Commits · 8cf5b6208cad807228a69d695e6521122d4b71da · mirror / ccan

17 Dec, 2012 3 commits

tal: don't automatically register cleanup function. · 8cf5b620

Rusty Russell authored Dec 17, 2012

It may interfere with other at_exit() calls, so let them call it manually.
Also, use memset to zero, which really does make valgrind notice any leaks.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

8cf5b620

configurator: fix single-arg case. · d21d629a

Rusty Russell authored Dec 17, 2012

In this case, CCAN_CFLAGS should be "".
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

d21d629a

tal: append the length property to the initial allocation. · 4814c4c7

Rusty Russell authored Dec 17, 2012

Since we never post-add a length property, we can save some cycles by
allocating it together with the object itself:

Before:
	$ ./samba-allocs talloc.dump --tal
	Tal time:                6058997-6215211(6102645)ns
	Tal_free time:           4791112-4968964(4824814)ns
	Single tal_free time:    3173647-3331584(3200183)ns
	$ ./samba-allocs talloc.dump --tal-size
	Virtual size = 9809920, RSS = 3952640

After:
	$ ./samba-allocs talloc.dump --tal
	Tal time:                5911994-6043808(5956914)ns
	Tal_free time:           4676026-4834598(4719908)ns
	Single tal_free time:    2888195-3067365(2922298)ns
	$ ./samba-allocs talloc.dump --tal-size
	Virtual size = 9809920, RSS = 3948544
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

4814c4c7

14 Dec, 2012 3 commits

tal: don't always allocate an array in the samba benchmark. · 3710d289

Rusty Russell authored Dec 15, 2012

Since we added the length property, this penalizes us with an extra
alloc.  Unfortunately, we can't tell from the talloc.dump which are
array tallocs and which are single tallocs, but grepping the sources
gives a rate of 1 in 27.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

3710d289

tal: fix skipping tal in samba-allocs benchmark. · dfaf896a
Rusty Russell authored Dec 15, 2012
```
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
```
dfaf896a
tal: fix compilation of speed benchmark. · 769352e9
Rusty Russell authored Dec 15, 2012
```
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
```
769352e9

07 Dec, 2012 1 commit
- ccanlint: fix --compiler and --cflags options to apply to _info files as well. · 0248fa2c
  Rusty Russell authored Dec 07, 2012
```
We weren't using the compiler and cflags options in tools/compile.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
```
  0248fa2c
03 Dec, 2012 27 commits

tools: manifest.c should use path_canon. · e0529e15
Rusty Russell authored Dec 03, 2012
```
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
```
e0529e15
ccanlint: remove redundant num_lines in struct ccan_file. · 5f551788
Rusty Russell authored Dec 03, 2012
```
We can use tal_count.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
```
5f551788

ccanlint: print module name properly for nested modules. · 756407b4

Rusty Russell authored Dec 03, 2012

eg:
	tal/str: Module's tests cover all the code (tests_coverage): PASS (+3/6)
	tal/str: Total score: 55/58
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

756407b4

tools: use tal/path instead of opencoding most paths. · 49a1a436
Rusty Russell authored Dec 03, 2012
```
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
```
49a1a436
tools: use tal/path instead of writing own path handlers. · 09d5cd70
Rusty Russell authored Dec 03, 2012
```
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
```
09d5cd70
tools: use tal instead of talloc. · dc8042b4
Rusty Russell authored Dec 03, 2012
```
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
```
dc8042b4

tools: use rbuf instead of grab_file. · aabf300e

Rusty Russell authored Dec 03, 2012

Slowly removing the talloc dependency.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

aabf300e

rbuf: new module. · 44b37ee8
Rusty Russell authored Dec 03, 2012
```
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
```
44b37ee8

tal/link: new module for reference-count style coding for tal. · 72ce8140

Rusty Russell authored Dec 03, 2012

Tests and example copied from talloc_link, which I also wrote.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

72ce8140

tal/str: add tal_append_fmt() and tal_append_vfmt() helpers. · e7ae27d6
Rusty Russell authored Dec 03, 2012
```
They are far too convenient, despite their inefficiency.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
```
e7ae27d6
tal/str: rename tal_asprintf/tal_vasprintf to tal_fmt/tal_vfmt. · 4710a928
Rusty Russell authored Dec 03, 2012
```
Shorter, sweeter, and less legacy.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
```
4710a928

tal/str: add tal_strcat(). · a5ebf314

Rusty Russell authored Dec 03, 2012

Useful for joining two strings.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

a5ebf314

tal: add tal_expand(). · 680aa496

Rusty Russell authored Dec 03, 2012

Helper to reallocate and append to an array.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

680aa496

tal/str: move tal string functions here from tal. · 5582b011

Rusty Russell authored Dec 03, 2012

They don't need anything internal to tal, they're just helpers.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

5582b011

tal/str: use tal_ prefix. · 15ed4f45

Rusty Russell authored Dec 03, 2012

It's short, and much clearer.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

15ed4f45

tal/str: make tal_count() work for strsplit. · 737dcc0f
Rusty Russell authored Dec 03, 2012
```
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
```
737dcc0f

tal: add tal_count() and length properties for arrays. · 83c75170

Rusty Russell authored Dec 03, 2012

Unlike talloc, we don't have a count on everything, just on array
allocations.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

83c75170

tal: adding or removing a notifier/destructor can be const. · 9a144114

Rusty Russell authored Dec 03, 2012

You don't need write access to the context to attach a destructor;
it's meta.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

9a144114

tal: optimize case where no (non-destructor) notifiers are used. · fce7c22f

Rusty Russell authored Dec 03, 2012

$ for i in `seq 10`; do ./samba-allocs talloc.dump --tal; done | stats.c --trim-outliers
Before:
Read 25998 nodes
Tal time:                2857988-2981158(2904324)ns
Tal_free time:           2417170-2582088(2470931)ns
Single tal_free time:    1551654-1609583(1575984)ns

After:
Read 25998 nodes
Tal time:                2780974-2901323(2839352)ns
Tal_free time:           2241725-2355309(2285165)ns
Single tal_free time:    1549908-1612485(1566609)ns
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

fce7c22f

tal: use notifier infrastructure for destructors. · 68b5c2c8
Rusty Russell authored Dec 03, 2012
```
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
```
68b5c2c8
tal: add general notifiers. · d73a2bcb
Rusty Russell authored Dec 03, 2012
```
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
```
d73a2bcb
tal: add del_destructor(). · ca6991d8
Rusty Russell authored Dec 03, 2012
```
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
```
ca6991d8

tal: check headers more carefully. · d73447c2

Rusty Russell authored Dec 03, 2012

We sanity check tal headers by ensuring that the pointers are in the
bounds of things we've allocated.  But the first one we check is the
prop ptr, which may also be a literal: this is_literal() dereferences
the pointer, which means we usually crash here if it's not a tal
object.

Move that last, and we have far more success with our sanity checking.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

d73447c2

tal: simplify. · 0f6d854a

Rusty Russell authored Dec 03, 2012

All the effort trying to keep the header size down to 2 pointers turns
out to be wasted.  In addition, getting the parent of a tal pointer is
now much faster.

./samba-allocs talloc.dump --tal-size:
Before:
	Virtual size = 9633792, RSS = 3952640
After:
	Virtual size = 9793536, RSS = 3948544

And we're much faster now, esp. on free.

./samba-allocs talloc.dump --tal:
Before:
	Tal time:                2718068ns
	Tal_free time:           3360258ns
	Single tal_free time:    1667412ns
After:
	Tal time:                2788287ns
	Tal_free time:           2290167ns
	Single tal_free time:    1566998ns
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>


Header from folded patch 'tal-fix-set-destroying-bit.patch':

tal: fix destroying bit usage.

Actually looking at a destroying contexts' parent didn't work, and a couple
of places didn't filter out the destroying bit.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

0f6d854a

ccanlint: handle nested modules when mentioned in examples. · 932d65dd

Rusty Russell authored Dec 03, 2012

If we see an ccan include line in an example, we add that to the
dependencies to try to build the example, but we didn't handle nested
modules.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

932d65dd

tools: don't abort on malformed documentation lines. · 6a4d4533

Rusty Russell authored Dec 03, 2012

ccanlint would abort with 'Malformed line 53' if there was a bad header.
That's very poor, and deeply unhelpful.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

6a4d4533

Makefile: score nested modules. · 2b7f1fca
Rusty Russell authored Dec 03, 2012
```
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
```
2b7f1fca

22 Nov, 2012 6 commits

tal: fix benchmark/Makefile · 07672941
Rusty Russell authored Nov 22, 2012
```
We need ccan/take here.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
```
07672941

tal/path: new module · 892f59bd

Rusty Russell authored Nov 22, 2012

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>


Header from folded patch 'path-talloc-take.patch':

tal/path: accept take() for arguments.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

892f59bd

configurator: test for HAVE_FCHDIR · 1322fd33
Rusty Russell authored Nov 22, 2012
```
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
```
1322fd33
tal/str: accept take() for arguments. · f103f73f
Rusty Russell authored Nov 22, 2012
```
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
```
f103f73f

tal/str: new module. · d873aaec

Rusty Russell authored Nov 22, 2012

Tal-enhanced string routines, copied from str_talloc (which I also wrote).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

d873aaec

alloc: move into antithread/alloc. · d379e0ae

Rusty Russell authored Nov 22, 2012

Our first nested module; easy because noone else relies on it.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>

d379e0ae