Instead of complaining that my authorship of the rewrite of
the mdev to use /sys/dev is totally gone from the git history
I bravely take credit by adding myself to the AUTHORS file
instead, he he.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
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>
User report:
or our board we setup eth0:0 on a 10.10.10.x/29 netwrok.
The problem is ip addr flush dev eth0:0 removes all ip addresses from
eth0. You can see this if you run
ip -stat -stat addr flush dev eth0:0
2: eth0 inet 172.27.105.10/22 brd 172.27.107.255 scope global eth0
valid_lft forever preferred_lft forever
2: eth0 inet 10.10.10.9/29 scope global eth0:0
valid_lft forever preferred_lft forever
2: eth0 inet6 fe80::a2f6:fdff:fe18:2b13/64 scope link
valid_lft forever preferred_lft forever
*** Round 1, deleting 3 addresses ***
*** Flush is complete after 1 round ***
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
A padding to align a message should not only be added between
different attributes of a netlink message, but also at the end of the
message to pad it to the correct size.
Without this patch the following command does not work and returns an
error code:
ip link add type nlmon
Without this ip from busybox sends this:
sendmsg(3, {msg_name={sa_family=AF_NETLINK, nl_pid=0, nl_groups=00000000},
msg_namelen=12, msg_iov=[{iov_base={{len=45, ...},
"\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\20\0\22\0\t\0\1nlmon"}, iov_len=45}],
msg_iovlen=1, msg_controllen=0, msg_flags=0}, 0) = 45
return value: 2
The normal ip utile from iproute2 sends this:
sendmsg(3, {msg_name={sa_family=AF_NETLINK, nl_pid=0, nl_groups=00000000},
msg_namelen=12, msg_iov=[{iov_base={{len=48, ...},
"\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\20\0\22\0\t\0\1nlmon\0\0\0"}, iov_len=48}],
msg_iovlen=1, msg_controllen=0, msg_flags=0}, 0) = 48
return value: 0
With this patch ip from busybox sends this:
sendmsg(3, {msg_name={sa_family=AF_NETLINK, nl_pid=0, nl_groups=00000000},
msg_namelen=12, msg_iov=[{iov_base={{len=48, ...},
"\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\20\0\22\0\t\0\1nlmon\0\0\0"}, iov_len=48}],
msg_iovlen=1, msg_controllen=0, msg_flags=0}, 0) = 48
return value: 0
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
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>
When using svlogd's processor functionality to run arbitrary commands
on log rotation, the line in the config is executed verbatim, i.e. the
exclamation mark is included.
For example, if the config file contains:
s100
!cat
then when it's time to rotate the log files after each 100 bytes, sh -c
"!cat" will be run, instead of sh -c "cat" as intended. The result is
svlogd logging
/bin/bash: !cat: command not found
svlogd: warning: processor failed, restart: /tmp/svlogd/
over and over again as it keeps attempting to execute the processor and
failing (unless you happen to have a "!cat" binary around :)).
Skipping the exclamation mark when performing the wstrdup() fixes the
issue.
Signed-off-by: Francis Rounds <francis.rounds@4bridgeworks.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
On some systems like Chromium OS, loading modules from non-verified
filesystems is denied. Only finit_module is allowed because an open
fd is passed which can be checked against a verified location.
Change the module loading code to first attempt finit_module and if
that fails for whatever reason, fall back to the existing logic.
On x86_64, this adds ~80 bytes to modutils/modutils.o and ~68 bytes
to modutils/modprobe-small.o.
Signed-off-by: Mike Frysinger <vapier@chromium.org>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
This is only necessary if we use stdout fd.
function old new delta
less_exit 32 51 +19
less_main 2540 2543 +3
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 2/0 up/down: 22/0) Total: 22 bytes
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Currently some new devices that have a bus but no class will
be missed by mdev coldplug device creation after boot. This
happens because mdev recursively searches /sys/class which will
by definition only find class devices.
Some important devices such as iio and gpiochip does not have
a class. But users will need them.
This switches from using /sys/class as the place to look for
devices to create to using /sys/dev where all char and block
devices are listed.
The subsystem lookup code that provide the G.subsystem
environment variable is changed from using the directory
name of the class device to instead dereference the
"subsystem" symlink for the device, and look at the last
element of the path of the symlink for the subsystem, which
will work with class devices and bus devices alike. (The new
bus-only devices only symlink to the /sys/bus/* hierarchy.)
We delete the legacy kernel v2.6.2x /sys/block device path
code as part of this change. It's too old to be kept alive.
Tested on kernel v4.6-rc2 with a bunch of devices, including
some IIO and gpiochip devices.
With a print inserted before make_device() the log looks
like so:
Create device from "/sys/dev/char/1:1", subsystem "mem"
Create device from "/sys/dev/char/1:2", subsystem "mem"
Create device from "/sys/dev/char/1:3", subsystem "mem"
Create device from "/sys/dev/char/1:5", subsystem "mem"
(...)
Create device from "/sys/dev/block/179:56", subsystem "block"
Create device from "/sys/dev/block/179:64", subsystem "block"
function old new delta
mdev_main 1388 1346 -42
dirAction 134 14 -120
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 0/2 up/down: 0/-162) Total: -162 bytes
Cc: Isaac Dunham <ibid.ag@gmail.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Jonathan Cameron <jic23@cam.ac.uk>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
If a non-starttls helper is in use, initial 220 response is processed by us,
not by helper.
Some servers consider us to be a spammer if we don't wait for it.
It is not in protocol, but it is a real-life problem.
The workaround in this patch is a magic envvar, $SMTP_ANTISPAM_DELAY:
...
-H 'PROG ARGS' Run connection helper. Examples:
openssl s_client -quiet -tls1 -starttls smtp -connect smtp.gmail.com:25
openssl s_client -quiet -tls1 -connect smtp.gmail.com:465
$SMTP_ANTISPAM_DELAY: seconds to wait after helper connect
...
By using it, people can tweak sendmail behavior even if sendmail invocation
is buried in some scripts.
function old new delta
packed_usage 30464 30497 +33
sendmail_main 1185 1206 +21
------------------------------------------------------------------------------
(add/remove: 0/0 grow/shrink: 2/0 up/down: 54/0) Total: 54 bytes
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Lauri Kasanen:
:: Over at TinyCore, we receive a huge number of questions of the type "I
:: got "short write", what does it mean?". Mostly for the rpi port and when
:: using bb wget.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
The udhcpc script may be used to setup fallback configuration (E.G. IPv4LL,
fixed IP address, ..) that also needs to be cleaned up on release (E.G.
when SIGUSR2 is called or on shutdown with -R), so unconditionally call
deconfig.
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Conditional rewrite can keep NUM_APPLETS.h mtime old,
this causes make to try to regenerate it at every invocation.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
"Total allocated sectors 2021315 greater than the maximum 2020356"
maximum what?
Turns out, that's the CHS size of the disk.
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
Before:
Device Boot Start End Blocks Id System
/dev/sdb1 * 1 998 255471+ 6 FAT16
What are "blocks"? What is that "+"?
How big is this partition?
Is start/end shown came from LBA fields or CHS fields?
Why are we torturing the user??
After:
Device Boot StartCHS EndCHS StartLBA EndLBA Sectors Size Id Type
/dev/sdb1 * 0,1,1 996,15,32 32 510974 510943 249M 6 FAT16
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
The filenames in docs/keep_data_small.txt are a little bit outdated.
It's better to change it to the current name.
decompress_unzip.c -> decompress_gunzip.c
(since commit 774bce8e8b)
libbb/messages.c -> libbb/ptr_to_globals.c
(since commit 574f2f4394)
Signed-off-by: Kang-Che Sung <explorer09@gmail.com>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
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>
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>
Some user managed to hit a race where iface is gone between SIOCGIFFLAGS
and SIOCSIFFLAGS (!). If SIOCSIFFLAGS fails, treat it the same as failed
SIOCGIFFLAGS
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
