Oleksij Rempel

Hm, looks like a bug in the driver. We have two places where socket is set to an error state, but no place to recover it: https://elixir.bootlin.com/linux/latest/source/net/can/j1939/socket.c#L1096 https://elixir.bootlin.com/linux/latest/source/net/can/j1939/socket.c#L1276 @marckleinebudde, @hartkopp,...

As counter part for the [j1939_sk_netdev_event_netdown](https://elixir.bootlin.com/linux/latest/C/ident/j1939_sk_netdev_event_netdown) - yes. For [j1939_sk_send_loop_abort](https://elixir.bootlin.com/linux/latest/C/ident/j1939_sk_send_loop_abort) it is a bit more difficult. It will not set the socket in to error state if error queue is...

@in-text are you able to detect re-attached bus by using error queue?

j1939cat is designed to abort on errors. You will need to write own code. Please note, there is no attach/detach "event". Just send errors.

Hi @xile273 , did you tried to use error queue? It will send you all packet notifications and prevent from blocking the socket.

Hm.. I tried to use isobusfs client and server, with following modification but was not able to reproduce issue with stalled socket: ``` @@ -288,8 +288,6 @@ static int isobusfs_cli_handle_events(struct...

@xile273 , there is one more application for testing, if you like: https://github.com/linux-can/can-utils/pull/524

Ok, thx. Now i can reproduce it. By the way, it is possible to reproduce it with following commands too: ``` dd if=/dev/zero of=test bs=1 count=10 j1939cat -i test can0:0x80...

@xile273 , can you please test following patch: https://github.com/olerem/linux-2.6/commit/a32c56908e9dee5933adf00dc9469ec195ea0095 If it works for you, can you please give me your Reported-by and Tested-by

@xile273 , thank you! Can you please tell more about what you are doing with the kernel j1939 stack. It is interesting for me how it is used and if...