See kernel commit 4b324126e0c6c3a5080ca3ec0981e8766ed6f1ee
----
Furthermore, notice that the initial checks:
if (!node->rb_left)
child = node->rb_right;
else if (!node->rb_right)
child = node->rb_left;
else
{
...
}
guarantee that old->rb_right is set in the final else branch, therefore
we can omit checking that again.
Signed-off-by: Wolfram Strepp <wstrepp@gmx.de>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
----
Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
See kernel commit 4c60117811171d867d4f27f17ea07d7419d45dae
----
There are two cases when a node, having 2 childs, is erased:
'normal case': the successor is not the right-hand-child of the node to be erased
'special case': the successor is the right-hand child of the node to be erased
Here some ascii-art, with following symbols (referring to the code):
O: node to be deleted
N: the successor of O
P: parent of N
C: child of N
L: some other node
normal case:
O N
/ \ / \
/ \ / \
L \ L \
/ \ P ----> / \ P
/ \ / \
/ /
N C
\ / \
\
C
/ \
special case:
O|P N
/ \ / \
/ \ / \
L \ L \
/ \ N ----> / C
\ / \
\
C
/ \
Notice that for the special case we don't have to reconnect C to N.
Signed-off-by: Wolfram Strepp <wstrepp@gmx.de>
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
----
Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
See kernel commit 16c047add3ceaf0ab882e3e094d1ec904d02312d
----
First, move some code around in order to make the next change more
obvious.
[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Wolfram Strepp <wstrepp@gmx.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
----
Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
See kernel commit 55a63998b8967615a15e2211ba0ff3a84a565824
----
Tfour 4 redundant if-conditions in function __rb_erase_color() in
lib/rbtree.c are removed.
In pseudo-source-code, the structure of the code is as follows:
if ((!A || B) && (!C || D)) {
.
.
.
} else {
if (!C || D) {//if this is true, it implies: (A == true) && (B == false)
if (A) {//hence this always evaluates to 'true'...
.
}
.
//at this point, C always becomes true, because of:
__rb_rotate_right/left();
//and:
other = parent->rb_right/left;
}
.
.
if (C) {//...and this too !
.
}
}
Signed-off-by: Wolfram Strepp <wstrepp@gmx.de>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Andrea Arcangeli <andrea@qumranet.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
----
Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
See kernel commit f4b477c47332367d35686bd2b808c2156b96d7c7
----
The 'rb_first()', 'rb_last()', 'rb_next()' and 'rb_prev()' calls
take a pointer to an RB node or RB root. They do not change the
pointed objects, so add a 'const' qualifier in order to make life
of the users of these functions easier.
Indeed, if I have my own constant pointer &const struct my_type *p,
and I call 'rb_next(&p->rb)', I get a GCC warning:
warning: passing argument 1 of ?~@~Xrb_next?~@~Y discards qualifiers from pointer target type
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
----
Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
alloca.h is not available on FreeBSD, use the default autoconf
function to check for it, there is a complete list[1] of what to
do for using alloca but let us see how far we get with this test.
Include netinet/in.h for the IPv4 and IPv6 socket address. Check
for dlopen in libraries and use this instead of linking -dl.
[1] http://www.gnu.org/s/hello/manual/autoconf/Particular-Functions.html
With the new generated headers for the CRC code, we need to include
the builddir as well in case we're building out of tree.
Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
This patch adds RB-tree based timers which scales better than the
previous list-based implementation.
It does not require any API changes. It breaks ABI because the
osmo_timer_list structure has changed though (to avoid this in
the future, we can put internal data in some private structure).
The following table summarizes the worst-case computational complexity
of this new implementation versus the previous one:
rb-tree list-based
------- ----------
calculate next timer to expire O(1) O(n)
insertion of new timer O(log n) O(n)
deletion of timer O(log n) O(1)
timer-fired scheduler O(log n) O(3n)
The most repeated cases are:
* the calculation of the next timer to expire, that happens in every
loop of our select function.
* the timer-fired scheduler execution.
This new implementation only loses in the deletion of timer scenario,
this happens because we may need to rebalance the tree after the
removal.
So I think there is some real gain if we have some situation in which
we have to handle lots of timers.
This patch adds red black trees implementation to libosmocore.
This data structure is very useful to search for elements in
ordered sets in O(log n) instead of O(n) that lists provide.
The first client of this code will be one follow up patch that
implements rbtree-based timer scheduler.
Instead of mixing together the GSM layer 1 interface and RSL interface
with the implementation of LAPD, the core function of LAPD is now
extracted from LAPDm. The core implementation is now in lapd_core.c
and lapd_core.h respectively.
The lapd_core.c implements exactly one datalink instance for one SAP.
The surrounding implementation "lapdm.c" codes/decodes the layer 2
headers and handles multiplexing and datalink instances, as well as
translates primitives from/to RSL layer.
lapd_core.c can now be used for other LAPD implementations. (ISDN/ABIS)
As it turns out, the other range were affected too, so Anreas wrote
the fix for them as well.
Written-by: Andreas.Eversberg <jolly@eversberg.eu>
Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
From the mail:
---
appended is another patch for fixing a bug in the calculation of the
frequency lists. This time the patch is for the "Range 256 format".
The problem is that the operand for the "smod" operation might be
negative, in this case the simplified version won't work as expected.
In the patch I introduced a separate function for "smod" which takes
care of the sign. I have not yet checked if the other formats are also
affected, this would be the case if the "smod" operand can be negative.
There might be other solutions to fix the problem without the need
for a separate function, however I have not thought further about it.
A test vector is the following frequency list ("Range 256 format",
first byte is the length):
09 8b 1c 83 8c 15 ef 02 2d 30
The correct ARFCNs are
569 571 576 578 586 608 712 715 719
The uncorrected version would instead return:
444 457 460 464 569 576 578 586 608
This means four ARFCNs are wrong which will cause problems if for
example the frequency list contains the ARFCNs for hopping.
----
Written-by: Dieter Spaar <spaar@mirider.augusta.de>
Signed-off-by: Sylvain Munaut <tnt@246tNt.com>
Sylvain Munaut
Holger Hans Peter Freyther
Harald Welte