Someone made some sort of mistake in the kernel serial driver, and added a
comparison against new_serial.xmit_fifo_size in the set_serial_info()
call. If it's 0, then that call fails. It's not setserial's fault. I'm
appending a unified diff against the kernel serial driver code.
Actually, xmit_fifo_size shouldn't be zero. If it's zero, then
something very bad is happening, and it used to cause a division by zero
error in rs_wait_until_sent. I've since added a safety check which
causes it to not kernel oops in that case, but the serial driver will
not function correctly if xmit_fifo_size is zero.
I'm told the plug and play code is setting xmit_fifo_size to be zero;
that's certainly wrong, but I haven't had a chance to try to track down
why the PNP code is doing that.
- Ted
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu