| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
already started
|
|
There is some issue on binding raw ICMPv6 socket immediately after assigning address on 3.14 (and probably later) kernel.
This causes ipv6_nd to fail to start.
This patch makes ipv6_nd to try to bind later if bind fails.
|
|
Previous version used linear list of vlan monitor enabled interfaces,
this caused huge overhead in packet receive path because for each
received packet it performed search on linear list to check if vlan monitor
is enabled on input device.
Current version binds internal structure to net_device->ml_priv so search is not performed more.
|
|
With unblocked signals log thread may catch SIGCHLD which causes zombie process.
|
|
If specified then accel-ppp will send tunnel type as string (pppoe,pptp,l2tp,ipoe).
|
|
auth modules
Passed strings was not nullterminated
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Before this time accel-ppp always worked in "ip unnumbered" mode.
This option may turn this mode off (default is on)
|
|
After introduction offer-delay accel-ppp stopped to answer NAK for REQUESTs if requested session is not existing
to give chance other servers to reply. But this causes that some broken clients enter to infinite REQUEST loop.
This patch corrects this issue, so if offer-delay is not configured (means that this is only server in the net) or
after receiving few REQUESTs accel-ppp send NAK for non-existing sessions.
|
|
logging purpose)
|
|
|
|
|
|
interface
|
|
|
|
|
|
First try to rename interface immediately and if it fails then do second try later just before bring interface up.
In case single-session=replace first try to rename interface may fail because first session may be active
|
|
|
|
Use separated thread to write messages to file instead of async io.
On mips platform async io does not work properly by some reason and may stuck logging
|
|
Check if there is space in previous chunk and append log data there.
This reduces memory usage and message fragmentation
|
|
Wait for previous session completely terminated before continuing authorization new session.
|
|
|
|
New options are applied to server option as server=...[,weight=x][,backup]
Weight is used for multi-server configurations (larger weight takes more users).
Backup is used to mark backup server f.e. server which will be used only if all other servers are failed.
|
|
By default list of IP address is serial at startup.
This option shuffles initial IP list so it becomes more random.
|
|
|
|
|
|
This new option will control how to represent Called-Station-ID.
Its value maybe ifname,mac or ifname:mac.
By default mac is used to keep compatibility with old versions.
Also old option ifname-in-sid is removed from documentation, but is still supported for compatibility.
|
|
|
|
There is no need to keep auth request structure in memory during whole session life time.
It wastes memory and file descriptors
|
|
|
|
Signed-off-by: Dmitry Kozlov <xeb@mail.ru>
|
|
table
Signed-off-by: Dmitry Kozlov <xeb@mail.ru>
|
|
Signed-off-by: Dmitry Kozlov <xeb@mail.ru>
|
|
Signed-off-by: Dmitry Kozlov <xeb@mail.ru>
|
|
Signed-off-by: Dmitry Kozlov <xeb@mail.ru>
|
|
Use unshare(CLONE_FILES) to create a local file descriptors namespace
for the SNMP thread. This is similar to what was done in bf5340
'net-snmp: run snmp in "special" thread', but without calling clone()
directly. So the net-snmp plugin keeps running in a regular pthread.
Signed-off-by: Guillaume Nault <g.nault@alphalink.fr>
|
|
This reverts commit bf5340291badcc631bd855bb924593cd445ecd26.
snmp_thread() needs to be run as regular thread to allow Valgrind
analysis. Using a different namespace for file desciptors will be
done later on using unshare().
Signed-off-by: Guillaume Nault <g.nault@alphalink.fr>
|
|
|
|
|
|
If accel-ppp is compiled with RADIUS support, but run without using it,
then pd->tmp_fname is NULL. This patch verifies that pd->tmp_fname is
valid before renaming the radattr file.
Signed-off-by: Guillaume Nault <g.nault@alphalink.fr>
|
|
of NAS-Port in Access-Request and Account-Request)
Signed-off-by: Dmitry Kozlov <xeb@mail.ru>
|
|
Signed-off-by: Dmitry Kozlov <xeb@mail.ru>
|
|
NET-SNMP is built on top of stale file descriptor multiplexing mechanism (select)
which prevents to use descriptors with number >= 1024.
Due to accel-ppp uses a lots of descriptors NET-SNMP easy overflows this limit and fails
and even may crash whole accel-ppp process.
This patch runs snmp in special thread which does not share file descritor table
with main process, so it can use less descriptors and live in __FD_SETSIZE limit.
Signed-off-by: Dmitry Kozlov <xeb@mail.ru>
|
