This is a warning coming up on Android (ARM) platform.
Clang 4.x creates a false positive here: the can_id is the first
(and aligned) value in the packed struct modattr, so it's always
aligned if the struct itself is aligned.
When using the -b/-B option to send received CAN frames to the brigde interface
the sending failed when processing CAN FD frames. This patch enables CAN FD on
the socket for the bridging interface.
https://github.com/linux-can/can-utils/issues/104
Reported-by: https://github.com/jm3lee
Signed-off-by: Oliver Hartkopp <socketcan@hartkopp.net>
The documentation stated that the CAN ID is assumed to be an extended CAN
identifier (29 bit ID / EFF) when "can_id and can_mask are both 8 digits".
The check for the CAN ID length to be 8 is common in other CAN utils
(e.g. cansend) but it has never been implemented in candump.
This patch adds that check for EFF CAN IDs and clarifies the documentation.
Thanks to Nick for pointing out this inconsistency!
Reported-by: Nick Elmschig <nick@ikerobotics.com>
Signed-off-by: Oliver Hartkopp <socketcan@hartkopp.net>
This option is needed for j1939 socket functionality test.
Connects wihout bind should not work.
Signed-off-by: Bastian Stender <bst@pengutronix.de>
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
This option is needed for j1939 socket functionality test.
Binding to a socket and rebinding to the socket with e.g. a different SA
should work.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Bastian Stender <bst@pengutronix.de>
In some cases we can get local TX echo after remote response
for same TXed message was already received.
For applications which relay on proper order of TXed and RXed
message this will be fatal.
With this patch the package generating side will test if TX and RX are
properly ordered.
Currently with following commands, this issue can be easily
reproducible:
canfdtest -g can0 &
while(true); do chrt 40 dd if=/dev/zero of=/dev/null count=10000; done
Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>
Scripts can be configured to handle different SIGnals and
return codes in different ways. In this case it is better
to forward received signal, instead of converting it to return 0.
Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>
For keeping backwards compatibility the configure switch
--enable-error-log is enabled by default. To disable error logging
use --enable-error-log=no.
Signed-off-by: Mark Jonas <mark.jonas@de.bosch.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
If you have an application stderr shall not be used by libsocketcan
you can disable all error logging by defining DISABLE_ERROR_LOG.
Signed-off-by: Leo Ruan <tingquan.ruan@cn.bosch.com>
Signed-off-by: Mark Jonas <mark.jonas@de.bosch.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
This commit adds an interface to get the statistics (64-bits) from
CAN interface. See more information from /usr/include/linux/if_link.h
where defines the struct rtnl_link_stats64.
Compare to the rtnl_link_stats, rtnl_link_stats64 was introduced
since linux kernel 2.6. After that the rtnl_link_stats is synchronous
with struct rtnl_link_stats64 by truncating each member from 64 bits
to 32 bits. Actually, the struct rtnl_link_stats is kept for
compatibility.
Signed-off-by: Leo Ruan <tingquan.ruan@cn.bosch.com>
Signed-off-by: Mark Jonas <mark.jonas@de.bosch.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
libsocketcan always retrieves link information from all interfaces
instead of only specified one. This commit fixes the issue by appending
an interface information to dump request netlink message.
See man 7 rtnetlink for detail description in section 'RTM_GETLINK'.
Signed-off-by: Leo Ruan <tingquan.ruan@cn.bosch.com>
Signed-off-by: Mark Jonas <mark.jonas@de.bosch.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Because max is the index of the last element in array tb, max+1
elements need to be initialized.
Signed-off-by: Leo Ruan <tingquan.ruan@cn.bosch.com>
Signed-off-by: Mark Jonas <mark.jonas@de.bosch.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
With commit 33a9249571 ("slcand: daemonise only once setup is complete")
the daemonize syscall is moved behind serial tty setup.
This patch moves the missing assignment of the loop variable slcand_running
to get it into the code context of the while() loop.
Signed-off-by: Oliver Hartkopp <socketcan@hartkopp.net>
Delay the daemon() call until the interface has been configured. This
simplifies scripts that wish to use the new interface immediately.
Signed-off-by: Mans Rullgard <mans@mansr.com>
This is a direct consequence of not allowing bind to any device
("3cb6a1204598 j1939: socket: bind() cleanups"), i.e. can_ifindex == 0.
This means before this change jacd would receive from any interface that
was opened by another program.
Without this patch jacd fails with:
acd: bind(): No such device
A more verbose call shows the missing interface:
$ jacd -v 1122334455667788
jacd: ready for can0:1122334455667788
- socket(PF_CAN, SOCK_DGRAM, CAN_J1939);
- setsockopt(, SOL_SOCKET, SO_BINDTODEVICE, can0, 4);
- setsockopt(, SOL_CAN_J1939, SO_J1939_FILTER, <filter>, 84);
- setsockopt(, SOL_CAN_J1939, SO_J1939_RECV_OWN, 1, 4);
- bind(, :fe,0ee00,1122334455667788, 24);
jacd: bind(): No such device
Signed-off-by: Bastian Stender <bst@pengutronix.de>
Marc Kleine-Budde
Oliver Hartkopp
Tomasz Wasilczyk
Oleksij Rempel
Nick Elmschig
Marc Kleine-Budde
Bastian Stender
Yegor Yefremov
Oleksij Rempel
Joel Winarske
Mark Jonas
Leo Ruan
Mans Rullgard
Frank Theile
Frank Theile
Tom Nicklin