Commit Graph

332 Commits

Author SHA1 Message Date
Jan Engelhardt
28ecf45548 pknock: redo inline hints
The compiler is usually smart enough to decide this on its own.
Some functions however do not warrant even the hint keyword.
2009-10-03 01:26:31 +02:00
Jan Engelhardt
58839b9b86 pknock: use bools for two-state ints 2009-10-03 01:26:31 +02:00
Jan Engelhardt
e436948ebd pknock: use explicit tests for non-bool types 2009-10-03 01:26:30 +02:00
Jan Engelhardt
79d54ebd7c pknock: split long instructions, esp. allocations in if() 2009-10-03 01:26:28 +02:00
Jan Engelhardt
d116000784 pknock: fix return values and memleak in has_secret
First, there is a memory leak - when the crypto functions fail,
hexresult was not freed. Second, in that error case, the return value
is simply passed up to the caller, ignoring the different meanings of
0 and 1 between crypto and pknock.
2009-09-30 01:19:59 +02:00
Jan Engelhardt
4b5f0b12d9 pknock: remove initialization of data where not needed
bss will be set to zero anyway.
For automatic variables, the preinit is not always needed
(gcc will warn about it).
2009-09-30 01:19:58 +02:00
Jan Engelhardt
c8dde526f2 pknock: use unsigned qualifiers where they are due 2009-09-30 01:19:39 +02:00
Jan Engelhardt
3058e42a21 pknock: remove pointless casts 2009-09-30 00:49:30 +02:00
Jan Engelhardt
40b30f7f27 pknock: add const qualifiers 2009-09-30 00:49:29 +02:00
Jan Engelhardt
aaad5f53ac pknock: remove pointless sizeof(char)
sizeof(char) is defined to be 1, anyway.
2009-09-30 00:09:07 +02:00
Jan Engelhardt
db1057158f pknock: return -ENXIO for problems during init
Some functions like the crypto init or proc_mkdir do not return an
error value. Replace the -1, which is actually EPERM, and looks a
little out of place, and use ENXIO instead. (After all, the requested
objects were not there after the request of initialization.)
2009-09-30 00:05:42 +02:00
Jan Rafaj
713c31e8f7 pknock: guard off crypto code from kernels before 2.6.19 2009-09-29 23:47:03 +02:00
Jan Engelhardt
e5ffb39750 pknock: rename DEBUGP to pk_debug and always enable
Since pr_debug depends on #if DEBUG anyway, there is no need to
contain DEBUGP/pk_debug within a further if. Rename from DEBUGP to
pk_debug, because DEBUGP is flagged by Xtables-addons (- and that's
good so, because often DEBUGPs should have been pr_debug instead).
2009-09-29 23:45:29 +02:00
Jan Rafaj
19067393b6 pknock: replace printks by pr_debug
The printk calls were protected using #if DEBUG anyway, so they can
just be changed to pr_debug. pr_debug also depends on #if DEBUG, and
takes care of the module prefix.
2009-09-29 23:45:29 +02:00
Jan Engelhardt
1de82a88a1 pknock: avoid accidental deletion of rules
If DEBUG was not defined, and no rule was effectively found, the
"rule" variable may still be non-NULL, and possibly cause deletion.
Therefore, always check for the rule having been found.
2009-09-29 23:45:29 +02:00
Jan Engelhardt
9928c864ab pknock: drop evil-tinygram warning messages
The duprintf macro then also becomes unneeded.
2009-09-29 23:45:28 +02:00
Jan Engelhardt
45515a0afd pknock: bump internal match revision number
Do this so that previous iptables installations do not mistakenly
operate with our new xt_pknock.
2009-09-29 23:45:28 +02:00
Jan Engelhardt
75f80fa8f8 pknock: move struct ipt_pknock_crypto to xt_pknock.c
It is only used in xt_pknock.c, and only once, so also make it an
anonymous struct.
2009-09-29 23:45:28 +02:00
Jan Engelhardt
d8dc72d151 pknock: only use size-fixated types
Types with variadic size are a no-no.
2009-09-29 23:36:53 +02:00
Jan Engelhardt
cee3d25d49 pknock: improve sizeof(struct ipt_pknock) by relayouting members
Stuffing struct holes as reported by pahole(1).
2009-09-29 19:36:14 +02:00
Jan Engelhardt
afbeb2fb47 pknock: improve sizeof(struct ipt_pknock) by using "even" field sizes
The field sizes were 33, which makes for lots of padding holes.
2009-09-29 19:36:14 +02:00
Jan Rafaj
c0dc0858de pknock: renaming of functions and prefix-name 2009-09-29 19:36:01 +02:00
Jan Rafaj
b04317576b pknock: add MODULE_ALIAS for pknock
Needed for autoloading.
2009-09-29 15:05:05 +02:00
Jan Rafaj
af2bbf0352 pknock: fit to Xtables-addons API 2009-09-29 15:03:45 +02:00
Jan Engelhardt
db9bb2778d pknock: import pknock trunk@463 2009-09-29 14:00:59 +02:00
Jan Engelhardt
674560d2d9 Merge branch 'psd' 2009-09-26 00:19:57 +02:00
Jan Engelhardt
7ab033e64c Update .gitignore files 2009-09-26 00:13:15 +02:00
Jan Engelhardt
71322e614c build: restore parallel build capability
It seems that ${MAKE} must be in the first expansion level.
2009-09-25 21:18:09 +02:00
Jan Engelhardt
46f9ddbf59 ACCOUNT: move manpage into ACCOUNT/ directory 2009-09-24 01:52:42 +02:00
Jan Engelhardt
e37985b527 ACCOUNT/userspace: use bool type 2009-09-24 01:48:16 +02:00
Jan Engelhardt
0ccd55abd9 ACCOUNT/userspace: remove pointless casts 2009-09-24 01:46:56 +02:00
Jan Engelhardt
60123a8c07 ACCOUNT/userspace: remove static zero initializers 2009-09-24 01:46:56 +02:00
Jan Engelhardt
504f1089fb ACCOUNT/userspace: simple reformat, use tabs instead of spaces 2009-09-24 01:46:55 +02:00
Jan Engelhardt
e119350000 ACCOUNT/userspace: remove trailing whitespace 2009-09-24 01:35:45 +02:00
Jan Rafaj
a260e3349f ACCOUNT: import userspace utilities 2009-09-24 01:27:11 +02:00
Jan Engelhardt
5cb36ed213 ACCOUNT: move xt_ACCOUNT into its own subdirectory
Will be adding related userspace programs shortly, and an extra
directory for the same reason it is for ipset/.
2009-09-24 01:00:15 +02:00
Jan Engelhardt
1a5b9f0e02 Merge branch 'master' into ACCOUNT 2009-09-24 01:00:10 +02:00
Jan Engelhardt
2c53605084 build: enable automake for extensions/ directory
Because we are likely to be having more userspace programs soon, and
reproducing manual makefiles is a bad idea, make extensions/ ready
for automake traversal. The build pattern now is:

1.  toplevel Makefile.am starts off with extensions/Makefile.am

2.  Makefile.am

    a.  builds programs in current directory

    b.  runs Kbuild
        - only from extensions/Makefile.am
          (so it does not get reinvoked from case 2cII)
        - Kbuild recurses on its own

    c.  runs Mbuild
        I.  builds iptables shared libraries in current directory
        II. runs Makefile.am in each subdir (goto step 2.)
2009-09-24 00:59:59 +02:00
Jan Engelhardt
537f25dbad build: use automake in ipset/ 2009-09-19 20:11:59 +02:00
Jan Rafaj
63a1e5329f ACCOUNT: add Kconfig file 2009-09-18 15:46:39 +02:00
Jan Engelhardt
b07434386b ipset: fix compile error on 2.6.31-rt
ip_set.c:42:8: warning: type defaults to 'int' in declaration of
'DECLARE_MUTEX'

2.6.31-rt has cleaned up the naming mess in semaphore.h. Without
making use of #if hackery, we can use struct semaphore as declaration
plus an explicit sema_init() call.
2009-09-12 02:41:36 +02:00
Jan Engelhardt
0c963dee06 ACCOUNT: fix pre-2.6.27 compilation error
xt_ACCOUNT.c:33:29: fatal error: linux/semaphore.h: No such file or
directory compilation terminated.
2009-09-12 02:27:26 +02:00
Jan Engelhardt
09bfb8496e condition: fix compile error on 2.6.31-rt
xt_condition.c:58:8: warning: type defaults to 'int' in declaration
of 'DECLARE_MUTEX'
2009-09-12 02:24:13 +02:00
Jan Engelhardt
9d1afc4268 ACCOUNT: fix compile error on realtime kernels
xt_ACCOUNT.c:1040: error: implicit declaration of function 'init_MUTEX'
2009-09-12 01:19:22 +02:00
Florian Westphal
342ccf62b2 psd: byteswap constant value instead
htons(CONSTANT) is computed at compile time.
reduces .text by 4 bytes on x86.

Signed-off-by: Florian Westphal <fwestphal@astaro.com>
2009-09-11 21:38:23 +02:00
Florian Westphal
adabd647b1 psd: use skb_header_pointer
Do not rely on tcphdr being in the linear area.

Signed-off-by: Florian Westphal <fwestphal@astaro.com>
2009-09-11 21:38:00 +02:00
Jan Engelhardt
bcda21a2b0 Merge branch 'psd' 2009-09-09 17:27:28 +02:00
Jan Engelhardt
4e6bc8af95 Merge branch 'ACCOUNT' 2009-09-09 17:26:00 +02:00
Jan Engelhardt
bd4be0d991 ACCOUNT: remove pointless casts 2009-09-05 04:39:00 +02:00
Jan Engelhardt
1f736c8203 ACCOUNT: remove static initializers to zero
These are not needed, .bss is zeroed.
2009-09-05 04:39:00 +02:00