B.A.T.M.A.N November 2011

b.a.t.m.a.n@lists.open-mesh.org

17 participants
45 discussions

[B.A.T.M.A.N.] [PATCH] batman-adv: add biggest_unsigned_int(x) macro
by Antonio Quartulli 11 Feb '12

11 Feb '12

in case of dynamic type variable, it could be needed to compute at compile time its maximal value. This macro helps in doing that for unsigned integer types Signed-off-by: Antonio Quartulli <ordex(a)autistici.org> --- main.h | 3 +++ 1 files changed, 3 insertions(+), 0 deletions(-) diff --git a/main.h b/main.h index 464439f..3c58235 100644 --- a/main.h +++ b/main.h @@ -210,6 +210,9 @@ static inline int compare_eth(const void *data1, const void *data2) /* Returns the smallest signed integer in two's complement with the sizeof x */ #define smallest_signed_int(x) (1u << (7u + 8u * (sizeof(x) - 1u))) +/* Returns the biggest unsigned integer with the sizeof x */ +#define biggest_unsigned_int(x) (~(x)0) + /* Checks if a sequence number x is a predecessor/successor of y. * they handle overflows/underflows and can correctly check for a * predecessor/successor unless the variable sequence number has grown by -- 1.7.3.4

3 4

[B.A.T.M.A.N.] [PATCH 00/11] bridge loop avoidance II
by Simon Wunderlich 26 Dec '11

26 Dec '11

This series of patches contains the changes for the redesigned bridge loop avoidance. The general concept is described in the wiki [1]. I've already performed a few testcases [2] which worked fine in my kvm environment. No crashes while running or unloading the extension either. This is the second version of the patchset, rebased on the master and including the changes according to reviews on the mailing list. The git repositories can be found on git.open-mesh.org: * blaII_dirty branch [3] has all the patches in development order * blaII branch [4] contains the patches squashed as sent the ml * blaII branch in batctl [5] contains the batctl changes Cheers, Simon [1] http://www.open-mesh.org/wiki/batman-adv/Bridge-loop-avoidance-II [2] http://www.open-mesh.org/wiki/batman-adv/Bridge-loop-avoidance-Testcases [3] http://git.open-mesh.org/?p=simon/batman-adv.git;a=shortlog;h=refs/heads/bl… [4] http://git.open-mesh.org/?p=simon/batman-adv.git;a=shortlog;h=refs/heads/bl… [5] http://git.open-mesh.org/?p=simon/batctl.git;a=shortlog;h=refs/heads/blaII --- The main changes to the previous version ([RFC]) are: * rebase on latest master * make bla sparse clean * check for hash_add() * check bla_get_backbone_gw() * shorten announce_mac[] * always use ETH_ALEN * fix debug message * don't change comment in batman_skb_recv() * split choose_claim/choose_backbone_gw * remove primary_if/own_orig address caching * fix forgotten backbone_gw_free_ref() and error checking * return 1 (handled) when there was an error in handle_claim() * unify bla_free() with the purge_* functions * call purge when (re)starting with a new primary interface * add compile option for bla Simon Wunderlich (11): batman-adv: remove old bridge loop avoidance code batman-adv: add basic bridge loop avoidance code batman-adv: make bridge loop avoidance switchable batman-adv: export claim tables through debugfs batman-adv: allow multiple entries in tt_global_entries batman-adv: don't let backbone gateways exchange tt entries batman-adv: add broadcast duplicate check batman-adv: drop STP over batman batman-adv: form groups in the bridge loop avoidance batman-adv: Update README and sysfs description batman-adv: add bridge loop avoidance compile option Makefile.kbuild | 2 + README | 28 +- bat_debugfs.c | 23 +- bat_sysfs.c | 8 +- bridge_loop_avoidance.c | 1592 +++++++++++++++++++++++++++++++++++++++++++++++ bridge_loop_avoidance.h | 49 ++ compat.c | 16 +- compat.h | 3 +- hard-interface.c | 18 +- main.c | 9 +- main.h | 9 +- originator.c | 3 +- packet.h | 16 + routing.c | 16 + soft-interface.c | 490 +-------------- soft-interface.h | 2 - sysfs-class-net-mesh | 9 + translation-table.c | 354 ++++++++--- types.h | 67 ++- 19 files changed, 2071 insertions(+), 643 deletions(-) create mode 100644 bridge_loop_avoidance.c create mode 100644 bridge_loop_avoidance.h -- 1.7.7.1

2 12

[B.A.T.M.A.N.] [RFC] batman-adv: switch routing algorithm at runtime
by Marek Lindner 06 Dec '11

06 Dec '11

Hi folks, after the previous discussions how to proceed with BATMAN V we came to the conclusion that a runtime switch for changing the routing algorithms would be the best solution. Hence, I drafted a couple of patches that I'd like to get some feedback on. Note, that these patches are not ready to be included. For instance, the rcu locking isn't clean. I'd like to focus your attention towards the following items: * How do we design the algorithm registration. At the moment batman_init() calls bat_iv_init() and possibly other algorithms in the future but this is less than ideal. I had hoped to find a mechanism which allows an init function to be declared and called inside of the respective routing algorithm files. These files can be compiled into the module or not. * The API and its implementation. It is highly likely that this API won't survive for very long as it has been designed to satisfy one routing algorithm only. However, changing it in the future should be easy. * Other ideas / comments ? Cheers, Marek

3 17

[B.A.T.M.A.N.] pull request: batman-adv 2011-11-26
by Marek Lindner 03 Dec '11

03 Dec '11

Hi David, the following 10 patches constitute the first batch I'd like to get the pulled into net-next-2.6/3.3. They're mostly uncritical fixes around the recently introduced tt code, some code refactoring, the kstrto update and the range check fix reported by Thomas Jarosch. Thanks, Marek The following changes since commit 1ea6b8f48918282bdca0b32a34095504ee65bab5: Linux 3.2-rc1 (2011-11-07 16:16:02 -0800) are available in the git repository at: git://git.open-mesh.org/linux-merge.git for_david Antonio Quartulli (5): batman-adv: tt_global_del_orig() has to print the correct message batman-adv: use orig_hash_find() instead of get_orig_node() in TT code batman-adv: fixed hash functions type to uint32_t instead of int batman-adv: linearise the tt_response skb only if needed batman-adv: check for tt_reponse packet real length Marek Lindner (1): batman-adv: refactoring gateway handling code Simon Wunderlich (2): batman-adv: directly write tt entries without buffering batman-adv: Fix range check for expected packets Sven Eckelmann (2): batman-adv: update internal version number batman-adv: Replace obsolete strict_strto<foo> with kstrto<foo> net/batman-adv/bat_sysfs.c | 4 +- net/batman-adv/bitarray.c | 2 +- net/batman-adv/gateway_client.c | 153 ++++++++++++++++++++++-------------- net/batman-adv/gateway_client.h | 5 +- net/batman-adv/gateway_common.c | 4 +- net/batman-adv/hash.c | 4 +- net/batman-adv/hash.h | 13 ++-- net/batman-adv/main.h | 2 +- net/batman-adv/originator.c | 13 ++- net/batman-adv/originator.h | 2 +- net/batman-adv/routing.c | 22 ++++-- net/batman-adv/soft-interface.c | 43 +++++++--- net/batman-adv/translation-table.c | 100 ++++++----------------- net/batman-adv/vis.c | 17 +++-- 14 files changed, 202 insertions(+), 182 deletions(-)

4 16

[B.A.T.M.A.N.] [patch] batman-adv: remove extra negation in gw_out_of_range()
by Dan Carpenter 29 Nov '11

29 Nov '11

There is a typo here where an extra '!' made the check to the opposite of what was intended. Signed-off-by: Dan Carpenter <dan.carpenter(a)oracle.com> diff --git a/net/batman-adv/gateway_client.c b/net/batman-adv/gateway_client.c index 9373a14..24403a7 100644 --- a/net/batman-adv/gateway_client.c +++ b/net/batman-adv/gateway_client.c @@ -695,7 +695,7 @@ bool gw_out_of_range(struct bat_priv *bat_priv, } neigh_old = find_router(bat_priv, orig_dst_node, NULL); - if (!!neigh_old) + if (!neigh_old) goto out; if (curr_tq_avg - neigh_old->tq_avg > GW_THRESHOLD)

2 1

[B.A.T.M.A.N.] [RFC] batman-adv: Move build compat stuff to Makefile
by Sven Eckelmann 26 Nov '11

26 Nov '11

Hi, I looked at the new patches from ordex and thought about using the Makefile from the kernel as the Makefile.kbuild in the standalone distribution. The Makefile is only the user friendly compatibility layer over the kbuild stuff. The problem is: The OpenWRT buildscript only uses the Makefile.kbuild (for a reason I don't know right now) and "adds" those extra includes for OpenWRT specific rules to the Makefile.kbuild (another part I never understood and only saw in the batman-adv makefiles). So maybe it is possible to discuss those different points and maybe find a solution for it. Things which are done here: * Import of kernel Makefile as Makefile.kbuild - Removes OpenWRT specific include of Rules.make - Removes place to manually set -DCONFIG_BATMAN_ADV_DEBUG - Removes compat.c build * Removal of the automatic detection of CPUs (see the discussion of the batctl patch) * Enabling of the batman-adv module build and adding of compat.o to the build through Makefile instead of Makefile.kbuild * Introducing the "new" way to enable optional features through `make CONFIG_BATMAN_ADV_FEATURE=y` * Fix of source revision "branding" when directly calling `make install` in the clean source directory Kind regards, Sven --- diff --git a/Makefile b/Makefile index 3374687..b4159ed 100644 --- a/Makefile +++ b/Makefile @@ -33,15 +33,19 @@ REVISION= $(shell if [ -d .git ]; then \ echo $$(git describe --always --dirty --match "v*" |sed 's/^v//' 2> /dev/null || echo "[unknown]"); \ fi) -NUM_CPUS = $(shell nproc 2> /dev/null || echo 1) - +CONFIG_BATMAN_ADV=m +batman-adv-y += compat.o +ifneq ($(REVISION),) +ccflags-y += -DSOURCE_VERSION=\"$(REVISION)\" +endif +ccflags-$(CONFIG_BATMAN_ADV_DEBUG) += -DCONFIG_BATMAN_ADV_DEBUG include $(PWD)/Makefile.kbuild all: - $(MAKE) -C $(KERNELPATH) REVISION=$(REVISION) M=$(PWD) PWD=$(PWD) -j $(NUM_CPUS) modules + $(MAKE) -C $(KERNELPATH) REVISION=$(REVISION) M=$(PWD) PWD=$(PWD) modules clean: $(MAKE) -C $(KERNELPATH) M=$(PWD) PWD=$(PWD) clean install: - $(MAKE) -C $(KERNELPATH) M=$(PWD) PWD=$(PWD) INSTALL_MOD_DIR=kernel/net/batman-adv/ modules_install + $(MAKE) -C $(KERNELPATH) REVISION=$(REVISION) M=$(PWD) PWD=$(PWD) INSTALL_MOD_DIR=kernel/net/batman-adv/ modules_install diff --git a/Makefile.kbuild b/Makefile.kbuild index bd7e93c..ce68611 100644 --- a/Makefile.kbuild +++ b/Makefile.kbuild @@ -18,20 +18,7 @@ # 02110-1301, USA # - - -# openwrt integration -ifeq ($(MAKING_MODULES),1) --include $(TOPDIR)/Rules.make -endif - -# ccflags-y += -DCONFIG_BATMAN_ADV_DEBUG - -ifneq ($(REVISION),) -ccflags-y += -DSOURCE_VERSION=\"$(REVISION)\" -endif - -obj-m += batman-adv.o +obj-$(CONFIG_BATMAN_ADV) += batman-adv.o batman-adv-y += bat_debugfs.o batman-adv-y += bat_iv_ogm.o batman-adv-y += bat_sysfs.o @@ -50,4 +37,3 @@ batman-adv-y += soft-interface.o batman-adv-y += translation-table.o batman-adv-y += unicast.o batman-adv-y += vis.o -batman-adv-y += compat.o diff --git a/README b/README index 7404c0a..ddc9569 100644 --- a/README +++ b/README @@ -207,10 +207,10 @@ times necessary to see more detail debug messages. This must be enabled when compiling the batman-adv module. When building bat- man-adv as part of kernel, use "make menuconfig" and enable the option "B.A.T.M.A.N. debugging". When compiling outside of the -kernel tree it is necessary to edit the file Makefile.kbuild and -uncomment the line +kernel tree it is necessary to enable it using the make option +CONFIG_BATMAN_ADV_DEBUG=y -#ccflags-y += -DCONFIG_BATMAN_ADV_DEBUG +# make CONFIG_BATMAN_ADV_DEBUG=y Those additional debug messages can be accessed using a special file in debugfs -- 1.7.7.1

3 7

[B.A.T.M.A.N.] [PATCHv4 0/7] DAT: Distributed ARP Table
by Antonio Quartulli 26 Nov '11

26 Nov '11

Hello people, **** This is the fourth version of this patchset. The whole code has been reviewed after some discussions on irc. Several bugs have been spotted and fixed. Thanks Marek, Sven and Simon. This should be a final version, unless some other mistakes/problems are find. This patchset assume that patch [PATCH] batman-adv: add biggest_unsigned_int(x) macro has already been applied **** as most of you may already know, last summer I've been working on the B.A.T.M.A.N.-Adv GSoC project named "DAT: Distributed ARP Table". For who wants to get deeper into the details of the project there are two links: - The GSoC proposal [1] - The DAT wikipage on open-mesh.org [2], with status and ideas description Just to recap: DAT is a distributes hash table meant to store ARP entries for fast lookup. In a normal scenario, whenever a node wants to communicate with another one, it first needs to issue a broadcast ARP request in order to retrieve its PHY/MAC address. In a sparse network a broadcast message could be lost several times before reaching the real destination so creating high latencies. With DAT, every ARP entries (a pair [IP addr, MAC addr]) is stored on a "computed" set of nodes, therefore in case of ARP request theses nodes can directly be contacted (in unicast) and the needed information can be quickly fetched. Cheers, Antonio [1] http://www.google-melange.com/gsoc/project/google/gsoc2011/ordex/4001 [2] http://www.open-mesh.org/wiki/batman-adv/GSOC2011_DAT

4 17

[B.A.T.M.A.N.] [PATCH] Remove useless condition
by Daniele Furlan 25 Nov '11

25 Nov '11

--- bat_iv_ogm.c | 3 +-- 1 files changed, 1 insertions(+), 2 deletions(-) diff --git a/bat_iv_ogm.c b/bat_iv_ogm.c index d60e1ba..e4bd54e 100644 --- a/bat_iv_ogm.c +++ b/bat_iv_ogm.c @@ -772,8 +772,7 @@ static int bat_ogm_calc_tq(struct orig_node *orig_node, /* if we have too few packets (too less data) we set tq_own to zero */ /* if we receive too few packets it is not considered bidirectional */ - if ((total_count < TQ_LOCAL_BIDRECT_SEND_MINIMUM) || - (neigh_rq_count < TQ_LOCAL_BIDRECT_RECV_MINIMUM)) + if (total_count < TQ_LOCAL_BIDRECT_SEND_MINIMUM) tq_own = 0; else /* neigh_node->real_packet_count is never zero as we -- 1.7.1

3 5

[B.A.T.M.A.N.] [PATCH] batman-adv: correct usage of TT_CLIENT_ROAM flag
by Antonio Quartulli 25 Nov '11

25 Nov '11

The TT_CLIENT_ROAM flag has to be set in tt_global_del() only if the latter has been invoked by the roaming_advertisement receiving routine. In case of changes update (normal operation done on OGM/TT_QUERY receiving) this doesn't have to happen. Moreover, a client which roamed to us has to be marked as ROAMED, not PENDING. This will make it to be kept for consistency purposes until the related change is received within the next OGM. Signed-off-by: Antonio Quartulli <ordex(a)autistici.org> --- translation-table.c | 12 ++++++++---- 1 files changed, 8 insertions(+), 4 deletions(-) diff --git a/translation-table.c b/translation-table.c index 379529d..633b35b 100644 --- a/translation-table.c +++ b/translation-table.c @@ -234,9 +234,11 @@ void tt_local_add(struct net_device *soft_iface, const uint8_t *addr, if (tt_global_entry) { /* This node is probably going to update its tt table */ tt_global_entry->orig_node->tt_poss_change = true; - /* The global entry has to be marked as PENDING and has to be + /* The global entry has to be marked as ROAMING and has to be * kept for consistency purpose */ - tt_global_entry->common.flags |= TT_CLIENT_PENDING; + tt_global_entry->common.flags |= TT_CLIENT_ROAM; + tt_global_entry->roam_at = jiffies; + send_roam_adv(bat_priv, tt_global_entry->common.addr, tt_global_entry->orig_node); } @@ -357,12 +359,14 @@ static void tt_local_set_pending(struct bat_priv *bat_priv, uint16_t flags) { tt_local_event(bat_priv, tt_local_entry->common.addr, - tt_local_entry->common.flags | flags); + TT_CLIENT_DEL); /* The local client has to be marked as "pending to be removed" but has * to be kept in the table in order to send it in a full table * response issued before the net ttvn increment (consistency check) */ tt_local_entry->common.flags |= TT_CLIENT_PENDING; + /* if this is a roaming deletion, mark the local entry accordingly */ + tt_local_entry->common.flags |= flags & TT_CLIENT_ROAM; } void tt_local_remove(struct bat_priv *bat_priv, const uint8_t *addr, @@ -1404,7 +1408,7 @@ static void _tt_update_changes(struct bat_priv *bat_priv, tt_global_del(bat_priv, orig_node, (tt_change + i)->addr, "tt removed by changes", - (tt_change + i)->flags & TT_CLIENT_ROAM); + false); else if (!tt_global_add(bat_priv, orig_node, (tt_change + i)->addr, ttvn, false, -- 1.7.3.4

2 1

Re: [B.A.T.M.A.N.] Connecting non-batman nodes to a mesh network
by Antonio Quartulli 24 Nov '11

24 Nov '11

On dom, nov 20, 2011 at 10:49:16 +0200, Panayot Daskalov wrote: > > Hi All, > > > > I want to ask for your advice about a batman problem I am facing. I have > > a mesh network of custom build embedded Linux devices. The devices have > > only network one interface and this is the Wi-Fi that is used by batman. > > So far the mesh is running OK, but at some point I need to be able to > > connect with a laptop (non-Linux) to the network and access some of the > > devices via its IP address. I tried simply connecting to the adhoc > > network and setting a static IP to my laptop. I have managed to do that, > > but I can't access the nodes from the mesh. > > > > I read the instruction for mixing with non-batman devices by bridging > > the bat0 interface with the "non-batman" interface, but as far as I > > understand that applies to systems with at least 2 network interfaces > > e.g one for batman nodes and one for the non-batman. So my question is - > > can I mix non-batman and batman nodes when I have only one network > > interface on the batman nodes? If yes, then how? > Hi Panayot, you are right: you need two "interfaces" in order to let clients connect to the mesh. However (if you want to use wifi interfaces only) you can also create "virtual interaces" using the same "real interface". This feature is available depending on the wifi driver you are using. If you are using madwifi or one of the new mac80211 drivers (not all support it), this technique should be feasiable without introducing many problems :) For instance, what driver are you using? Cheers, -- Antonio Quartulli ..each of us alone is worth nothing.. Ernesto "Che" Guevara ☭

3 6

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

B.A.T.M.A.N November 2011