From: NIIBE Y. <gn...@ch...> - 2000-05-10 00:22:50
|
pi...@li... wrote: > As some of you know already, I'm working with a 7707-based (little > endian) board from Hypercom, equipped with 16M of RAM and a "debug > sub-board" from Hitachi. This debug board is what I use to download the > kernel and the initrd filesystem to the board (I don't have a working > flash driver yet so I use a ramdisk). Also, the debug board features 8 > LEDs and a 8-digit 7-segment display that one can use to help debugging > without console or debugger. I'm not sure what's 7707 is. It seems for me that it's slower version of 7709. Would it be better to have config option for 7707/7709/7709A? Is there any footprint (or fingerprint, whatever we say) to distingush 7707 on runtime? > So, I thought I might contribute the modifications I have done to the > kernel to support that specific board, as well as other "niceties". The > patch at the end of this message is against the 2.3.99-pre7 kernel that > is in the linuxsh CVS on sourceforge (tagged > "L2_3_99_PRE7-6_AFTER: 1.4"). It provides : Thanks for the patch. For now, I consider about: > - drivers/char/sh_sci.c and drivers/char/sh_sci.h have been modified for > a bus clock of 30Mhz and take the clock source of the SCI into account, > so that it is possible to use the full range of baudrates from 110 bps > to 115200 bps. Also, "serial_console_setup()" now sets the baudrate > before exiting, which is required if the kernel uses the SCI as initial > console. TODO : set the bus clock automagically, or at least have an > config option. Right now, it is hard-coded (yuk). There's many ways (mode) to set clock for SuperH. I'm not fully understand about CPG, and I'm not sure, which is the one we can find on runtime, which is the one we should provide as config option. I think that we should provide better support of CPG to get the information of module clock, and place it in arch/sh/kernel/time.c. And use the value in the driver. Thought? -- |