Annotation of /tags/mkinitrd-6_3_1/busybox/TODO

Busybox TODO

Harvest patches from
http://git.openembedded.org/cgit.cgi/openembedded/tree/recipes/busybox/

Stuff that needs to be done.  This is organized by who plans to get around to
doing it eventually, but that doesn't mean they "own" the item.  If you want to
do one of these bounce an email off the person it's listed under to see if they
have any suggestions how they plan to go about it, and to minimize conflicts
between your work and theirs.  But otherwise, all of these are fair game.

Rob Landley suggested this:
  Implement bb_realpath() that can handle NULL on non-glibc.

  sh
    The command shell situation is a mess.  We have two different
    shells that don't really share any code, and the "standalone shell" doesn't
    work all that well (especially not in a chroot environment), due to apps not
    being reentrant.

  Do a SUSv3 audit
    Look at the full Single Unix Specification version 3 (available online at
    "http://www.opengroup.org/onlinepubs/009695399/nfindex.html") and
    figure out which of our apps are compliant, and what we're missing that
    we might actually care about.

    Even better would be some kind of automated compliance test harness that
    exercises each command line option and the various corner cases.

  Internationalization
    How much internationalization should we do?

    The low hanging fruit is UTF-8 character set support.  We should do this.
    See TODO_unicode file.

    We also have lots of hardwired english text messages.  Consolidating this
    into some kind of message table not only makes translation easier, but
    also allows us to consolidate redundant (or close) strings.

    We probably don't want to be bloated with locale support.  (Not unless we
    can cleanly export it from our underlying C library without having to
    concern ourselves with it directly.  Perhaps a few specific things like a
    config option for "date" are low hanging fruit here?)

    What level should things happen at?  How much do we care about
    internationalizing the text console when X11 and xterms are so much better
    at it?  (There's some infrastructure here we don't implement: The
    "unicode_start" and "unicode_stop" shell scripts need "vt-is-UTF8" and a
    --unicode option to loadkeys.  That implies a real loadkeys/dumpkeys
    implementation to replace loadkmap/dumpkmap.  Plus messing with console font
    loading.  Is it worth it, or do we just say "use X"?)

  Individual compilation of applets.
    It would be nice if busybox had the option to compile to individual applets,
    for people who want an alternate implementation less bloated than the gnu
    utils (or simply with less political baggage), but without it being one big
    executable.

    Turning libbb into a real dll is another possibility, especially if libbb
    could export some of the other library interfaces we've already more or less
    got the code for (like zlib).

  buildroot - Make a "dogfood" option
    Busybox 1.1 will be capable of replacing most gnu packages for real world
    use, such as developing software or in a live CD.  It needs wider testing.

    Busybox should now be able to replace bzip2, coreutils, e2fsprogs, file,
    findutils, gawk, grep, inetutils, less, modutils, net-tools, patch, procps,
    sed, shadow, sysklogd, sysvinit, tar, util-linux, and vim.  The resulting
    system should be self-hosting (I.E. able to rebuild itself from source
    code).  This means it would need (at least) binutils, gcc, and make, or
    equivalents.

    It would be a good "eating our own dogfood" test if buildroot had the option
    of using a "make allyesconfig" busybox instead of the all of the above
    packages.  Anything that's wrong with the resulting system, we can fix.  (It
    would be nice to be able to upgrade busybox to be able to replace bash and
    diffutils as well, but we're not there yet.)

    One example of an existing system that does this already is Firmware Linux:
      http://www.landley.net/code/firmware

  initramfs
    Busybox should have a sample initramfs build script.  This depends on
    bbsh, mdev, and switch_root.

  mkdep
    Write a mkdep that doesn't segfault if there's a directory it doesn't
    have permission to read, isn't based on manually editing the output of
    lexx and yacc, doesn't make such a mess under include/config, etc.

  Group globals into unions of structures.
    Go through and turn all the global and static variables into structures,
    and have all those structures be in a big union shared between processes,
    so busybox uses less bss.  (This is a big win on nommu machines.)  See
    sed.c and mdev.c for examples.

  Go through bugs.busybox.net and close out all of that somehow.
    This one's open to everybody, but I'll wind up doing it...

Bernhard Reutner-Fischer <busybox@busybox.net> suggests to look at these:
  New debug options:
    -Wlarger-than-127
    Cleanup any big users
  Collate BUFSIZ IOBUF_SIZE MY_BUF_SIZE PIPE_PROGRESS_SIZE BUFSIZE PIPESIZE
    make bb_common_bufsiz1 configurable, size wise.
    make pipesize configurable, size wise.
    Use bb_common_bufsiz1 throughout applets!

As yet unclaimed:

----
diff
  Make sure we handle empty files properly:
    From the patch man page:

    you can remove a file by sending out a context diff that compares
    the file to be deleted with an empty file dated the Epoch.  The
    file will be removed unless patch is conforming to POSIX and the
    -E or --remove-empty-files option is not given.
---
patch
  Should have simple fuzz factor support to apply patches at an offset which
  shouldn't take up too much space.

  And while we're at it, a new patch filename quoting format is apparently
  coming soon:  http://marc.theaimsgroup.com/?l=git&m=112927316408690&w=2
---
stty / catv
  stty's visible() function and catv's guts are identical. Merge them into
  an appropriate libbb function.
---
struct suffix_mult
  Several duplicate users of: grep -r "1024\*1024" * -B2 -A1
  Merge to a single size_suffixes[] in libbb.
  Users: head tail od_bloaty hexdump and (partially as it wouldn't hurt) svlogd
---
tail
  ./busybox tail -f foo.c~ TODO
  should not print fmt=header_fmt for subsequent date >> TODO; i.e. only
  fmt+ if another (not the current) file did change

Architectural issues:

bb_close() with fsync()
  We should have a bb_close() in place of normal close, with a CONFIG_ option
  to not just check the return value of close() for an error, but fsync().
  Close can't reliably report anything useful because if write() accepted the
  data then it either went out to the network or it's in cache or a pipe
  buffer.  Either way, there's no guarantee it'll make it to its final
  destination before close() gets called, so there's no guarantee that any
  error will be reported.

  You need to call fsync() if you care about errors that occur after write(),
  but that can have a big performance impact.  So make it a config option.
---
Unify archivers
  Lots of archivers have the same general infrastructure.  The directory
  traversal code should be factored out, and the guts of each archiver could
  be some setup code and a series of callbacks for "add this file",
  "add this directory", "add this symlink" and so on.

  This could clean up tar and zip, and make it cheaper to add cpio and ar
  write support, and possibly even cheaply add things like mkisofs or
  mksquashfs someday, if they become relevant.
---
Text buffer support.
  Several existing applets (sort, vi, less...) read
  a whole file into memory and act on it.  Use open_read_close().
---
Memory Allocation
  We have a CONFIG_BUFFER mechanism that lets us select whether to do memory
  allocation on the stack or the heap.  Unfortunately, we're not using it much.
  We need to audit our memory allocations and turn a lot of malloc/free calls
  into RESERVE_CONFIG_BUFFER/RELEASE_CONFIG_BUFFER.
  For a start, see e.g. make EXTRA_CFLAGS=-Wlarger-than-64

  And while we're at it, many of the CONFIG_FEATURE_CLEAN_UP #ifdefs will be
  optimized out by the compiler in the stack allocation case (since there's no
  free for an alloca()), and this means that various cleanup loops that just
  call free might also be optimized out by the compiler if written right, so
  we can yank those #ifdefs too, and generally clean up the code.
---
FEATURE_CLEAN_UP
  This is more an unresolved issue than a to-do item.  More thought is needed.

  Normally we rely on exit() to free memory, close files and unmap segments
  for us.  This makes most calls to free(), close(), and unmap() optional in
  busybox applets that don't intend to run for very long, and optional stuff
  can be omitted to save size.

  The idea was raised that we could simulate fork/exit with setjmp/longjmp
  for _really_ brainless embedded systems, or speed up the standalone shell
  by not forking.  Doing so would require a reliable FEATURE_CLEAN_UP.
  Unfortunately, this isn't as easy as it sounds.

  The problem is, lots of things exit(), sometimes unexpectedly (xmalloc())
  and sometimes reliably (bb_perror_msg_and_die() or show_usage()).  This
  jumps out of the normal flow control and bypasses any cleanup code we
  put at the end of our applets.

  It's possible to add hooks to libbb functions like xmalloc() and xopen()
  to add their entries to a linked list, which could be traversed and
  freed/closed automatically.  (This would need to be able to free just the
  entries after a checkpoint to be usable for a forkless standalone shell.
  You don't want to free the shell's own resources.)

  Right now, FEATURE_CLEAN_UP is more or less a debugging aid, to make things
  like valgrind happy.  It's also documentation of _what_ we're trusting
  exit() to clean up for us.  But new infrastructure to auto-free stuff would
  render the existing FEATURE_CLEAN_UP code redundant.

  For right now, exit() handles it just fine.


Minor stuff:
  watchdog.c could autodetect the timer duration via:
    if(!ioctl (fd, WDIOC_GETTIMEOUT, &tmo)) timer_duration = 1 + (tmo / 2);
  Unfortunately, that needs linux/watchdog.h and that contains unfiltered
  kernel types on some distros, which breaks the build.
---
  use bb_error_msg where appropriate: See
  egrep "(printf.*\([[:space:]]*(stderr|2)|[^_]write.*\([[:space:]]*(stderr|2))"
---
  use bb_perror_msg where appropriate: See
  egrep "[^_]perror"
---
  possible code duplication ingroup() and is_a_group_member()
---
  Move __get_hz() to a better place and (re)use it in route.c, ash.c
---
  See grep -r strtod
  Alot of duplication that wants cleanup.
---
  in_ether duplicated in network/{interface,ifconfig}.c
---
  unify progress_meter. wget, flash_eraseall, pipe_progress, fbsplash, setfiles.
---
  support start-stop-daemon -d <chdir-path>
---
vdprintf() -> similar sized functionality
---

(TODO list after discussion 11.05.2009)

* shrink tc/brctl/ip
  tc/brctl seem like fairly large things to try and tackle in your timeframe,
  and i think people have posted attempts in the past. Adding additional
  options to ip though seems reasonable.

* add tests for some applets

* implement POSIX utilities and audit them for POSIX conformance. then
  audit them for GNU conformance. then document all your findings in a new
  doc/conformance.txt file while perhaps implementing some of the missing
  features.
  you can find the latest POSIX documentation (1003.1-2008) here:
  http://www.opengroup.org/onlinepubs/9699919799/
  and the complete list of all utilities that POSIX covers:
  http://www.opengroup.org/onlinepubs/9699919799/idx/utilities.html
  The first step would to generate a file/matrix what is already archived
  (also IPV6)

* implement 'at'

* rpcbind (former portmap) or equivalent
  so that we don't have to use -o nolock on nfs mounts

* check IPV6 compliance

* generate a mini example using kernel+busybox only (+libc) for example

* more support for advanced linux 2.6.x features, see: iotop
  most likely there is more

* even more support for statistics: mpstat, iostat, powertop....
1	niro	532	Busybox TODO
2
3	niro	1123	Harvest patches from
4			http://git.openembedded.org/cgit.cgi/openembedded/tree/recipes/busybox/
5
6	niro	532	Stuff that needs to be done. This is organized by who plans to get around to
7			doing it eventually, but that doesn't mean they "own" the item. If you want to
8			do one of these bounce an email off the person it's listed under to see if they
9			have any suggestions how they plan to go about it, and to minimize conflicts
10			between your work and theirs. But otherwise, all of these are fair game.
11
12	niro	984	Rob Landley suggested this:
13			Implement bb_realpath() that can handle NULL on non-glibc.
14	niro	532
15			sh
16	niro	984	The command shell situation is a mess. We have two different
17	niro	532	shells that don't really share any code, and the "standalone shell" doesn't
18			work all that well (especially not in a chroot environment), due to apps not
19	niro	816	being reentrant.
20	niro	984
21	niro	532	Do a SUSv3 audit
22			Look at the full Single Unix Specification version 3 (available online at
23			"http://www.opengroup.org/onlinepubs/009695399/nfindex.html") and
24			figure out which of our apps are compliant, and what we're missing that
25			we might actually care about.
26
27			Even better would be some kind of automated compliance test harness that
28			exercises each command line option and the various corner cases.
29	niro	984
30	niro	532	Internationalization
31			How much internationalization should we do?
32
33			The low hanging fruit is UTF-8 character set support. We should do this.
34	niro	1123	See TODO_unicode file.
35	niro	532
36			We also have lots of hardwired english text messages. Consolidating this
37			into some kind of message table not only makes translation easier, but
38			also allows us to consolidate redundant (or close) strings.
39
40			We probably don't want to be bloated with locale support. (Not unless we
41			can cleanly export it from our underlying C library without having to
42			concern ourselves with it directly. Perhaps a few specific things like a
43			config option for "date" are low hanging fruit here?)
44
45			What level should things happen at? How much do we care about
46			internationalizing the text console when X11 and xterms are so much better
47			at it? (There's some infrastructure here we don't implement: The
48			"unicode_start" and "unicode_stop" shell scripts need "vt-is-UTF8" and a
49			--unicode option to loadkeys. That implies a real loadkeys/dumpkeys
50			implementation to replace loadkmap/dumpkmap. Plus messing with console font
51			loading. Is it worth it, or do we just say "use X"?)
52
53			Individual compilation of applets.
54			It would be nice if busybox had the option to compile to individual applets,
55			for people who want an alternate implementation less bloated than the gnu
56			utils (or simply with less political baggage), but without it being one big
57			executable.
58
59			Turning libbb into a real dll is another possibility, especially if libbb
60			could export some of the other library interfaces we've already more or less
61			got the code for (like zlib).
62	niro	1123
63	niro	532	buildroot - Make a "dogfood" option
64			Busybox 1.1 will be capable of replacing most gnu packages for real world
65			use, such as developing software or in a live CD. It needs wider testing.
66
67			Busybox should now be able to replace bzip2, coreutils, e2fsprogs, file,
68			findutils, gawk, grep, inetutils, less, modutils, net-tools, patch, procps,
69			sed, shadow, sysklogd, sysvinit, tar, util-linux, and vim. The resulting
70			system should be self-hosting (I.E. able to rebuild itself from source
71			code). This means it would need (at least) binutils, gcc, and make, or
72			equivalents.
73
74			It would be a good "eating our own dogfood" test if buildroot had the option
75			of using a "make allyesconfig" busybox instead of the all of the above
76			packages. Anything that's wrong with the resulting system, we can fix. (It
77			would be nice to be able to upgrade busybox to be able to replace bash and
78			diffutils as well, but we're not there yet.)
79
80			One example of an existing system that does this already is Firmware Linux:
81			http://www.landley.net/code/firmware
82	niro	1123
83	niro	532	initramfs
84			Busybox should have a sample initramfs build script. This depends on
85			bbsh, mdev, and switch_root.
86	niro	1123
87	niro	532	mkdep
88			Write a mkdep that doesn't segfault if there's a directory it doesn't
89			have permission to read, isn't based on manually editing the output of
90			lexx and yacc, doesn't make such a mess under include/config, etc.
91	niro	1123
92	niro	532	Group globals into unions of structures.
93			Go through and turn all the global and static variables into structures,
94			and have all those structures be in a big union shared between processes,
95			so busybox uses less bss. (This is a big win on nommu machines.) See
96			sed.c and mdev.c for examples.
97	niro	1123
98	niro	532	Go through bugs.busybox.net and close out all of that somehow.
99			This one's open to everybody, but I'll wind up doing it...
100
101	niro	816	Bernhard Reutner-Fischer <busybox@busybox.net> suggests to look at these:
102	niro	532	New debug options:
103			-Wlarger-than-127
104			Cleanup any big users
105			Collate BUFSIZ IOBUF_SIZE MY_BUF_SIZE PIPE_PROGRESS_SIZE BUFSIZE PIPESIZE
106			make bb_common_bufsiz1 configurable, size wise.
107			make pipesize configurable, size wise.
108			Use bb_common_bufsiz1 throughout applets!
109
110			As yet unclaimed:
111
112			----
113			diff
114			Make sure we handle empty files properly:
115			From the patch man page:
116
117			you can remove a file by sending out a context diff that compares
118			the file to be deleted with an empty file dated the Epoch. The
119			file will be removed unless patch is conforming to POSIX and the
120			-E or --remove-empty-files option is not given.
121			---
122			patch
123			Should have simple fuzz factor support to apply patches at an offset which
124			shouldn't take up too much space.
125
126			And while we're at it, a new patch filename quoting format is apparently
127			coming soon: http://marc.theaimsgroup.com/?l=git&m=112927316408690&w=2
128			---
129			stty / catv
130			stty's visible() function and catv's guts are identical. Merge them into
131			an appropriate libbb function.
132	niro	816	---
133			struct suffix_mult
134			Several duplicate users of: grep -r "1024\1024" -B2 -A1
135			Merge to a single size_suffixes[] in libbb.
136			Users: head tail od_bloaty hexdump and (partially as it wouldn't hurt) svlogd
137			---
138			tail
139			./busybox tail -f foo.c~ TODO
140			should not print fmt=header_fmt for subsequent date >> TODO; i.e. only
141			fmt+ if another (not the current) file did change
142	niro	532
143			Architectural issues:
144
145			bb_close() with fsync()
146			We should have a bb_close() in place of normal close, with a CONFIG_ option
147			to not just check the return value of close() for an error, but fsync().
148			Close can't reliably report anything useful because if write() accepted the
149			data then it either went out to the network or it's in cache or a pipe
150			buffer. Either way, there's no guarantee it'll make it to its final
151			destination before close() gets called, so there's no guarantee that any
152			error will be reported.
153
154			You need to call fsync() if you care about errors that occur after write(),
155			but that can have a big performance impact. So make it a config option.
156			---
157			Unify archivers
158			Lots of archivers have the same general infrastructure. The directory
159			traversal code should be factored out, and the guts of each archiver could
160			be some setup code and a series of callbacks for "add this file",
161			"add this directory", "add this symlink" and so on.
162
163			This could clean up tar and zip, and make it cheaper to add cpio and ar
164			write support, and possibly even cheaply add things like mkisofs or
165			mksquashfs someday, if they become relevant.
166			---
167			Text buffer support.
168			Several existing applets (sort, vi, less...) read
169	niro	984	a whole file into memory and act on it. Use open_read_close().
170	niro	532	---
171			Memory Allocation
172			We have a CONFIG_BUFFER mechanism that lets us select whether to do memory
173			allocation on the stack or the heap. Unfortunately, we're not using it much.
174			We need to audit our memory allocations and turn a lot of malloc/free calls
175			into RESERVE_CONFIG_BUFFER/RELEASE_CONFIG_BUFFER.
176			For a start, see e.g. make EXTRA_CFLAGS=-Wlarger-than-64
177
178			And while we're at it, many of the CONFIG_FEATURE_CLEAN_UP #ifdefs will be
179			optimized out by the compiler in the stack allocation case (since there's no
180			free for an alloca()), and this means that various cleanup loops that just
181			call free might also be optimized out by the compiler if written right, so
182			we can yank those #ifdefs too, and generally clean up the code.
183			---
184			FEATURE_CLEAN_UP
185			This is more an unresolved issue than a to-do item. More thought is needed.
186
187	niro	984	Normally we rely on exit() to free memory, close files and unmap segments
188	niro	532	for us. This makes most calls to free(), close(), and unmap() optional in
189			busybox applets that don't intend to run for very long, and optional stuff
190			can be omitted to save size.
191
192			The idea was raised that we could simulate fork/exit with setjmp/longjmp
193			for _really_ brainless embedded systems, or speed up the standalone shell
194			by not forking. Doing so would require a reliable FEATURE_CLEAN_UP.
195			Unfortunately, this isn't as easy as it sounds.
196
197			The problem is, lots of things exit(), sometimes unexpectedly (xmalloc())
198			and sometimes reliably (bb_perror_msg_and_die() or show_usage()). This
199			jumps out of the normal flow control and bypasses any cleanup code we
200			put at the end of our applets.
201
202			It's possible to add hooks to libbb functions like xmalloc() and xopen()
203			to add their entries to a linked list, which could be traversed and
204			freed/closed automatically. (This would need to be able to free just the
205			entries after a checkpoint to be usable for a forkless standalone shell.
206			You don't want to free the shell's own resources.)
207
208			Right now, FEATURE_CLEAN_UP is more or less a debugging aid, to make things
209			like valgrind happy. It's also documentation of _what_ we're trusting
210			exit() to clean up for us. But new infrastructure to auto-free stuff would
211			render the existing FEATURE_CLEAN_UP code redundant.
212
213			For right now, exit() handles it just fine.
214
215
216			Minor stuff:
217			watchdog.c could autodetect the timer duration via:
218			if(!ioctl (fd, WDIOC_GETTIMEOUT, &tmo)) timer_duration = 1 + (tmo / 2);
219			Unfortunately, that needs linux/watchdog.h and that contains unfiltered
220			kernel types on some distros, which breaks the build.
221			---
222			use bb_error_msg where appropriate: See
223			egrep "(printf.\([[:space:]](stderr\|2)\|[^_]write.\([[:space:]](stderr\|2))"
224			---
225			use bb_perror_msg where appropriate: See
226			egrep "[^_]perror"
227			---
228			possible code duplication ingroup() and is_a_group_member()
229			---
230	niro	984	Move __get_hz() to a better place and (re)use it in route.c, ash.c
231	niro	532	---
232	niro	816	See grep -r strtod
233			Alot of duplication that wants cleanup.
234			---
235			in_ether duplicated in network/{interface,ifconfig}.c
236			---
237	niro	984	unify progress_meter. wget, flash_eraseall, pipe_progress, fbsplash, setfiles.
238			---
239	niro	1123	support start-stop-daemon -d <chdir-path>
240	niro	532	---
241			vdprintf() -> similar sized functionality
242			---
243	niro	984
244			(TODO list after discussion 11.05.2009)
245
246			* shrink tc/brctl/ip
247			tc/brctl seem like fairly large things to try and tackle in your timeframe,
248			and i think people have posted attempts in the past. Adding additional
249			options to ip though seems reasonable.
250
251			* add tests for some applets
252
253			* implement POSIX utilities and audit them for POSIX conformance. then
254			audit them for GNU conformance. then document all your findings in a new
255			doc/conformance.txt file while perhaps implementing some of the missing
256			features.
257			you can find the latest POSIX documentation (1003.1-2008) here:
258			http://www.opengroup.org/onlinepubs/9699919799/
259			and the complete list of all utilities that POSIX covers:
260			http://www.opengroup.org/onlinepubs/9699919799/idx/utilities.html
261			The first step would to generate a file/matrix what is already archived
262			(also IPV6)
263
264			* implement 'at'
265
266			* rpcbind (former portmap) or equivalent
267			so that we don't have to use -o nolock on nfs mounts
268
269			* check IPV6 compliance
270
271			* generate a mini example using kernel+busybox only (+libc) for example
272
273			* more support for advanced linux 2.6.x features, see: iotop
274			most likely there is more
275
276			* even more support for statistics: mpstat, iostat, powertop....