It is not fixed in actual version. It is required to fix here: https://github.com/accel-ppp/accel-ppp/blob/2b865db72bc2ddc6411950d72f1c23e8ef115b8a/accel-pppd/ppp/ipcp_opt_ipaddr.c#L100 , https://github.com/accel-ppp/accel-ppp/blob/2b865db72bc2ddc6411950d72f1c23e8ef115b8a/accel-pppd/ppp/ipv6cp_opt_intfid.c#L209 , https://github.com/accel-ppp/accel-ppp/blob/2b865db72bc2ddc6411950d72f1c23e8ef115b8a/accel-pppd/ctrl/ipoe/ipoe.c#L668
Apr 16 2023
Feb 16 2023
Jan 30 2023
Dec 22 2022
Dec 21 2022
Dec 20 2022
Dec 15 2022
Dec 8 2022
Dec 2 2022
Nov 18 2022
Attached the patch i made for the version 1.12.0-202-g28fe4de. It will use Calling Number/Called Number as RADIUS Calling-Station-Id/Called-Station-Id in case they are present in L2TP ICRQ, otherwise old beahvior is used (LAC/LNS IP addresses).
Nov 16 2022
Nov 15 2022
Oct 23 2022
UPD:// CMake Error looks pretty strange. It get info by the command git describe --tags --always
From your repo it return 6e5f998 because you clone only single branch.
root@debian11:/opt/accel-ppp/build# git describe --tags --always 6e5f998
So, maybe we have to improve generate version functionality also
Hi @v.huti , I see some cmake Errors
root@debian11:/opt/accel-ppp/build# cmake -DBUILD_IPOE_DRIVER=TRUE -DBUILD_VLAN_MON_DRIVER=TRUE -DCMAKE_INSTALL_PREFIX=/usr -DKDIR=/usr/src/linux-headers-`uname -r` -DLUA=TRUE -DCPACK_TYPE=Debian11 .. -- The C compiler identification is GNU 10.2.1 -- Detecting C compiler ABI info -- Detecting C compiler ABI info - done -- Check for working C compiler: /usr/bin/cc - skipped -- Detecting C compile features -- Detecting C compile features - done -- 'x86_64' CMake Error at cmake/cpack.cmake:5 (list): list index: 1 out of range (-1, 0) Call Stack (most recent call first): CMakeLists.txt:49 (include)
Oct 22 2022
In general, assuming anything about data alignment is a bad practice that leads to very obscure errors. Taking the address of a packed member is dangerous since the reduced alignment of the pointee is lost. If the pointer value is dereferenced, this can lead to memory alignment faults in some architectures.
Oct 12 2022
Oct 5 2022
I think it is not implemented yet, but is is good feature request
Oct 3 2022
Sep 16 2022
just to discuss the possibilities.
I see Calling-Number/Called-Numer AVPs can be taken from attr->val.octets doing something within l2tp_recv_ICRQ procedure:
Sep 5 2022
Aug 15 2022
This cmake gives working deb on VyOS on Debian11 machine:
cmake -DCMAKE_BUILD_TYPE=Debug \
-DCMAKE_C_FLAGS='-g -O0' \ -DBUILD_IPOE_DRIVER=TRUE \ -DBUILD_VLAN_MON_DRIVER=TRUE \ -DKDIR=/usr/src/linux-headers-`uname -r` \ -DCMAKE_INSTALL_PREFIX=/usr \ -DLUA=TRUE \ -DLUA=5.3 \ -DCPACK_TYPE=Debian11 ..
cpack -G DEB
i have testsed last version accel. its ok
Is there any recommendations on compiling and embedding newly compiled version of accel-ppp to vyos 1.4 ?
I mean how to install newly built package into VyOS 1.4.
The deb built on Debian11 and installed with dpkg on VyOS does not start for some reason with no output.
Aug 4 2022
В логе должны отображаться команды полученые при accel-cmd *.
Aug 2 2022
patch under test
diff --git a/accel-pppd/ctrl/sstp/sstp.c b/accel-pppd/ctrl/sstp/sstp.c index f17db43..40c6ad9 100644 --- a/accel-pppd/ctrl/sstp/sstp.c +++ b/accel-pppd/ctrl/sstp/sstp.c @@ -2415,12 +2415,12 @@ static int sstp_connect(struct triton_md_handler_t *h)
Jul 28 2022
Is there any chance to fix that ?
Jul 26 2022
wget https://github.com/accel-ppp/accel-ppp/archive/refs/heads/master.tar.gz -O accel-ppp-master.tar.gz
tar -xvf accel-ppp-master.tar.gz
cmake -DCMAKE_INSTALL_PREFIX=/usr -DKDIR=/usr/src/linux-headers-uname -r -DRADIUS=TRUE -DSHAPER=TRUE -DLOG_PGSQL=FALSE -DLUA=TRUE -DBUILD_IPOE_DRIVER=FALSE -DCPACK_TYPE=Debian10 -DBUILD_VLAN_MON_DRIVER=FALSE -DCMAKE_BUILD_TYPE=Debug -DCMAKE_C_FLAGS='-g -O0' ..
make -j $CPU_COUNT.
cpack -G DEB
Add coredump file
Jun 20 2022
May 20 2022
May 13 2022
Apr 27 2022
Apr 21 2022
Any one for help to make this ?
Mar 31 2022
Mar 10 2022
Feb 17 2022
Jan 30 2022
Jan 26 2022
Jan 9 2022
Jan 5 2022
also, we have to check driver performance when ipoe_lookup_hwaddr is used for many sessions (many sessions in vrf)
It will not work if there in the conditions below:
- two sessions in L2 ipoe shared mode
- the same interface (e.g. eth0)
- first session without vrf
- second session with vrf
- the same ip.src for both sessions
additional information and patches how it should works
Jan 4 2022
But for Service name i need your help to extract from PADI or PADO
Jan 3 2022
i will try to make patch for this.
many of use case send to radius clear mac for user and mac for machine and not need to send interface
this with interface and service will use only fkr cli interface to show local in system .
I use this one: show sessions match calling-sid vlan123
calling-sid/called-sid send to radius as info for user
separate column need only for statistic local in cli when statistic or only to see user per interface
why do you want to have "uplink interface" as a separate column? why calling-sid/called-sid with interface name is not enough?
Yes this list interface here :
regarding to interface:
yes check in man/doc but calling-sid or called-sid not get interface .
You can include "uplinkif" to calling-sid or called-sid, it is already implemented. Please check man/doc related to this feature
Dec 24 2021
Related task https://phabricator.accel-ppp.org/T56
Dec 22 2021
Dec 21 2021
Thank you @svlobanov .
Added possibility to define netlink buffers
[common] nl-snd-buffer=32768 nl-rcv-buffer=1048576
Dec 20 2021
Dec 17 2021
Unfortunately, I can not reproduce this issue, please reopen this ticket when this issue will appear again.
no in this case work with latest master and after this segfault not see this problem.
may be is network problem send wrong package
Hi @micron , did you modify the accel-ppp code by yourself, exactly functions related to Service-Name?
Hi @slima , do you have a chance to test on the fresh accel-ppp version from the master branch?
please check this patch https://github.com/accel-ppp/accel-ppp/pull/31
Should be already fixed by https://github.com/accel-ppp/accel-ppp/commit/ff91c7376f2a6200be7ea90d36411e1acb4cd9c8
Let me know if you will face a similar backtrace with the fresh accel-ppp version.
@djan I believe there is no interesting stat from the connlimit module. Maybe you think that it works in a bit different way
Dec 16 2021
Спасибо, установил. Неделя полёт нормальный.