Commit Graph

106 Commits

Author SHA1 Message Date
Warner Losh
413a368c90 rm -rf can fail sometimes with an error from fts_read. Make it honor
fflag to ignore fts_read errors, but stop deleting from that directory
because no further progress can be made.

When building a kernel with a high -j value on a high core count
machine, during the cleanobj phase we can wind up doing multiple rm
-rf at the same time for modules that have subdirectories. This
exposed this race (sometimes) as fts_read can return an error if the
directory is removed by another rm -rf. Since the intent of the -f
flag was to ignore errors, even if this was a bug in fts_read, we
should ignore the error like we've been instructed to do.
2014-07-07 23:21:20 +00:00
Joel Dahl
cf20d40493 mdoc: remove superfluous paragraph macros. 2014-06-15 10:43:12 +00:00
Eitan Adler
4c99904b9a Add two more 'static' qualifiers 2013-04-26 17:56:35 +00:00
Eitan Adler
9d6d3f96d0 Take some improvements from DragonFlyBSD:
- add const where appropriate
	- add static where appropriate
	- fix a whitespace issues

Reviewed by:	brooks
Obtained from:	DragonFlyBSD
MFC After:	1 week
2013-04-26 17:45:40 +00:00
Eitan Adler
d4319e7433 Add -x option to avoid crossing mount points when removing a hierarchy.
Discussed on:	-hackers
Inspired by:	DragonflyBSD
MFC After:	1 week
2013-04-26 17:45:37 +00:00
Joel Dahl
e9e3a1ab1f Add a few examples.
Obtained from:	OpenBSD
2013-03-15 20:12:54 +00:00
Matthew D Fleming
6db1a7f11e Fix bin/ build with a 64-bit ino_t.
Original code by:	Gleb Kurtsou
2012-09-27 23:31:12 +00:00
Xin LI
930e323894 Polish previous revision: if the fts_* routines have lstat()'ed the
directory entry then use the struct stat from that instead of doing
it again, and skip the rm_overwrite() call if fts_read() indicated
that the entry couldn't be a regular file.

Obtained from:	OpenBSD
MFC after:	1 week
2012-06-20 21:10:38 +00:00
Kevin Lo
4ec1405b39 Fix potential symlink race condition in "rm -P" by adding a check
that the file we have opened is the one we expected.  Also open in
non-blocking mode to avoid a potential hang with FIFOs.

Obtained from:	NetBSD via OpenBSD
2012-06-20 02:21:53 +00:00
Ed Schouten
f9d4afb439 Put some static keywords in the source code.
For these simple utilities, it doesn't harm to make all global variables
static. In fact, this allows the compiler to perform better forms of
optimisation and analysis.
2011-10-31 08:59:17 +00:00
Jilles Tjoelker
0cf90cd1d3 bin: Prefer strrchr() to rindex().
This removes the last index/rindex usage from /bin.
2011-03-15 22:22:11 +00:00
Ulrich Spörlein
ec1e236534 Fix typo and grammar nit
Submitted by:	arundel
MFC after:	7 days (or when the bikeshed has abated)
2010-10-31 19:16:54 +00:00
Ulrich Spörlein
c2c9a54067 Elaborate some more on the non-security implications of using -P
Submitted by:	delphij
Discussion at:	svn-src-all
2010-10-31 09:21:27 +00:00
Dag-Erling Smørgrav
1f9d174116 Language cleanup. 2010-10-27 18:48:23 +00:00
Ulrich Spörlein
a3800f8f0e rm(1): clarify that -P works only when blocks are updated in-place
Suggested by:	pjd, ivoras, arundel
2010-10-08 15:20:20 +00:00
Xin LI
873ddec3fb Clarify the combination effect of -P and -f to make it clear.
Submitted by:	arundel
MFC after:	2 weeks
2010-10-04 06:17:45 +00:00
Jilles Tjoelker
6911f596a1 rm, find -delete: fix removing symlinks with uchg/uappnd set.
Formerly, this tried to clear the flags on the symlink's target
instead of the symlink itself.

As before, this only happens for root or for the unlink(1) variant of rm.

PR:		bin/111226 (part of)
Submitted by:	Martin Kammerhofer
Approved by:	ed (mentor)
MFC after:	3 weeks
2009-05-30 10:42:19 +00:00
Warner Losh
b5260db685 Implement ^T support for rm: now it will report the next file it
removes when you hit ^T.  This is similar to what's done for cp.  The
signal handler and type definitions for "info" were borrowed directly
from cp.
2009-04-29 18:08:18 +00:00
Ruslan Ermilov
88b0915eb1 Give a hint to the reader as to what the "whiteout" actually means. 2006-12-26 13:57:40 +00:00
Ruslan Ermilov
ca3f81d885 Fix markup, add the EXIT STATUS section. 2006-12-26 13:44:41 +00:00
Xin LI
86da4a5eea Correct a security issue introduced in previous commit:
instead of removing the file and issue a warning about
the removal, do not do any operation at all in case -P
is specified when the dinode has hard links.

With -f and -P specified together, we assume that the
user wants rm to overwrite the contents of the file
and remove it (destroy the contents of file but leave
its hard links as is).

The reason of doing it this way is that, in case where
a hard link is created by a malicious user (currently
this is permitted even if the user has no access to the
file).  Losing the link can potentially mean that the
actual owner would lose control completely to the user
who wants to obtain access in a future day.

Discussed with:	Peter Jermey
2006-10-31 02:22:36 +00:00
Xin LI
0b6f55b77c Be more reasonable when overwrite mode is specified while there
is hard links.  Overwritting when links > 1 would cause data
loss, which is usually undesired.

Inspired by:	discussion on -hackers@
Suggested by:	elessar at bsdforen de
Obtained from:	OpenBSD
2006-10-30 03:32:09 +00:00
Maxim Konovalov
da29c6560b o Backout rev. 1.55. Don't waste cpu cycles for bzero(), do not
call chflags() for whiteouted files.

Prodded by:	ru
2006-10-18 13:16:06 +00:00
Maxim Konovalov
cf1db7b34f o Zero out struct stat before usage. lstat(2) can fail and
leave garbage there which will break -W code path.

PR:		bin/84569
Submitted by:	Igor
MFC after:	2 weeks
2006-10-18 08:22:33 +00:00
Maxim Konovalov
089418247a o Be pedantic and do fts_close() when done.
PR:		bin/95292
Submitted by:	Charles Hardin
Obtained from:	NetBSD via OpenBSD, PR
2006-04-15 09:26:23 +00:00
Ruslan Ermilov
8af1113166 -mdoc sweep. 2005-11-17 12:15:23 +00:00
Doug Barton
0b5efa8ad4 Give .Dd a tummy rub, forgotten on my last commit. 2005-09-30 02:12:15 +00:00
Doug Barton
a5f6295013 Handle the case where the -P flag is specified for a read-only file
earlier, and more gracefully. Previously, this combination would be
ignored early in the code where permissions are tested and fail later
with a very unhelpful "permission denied" error.

Instead, test for this flag in the same block that generates the
"override?" messages for read-only files, but instead of trying
to guess what the user has in mind, generate an error and exit.

Update the man page to reflect this new behavior.

Not objected to by:	freebsd-hackers@
2005-09-29 20:40:29 +00:00
Ruslan Ermilov
a89237ae95 Sync program's usage() with manpage's SYNOPSIS. 2005-02-09 17:37:39 +00:00
Xin LI
c49d3c9bac Remove unnecessary SRCS= where could be guessed directly by our
bsd.*.mk infrasture.

Obtained from:	ru
2005-01-27 14:52:47 +00:00
Ruslan Ermilov
6fca4c7c3f Add the new standard EXIT STATUS section where appropriate.
Sort standard sections in the (documented) preferred order.
2005-01-16 16:41:59 +00:00
Warner Losh
9ddb49cbe4 /*- or .\"- or #- to begin license clauses. 2005-01-10 08:39:26 +00:00
Jordan K. Hubbard
de3abdfaf4 UNIX conformance: If -r -f on non-existent directory, don't emit error. 2004-11-13 04:07:01 +00:00
Xin LI
164d3f0bdc Correct a typo and make the documentation more mdoc(7) complaint.
Pointed out by:	dd
Approved by:	murray (mentor)
2004-11-01 16:52:34 +00:00
Xin LI
24c0f7385b Add -I, an option that asks for confirmation once if recursively
removing directories or if more than 3 files are listed in the
command line.

This feature is intended to provide a safe net but not being too
annoying like having "rm -i" for every deleting operations, and
is generally good for both newbies and power users, preventing
them from being so easily run into ``rm -rf /'', ``rm -rf *''
and so forth.

Originally implemented by Matthew Dillon for DragonFly, plus
some improvements done by various DragonFly contributors.

Approved by:	murray (mentor; the original dillon's version)
Discussed with:	des
Obtained from:	DragonFly's bin/rm/
		rm.c rev. 1.4 - 1.8
		rm.1 rev. 1.3 - 1.4
MFC After:	1 month
2004-10-28 08:25:30 +00:00
Dag-Erling Smørgrav
3f91ab9262 The previous commit added code to rm(1) to warn about and remove any
occurrences of "/" in the argument list.  This corresponds to Enhancement
Request Number 5 in the Austin Group TC2 Aardvark's XCU Defects Report
(<URL:http://www.opengroup.org/austin/aardvark/finaltext/xcubug.txt>).
Further discussion is available in the Austin Group mailing list archives
(<URL:http://www.opengroup.org/austin/mailarchives/>, "Defect in XCU rm")
and for Austin Group members, in the Austin Group Interpretations archive
(<URL:http://www.opengroup.org/austin/interps/>, AI-019)

This commit makes that check conditional on !POSIXLY_CORRECT, since it
is not strictly correct according to the current version of the standard
(but is expected to be correct according to the next version, and has
already been adopted by Solaris).
2004-10-04 19:24:28 +00:00
Dag-Erling Smørgrav
3a667dafcc Markup fixes.
Pointed out by:	ru
2004-10-04 19:03:44 +00:00
Dag-Erling Smørgrav
68ef5f71b0 Find out how flame-proof my underwear really is. 2004-10-04 11:26:01 +00:00
Dag-Erling Smørgrav
c335b1ecdb Whitespace cleanup. 2004-10-04 11:14:12 +00:00
Ruslan Ermilov
eccea571a7 Mechanically kill hard sentence breaks. 2004-07-02 21:04:19 +00:00
Mark Murray
6195fb4102 Remove clause 3 from the UCB licenses.
OK'ed by:	imp, core
2004-04-06 20:06:54 +00:00
Jun Kuriyama
b800f53d91 o Fix a style bug and poor wording in comment.
o When fts_read() cannot stat the file, it can't be unlinked.  At
  that case, don't display error message when -f flag is used.

Obtained from:	bde
PR:		kern/16815, bin/35842
Reported by:	kuriyama, Aleksandr A. Babaylov <.@babolo.ru>
2004-01-01 10:26:43 +00:00
Guido van Rooij
f3761deee3 When the P flag is set (i.e. Overwrite regular files before deleting them),
do only unlink the file if we could indeed overwrite the file.
Old behaviour: rm -P /tmp/foo (foo mode 0444) would NOT overwrite foo,
but still delete it (with a warning: rm: foo: Permission denied)
New behaviour: Just the EPERM warning, but no deletion

Reviewed by:	bde
2003-11-10 09:40:18 +00:00
Bruce Evans
8f1f433834 1. Fixed leakage of a file descriptor for every non-fatal failure in
rm_overwrite() (for rm -P).

2. Print the file name in the error message for (fatal) malloc() failures
   in rm_overwrite().  I first thought that malloc() failures should be
   non-fatal since they don't prevent proceeding the the next file, but
   making them non-fatal would normally give too much output for rm -Pr
   on a large tree in the unlikely event that even one occurs, since the
   malloc()ed amounts are usually the same.  Just print the file name since
   the malloc()ed amounts are not always the same and it doesn't hurt to
   know where rm was when it quit.

Submitted by:	guido ((1) and original version of (2))
2003-11-08 09:55:16 +00:00
David E. O'Brien
09a80d4867 Quiet warnings about copyright[]. 2003-05-01 16:58:57 +00:00
Tom Rhodes
02b911053d Cross ref chflags(1).
Suggested by:	Craig Carey <research@ijs.co.nz) on -doc.
2003-02-04 20:38:51 +00:00
Tom Rhodes
7d971bbf29 s/filesystem/file system/ as discussed on -developers 2002-08-21 17:32:44 +00:00
Sheldon Hearn
ced4deb156 Revert previous delta, which is not required with rev 1.5 of
src/contrib/gcc/c-format.c .
2002-07-15 12:08:21 +00:00
Tim J. Robbins
90833c99de Complain if more than one file argument is given to unlink(1) like we did
before I made unlink use getopt().
2002-07-12 07:20:20 +00:00
Sheldon Hearn
a3ee32f3f7 Set WFORMAT=0, because our compiler currently complains about NULL
arg 2 to err() and friends, and warnings are promoted to errors.

This allows the following revisions to be reverted:

	rev 1.39	src/bin/cp/cp.c
	rev 1.26	src/bin/chmod/chmod.c
	rev 1.40	src/bin/rm/rm.c

The following revisions can already be reverted, because they were
already covered by WFORMAT=0:

	rev 1.8		src/bin/ls/lomac.c
	rev 1.63	src/bin/ls/ls.c
	rev 1.8		src/bin/ps/lomac.c
	rev 1.34	src/bin/rcp/rcp.c
2002-07-11 08:45:38 +00:00