diff options
Diffstat (limited to 'c/src/lib/libbsp/m68k/mvme147/README')
-rw-r--r-- | c/src/lib/libbsp/m68k/mvme147/README | 86 |
1 files changed, 86 insertions, 0 deletions
diff --git a/c/src/lib/libbsp/m68k/mvme147/README b/c/src/lib/libbsp/m68k/mvme147/README new file mode 100644 index 0000000000..b7952ad9cb --- /dev/null +++ b/c/src/lib/libbsp/m68k/mvme147/README @@ -0,0 +1,86 @@ +# +# $Id$ +# + +Notes about the MVME147 bsp + +MVME147 port for TNI - Telecom Bretagne +by Dominique LE CAMPION (Dominique.LECAMPION@enst-bretagne.fr) +May 1996 + + +This bsp is essentially based on the mvme136 bsp. + +Summary of the modifications that were made: + + * include + + - bsp.h + Peripheral Channel Controller memory mapping + Z8530 memory mapping + + * startup + + - bspstart.c + main () setup for VME roundrobin mode + setup for the PCC interrupt vector base + - bspclean.c + bsp_cleanup () disable timer 1 & 2 interruptions + - linkcmds set the RAM start (0x5000) and size (4Meg - 0x5000) + - setvec.c unchanged + - sbrk.c unchanged + + * console + + - console.c taken from the dmv152 bsp (Zilog Z8530) + with no modification + + * clock + + - ckinit.c entirely rewritten for the PCC tick timer 2 + + * timer + + - timerisr.s and timer.c + entirely rewritten for the PCC tick timer 1 + now gives results un 6.25 us units (mininum timer delay, + suprising big grain) + + * times + + - updated results for the mvme147 (beware of the 6.25 us grain) + + * Makefiles + + - compilation of shmsupp simply removed + + +To be done: + + * add VMEchip memory mapping to include/bsp.h + + * update the overheads in coverhead.h + + * add support for serila ports 2,3 and 4. + +Other notes: + + * There is no MP support (no more shmsupp) because I have no + experience of the VME bus. The mvme136 shared memory support + does not seem applicable on the VMEchip of the mvme147, so + I don't know where to start. Suggestions are welcome. + + * All the timing tests and sp tests have been run except tmoverhd. + The test hangs during the pause (where the task should be suspended + until a return). Maybe the rtems_initialize_executive is no more + reentrant with this bsp. + +Future work: + + * Add gdb serial remote support. + + * Shared memory support (I don't really need it, but I can do + it if it's simple). + + * Message passing on VME bus, with Ada 95 annex E (distributed + systems) in mind. |