Re: [PATCH] can: isotp: implement N_Ar timeout handling for FC transmission

[yewentian395]

On 24.08.26 20:03, Oliver Hartkopp wrote:
> Nobody cares about so->frame_txtime and therefore I will not add any
> extra complexity to check for Ar/As timeouts which have no real world
> effect.

Thanks for the detailed explanation. The simplified model makes sense
given that no real-world user depends on strict N_Ar confirmation.

> Btw. while double-checking the code I have seen, that I was missing the
> addition of so->frame_txtime when starting the rxtimer above.
>
> I'll mention you with a Reported-by tag then.

Thank you, appreciated.

> Please also note that your patch would not apply on the latest upstream
> isotp code, which does not contain "*(u32 *)cf->data" anymore.

Noted. I was working against an older tree. Will rebase to latest
upstream for any future submissions.

Best regards,
wentian