Aktueller Inhalt von poruid

@JD, did you disable AVM_PA and GENERIC_CONNTRACK in the kernel config? P

When having both nf_contrack* and AVM's generic conntracking modules loaded the system log show many messages like: and connections to the Internet become slow and http requests yield incomplete pages. It seems some IP packages are lost. It seems that a statefull firewall, i.e. a filrewall...

A question: does anyone know what happens when CONFIG_AVM_PA and CONFIG_GENERIC_CONNTRACK aren't set in the kernels .config? Especially, I would like to know what problems it may cause. (because my FB7270-v2 is deployed in a production environment doing tests myself is troublesome).

Risking that I write some stuff obvious to many, I try to reason here what the impact of the AVM closed source modules is. Kernel modules can only be loaded if the used kernel was build with a config that enabled these modules, otherwise the kernel lacks the needed module stubs, and also...

Strange, because one would assume that modules.deb is build with the kernel for which the local .config (of the kernel source/build dir) is decisive. On top of that the freetz .config flags are needed to make fwmod include the module in the image. My router is a 7270v2 (currently AVME/freetz...

The 7270 trunk has the same problem. Work-around: select the module in the kernel config (make kernel-menuconfig) and also, when missing, add the needed flags to freetz .config (add the lines "FREETZ_MODULE_xt_state=y" and "FREETZ_LIB_libxt_state=y")

Compilation fails, struct undefined etc. Seems a bit more complicated and also whether the half-way implementation would work is uncertain. When one looks at source/fritzbox/7390_5.01/GPL-release_kernel/linux/drivers/net/ifxmips_switch_api/ifx_ethsw_tantosG_api.c which also has the PSB 6970...

In linux-2.6.32/drivers/net/avm_cpmac/cpphy_adm6996.c larger parts of the code that seems to handle 802.1q tag has been surrounded with #if 0 ... #endif. Two questions: - has anyone tried what happens when some of these #if 0 pre-processor statements are replaced with #if 1? - has anynone a clue...

The make of MDEV is broken because root/etc/device.table has been moved. change 5717 because of ticket 1018 rebased that file. The file device.table is now part of the mod package. How must this build-time depency on a file of an other package be handled? I've created ticket 1051 for this...

On a *NIX (cygwin, linux, *bsd, osx etc) commandline type: man 5 crontab

Good question. I wouldn't know the answer, but the code of sys_dlopen() has been patched with make/samba/patches/130-update_avm.patch which was introduced in changeset 2939 "samba: Add patch (sync with 04.67 Open Source Package)" by oliver. The patch was not associated to any bug number. So...

Received an email from AVM:

Here a simple script that adds IP addresses of foreign hosts that have failed to authenticate with dropbear to the iptables firewall so that it drops packets from those sources. The script should be run at regular intervals from crontab. The script may serve as a sample...

and MDEV? Please consider the MDEV hotplug chain as a possible means to implement your ideas. Its better structure and design seems well suited for your plans. The current mdev implementation has a special directory, /lib/mdev/interface, that is to contains scripts that provide a AVM...

A quick inspection of the revisions of the mdev sources learned that mdev hasn't changed for quite some time except for the recent addition of support for the 7240 in r3651. So that leaves a change in the AVM firmware as probable cause. I cannot imagine that it has changed the hotplug interface...