On Tue, Mar 08, 2011 at 11:46:47AM +0800, Shawn Guo wrote:
On Tue, Mar 08, 2011 at 11:31:07AM +0800, Jason Hui wrote:
Hi, Shawn,
On Tue, Mar 8, 2011 at 11:11 AM, Shawn Guo shawn.guo@freescale.com wrote:
On Tue, Mar 08, 2011 at 10:20:33AM +0800, Jason Hui wrote:
Hi, Shawn,
On Mon, Mar 7, 2011 at 10:51 PM, Shawn Guo shawn.guo@freescale.com wrote:
Aha, Jason. Grant and I put many comments on v1 of this patch, but it seems that v2 is identical to v1, except adding Rob Herring sign-off.
Please compare with v1 and say it once again, It should include the
Oops! Sorry, I messed up with my local branches last night, probably because it was late :)
grant and your comments except the mxctty1 due to I think that it's not one issue, So, I did not take your comments.
I still think it's an issue that the dt kernel needs a different console cmdline from the non-dt one.
In fact, this is user defined, I think But why we need mxctty1 in dt here, the reason is that,
uart3 will be probed first and then uart0, uart1, which lead to mxctty0 ---> uart3,
mxctty1---> uart0 and mxctty2 --->uart1.
The debug uart of mx51 babbage board is uart0, so we need define mxctty1 in the kernel command line.
We can put spba@70000000 at the back of aips@73f00000, but it will make the memory address not sorted from lower to higher.
So we probably need one more number in 'reg' to tell the port id?
As discussed in my other reply; it is a bad idea to try and encoded the port id into a device node. Use /aliases instead.
g.