busybox/libbb/Config.src
Denys Vlasenko 6472ac9428 libbb/sha256: optional x86 hardware accelerated hashing
64 bit:
function                                             old     new   delta
sha256_process_block64_shaNI                           -     730    +730
.rodata                                           108314  108586    +272
sha256_begin                                          31      83     +52
------------------------------------------------------------------------------
(add/remove: 5/1 grow/shrink: 2/0 up/down: 1055/-1)          Total: 1054 bytes

32 bit:
function                                             old     new   delta
sha256_process_block64_shaNI                           -     747    +747
.rodata                                           104318  104590    +272
sha256_begin                                          29      84     +55
------------------------------------------------------------------------------
(add/remove: 5/1 grow/shrink: 2/0 up/down: 1075/-1)          Total: 1074 bytes

Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>
2022-02-03 14:58:02 +01:00

371 lines
11 KiB
Plaintext

#
# For a description of the syntax of this configuration file,
# see docs/Kconfig-language.txt.
#
comment "Library Tuning"
INSERT
choice
prompt "Buffer allocation policy"
default FEATURE_BUFFERS_USE_MALLOC
help
There are 3 ways busybox can handle buffer allocations:
- Use malloc. This costs code size for the call to xmalloc.
- Put them on stack. For some very small machines with limited stack
space, this can be deadly. For most folks, this works just fine.
- Put them in BSS. This works beautifully for computers with a real
MMU (and OS support), but wastes runtime RAM for uCLinux. This
behavior was the only one available for versions 0.48 and earlier.
config FEATURE_BUFFERS_USE_MALLOC
bool "Allocate with Malloc"
config FEATURE_BUFFERS_GO_ON_STACK
bool "Allocate on the Stack"
config FEATURE_BUFFERS_GO_IN_BSS
bool "Allocate in the .bss section"
endchoice
config PASSWORD_MINLEN
int "Minimum password length"
default 6
range 5 32
help
Minimum allowable password length.
config MD5_SMALL
int "MD5: Trade bytes for speed (0:fast, 3:slow)"
default 1 # all "fast or small" options default to small
range 0 3
help
Trade binary size versus speed for the md5 algorithm.
Approximate values running uClibc and hashing
linux-2.4.4.tar.bz2 were:
value user times (sec) text size (386)
0 (fastest) 1.1 6144
1 1.4 5392
2 3.0 5088
3 (smallest) 5.1 4912
config SHA1_SMALL
int "SHA1: Trade bytes for speed (0:fast, 3:slow)"
default 3 # all "fast or small" options default to small
range 0 3
help
Trade binary size versus speed for the sha1 algorithm.
throughput MB/s size of sha1_process_block64
value 486 x86-64 486 x86-64
0 367 375 3657 3502
1 224 229 654 732
2,3 200 195 358 380
config SHA1_HWACCEL
bool "SHA1: Use hardware accelerated instructions if possible"
default y
help
On x86, this adds ~590 bytes of code. Throughput
is about twice as fast as fully-unrolled generic code.
config SHA256_HWACCEL
bool "SHA256: Use hardware accelerated instructions if possible"
default y
help
On x86, this adds ~1k bytes of code.
config SHA3_SMALL
int "SHA3: Trade bytes for speed (0:fast, 1:slow)"
default 1 # all "fast or small" options default to small
range 0 1
help
Trade binary size versus speed for the sha3 algorithm.
SHA3_SMALL=0 compared to SHA3_SMALL=1 (approximate):
64-bit x86: +270 bytes of code, 45% faster
32-bit x86: +450 bytes of code, 75% faster
config FEATURE_NON_POSIX_CP
bool "Non-POSIX, but safer, copying to special nodes"
default y
help
With this option, "cp file symlink" will delete symlink
and create a regular file. This does not conform to POSIX,
but prevents a symlink attack.
Similarly, "cp file device" will not send file's data
to the device. (To do that, use "cat file >device")
config FEATURE_VERBOSE_CP_MESSAGE
bool "Give more precise messages when copy fails (cp, mv etc)"
default n
help
Error messages with this feature enabled:
$ cp file /does_not_exist/file
cp: cannot create '/does_not_exist/file': Path does not exist
$ cp file /vmlinuz/file
cp: cannot stat '/vmlinuz/file': Path has non-directory component
If this feature is not enabled, they will be, respectively:
cp: cannot create '/does_not_exist/file': No such file or directory
cp: cannot stat '/vmlinuz/file': Not a directory
This will cost you ~60 bytes.
config FEATURE_USE_SENDFILE
bool "Use sendfile system call"
default y
help
When enabled, busybox will use the kernel sendfile() function
instead of read/write loops to copy data between file descriptors
(for example, cp command does this a lot).
If sendfile() doesn't work, copying code falls back to read/write
loop. sendfile() was originally implemented for faster I/O
from files to sockets, but since Linux 2.6.33 it was extended
to work for many more file types.
config FEATURE_COPYBUF_KB
int "Copy buffer size, in kilobytes"
range 1 1024
default 4
help
Size of buffer used by cp, mv, install, wget etc.
Buffers which are 4 kb or less will be allocated on stack.
Bigger buffers will be allocated with mmap, with fallback to 4 kb
stack buffer if mmap fails.
config MONOTONIC_SYSCALL
bool "Use clock_gettime(CLOCK_MONOTONIC) syscall"
default y
help
Use clock_gettime(CLOCK_MONOTONIC) syscall for measuring
time intervals (time, ping, traceroute etc need this).
Probably requires Linux 2.6+. If not selected, gettimeofday
will be used instead (which gives wrong results if date/time
is reset).
config IOCTL_HEX2STR_ERROR
bool "Use ioctl names rather than hex values in error messages"
default y
help
Use ioctl names rather than hex values in error messages
(e.g. VT_DISALLOCATE rather than 0x5608). If disabled this
saves about 1400 bytes.
config FEATURE_EDITING
bool "Command line editing"
default y
help
Enable line editing (mainly for shell command line).
config FEATURE_EDITING_MAX_LEN
int "Maximum length of input"
range 128 8192
default 1024
depends on FEATURE_EDITING
help
Line editing code uses on-stack buffers for storage.
You may want to decrease this parameter if your target machine
benefits from smaller stack usage.
config FEATURE_EDITING_VI
bool "vi-style line editing commands"
default n
depends on FEATURE_EDITING
help
Enable vi-style line editing. In shells, this mode can be
turned on and off with "set -o vi" and "set +o vi".
config FEATURE_EDITING_HISTORY
int "History size"
# Don't allow way too big values here, code uses fixed "char *history[N]" struct member
range 0 9999
default 255
depends on FEATURE_EDITING
help
Specify command history size (0 - disable).
config FEATURE_EDITING_SAVEHISTORY
bool "History saving"
default y
depends on FEATURE_EDITING
help
Enable history saving in shells.
config FEATURE_EDITING_SAVE_ON_EXIT
bool "Save history on shell exit, not after every command"
default n
depends on FEATURE_EDITING_SAVEHISTORY
help
Save history on shell exit, not after every command.
config FEATURE_REVERSE_SEARCH
bool "Reverse history search"
default y
depends on FEATURE_EDITING
help
Enable readline-like Ctrl-R combination for reverse history search.
Increases code by about 0.5k.
config FEATURE_TAB_COMPLETION
bool "Tab completion"
default y
depends on FEATURE_EDITING
config FEATURE_USERNAME_COMPLETION
bool "Username completion"
default y
depends on FEATURE_TAB_COMPLETION
config FEATURE_EDITING_FANCY_PROMPT
bool "Fancy shell prompts"
default y
depends on FEATURE_EDITING
help
Setting this option allows for prompts to use things like \w and
\$ and escape codes.
config FEATURE_EDITING_WINCH
bool "Enable automatic tracking of window size changes"
default y
depends on FEATURE_EDITING
config FEATURE_EDITING_ASK_TERMINAL
bool "Query cursor position from terminal"
default n
depends on FEATURE_EDITING
help
Allow usage of "ESC [ 6 n" sequence. Terminal answers back with
current cursor position. This information is used to make line
editing more robust in some cases.
If you are not sure whether your terminals respond to this code
correctly, or want to save on code size (about 400 bytes),
then do not turn this option on.
config LOCALE_SUPPORT
bool "Enable locale support (system needs locale for this to work)"
default n
help
Enable this if your system has locale support and you would like
busybox to support locale settings.
config UNICODE_SUPPORT
bool "Support Unicode"
default y
help
This makes various applets aware that one byte is not
one character on screen.
Busybox aims to eventually work correctly with Unicode displays.
Any older encodings are not guaranteed to work.
Probably by the time when busybox will be fully Unicode-clean,
other encodings will be mainly of historic interest.
config UNICODE_USING_LOCALE
bool "Use libc routines for Unicode (else uses internal ones)"
default n
depends on UNICODE_SUPPORT && LOCALE_SUPPORT
help
With this option on, Unicode support is implemented using libc
routines. Otherwise, internal implementation is used.
Internal implementation is smaller.
config FEATURE_CHECK_UNICODE_IN_ENV
bool "Check $LC_ALL, $LC_CTYPE and $LANG environment variables"
default n
depends on UNICODE_SUPPORT && !UNICODE_USING_LOCALE
help
With this option on, Unicode support is activated
only if locale-related variables have the value of the form
"xxxx.utf8"
Otherwise, Unicode support will be always enabled and active.
config SUBST_WCHAR
int "Character code to substitute unprintable characters with"
depends on UNICODE_SUPPORT
default 63
help
Typical values are 63 for '?' (works with any output device),
30 for ASCII substitute control code,
65533 (0xfffd) for Unicode replacement character.
config LAST_SUPPORTED_WCHAR
int "Range of supported Unicode characters"
depends on UNICODE_SUPPORT
default 767
help
Any character with Unicode value bigger than this is assumed
to be non-printable on output device. Many applets replace
such characters with substitution character.
The idea is that many valid printable Unicode chars
nevertheless are not displayed correctly. Think about
combining charachers, double-wide hieroglyphs, obscure
characters in dozens of ancient scripts...
Many terminals, terminal emulators, xterms etc will fail
to handle them correctly. Choose the smallest value
which suits your needs.
Typical values are:
126 - ASCII only
767 (0x2ff) - there are no combining chars in [0..767] range
(the range includes Latin 1, Latin Ext. A and B),
code is ~700 bytes smaller for this case.
4351 (0x10ff) - there are no double-wide chars in [0..4351] range,
code is ~300 bytes smaller for this case.
12799 (0x31ff) - nearly all non-ideographic characters are
available in [0..12799] range, including
East Asian scripts like katakana, hiragana, hangul,
bopomofo...
0 - off, any valid printable Unicode character will be printed.
config UNICODE_COMBINING_WCHARS
bool "Allow zero-width Unicode characters on output"
default n
depends on UNICODE_SUPPORT
help
With this option off, any Unicode char with width of 0
is substituted on output.
config UNICODE_WIDE_WCHARS
bool "Allow wide Unicode characters on output"
default n
depends on UNICODE_SUPPORT
help
With this option off, any Unicode char with width > 1
is substituted on output.
config UNICODE_BIDI_SUPPORT
bool "Bidirectional character-aware line input"
default n
depends on UNICODE_SUPPORT && !UNICODE_USING_LOCALE
help
With this option on, right-to-left Unicode characters
are treated differently on input (e.g. cursor movement).
config UNICODE_NEUTRAL_TABLE
bool "In bidi input, support non-ASCII neutral chars too"
default n
depends on UNICODE_BIDI_SUPPORT
help
In most cases it's enough to treat only ASCII non-letters
(i.e. punctuation, numbers and space) as characters
with neutral directionality.
With this option on, more extensive (and bigger) table
of neutral chars will be used.
config UNICODE_PRESERVE_BROKEN
bool "Make it possible to enter sequences of chars which are not Unicode"
default n
depends on UNICODE_SUPPORT
help
With this option on, on line-editing input (such as used by shells)
invalid UTF-8 bytes are not substituted with the selected
substitution character.
For example, this means that entering 'l', 's', ' ', 0xff, [Enter]
at shell prompt will list file named 0xff (single char name
with char value 255), not file named '?'.