PR 415/bsps
* pci/detect_raven_bridge.c, startup/bspstart.c:
This patch reverts enabling MCP interrupts for the motorola_powerpc
BSPs. REASON: pci config space scanning (as e.g. done by the DEC
Ethernet driver) would raise machine check exceptions.
PR 428/bsps
PR 432/bsps
* bootloader/pci.c: Re-instated code that prevents remapping small
IO regions, which if remapped would cause i8259 registers to move
out from under the #define'd base addresses.
* startup/bspstart.c: Reduced BAT2 PCI memory allocation to 256
megs, I incorrectly had extended it which would cause problems with
PCI devices that defined prefetchable memory.
PR 433/tools
* mips-stub.c: Patch from C. M. Heard <heard@pobox.com> described as:
The purpose of the patch is to make the stub ignore all
characters except '+' (ack) or '-' (nak) when it has transmitted a
packet and expects an acknowledgment. Without the patch,
putpacket() will retransmit each time it receives anything other
than '+' when expecting an ack, and that, bitter experience has
found, frequently results in the stub failing to sync up when
initially attaching to gdb.
I think that rtems users will notice significant improvement in
the of the mips stub if this patch is incorporated in the
distribution; certainly I did in my stand-aline environment. I
should hasten to add that it is only an improvement and not a 100%
cure for synchronization problems because '+' characters can (and
do) during as message data, and it's still possible for these to
be mistaken as acks. I have no idea how to fix that short of
redesigning the gdb-remote protocol.
PR 416/bsps
* ppc403/ictrl/ictrl.c (ictrl_isr): We acknolwegde the interrupt in
interrupt controller (clr_exisr(mask)) before calling the interrupt
handler that will acnowledge the interrupt source. This results in
the interrupt beeing seen a second time by the interrupt controller.
Reported and fixed by El Kolli Yacine <yacine.elkolli@crf.canon.fr>.
PR414 - Fixed the global interupt vector register.
Linker command file to have the boot code in the first 8K.
The chip select remap needed to be volatile.
PR 405/bsps
* bootloader/pci.c: Added support for configuring devices for pci
busses > 0
* pci/pci.c, pci/pci.h: Added FixupPCI() to store vectors in the
INTERRUPT_LINE register of pci devices any # of hops away
from the host processor.
* motorola/motorola.c, motorola/motorola.h: Added interrupt
routing tables in support of FixupPCI. This is board-specific,
each board will have to supply information for FixupPCI() to do
anything for it.
* startup/bspstart.c: Extended bat2 to cover entire PCI address space.
* irq/irq.c, irq/irq.h: Added support for shared interrupts.
Existing single hander vectors are undisturbed, a new function
added to allow adding/removing handlers from a vector.
PR 415/bsps
* include/bsp.h, pci/detect_raven_bridge.c, startup/bspstart.c:
Support enabling MCP exceptions on the host bridge. This can
be used for memory probing on the VME bus.
PR 380/bsps
* vmeUniverse/vmeUniverse.c, vmeUniverse/vmeUniverse.h:
make printk format strings compliant with libcpu/printk.
minor fixes and lazy init bugfix.
PR 379/bsps
* console/polled_io.c: libcpu provides 'printk' already.
Therefore, the implementation in this file was removed (still
used for the bootloader, though). It now provides BSP_output_char()
for libcpu's printk().
* console/uart.c, console/uart.h: BSP_output_char_via_serial()
prototype changed to match the BSP_output_char_function_type. Note
that the motorola BSPs use polled-io for the output_char routine, not
the uart.c version. The latter can be used be other BSPs however
(e.g. SVGM).
* console/console.c, console/consoleIo.h, console/polled_io.c,
irq/irq_init.c, openpic/openpic.c, pci/detect_raven_bridge.c:
Unfortunately, the supported 'printk' format string subset of the
polled-io and libcpu implementations are different - hence, a few
format strings in the ppc/shared BSP were changed.
PR 368/filesystems
* startup/ldsegs.S: Increase alignment.
* Makefile.am, configure.ac, include/bsp.h, wrapup/Makefile.am:
Added IDE supporting infrastructure.
* ide/Makefile.am, ide/ide.c, ide/idecfg.c: New files. Added
BSP support for libchip standard ide driver.
PR 360/bsps
* irq/irq.c, irq/irq_init.c, openpic/openpic.c, openpic/openpic.h:
BSP code had set the task priority register multiple times of
the OpenPIC instead of setting the individual source priorities.
This patch adds openpic_get_source_priority() and
openpic_set_source_priority() calls and lets IRQ management code
use them.
PR 349/bsps
* irq/irq.c, vme/VME.h, vme/vmeconfig.c: Fixes library dependency
on the vmeUniverse driver. It is now possible to use the ppc/shared/irq
code on non-VME BSPs without triggering linkage of the vmeUniverse
driver.
PR 356/bsps
This patch makes RTEMS/powerpc/shared EABI compliant.
* irq/irq_init.c, start/Makefile.am, start/start.S, startup/bspstart.c,
startup/pgtbl_setup.c, vectors/vectors.h, vectors/vectors_init.c:
zero_bss() should clear not only bss but sbss and sbss2
also (this is probably a bugfix, as sbss/sbss2 are
probably used even if -msdata=eabi is not specified).
* start/rtems_crti.S: New file which must
be linked immediately after ecrti.o. rtems_crti.o contains
a code snippet who lets __init() return immediately. Also,
a new entry point '_init' to the C++ Ctors is generated
for use by the RTEMS Thread_Handler.
PR 356/bsps
* bsp_specs, start/Makefile.am, startup/linkcmds
This patch makes RTEMS/powerpc/shared EABI compliant.
Declare all linker-script defined addresses as
variables of unknown size and not as short objects
like 'int', 'void*' etc.
This prevents the compiler from generating a relocation
who assumes the existence of an __rtems_end object in
one of the short data areas (which would result in a
linker error).
PR 349/bsps
* mpc6xx/exceptions/raw_exception.c, mpc6xx/mmu/bat.c,
mpc6xx/mmu/pte121.c, shared/include/cpuIdent.c,
shared/include/cpuIdent.h, shared/src/Makefile.am, shared/src/stack.c,
shared/src/stackTrace.h, powerpc/registers.h:
- undo improper 'fix' who broke mpc604r identification
- fix: 7400 identification PVR value was wrong
- enhance 'setdbat()' to switch OFF a given BAT if called with 0 size
- fix: page table support bugfix
- enhancement: provide routines to take and print stack trace
snapshots
- add definitions for HID1 and DABR SPRs
Cherrypick from master 2003-02-20 22:07:22 UTC Joel Sherrill <joel.sherrill@OARcorp.com> '2003-02-20 Till Straumann <strauman@slac.stanford.edu>':
c/src/lib/libcpu/powerpc/shared/src/stack.c
c/src/lib/libcpu/powerpc/shared/src/stackTrace.h
PR 349/bsps
* startup/Makefile.am, startup/linkcmds startup/Makefile.am,
startup/bspstart.c, startup/pgtbl_activate.c, startup/pgtbl_setup.c:
Let the powerpc/shared (+derived) BSPs use pagetable support
from libcpu.
PR 349/bsps
* startup/Makefile.am, startup/linkcmds startup/Makefile.am,
startup/bspstart.c, startup/pgtbl_activate.c, startup/pgtbl_setup.c:
Let the powerpc/shared (+derived) BSPs use pagetable support
from libcpu.
PR 349/bsps
* Makefile.am, configure.ac: Add glue to the powerpc/shared BSP
to use the vmeUniverse VME-PCI bridge driver.
* vme/.cvsignore, vme/Makefile.am, wrapup/Makefile.am: New files.
PR 349/bsps
* openpic/Makefile.am, pci/Makefile.am:
- install the 'openpic.h' and 'pci.h' headers
- avoid a name clash by renaming 'vsprintf' & friends
to 'k_vsprintf' etc.
- let 'rtemsReboot' print a stack trace (in case an 'assert' failed)
- irq.c:
* fix: remove implicit assumption that ISA interrupt vectors start
at 0
* add BSP hook to let a VME interrupt acknowledge the PCI/openpic
IRQ. (SEE ALSO THE ppc-sharedbsp-vme.diff patch)
- fix: EABI compliance; isr/exception lowlevel handlers must not
clobber R2 or R13
- fix: openpic_init was called with the polarity/senses tables
swapped (fixed wrong order of arguments)
- cosmetics: use new _read_SPRXX() _write_SPRXX() macros
PR 349/bsps
* console/polled_io.c, console/reboot.c, irq/irq.c, irq/irq_asm.S,
irq/irq_init.c, openpic/Makefile.am, startup/linkcmds,
vectors/vectors.S, vectors/vectors_init.c:
- install the 'openpic.h' and 'pci.h' headers
- avoid a name clash by renaming 'vsprintf' & friends
to 'k_vsprintf' etc.
- let 'rtemsReboot' print a stack trace (in case an 'assert' failed)
- irq.c:
* fix: remove implicit assumption that ISA interrupt vectors start
at 0
* add BSP hook to let a VME interrupt acknowledge the PCI/openpic
IRQ. (SEE ALSO THE ppc-sharedbsp-vme.diff patch)
- fix: EABI compliance; isr/exception lowlevel handlers must not
clobber R2 or R13
- fix: openpic_init was called with the polarity/senses tables
swapped (fixed wrong order of arguments)
- cosmetics: use new _read_SPRXX() _write_SPRXX() macros
PR 349/bsps
* startup/pgtbl_setup.c, startup/pgtbl_activate.c: New files
* startup/Makefile.am, startup/bspstart.c, startup/linkcmds:
Let the powerpc/shared (+derived) BSPs use pagetable support
from libcpu.
PR 349/bsps
* vme/.cvsignore, vme/Makefile.am, vme/VME.h, vme/VMEConfig.h,
vme/vmeconfig.c: Add glue to the powerpc/shared BSP to use
the vmeUniverse VME-PCI bridge driver.
PR 349/bsps
* console/console.c, console/uart.c, console/uart.h: implement
IOCTLs for the serial (UART) console to install/retrieve a BREAK-IRQ
callback. The callback routine (if installed) is invoked from the
UART ISR when a BREAK interrupt is detected. This can be used
e.g. to enforce a "hotkey" reboot a la vxWorks Ctrl-X (although we
use the serial line break condition) NOTE: The callback runs in
ISR context.
