On Thu, Dec 3, 2015 at 10:35 PM, David Mandala email@example.com wrote:
On 12/3/15 4:32 PM, Kevin Hilman wrote:
Grant Likely firstname.lastname@example.org writes:
On 3 Dec 2015 07:46, "Victor Chong" email@example.com wrote:
On Thursday, December 3, 2015, Kevin Hilman firstname.lastname@example.org wrote:
aaaaah! I wasn't expecting partial output on UART0, then on UART3.
Yea.. the partial output can usually be ignored for the most part unless
you're working on those components directly.
After setting the right jumper settings on the UART board, I'm now seeing the BL1 on UART0 and the rest on UART3 (a.k.a. UART1 on the UART board... which is also confusing.)
It's more confusing than that. The UART board gives you UART2 & UART3. UART0 isn't available to it.
I suspect it's likely that hikey will not be the only one to have SoC UART numbering different that the UART board numbering. Maybe the UART board options should have different names, like UART_A and UART_B ?
It's the reason why we have UDEV rules now making the two UARTs on the LS connector /dev/tty96B0 and /dev/tty96B1 get rid of the confusion and make it MUCH easier to document things.
First, that only helps userspace. When I go add "console=tty96B0" and it doesn't work...
Second, now we need udev rules for every board. That will be fun when we have 96 of them.
We already have big enough problem with tty naming being: tty<some random soc naming>n. Ask the OMAP folks about how fun changing names was. IMO, we should move towards everything being ttySn.
We have aliases in DT to provide consistent numbering. We should use that so LS connector UARTs are consistently numbered. Alternatively, we could use "label" properties to provide standard labels. That is what label is for: providing a human readable label.
Anything in userspace that is 96boards specific is a flawed solution. These problems have nothing to do with 96boards, but are shortcomings in the kernel interfaces. Lets fix those and not be working around them.