode):
return
if (st.st_ino, st.st_dev) in skip_inodes:
@ -664,17 +721,20 @@ def _process(self, *, path, parent_fd=None, name=None,
# but we WILL save the mountpoint directory (or more precise: the root
# directory of the mounted filesystem that shadows the mountpoint dir).
recurse = restrict_dev is None or st.st_dev == restrict_dev
status = None
if self.exclude_nodump:
# Ignore if nodump flag is set
with backup_io('flags'):
if get_flags(path=path, st=st) & stat.UF_NODUMP:
self.print_file_status('x', path)
return
if stat.S_ISREG(st.st_mode):
if not dry_run:
status = fso.process_file(path=path, parent_fd=parent_fd, name=name, st=st, cache=cache)
elif stat.S_ISDIR(st.st_mode):
if not stat.S_ISDIR(st.st_mode):
# directories cannot go in this branch because they can be excluded based on tag
# files they might contain
status = self._process_any(path=path, parent_fd=parent_fd, name=name, st=st, fso=fso, cache=cache,
read_special=read_special, dry_run=dry_run)
else:
with OsOpen(path=path, parent_fd=parent_fd, name=name, flags=flags_dir,
noatime=True, op='dir_open') as child_fd:
# child_fd is None for directories on windows, in that case a race condition check is not possible.
@ -688,90 +748,36 @@ def _process(self, *, path, parent_fd=None, name=None,
# returning (we do not need to archive or recurse into tagged directories), see #3991:
if not recurse_excluded_dir:
if keep_exclude_tags and not dry_run:
fso.process_dir(path=path, fd=child_fd, st=st)
fso.process_dir_with_fd(path=path, fd=child_fd, st=st)
for tag_name in tag_names:
tag_path = os.path.join(path, tag_name)
self._process(path=tag_path, parent_fd=child_fd, name=tag_name,
fso=fso, cache=cache, matcher=matcher,
exclude_caches=exclude_caches, exclude_if_present=exclude_if_present,
keep_exclude_tags=keep_exclude_tags, skip_inodes=skip_inodes,
restrict_dev=restrict_dev, read_special=read_special, dry_run=dry_run)
self._rec_walk(
path=tag_path, parent_fd=child_fd, name=tag_name, fso=fso, cache=cache,
matcher=matcher, exclude_caches=exclude_caches, exclude_if_present=exclude_if_present,
keep_exclude_tags=keep_exclude_tags, skip_inodes=skip_inodes,
restrict_dev=restrict_dev, read_special=read_special, dry_run=dry_run)
self.print_file_status('x', path)
return
if not dry_run:
if not recurse_excluded_dir:
status = fso.process_dir(path=path, fd=child_fd, st=st)
if not recurse_excluded_dir and not dry_run:
status = fso.process_dir_with_fd(path=path, fd=child_fd, st=st)
if recurse:
with backup_io('scandir'):
entries = helpers.scandir_inorder(path=path, fd=child_fd)
for dirent in entries:
normpath = os.path.normpath(os.path.join(path, dirent.name))
self._process(path=normpath, parent_fd=child_fd, name=dirent.name,
fso=fso, cache=cache, matcher=matcher,
exclude_caches=exclude_caches, exclude_if_present=exclude_if_present,
keep_exclude_tags=keep_exclude_tags, skip_inodes=skip_inodes,
restrict_dev=restrict_dev, read_special=read_special, dry_run=dry_run)
elif stat.S_ISLNK(st.st_mode):
if not dry_run:
if not read_special:
status = fso.process_symlink(path=path, parent_fd=parent_fd, name=name, st=st)
else:
try:
st_target = os.stat(name, dir_fd=parent_fd, follow_symlinks=True)
except OSError:
special = False
else:
special = is_special(st_target.st_mode)
if special:
status = fso.process_file(path=path, parent_fd=parent_fd, name=name, st=st_target,
cache=cache, flags=flags_special_follow)
else:
status = fso.process_symlink(path=path, parent_fd=parent_fd, name=name, st=st)
elif stat.S_ISFIFO(st.st_mode):
if not dry_run:
if not read_special:
status = fso.process_fifo(path=path, parent_fd=parent_fd, name=name, st=st)
else:
status = fso.process_file(path=path, parent_fd=parent_fd, name=name, st=st,
cache=cache, flags=flags_special)
elif stat.S_ISCHR(st.st_mode):
if not dry_run:
if not read_special:
status = fso.process_dev(path=path, parent_fd=parent_fd, name=name, st=st, dev_type='c')
else:
status = fso.process_file(path=path, parent_fd=parent_fd, name=name, st=st,
cache=cache, flags=flags_special)
elif stat.S_ISBLK(st.st_mode):
if not dry_run:
if not read_special:
status = fso.process_dev(path=path, parent_fd=parent_fd, name=name, st=st, dev_type='b')
else:
status = fso.process_file(path=path, parent_fd=parent_fd, name=name, st=st,
cache=cache, flags=flags_special)
elif stat.S_ISSOCK(st.st_mode):
# Ignore unix sockets
return
elif stat.S_ISDOOR(st.st_mode):
# Ignore Solaris doors
return
elif stat.S_ISPORT(st.st_mode):
# Ignore Solaris event ports
return
else:
self.print_warning('Unknown file type: %s', path)
return
self._rec_walk(
path=normpath, parent_fd=child_fd, name=dirent.name, fso=fso, cache=cache, matcher=matcher,
exclude_caches=exclude_caches, exclude_if_present=exclude_if_present,
keep_exclude_tags=keep_exclude_tags, skip_inodes=skip_inodes, restrict_dev=restrict_dev,
read_special=read_special, dry_run=dry_run)
except (BackupOSError, BackupError) as e:
self.print_warning('%s: %s', path, e)
status = 'E'
if status == 'C':
self.print_warning('%s: file changed while we backed it up', path)
# Status output
if status is None:
if not dry_run:
status = '?' # need to add a status code somewhere
else:
status = '-' # dry run, item was not backed up
status = '?' # need to add a status code somewhere
if not recurse_excluded_dir:
self.print_file_status(status, path)
@ -4592,7 +4598,7 @@ def sig_info_handler(sig_no, stack): # pragma: no cover
with signal_handler(sig_no, signal.SIG_IGN):
for frame in inspect.getouterframes(stack):
func, loc = frame[3], frame[0].f_locals
if func in ('process_file', '_process', ): # create op
if func in ('process_file', '_rec_walk', ): # create op
path = loc['path']
try:
pos = loc['fd'].tell()