Hi,
On 04/04/18 16:59, Dave Ellis wrote:
After that the instructions mention using openggsn.
I'm not sure if
there is further configuration required for the kernel gtp module - do I
need to create a tunnel? Or should this be done by osmo-ggsn as it did
for the gtpu-mode tun setting?
I'm not sure here since I didn't find the opportunity to use the gtp
kernel module myself yet. Looking at code in osmo-ggsn gtp-kernel.c,
gtp_kernel_init() calls gtp_dev_create() which seems to create a GTP
tunnel device, so osmo-ggsn seems to be creating the device.
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
<0002> ggsn.c:203 APN(S1U): Starting
<0002> ggsn.c:281 APN(S1U): Opening Kernel GTP device gtp0
<0002> gtp-kernel.c:75 Initialized GTP kernel mode (genl ID is 30)
Program received signal SIGSEGV, Segmentation fault.
gtp_kernel_init (gsn=0x0, devname=0x6a6fe0 "gtp0",
prefix=prefix@entry=0x6a76e0, ipup=0x0) at gtp-kernel.c:94
94if (gtp_dev_create(-1, devname, gsn->fd0, gsn->fd1u) < 0) {
This looks like a bug in the code (gsn=NULL). Please open a ticket in
redmine in OsmoGGSN project to remember to have a look at it. Also
compiling everything with CFLAGS="-g -O0", then running under gdb and
providing a full backtrace may help pinpoint the issue quicker.
Feel free to open a new task also about the bad performance when running
without gtp kernel support and providing some performance figures, and
if possible a linux perf output of the process.
--
- Pau Espin Pedrol <pespin(a)sysmocom.de>
http://www.sysmocom.de/
=======================================================================
* sysmocom - systems for mobile communications GmbH
* Alt-Moabit 93
* 10559 Berlin, Germany
* Sitz / Registered office: Berlin, HRB 134158 B
* Geschaeftsfuehrer / Managing Director: Harald Welte