Skip to contentSkip to conversation

-tiny tree for small systems (2.6.0-test11)

ANNOUNCE Matt Mackall 4 participants 7 messages Dec 12, 2003 Scheduler

lore
Share & tools
lore
#0 lore raw

This is the first release of a new kernel tree dubbed '-tiny' (someone already took -mm). The aim of this tree is to collect patches that reduce kernel disk and memory footprint as well as tools for working on small systems, an area Linux mainstream has been moving away from since Linus got a real job. Target users are things like embedded systems, small or legacy desktop folks, and handhelds.

To get the ball rolling, I've thrown in about 50 patches that trim various bits of the kernel, almost all configurable, and a fair number may eventually be appropriate for mainline. All the config options are currently thrown under CONFIG_EMBEDDED and many of the minor tweaks are covered under a set of config options called CONFIG_CORE_SMALL, CONFIG_NET_SMALL, and CONFIG_CONSOLE_SMALL.

Nifty things I've included:
- building with -Os
- 4k process stacks (via -wli)
- configurable removal of printk, BUG, and panic() strings
- configurable HZ
- configurable support for vm86, core dumps, kcore, sysfs, aio, etc.
- a very nice kmalloc auditing system via /proc/kmalloc
- auditing of bootmem usage
- a system for counting inline instantiations
- my netpoll/netconsole patches
- my drivers/char/random fixups

Some items on my todo list:
- borrow kgdb from -mm
- merge my netpoll-based kgdb-over-ethernet
- pageable kernel memory for a select subset of the kernel
- reduced functionality vt for small gui systems
- generic lookup/hash management code

How small is -tiny? It's hard to quantify as it's all configurable and some functionality is more important than others, but my current test config has full IPv4 net stack and most other important functionality and will boot comfortably on a 4M x86 box with about 2M free+buffers+cache.

Bug reports, suggestions, and patch submissions are welcome!

The patch, currently against 2.6.0-test11, can be found at:

http://selenic.com/tiny/

Full patch list:

# netpoll
#
netpoll-core.patch
tg3-poll.patch
netconsole.patch
#
# random
#
debug-cleanup.patch
fix-random-wait.patch
kill-pool-resize.patch
kill-getstate.patch
pool-rename.patch
pool-init.patch
kill-extract-secondary.patch
bug-on-grb.patch
extract-min-max.patch
debit-entropy.patch
reserved-in-struct.patch
nonblocking-pool.patch
kill-extract-state.patch
kill-md5.patch
kill-sha-variants.patch
random-waitqueue.patch
shrink-random.patch
ln-to-fls.patch
random-sched-clock.patch
kill-rotate.patch
kill-batching.patch
random-credit.patch
tiny-syncookie.patch
static-rnd-pools.patch
#
# akpm
#
gcc-Os-if-embedded.patch
fix-sqrt.patch
scale-min_free_kbytes.patch
config_spinline.patch
#
# new
# 
tiny-extraversion.patch
core-small.patch
config-net-small.patch
console-small.patch
deprecate-inline.patch
audit-bootmem.patch
kmalloc-accounting.patch
change-hz.patch
remove-vm86.patch
ide-hwif.patch
tiny-crc.patch
no-doublefault.patch
nosysfs.patch
tg3-oops.patch
hash-sizes.patch
ether-queue.patch
cache_defer_hash.patch
con_buf.patch
serial-pci.patch
vprintk.patch
kill-ext2-error-buff.patch
kill-ext3-error-buff.patch
kill-printk.patch
tiny-bug.patch
tiny-panic.patch
rolled-md4.patch
kill-ext3-md4.patch
no-elf-core.patch
no-kcore.patch
max-swapfiles.patch
user-hash.patch
futex-queues.patch
proc_alloc_map.patch
ldiscs.patch
tvec_bases.patch
dmi_blacklist.patch
no-translations.patch
unix_socket_table.patch
bh_wait_queue_heads.patch
major_names.patch
chrdevs.patch
pid-max.patch
roll-inflate-crc.patch
no-aio.patch
max_user_rt_prio.patch
ethtool.patch
4k-stacks.patch
Signature · 1 line
Matt Mackall : http://www.selenic.com : Linux development and consulting
#1 lore raw
Quoted 6 lines · Matt Mackall

To get the ball rolling, I've thrown in about 50 patches that trim various bits of the kernel, almost all configurable, and a fair number may eventually be appropriate for mainline. All the config options are currently thrown under CONFIG_EMBEDDED and many of the minor tweaks are covered under a set of config options called CONFIG_CORE_SMALL, CONFIG_NET_SMALL, and CONFIG_CONSOLE_SMALL.

A small trimming for you:

include/linux/sched.h +0 -1
diff -prauN linux-2.6.0-test11/include/linux/sched.h wli-2.6.0-test11-30/include/linux/sched.h
@@ -205,7 +208,6 @@ struct mm_struct {
 	unsigned long rss, total_vm, locked_vm;
 	unsigned long def_flags;
 	cpumask_t cpu_vm_mask;
-	unsigned long swap_address;
 
 	unsigned long saved_auxv[40]; /* for /proc/PID/auxv */
 
#2 lore raw
In reply to Matt Mackall
Quoted 14 lines · Matt Mackall

This is the first release of a new kernel tree dubbed '-tiny' (someone already took -mm). The aim of this tree is to collect patches that reduce kernel disk and memory footprint as well as tools for working on small systems, an area Linux mainstream has been moving away from since Linus got a real job. Target users are things like embedded systems, small or legacy desktop folks, and handhelds.

To get the ball rolling, I've thrown in about 50 patches that trim various bits of the kernel, almost all configurable, and a fair number may eventually be appropriate for mainline. All the config options are currently thrown under CONFIG_EMBEDDED and many of the minor tweaks are covered under a set of config options called CONFIG_CORE_SMALL, CONFIG_NET_SMALL, and CONFIG_CONSOLE_SMALL.

Quoted 11 lines

Nifty things I've included:
- building with -Os
- 4k process stacks (via -wli)
- configurable removal of printk, BUG, and panic() strings
- configurable HZ
- configurable support for vm86, core dumps, kcore, sysfs, aio, etc.
- a very nice kmalloc auditing system via /proc/kmalloc
- auditing of bootmem usage
- a system for counting inline instantiations
- my netpoll/netconsole patches
- my drivers/char/random fixups

I'd like to suggest you check out the "tweaks" idea I tossed out here:
http://www.ussg.iu.edu/hypermail/linux/kernel/0211.0/2229.html
If this sounds interesting, I've got a version of the patch (albeit old and not applying directly right now I bet) that moved things into header files and got all of the dependancy stuff correct except for the initial run (so I think I was forcing an update with any make invocation, but there were no spurious recompiles).

Signature · 2 lines
Tom Rini
http://gate.crashing.org/~trini/
#3 lore raw
Quoted 28 lines · Tom Rini
Quoted 14 lines · Matt Mackall

This is the first release of a new kernel tree dubbed '-tiny' (someone already took -mm). The aim of this tree is to collect patches that reduce kernel disk and memory footprint as well as tools for working on small systems, an area Linux mainstream has been moving away from since Linus got a real job. Target users are things like embedded systems, small or legacy desktop folks, and handhelds.

To get the ball rolling, I've thrown in about 50 patches that trim various bits of the kernel, almost all configurable, and a fair number may eventually be appropriate for mainline. All the config options are currently thrown under CONFIG_EMBEDDED and many of the minor tweaks are covered under a set of config options called CONFIG_CORE_SMALL, CONFIG_NET_SMALL, and CONFIG_CONSOLE_SMALL.

Quoted 11 lines

Nifty things I've included:
- building with -Os
- 4k process stacks (via -wli)
- configurable removal of printk, BUG, and panic() strings
- configurable HZ
- configurable support for vm86, core dumps, kcore, sysfs, aio, etc.
- a very nice kmalloc auditing system via /proc/kmalloc
- auditing of bootmem usage
- a system for counting inline instantiations
- my netpoll/netconsole patches
- my drivers/char/random fixups

Quoted 7 lines

I'd like to suggest you check out the "tweaks" idea I tossed out here:
http://www.ussg.iu.edu/hypermail/linux/kernel/0211.0/2229.html
If this sounds interesting, I've got a version of the patch (albeit old and not applying directly right now I bet) that moved things into header files and got all of the dependancy stuff correct except for the initial run (so I think I was forcing an update with any make invocation, but there were no spurious recompiles).

Looks cool. My only worry is that to do it right, it has to make some fairly sweeping changes. I'm trying to keep the stuff in -tiny fairly small and independent so that stuff can be cherry-picked, but if we can get a consensus that "tweaks" is a good direction for mainline, it might prove useful for some of the stuff I'm doing with CONFIG_CORE_SMALL and friends now.

Signature · 1 line
Matt Mackall : http://www.selenic.com : Linux development and consulting
#4 lore raw
Quoted 37 lines · Matt Mackall
Quoted 28 lines · Tom Rini
Quoted 14 lines · Matt Mackall

This is the first release of a new kernel tree dubbed '-tiny' (someone already took -mm). The aim of this tree is to collect patches that reduce kernel disk and memory footprint as well as tools for working on small systems, an area Linux mainstream has been moving away from since Linus got a real job. Target users are things like embedded systems, small or legacy desktop folks, and handhelds.

To get the ball rolling, I've thrown in about 50 patches that trim various bits of the kernel, almost all configurable, and a fair number may eventually be appropriate for mainline. All the config options are currently thrown under CONFIG_EMBEDDED and many of the minor tweaks are covered under a set of config options called CONFIG_CORE_SMALL, CONFIG_NET_SMALL, and CONFIG_CONSOLE_SMALL.

Quoted 11 lines

Nifty things I've included:
- building with -Os
- 4k process stacks (via -wli)
- configurable removal of printk, BUG, and panic() strings
- configurable HZ
- configurable support for vm86, core dumps, kcore, sysfs, aio, etc.
- a very nice kmalloc auditing system via /proc/kmalloc
- auditing of bootmem usage
- a system for counting inline instantiations
- my netpoll/netconsole patches
- my drivers/char/random fixups

Quoted 7 lines

I'd like to suggest you check out the "tweaks" idea I tossed out here:
http://www.ussg.iu.edu/hypermail/linux/kernel/0211.0/2229.html
If this sounds interesting, I've got a version of the patch (albeit old and not applying directly right now I bet) that moved things into header files and got all of the dependancy stuff correct except for the initial run (so I think I was forcing an update with any make invocation, but there were no spurious recompiles).

Quoted 6 lines

Looks cool. My only worry is that to do it right, it has to make some fairly sweeping changes. I'm trying to keep the stuff in -tiny fairly small and independent so that stuff can be cherry-picked, but if we can get a consensus that "tweaks" is a good direction for mainline, it might prove useful for some of the stuff I'm doing with CONFIG_CORE_SMALL and friends now.

Well part of the problem that came up when this was brought up during 2.5 is that adding a whole bunch of CONFIG options for things Just Won't Happen (too complex, PITA, etc). OTOH however, lots of stuff like that keeps getting in.

Signature · 2 lines
Tom Rini
http://gate.crashing.org/~trini/
#5 lore raw

In article <[email protected]>, Tom Rini <[email protected]> wrote:

| Well part of the problem that came up when this was brought up during
| 2.5 is that adding a whole bunch of CONFIG options for things Just Won't
| Happen (too complex, PITA, etc).  OTOH however, lots of stuff like that
| keeps getting in.

But now we have a config section for disabling things which may not be needed in embedded systems. So there's a fair chance that tweaks and such can be accepted as long as they're in the area where most will never go.

Signature · 3 lines
bill davidsen <[email protected]>
  CTO, TMR Associates, Inc
Doing interesting things with little computers since 1979.
#6 lore raw
Quoted 8 lines · bill davidsen

In article <[email protected]>, Tom Rini <[email protected]> wrote:

| Well part of the problem that came up when this was brought up during
| 2.5 is that adding a whole bunch of CONFIG options for things Just Won't
| Happen (too complex, PITA, etc).  OTOH however, lots of stuff like that
| keeps getting in.
Quoted 4 lines

But now we have a config section for disabling things which may not be needed in embedded systems. So there's a fair chance that tweaks and such can be accepted as long as they're in the area where most will never go.

That was the case at the time as well.

Signature · 2 lines
Tom Rini
http://gate.crashing.org/~trini/