CVE-2026-72124

In the Linux kernel, the following vulnerability has been resolved:

can: isotp: serialize TX state transitions under so->rx_lock

The TX state machine (so->tx.state) is driven from three contexts: sendmsg() claiming and progressing a transfer, the RX path consuming Flow Control/echo frames, and two hrtimers timing out a stalled transfer. Mixing a lock-free cmpxchg() claim in sendmsg() with hrtimercancel() calls made under so->rxlock elsewhere left windows where a frame or timer callback could act on a state that had already moved on, corrupting an unrelated transfer.

so->rxlock now covers the full lifecycle of a TX claim: sendmsg() takes it to check so->tx.state is ISOTPIDLE, switch it to ISOTPSENDING, bump so->txgen and drain the previous transfer's timers - all as one critical section. isotprcvfc()/isotprcvcf() already run under this lock via isotprcv(), and isotprcv_echo() now takes it itself, so none of them can ever observe a transfer mid-claim. This also means a transfer can no longer be handed to sendmsg()'s cleanup paths (signal or send error) while another thread is concurrently claiming or finishing it, so those paths can cancel timers and reset the state unconditionally.

isotprelease() claims the socket the same way, so a racing sendmsg() sees a consistent ISOTPSHUTDOWN and skips arming its timer or sending.

Only the hrtimer callbacks stay outside so->rxlock, since they run under so->rxlock's cancellation elsewhere and taking it themselves would deadlock. so->tx_gen lets them recognize whether the transfer they timed out is still the one currently active, so they don't report an error against a transfer that has since completed or been superseded.