On 2024-12-02, Jiri Slaby <jirislaby@xxxxxxxxxx> wrote: > I am still asking why do you want to wait for the TX machinery at the > *end* (for the last 64 B of the 640 B line) of transmission at all? It > occurs to me as wasted cycles. The printk-framework has always expected that when console->write() returns, the data has been flushed out of the hardware. I am guessing because it is easiest to avoid possible data loss, for example, due to suspending hardware. If you want me to change the current behavior, I can do that in a separate patch. I would like this patch to only be about fixing the FIFO timeout issue. John
