wd@denx.de wrote on Wednesday, July 19, 2006 2:01 PM:
In message 47F3F98010FF784EBEE6526EAAB078D1024F812D@tq-mailsrv.tq-net.de you wrote:
Change POST behaviour for TQM5200 boards Patch by Martin Krause, 8 Mar 2006
Now a POST memory test is only performed on STK52XX base boards (and there only if pin PSC6_3 is read as "1" during boot time). On all other TQM5200 base boards no memory test is performed during POST (regardless of the state of pin PSC6_3).
Please explain the rationale behind this change. Also explain how the user can influence the PSC6_3 port reading.
Current behaviour for all TQM5200 configurations (without patch):
If CONFIG_POST is defined, The pins PSC6_1 and PSC6_3 are configured as inputs. The state of PSC6_3 is read in in the function post_hotkeys_pressed(). If PSC6_3 is read as "1", a memory test is performed during POST. This test increases boot time a lot. If PSC6_3 is read as "0", the memory test is skipped.
The reason for this was to have an easy way to switch the POST memory test on and off during boot time. On the Evaluation Kit for the TQM5200 (the STK52xx) a Jumper could be used to pull PSC6_3 high or low and so turn the test on and off.
On base boards other than the STK52xx this behaviour could conclict with the designated usage of the pin PSC6_3 on this board.
New behaviour _with_ the patch:
With the patch, PSC6_3 is now only checked (and configured) if CONFIG_STK52XX is defined. So for TQM5200 base boards other than the STK52xx the pin is not configured, not checked, and no memory test is performed.
Regards, Martin