Note that adding UART info to the coreboot table is still guarded by CONSOLE_SERIAL, not sure if that's intentional (conceptually, since this enables passing UART info to the kernel via ACPI, maybe it should also enable passing UART info to payloads?).

Also, there is a lot of SoC and mainboard code using CONFIG(CONSOLE_SERIAL) to check for things that should really be guarded by ENABLE_UART instead after this. Have you considered that it might be easier to just add an option that makes DEFAULT_CONSOLE_LOGLEVEL -1 instead? That would also allow users on systems that allow runtime-configured log levels to turn it up later if desired, which might be nice. (Alternatively, for a production Chromebook, I think setting it at BIOS_EMERG would be fine as well since that should essentially never output anything unless the machine hangs anyway.)

View Change

3 comments:

• File src/drivers/uart/Kconfig:

  • Patch Set #1, Line 4: config ENABLE_UART

    nit: why not just make this option directly menuconfig-visible? I think it would have the same effect.

  • Patch Set #1, Line 12: The primary UART has previously only been set up when the serial console

    nit: I think explaining how older versions used to work is confusing and somewhat irrelevant here... I'd just explain what it does.

• File src/include/console/uart.h:

  • Patch Set #1, Line 70: #if CONFIG(ENABLE_UART)

    This is changing the effect for (ENV_SMM && !CONFIG(DEBUG_SMI)). Not sure if that's intentional or whether it matters in practice.

To view, visit change 33380. To unsubscribe, or for help writing mail filters, visit settings.