[RTEMS Project] #3849: Fix PSIM memory map

[RTEMS trac]

#3849: Fix PSIM memory map
-----------------------------+---------------------
 Reporter:  Sebastian Huber  |       Owner:  (none)
     Type:  defect           |      Status:  new
 Priority:  normal           |   Milestone:  5.1
Component:  arch/powerpc     |     Version:  5
 Severity:  normal           |  Resolution:
 Keywords:                   |  Blocked By:
 Blocking:                   |
-----------------------------+---------------------

Comment (by Joel Sherrill):

 The code posted here misses the most important part -- the simulator
 device tree (https://git.rtems.org/rtems-
 tools/tree/tester/rtems/testing/bsps/psim-device-tree). All of the device
 addresses start with 0x0c000000 which was OK when the RAM was smaller. But
 increasing the RAM to 256MB (0x10000000) resulted in the RAM overlapping
 the devices. The simplest option I see is to change the leading 0x0c... to
 0xfc... in all device addresses and avoid the conflict.

 As best I can tell, the set of changes is pretty small:

 * Change rtems-tools/tester/testing/bsps/psim-device-tree to use 0xFC for
 all devices. This reconfigures the "hardware"
 * Change rtems/bsps/powerpc/psim/linkcmds to change new base address of of
 all the devices which are at the symbol PSIM
 * Delete rtems/bsps/powerpc/start/device-tree and update the BSP README to
 note where the device tree file is. Also update the README to note that
 you don't need a modified gdb since the modifications have been in since
 the last century and no one knows the papyrus BSP anymore. I can't even
 remember it.
 * Change rtems/bsps/start/bspstart.c to fix 0x0c to 0xfc in the PCI
 setdbat calls. Note: I don't remember how a base address is broken down to
 be passed in

 I think that's it.

--
Ticket URL: <http://devel.rtems.org/ticket/3849#comment:1>
RTEMS Project <http://www.rtems.org/>
RTEMS Project