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

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).

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)

Hi @Dimka88! The warnings are fixed in the following pull request:
https://github.com/accel-ppp/accel-ppp/pull/64
From my understanding, the situation is next:

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.

I think it is not implemented yet, but is is good feature request

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:

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 ..

make
cpack -G DEB

i have testsed last version accel. its ok

Hi.
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.

В логе должны отображаться команды полученые при accel-cmd *.
Необходимо указывать

[cli]
versobe=2

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)

Is there any chance to fix that ?

build command
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
cd accel-ppp-master
mkdir build
cd build
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

added

Hi
Any one for help to make this ?

Hi Dimka88!

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:

1. two sessions in L2 ipoe shared mode
2. the same interface (e.g. eth0)
3. first session without vrf
4. second session with vrf
5. the same ip.src for both sessions

additional information and patches how it should works

But for Service name i need your help to extract from PADI or PADO

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?

Hi @svlobanov
Yes this list interface here :

@micron
regarding to interface:

ifname-in-sid=called-sid|calling-sid|both:

Hi
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

Related task https://phabricator.accel-ppp.org/T56

Proposed patch https://github.com/DmitriyEshenko/accel-ppp/commit/e7b3e53fb34e7e3b59b367a9fd777edc0cc3bd97

Thank you @svlobanov .
Added possibility to define netlink buffers

[common]
nl-snd-buffer=32768
nl-rcv-buffer=1048576

PR https://github.com/accel-ppp/accel-ppp/pull/32

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.

Should be aready fixed. Related tasks: T46, T43

@djan I believe there is no interesting stat from the connlimit module. Maybe you think that it works in a bit different way

PR https://github.com/accel-ppp/accel-ppp/pull/30

Спасибо, установил. Неделя полёт нормальный.