Commit Graph

1759 Commits

Author SHA1 Message Date
Denys Vlasenko
60ca834358 ash: [MEMALLOC] Add pushstackmark
Upstream commit:

    Author: Herbert Xu <herbert@gondor.apana.org.au>
    Date:   Sat Oct 6 00:45:52 2007 +0800

    [MEMALLOC] Add pushstackmark

    This patch gets rid of the stack mark tracking hack by allocating a little
    bit of stack memory if we're at risk of planting a stack mark which may be
    grown later.  To do this a new function pushstackmark is added which lets
    the user pick a bigger amount to allocate since some users do that anyway
    after setting a stack mark.

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-30 11:21:21 +02:00
Denys Vlasenko
08755f9bcb hush: fix typo in comment
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-30 02:02:25 +02:00
Denys Vlasenko
4074d49557 hush: fix interactive input handling of backslash+newline
function                                             old     new   delta
fgetc_interactive                                      -     258    +258
i_peek_and_eat_bkslash_nl                             43      93     +50
static_peek2                                           7       -      -7
setup_string_in_str                                   46      39      -7
setup_file_in_str                                     47      40      -7
file_peek                                             72      52     -20
expand_vars_to_list                                 1167    1143     -24
file_peek2                                            74       -     -74
file_get                                             326      65    -261
------------------------------------------------------------------------------
(add/remove: 1/2 grow/shrink: 1/5 up/down: 308/-400)          Total: -92 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-30 01:49:53 +02:00
Denys Vlasenko
0448c55cc8 hush: G.user_input_buf[] is needed only if line editing is enabled
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-29 20:25:44 +02:00
Denys Vlasenko
7a24e8ffeb shell testsuite: add trailing newline to var_unbackslash1.tests
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-29 20:02:53 +02:00
Denys Vlasenko
46e6498b2a hush: speed optimizations
Make o_addchr() faster: do not call o_grow_by() each time.
Create i_getch_and_eat_bkslash_nl(), use it instead of peek+getch pair.

function                                             old     new   delta
o_addchr                                              42      54     +12
parse_dollar                                         761     771     +10
o_grow_by                                             48      37     -11
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 3/1 up/down: 24/-11)             Total: 11 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-29 19:50:55 +02:00
Denys Vlasenko
657086a3dc hush: fix handling of by backslash-newline in $((arith)) and $(cmd)
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-29 18:07:42 +02:00
Denys Vlasenko
d17a91db6e hush: rework input char buffering to allow more than one-deep peek
...this time with actual hush.c changes too :)

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-29 18:02:37 +02:00
Denys Vlasenko
459293b1c5 ash: fix arithmetic closing )) split by backslash-newline
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-29 17:58:58 +02:00
Denys Vlasenko
73c3e074df ash: [PARSER] Handle backslash newlines properly after dollar sign
Fixes var_unbackslash1.tests failure.

Upstream commit:

    [PARSER] Handle backslash newlines properly after dollar sign

    On Tue, Aug 26, 2014 at 12:34:42PM +0000, Eric Blake wrote:
    > On 08/26/2014 06:15 AM, Oleg Bulatov wrote:
    > > While playing with sh generators I found that dash and bash have different
    > > interpretations for <slash><newline> sequence.
    > >
    > > $ dash -c 'EDIT=xxx; echo $EDIT\
    > >> OR'
    > > xxxOR
    >
    > Buggy.
    > >
    > > $ dash -c 'echo "$\
    > > (pwd)"'
    > > $(pwd)
    > >
    > > Is it undefined behaviour in POSIX?
    >
    > No, it's well-defined, and dash is buggy.
    ...

    I agree.  This patch should resolve this problem and similar ones
    affecting blackslash newlines after we encounter a dollar sign.

    Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-29 17:17:04 +02:00
Denys Vlasenko
8286513838 hush: rework input char buffering to allow more than one-deep peek
This fixes backslash+newline continuation in
	$VAR\
	NAME
construct. (ash has a bug there as well).

function                                             old     new   delta
file_peek2                                             -      74     +74
parse_dollar                                         746     773     +27
expand_vars_to_list                                 1143    1167     +24
setup_string_in_str                                   32      46     +14
setup_file_in_str                                     33      47     +14
file_get                                             264     278     +14
static_peek2                                           -       7      +7
file_peek                                             91      72     -19
------------------------------------------------------------------------------
(add/remove: 2/0 grow/shrink: 5/1 up/down: 174/-19)           Total: 155 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-29 16:59:06 +02:00
Denys Vlasenko
3b4d04b77e ash: input: Allow two consecutive calls to pungetc
Upstream commit:

    input: Allow two consecutive calls to pungetc

    The commit ef91d3d6a4c39421fd3a391e02cd82f9f3aee4a8 ([PARSER]
    Handle backslash newlines properly after dollar sign) created
    cases where we make two consecutive calls to pungetc.  As we
    don't explicitly support that there are corner cases where you
    end up with garbage input leading to undefined behaviour.

    This patch adds explicit support for two consecutive calls to
    pungetc.

    Reported-by: Jilles Tjoelker <jilles@stack.nl>
    Reported-by: Juergen Daubert <jue@jue.li>
    Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

In bbox case, bashism >& may need two pungetc() too.

function                                             old     new   delta
pgetc                                                514     555     +41
pushstring                                           114     144     +30
basepf                                                52      76     +24
popstring                                            134     151     +17
parse_command                                       1584    1585      +1
pungetc                                               12       9      -3
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 5/1 up/down: 113/-3)            Total: 110 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-29 02:11:19 +02:00
Denys Vlasenko
78c9c736ab hush: fix 'eval ""' handling
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-29 01:44:17 +02:00
Denys Vlasenko
992e0ff7e9 hush: fix ". EMPTY_LINE" not setting $? to 0
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-29 01:27:09 +02:00
Denys Vlasenko
46a45ce02f ash: jobs: Don't attempt to access job table for job %0
Upstream commit:

    jobs: Don't attempt to access job table for job %0

    If job %0 is (mistakenly) specified, an out-of-bounds access to the
    jobtab occurs in function getjob() if num = 0:

            jp = jobtab + 0 - 1

    Fix this by checking that the job number is larger than 0 before
    accessing the jobtab.

    Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
    Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-29 01:10:08 +02:00
Denys Vlasenko
928e2a7ef4 ash: [EVAL] Make eval with empty arguments return 0
This is a backport of upstream commit:

    [EVAL] Make eval with empty arguments return 0

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-29 00:30:31 +02:00
Denys Vlasenko
8e2bc47d62 ash: [EVAL] Fix use-after-free in dotrap/evalstring
From upstream:

    [EVAL] Fix use-after-free in dotrap/evalstring

    The function dotrap calls evalstring using the stored trap string.
    If evalstring then unsets that exact trap string then we will end
    up using freed memory.

    This patch fixes it by making evalstring always duplicate the string
    before using it.

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-28 23:02:57 +02:00
Denys Vlasenko
7ee7c6fc20 ash: Remove unused EV_BACKCMD flag
The original ash defered forking commands in backquotes so builtins
    could be run in the same context as the shell.  This behavior was
    controlled using the EV_BACKCMD to evaltree.

    Unfortunately, as Matthias Scheler noticed in 1999 (NetBSD PR/7814),
    the result was counterintuitive; for example, echo "`cd /`" would
    change the cwd.  So ash 0.3.5 left out that optimization.

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-28 19:44:16 +02:00
Denys Vlasenko
eb17b6f6c9 ash: eval: Return status in eval functions
Backported from dash:

    eval: Return status in eval functions

    The exit status is currently clobbered too early for case statements
    and loops.  This patch fixes it by making the eval functions return
    the current exit status and setting them in one place -- evaltree.

    Harald van Dijk pointed out a number of bugs in the original patch.

function                                             old     new   delta
evalcommand                                         1226    1242     +16
cmdloop                                              383     398     +15
evalfor                                              223     227      +4
evalcase                                             271     275      +4
localcmd                                             348     350      +2
evaltreenr                                           927     928      +1
evaltree                                             927     928      +1
evalsubshell                                         150     151      +1
evalpipe                                             356     357      +1
parse_command                                       1585    1584      -1
evalloop                                             177     164     -13
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 9/2 up/down: 45/-14)             Total: 31 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-28 19:41:57 +02:00
Denys Vlasenko
557482c1cb ash: in heredoc code, fix access past the end of allocated memory. Closes 9276
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-25 21:24:04 +02:00
Denys Vlasenko
13f20919b2 ash: fix handling of NULs in $'abc\000def\x00asd'. Closes 9286
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-25 20:54:25 +02:00
Denys Vlasenko
b3f29b452a ash: use glob() from libc
Adapted from dash.
The "homegrown" glob code is retained (ifdef'ed out).
This changes was inspired by bug 9261, which detected out-of bounds use of heap
for 2098 byte long name in the "homegrown" code. This is still not fixed...

function                                             old     new   delta
expandarg                                            960     982     +22
static.syntax_index_table                             26      25      -1
static.spec_symbls                                    27      26      -1
static.metachars                                       4       -      -4
addfname                                              42       -     -42
msort                                                126       -    -126
expmeta                                              528       -    -528
------------------------------------------------------------------------------
(add/remove: 0/4 grow/shrink: 1/2 up/down: 22/-702)          Total: -680 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-21 16:25:58 +02:00
Denys Vlasenko
244fdd45c7 ash: fix handling of bashism $'xxx' with high-bit chars. Closes 9236
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-20 17:04:09 +02:00
Denys Vlasenko
d6a37d86ba hush: document better where bad redirect syntax is detected
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-20 16:22:24 +02:00
Rostislav Skudnov
204c7fb229 ash: exit after subshell error when errexit option is set
When "set -e" option is on, shell must exit when any command fails,
including compound commands of the form (compound-list) executed in a
subshell. Bash and dash shells have this behaviour.

Also add a corresponding testcase.

Signed-off-by: Rostislav Skudnov <rostislav@tuxera.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-17 23:28:23 +02:00
Denys Vlasenko
7373759947 fix "aloc" -> "alloc" typos
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-17 20:58:22 +02:00
Denys Vlasenko
7bc3d39695 ash: add a FIXME for bug 9246
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-17 20:53:47 +02:00
Denys Vlasenko
ef15970d7e *: placate some compile warnings on OSX
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-09-01 11:16:22 +02:00
Denys Vlasenko
aa3576a29b hush: fix "redirects can close script fd" bug
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-08-22 19:54:12 +02:00
Denys Vlasenko
7b25b1c5b2 hush: do not leak script fds into NOEXEC children
We set all opened script fds to CLOEXEC, thus making then go away
after fork+exec.
Unfortunately, CLOFORK does not exist. NOEXEC children will still see those fds open.

For one, "ls" applet is NOEXEC. Therefore running "ls -l /proc/self/fd"
in a script from standalone shell shows this:

lrwx------    1 root     root            64 Aug 20 15:17 0 -> /dev/pts/3
lrwx------    1 root     root            64 Aug 20 15:17 1 -> /dev/pts/3
lrwx------    1 root     root            64 Aug 20 15:17 2 -> /dev/pts/3
lr-x------    1 root     root            64 Aug 20 15:17 3 -> /path/to/top/level/script
lr-x------    1 root     root            64 Aug 20 15:17 4 -> /path/to/sourced/SCRIPT1
...

with as many open fds as there are ". SCRIPTn" nest levels.
Fix it by closing these fds after fork (only for NOEXEC children).

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-08-20 15:58:34 +02:00
Denys Vlasenko
869994cf4f hush: bit better comments in redirect code. No logic changes
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-08-20 15:16:00 +02:00
Denys Vlasenko
e9abe75fda hush: cmd and arithmetic also need the fix for FILE rewind
Discovered by running testsuite with a newest glibc

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-08-19 20:15:26 +02:00
Denys Vlasenko
215b0ca6e4 hush: fix a bug in FEATURE_SH_STANDALONE=y config. Closes 9186
Run this in a "sh SCRIPT":

sha256sum /dev/null
echo END

sha256sum is a NOEXEC applet. It runs in a forked child. Then child exit()s.
By this time, entire script is read, and buffered in a FILE object
from fopen("SCRIPT"). But fgetc() did not consume entire input.
exit() lseeks back by -9 bytes, from <eof> to 'e' in 'echo'.
(this may be libc-specific).
This change of fd position *is shared with the parent*!

Now parent can read more, and it thinks there is another "echo END".
End result: two "echo END"s are run.

Fix this by _exit()ing instead.

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-08-19 18:43:06 +02:00
Denys Vlasenko
f8ddbe1ccc ash: fix handling of ${VAR: -2}
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-07-25 03:56:00 +02:00
Denys Vlasenko
0fb0045aa9 config: disentangle PREFER_APPLETS from SH_STANDALONE and SH_NOFORK
On user request.
I thought enabling/disabling them all together is more consistent.
Evidently, some people do want them to be separately selectable.

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-07-22 18:48:38 +02:00
Denys Vlasenko
49117b4800 hush: fix a possible bug
Not sure this was actually a triggerable bug, but the code looked flaky.

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-07-21 14:42:13 +02:00
Denys Vlasenko
e5814a5a42 ash: do not leave SIGQUIT ignored on "exec CMD"
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-07-16 18:33:55 +02:00
Denys Vlasenko
3e134ebf6a *: slap on a few ALIGN1/2s where appropriate
The result of looking at "grep -F -B2 '*fill*' busybox_unstripped.map"

   text	   data	    bss	    dec	    hex	filename
 829901	   4086	   1904	 835891	  cc133	busybox_before
 829665	   4086	   1904	 835655	  cc047	busybox

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-04-22 18:09:21 +02:00
Ron Yorston
84ba50c32f ash: bash-compatible $'...' shouldn't expand in double quotes
Bash doesn't expand its $'...' construct in double quotes:

   $ echo "$'a\tb'"
   $'a\tb'

Change BusyBox ash to do the same.  This also fixes a problem with
here documents where BusyBox ash gave an incorrect result for:

   $ cat <<EOF
   > '$'
   > EOF
   '$'

Reported-by: Timo Teras <timo.teras@iki.fi>
Signed-off-by: Ron Yorston <rmy@pobox.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-04-15 22:16:46 +02:00
Ron Yorston
2b91958dff Rewrite iteration through applet names to save a few bytes
function                                             old     new   delta
run_applet_and_exit                                  758     755      -3
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 0/1 up/down: 0/-3)               Total: -3 bytes

In standalone shell mode the saving increases to 17 bytes.

Signed-off-by: Ron Yorston <rmy@pobox.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-04-15 22:13:51 +02:00
Mike Frysinger
b9b7aa1910 ash_test: ignore generated files
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
2016-03-22 18:15:24 -04:00
Mike Frysinger
787972f60b ash_test: printenv: fix missing includes
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
2016-03-22 18:15:14 -04:00
Ron Yorston
3e3bfb896e ash: fix corruption of ${#var} if $var contains UTF-8 characters
As reported in bug 8506:

   $ X=abcdÉfghÍjklmnÓpqrstÚvwcyz
   $ echo ${#X}
   abcd26

The result should be 26.

This regression was introduced by:

   <d68d1fb> 2015-05-18 [Ron Yorston]  ash: code shrink around varvalue

The length in characters was being used to discard the contents of
the variable instead of the length in bytes.

URL: https://bugs.busybox.net/8506
Reported-by: Martijn Dekker <martijn@inlv.org>
Signed-off-by: Ron Yorston <rmy@pobox.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
2016-03-22 18:19:29 -04:00
Kylie McClain
40eea690c7 Fix compiling with musl's utmp stubs
This patch fixes compiling busybox with FEATURE_UTMP and _WTMP enabled.
musl, while not really support utmp/wtmp, provides stub functions, as well
as variables such as _PATH_UTMP, so that programs using utmp or wtmp can
still compile fine.

My reasoning for this patch is that on Exherbo, I'm currently trying to get
us to be able to use the same busybox config file for both glibc and musl
systems, using utmp/wtmp on systems that support it, and using the stubs
on musl without needing two different configs.

As of latest musl git, it provides all utmp functions needed; 1.1.12 doesn't,
but I sent a patch to Rich to add the utmp{,x}name functions expected to
exist, which was merged into musl upstream.

Signed-off-by: Kylie McClain <somasissounds@gmail.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-02-01 01:36:05 +01:00
Cristian Ionescu-Idbohrn
c2a2625bcb ash: suppress a compilation warning
Reported by gcc (Debian 5.3.1-4) 5.3.1 20151219

shell/ash.c: In function 'evaltree':
shell/ash.c:8432:19: warning: logical not is only applied to the left hand side of comparison

Signed-off-by: Cristian Ionescu-Idbohrn <cristian.ionescu-idbohrn@axis.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2016-01-02 00:52:29 +01:00
Ron Yorston
95ebcf79ff ash: add support for bash 'function' keyword
Where the POSIX shell allows functions to be defined as:

   name () compound-command [ redirections ]

bash adds the alternative syntax:

   function name [()] compound-command [ redirections ]

Implement this in ash's bash compatibility mode.  Most compound
commands work (for/while/until/if/case/[[]]/{}); one exception is:

   function f (echo "no way!")

The other two variants work:

   f() (echo "ok")
   function f() (echo "also ok")

function                                             old     new   delta
parse_command                                       1555    1744    +189
tokname_array                                        232     240      +8
.rodata                                           155612  155566     -46
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 2/1 up/down: 197/-46)           Total: 151 bytes

Signed-off-by: Ron Yorston <rmy@pobox.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2015-11-04 19:30:24 +01:00
Denys Vlasenko
bc9bee01f3 hush-misc/func_args1.tests: remove "UNFIXED BUG", it does not fail
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2015-11-04 14:50:19 +01:00
Ron Yorston
85405c80a2 ash: copy function tests from hush testsuite
Signed-off-by: Ron Yorston <rmy@pobox.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2015-11-04 14:48:16 +01:00
Ron Yorston
95650a86d1 ash: allow popredir to be called if the stack is empty
If /tmp/test.sh is a script that tries to run a second script which
happens to be non-executable this:

   command . /tmp/test.sh

causes a seg fault.

This is because clearredir is called in the error path to clear the
stack of redirections.  The normal path then calls popredir, but popredir
fails when the stack is empty.

Reported-by: Bastian Bittorf <bittorf@bluebottle.com>
Signed-off-by: Ron Yorston <rmy@pobox.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2015-10-30 22:23:13 +01:00
Ron Yorston
8c55dc79a7 ash: fix EXEXEC status clobbering
evalcommand always clobbers the exit status in case of an EXEXEC
which means that exec always fails with exit status 2 regardless
of what it actually returns.

This patch adds the missing check for EXEXEC so that the correct
exit status is preserved.  It causes the test ash-misc/exec.tests
to succeed.

Based on commit 7f68426 in dash git, by Herbert Xu.

Signed-off-by: Ron Yorston <rmy@pobox.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2015-10-30 22:19:48 +01:00