Seeking help diagnosing serial port failure to transmit (while receive works)

Seeking help diagnosing serial port failure to transmit (while receive works)

[Forest]

Hi, folks.

I'm seeing puzzling behavior from a recent AMD motherboard serial port:
It receives data, but doesn't appear to send, and doesn't report any
errors. Does this problem ring a bell for anyone here?

dmesg reports this:
Serial: 8250/16550 driver, 4 ports, IRQ sharing enabled
00:04: ttyS0 at I/O 0x3f8 (irq = 4, base_baud = 115200) is a 16550A

The /proc/tty/driver/serial tx count rises as I type in picocom, but the
terminal at the other end of a null-modem cable (a laptop with a
USB/serial adapter) doesn't see the typed characters. The other
direction of the same connection works just fine.

Things that have not helped:
- Replacing the motherboard's cables and connectors
- Replacing the null modem adapter with a different model
- Replacing the remote USB/serial adapter with a different model
- Switching picocom's flow control between RTS/CTS, xon/xoff, none
- Changing the serial port's IRQ and address in the BIOS
- Changing between 9600 and 115200 bps

I'm approaching my wit's end. Could this be a known problem with the
kernel's serial driver and the B650 chipset (socket AM5), or with recent
AGESA versions?

Thanks for your thoughts.

Re: Seeking help diagnosing serial port failure to transmit (while receive works)

[Greg KH]

On Mon, Feb 23, 2026 at 03:19:03PM -0800, Forest wrote:
> Hi, folks.
> 
> I'm seeing puzzling behavior from a recent AMD motherboard serial port:
> It receives data, but doesn't appear to send, and doesn't report any
> errors. Does this problem ring a bell for anyone here?
> 
> dmesg reports this:
> Serial: 8250/16550 driver, 4 ports, IRQ sharing enabled
> 00:04: ttyS0 at I/O 0x3f8 (irq = 4, base_baud = 115200) is a 16550A
> 
> The /proc/tty/driver/serial tx count rises as I type in picocom, but the
> terminal at the other end of a null-modem cable (a laptop with a
> USB/serial adapter) doesn't see the typed characters. The other
> direction of the same connection works just fine.
> 
> Things that have not helped:
> - Replacing the motherboard's cables and connectors
> - Replacing the null modem adapter with a different model
> - Replacing the remote USB/serial adapter with a different model
> - Switching picocom's flow control between RTS/CTS, xon/xoff, none
> - Changing the serial port's IRQ and address in the BIOS
> - Changing between 9600 and 115200 bps
> 
> I'm approaching my wit's end. Could this be a known problem with the
> kernel's serial driver and the B650 chipset (socket AM5), or with recent
> AGESA versions?

Based on experience, it's always flow control being set somewhere (on
one side or the other), so watch out for that.

good luck!

greg k-h