12 years ago[POWERPC] 4xx: Add AMCC Kilauea eval board support to platforms/40x
Stefan Roese [Thu, 11 Oct 2007 12:08:21 +0000]
[POWERPC] 4xx: Add AMCC Kilauea eval board support to platforms/40x

This patch adds basic support for the new 405EX and the AMCC eval board
Kilauea to arch/powerpc.

Signed-off-by: Stefan Roese <sr@denx.de>
Signed-off-by: Josh Boyer <jwboyer@linux.vnet.ibm.com>

12 years ago[POWERPC] 4xx: Add AMCC 405EX support to cputable.c
Stefan Roese [Thu, 11 Oct 2007 12:08:14 +0000]
[POWERPC] 4xx: Add AMCC 405EX support to cputable.c

Signed-off-by: Stefan Roese <sr@denx.de>
Signed-off-by: Josh Boyer <jwboyer@linux.vnet.ibm.com>

12 years agoMerge branch 'virtex-for-2.6.24' of git://git.secretlab.ca/git/linux-2.6-virtex into...
Josh Boyer [Thu, 11 Oct 2007 12:45:20 +0000]
Merge branch 'virtex-for-2.6.24' of git://git.secretlab.ca/git/linux-2.6-virtex into for-2.6.24-4xx

12 years ago[POWERPC] Make clockevents work on PPC601 processors
Paul Mackerras [Thu, 11 Oct 2007 11:46:45 +0000]
[POWERPC] Make clockevents work on PPC601 processors

In testing the new clocksource and clockevent code on a PPC601
processor, I discovered that the clockevent multiplier value for the
decrementer clockevent was overflowing.  Because the RTCL register in
the 601 effectively counts at 1GHz (it doesn't actually, but it
increases by 128 every 128ns), and the shift value was 32, that meant
the multiplier value had to be 2^32, which won't fit in an unsigned
long on 32-bit.  The same problem would arise on any platform where
the timebase frequency was 1GHz or more (not that we actually have any
such machines today).

This fixes it by reducing the shift value to 16.  Doing the
calculations with a resolution of 2^-16 nanoseconds (15 femtoseconds)
should be quite adequate.  :)

Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] Prevent decrementer clockevents from firing early
Paul Mackerras [Mon, 8 Oct 2007 23:59:17 +0000]
[POWERPC] Prevent decrementer clockevents from firing early

On old powermacs, we sometimes set the decrementer to 1 in order to
trigger a decrementer interrupt, which we use to handle an interrupt
that was pending at the time when it was re-enabled.  This was causing
the decrementer clock event device to call the event function for the
next event early, which was causing problems when high-res timers were
not enabled.

This fixes the problem by recording the timebase value at which the
next event should occur, and checking the current timebase against the
recorded value in timer_interrupt.  If it isn't time for the next
event, it just reprograms the decrementer and returns.

This also subtracts 1 from the value stored into the decrementer,
which is appropriate because the decrementer interrupts on the
transition from 0 to -1, not when the decrementer reaches 0.

Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] Fix performance monitor on machines with logical PVR
Paul Mackerras [Thu, 4 Oct 2007 04:18:01 +0000]
[POWERPC] Fix performance monitor on machines with logical PVR

Some IBM machines supply a "logical" PVR (processor version register)
value in the device tree in the cpu nodes rather than the real PVR.
This is used for instance to indicate that the processors in a POWER6
partition have been configured by the hypervisor to run in POWER5+
mode rather than POWER6 mode.  To cope with this, we call identify_cpu
a second time with the logical PVR value (the first call is with the
real PVR value in the very early setup code).

However, POWER5+ machines can also supply a logical PVR value, and use
the same value (the value that indicates a v2.04 architecture
compliant processor).  This causes problems for code that uses the
performance monitor (such as oprofile), because the PMU registers are
different in POWER6 (even in POWER5+ mode) from the real POWER5+.

This change works around this problem by taking out the PMU
information from the cputable entries for the logical PVR values, and
changing identify_cpu so that the second call to it won't overwrite
the PMU information that was established by the first call (the one
with the real PVR), but does update the other fields.  Specifically,
if the cputable entry for the logical PVR value has num_pmcs == 0,
none of the PMU-related fields get used.

So that we can create a mixed cputable entry, we now make cur_cpu_spec
point to a single static struct cpu_spec, and copy stuff from
cpu_specs[i] into it.  This has the side-effect that we can now make
cpu_specs[] be initdata.

Ultimately it would be good to move the PMU-related fields out to a
separate structure, pointed to by the cputable entries, and change
identify_cpu so that it saves the PMU info pointer, copies the whole
structure, and restores the PMU info pointer, rather than identify_cpu
having to list all the fields that are *not* PMU-related.

Signed-off-by: Paul Mackerras <paulus@samba.org>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

12 years ago[POWERPC] Don't enable cpu hotplug on pSeries machines with MPIC
Olof Johansson [Wed, 10 Oct 2007 00:38:24 +0000]
[POWERPC] Don't enable cpu hotplug on pSeries machines with MPIC

Don't allow cpu hotplug on systems lacking XICS interrupt controller
(i.e. with an MPIC interrupt controller), since the current pSeries
platform code is hardcoded for XICS.

This works around the bug reported by Paul Mackerras where the
disable_nonboot_cpus() call recently added to the shutdown path will
cause an oops on older pSeries machines.

Signed-off-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] Move of_platform_driver initialisations: arch/powerpc
Stephen Rothwell [Thu, 11 Oct 2007 05:19:03 +0000]
[POWERPC] Move of_platform_driver initialisations: arch/powerpc

We no longer initialise the name and owner fields of the
of_platform_driver, but use the fields of the embedded device_driver's
name field instead.

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Acked-by: Olof Johansson <olof@lixom.net>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] iSeries: Move viodasd probing
Stephen Rothwell [Thu, 11 Oct 2007 04:59:54 +0000]
[POWERPC] iSeries: Move viodasd probing

This way we only have entries in the device tree for disks that actually
exist.  A slight complication is that disks may be attached to LPARs
at runtime.

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Acked-by: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] iSeries: Move detection of virtual tapes
Stephen Rothwell [Thu, 11 Oct 2007 04:58:31 +0000]
[POWERPC] iSeries: Move detection of virtual tapes

Now we will only have entries in the device tree for the actual existing
devices (including their OS/400 properties).  This way viotape.c gets
all the information about the devices from the device tree.

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] iSeries: Move detection of virtual cdroms
Stephen Rothwell [Thu, 11 Oct 2007 04:57:26 +0000]
[POWERPC] iSeries: Move detection of virtual cdroms

Now we will only have entries in the device tree for the actual existing
devices (including their OS/400 properties).  This way viocd.c gets all
the information about the devices from the device tree.

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Acked-by: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] Remove more iSeries-specific stuff from vio.c
Stephen Rothwell [Thu, 11 Oct 2007 04:55:02 +0000]
[POWERPC] Remove more iSeries-specific stuff from vio.c

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] Remove iSeries_vio_dev
Stephen Rothwell [Thu, 11 Oct 2007 04:53:32 +0000]
[POWERPC] Remove iSeries_vio_dev

It was only being used to carry around dma_iommu_ops and vio_iommu_table
which we can use directly instead.  This also means that vio_bus_device
doesn't need to refer to them either.

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Acked-by: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] iSeries: Simplify viocd initialisation
Stephen Rothwell [Thu, 11 Oct 2007 04:50:55 +0000]
[POWERPC] iSeries: Simplify viocd initialisation

We don't need to keep a lump of dma coherent memory around for the life
of the module.

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Acked-by: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] Clean up vio.h
Stephen Rothwell [Thu, 11 Oct 2007 04:48:24 +0000]
[POWERPC] Clean up vio.h

Remove vio_dma_ops declaration (since it no longer exists) and some
unused fields from struct vio_driver.

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] Platforms shouldn't mess with ROOT_DEV
Grant Likely [Wed, 10 Oct 2007 18:48:28 +0000]
[POWERPC] Platforms shouldn't mess with ROOT_DEV

There is no good reason for board platform code to mess with the
ROOT_DEV.  Remove it from all in-tree platforms except powermac.

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] Remove empty ppc_md.setup_arch hooks
Grant Likely [Wed, 10 Oct 2007 18:48:23 +0000]
[POWERPC] Remove empty ppc_md.setup_arch hooks

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Acked-by: Josh Boyer <jwboyer@linux.vnet.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] Only call ppc_md.setup_arch() if it is provided
Grant Likely [Wed, 10 Oct 2007 18:48:18 +0000]
[POWERPC] Only call ppc_md.setup_arch() if it is provided

This allows platforms which don't have anything to do at setup_arch time
(like a bunch of the 4xx platforms) to eliminate an empty setup_arch hook.

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] Disable vDSO support for ARCH=ppc where it's not implemented
Wolfgang Denk [Tue, 9 Oct 2007 22:36:18 +0000]
[POWERPC] Disable vDSO support for ARCH=ppc where it's not implemented

Signed-off-by: Wolfgang Denk <wd@denx.de>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] Align the sys_call_table
Stephen Rothwell [Tue, 9 Oct 2007 07:03:57 +0000]
[POWERPC] Align the sys_call_table

Our _GLOBAL macro does a ".align 2" so the alignment is fine for 32
bit, but on 64 bit it is possible for it to end up only 4 byte aligned.
I don't know if it matters, but it can't hurt to 8 byte align it.

It also means that when we build with --emit_relocs, none of our 64 bit
relocations are to misaligned places.

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] Uartlite: bootwrapper bug fix, getc loops forever
Grant Likely [Thu, 4 Oct 2007 21:44:54 +0000]
[POWERPC] Uartlite: bootwrapper bug fix, getc loops forever

Fixes inverted logic in uartlite_getc

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>

12 years ago[POWERPC] Don't build arch/powerpc/sysdev/dcr.c for ARCH=ppc kernels
Grant Likely [Thu, 4 Oct 2007 21:44:51 +0000]
[POWERPC] Don't build arch/powerpc/sysdev/dcr.c for ARCH=ppc kernels

dcr.c is an arch/powerpc only thing.  Compiling ppc405 arch/ppc kernels
throws warnings without this change.

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>

12 years ago[POWERPC] Virtex: Fix URL for Xilinx Virtex support in MAINTAINERS
Grant Likely [Thu, 4 Oct 2007 05:24:52 +0000]
[POWERPC] Virtex: Fix URL for Xilinx Virtex support in MAINTAINERS

Change URL in MAINTAINERS to a more relevant one.

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>

12 years ago[POWERPC] XilinxFB: sparse fixes
Grant Likely [Thu, 4 Oct 2007 16:48:37 +0000]
[POWERPC] XilinxFB: sparse fixes

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>

12 years ago[POWERPC] XilinxFB: Make missing pdata structure non-fatal
Grant Likely [Thu, 4 Oct 2007 21:44:52 +0000]
[POWERPC] XilinxFB: Make missing pdata structure non-fatal

Missing pdata structure is not a fatal error.  The device can still be
initialized without it.

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>

12 years ago[POWERPC] XilinxFB: add of_platform bus binding
Grant Likely [Thu, 4 Oct 2007 16:48:37 +0000]
[POWERPC] XilinxFB: add of_platform bus binding

Adds the of_platform bus binding to the xilinxfb driver.  Needed to
use framebuffer devices described in the OF device tree (used by
arch/powerpc).

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Acked-by: Andrei Konovalov <akonovalov@ru.mvista.com>

12 years ago[POWERPC] XilinxFB: cleanup platform_bus binding to use platform bus API.
Grant Likely [Thu, 4 Oct 2007 16:48:37 +0000]
[POWERPC] XilinxFB: cleanup platform_bus binding to use platform bus API.

Change the platform bus binding to make use of the established
platform_bus API.

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Acked-by: Andrei Konovalov <akonovalov@ru.mvista.com>

12 years ago[POWERPC] XilinxFB: Split device setup from bus binding
Grant Likely [Thu, 4 Oct 2007 16:48:37 +0000]
[POWERPC] XilinxFB: Split device setup from bus binding

Split the device setup code away from the platform bus binding.  This is
in preparation for adding the of_platform bus binding to this driver and
most of the setup code is common between the two busses.

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Acked-by: Andrei Konovalov <akonovalov@ru.mvista.com>

12 years ago[POWERPC] XilinxFB: rename failout labels to reflect failure
Grant Likely [Thu, 4 Oct 2007 16:48:37 +0000]
[POWERPC] XilinxFB: rename failout labels to reflect failure

Labels and gotos are used in xilinxfb_assign to unwind allocations
on device registration failures.  Rename the labels to reflect the
error which occured.  This change is being made to make it easier
to add new failout paths (which occurs in a subsuquent patch) and
to make reviewing the failout path easier.

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Acked-by: Andrei Konovalov <akonovalov@ru.mvista.com>

12 years ago[POWERPC] XilinxFB: Replace calls to printk with dev_dbg, dev_err, etc.
Grant Likely [Thu, 4 Oct 2007 16:48:36 +0000]
[POWERPC] XilinxFB: Replace calls to printk with dev_dbg, dev_err, etc.

The dev_dbg, dev_err, etc functions provide more context that plain
vanilla printk which is useful for debugging.  Where appropriate,
change printk calls to the appropriate dev_*() call.

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Acked-by: Andrei Konovalov <akonovalov@ru.mvista.com>

12 years ago[POWERPC] XilinxFB: add banner output to probe routine when DEBUG is defined
Grant Likely [Thu, 4 Oct 2007 16:48:36 +0000]
[POWERPC] XilinxFB: add banner output to probe routine when DEBUG is defined

Debug support: when DEBUG is defined, output relevant details to the
log about the framebuffer registration.

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Acked-by: Andrei Konovalov <akonovalov@ru.mvista.com>

12 years ago[POWERPC] Remove redundant reference to non-existent CONFIG_BOOTIMG
Robert P. J. Day [Sun, 7 Oct 2007 11:34:57 +0000]
[POWERPC] Remove redundant reference to non-existent CONFIG_BOOTIMG

There is no BOOTIMG Kconfig variable, not to mention that there is no
include/linux/bootimg.h header file.

Signed-off-by: Robert P. J. Day <rpjday@mindspring.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] PS3: Add os-area database routines
Geoff Levand [Tue, 9 Oct 2007 01:07:24 +0000]
[POWERPC] PS3: Add os-area database routines

Add support for a simple tagged database in the PS3 flash rom
os-area.  The database allows the flash rom os-area to be shared
between a bootloader and installed operating systems.   The
application ps3-flash-util or the library libps3-utils from the
ps3-utils package can be used for userspace database operations.

The latest ps3-utils package is available here:

  git://git.kernel.org/pub/scm/linux/kernel/git/geoff/ps3-utils.git

Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] PS3: Save os-area params to device tree
Geoff Levand [Sat, 6 Oct 2007 21:35:47 +0000]
[POWERPC] PS3: Save os-area params to device tree

Add the PS3 os-area startup params to the device tree.  This allows
a second stage kernel loaded with kexec to use these values.

Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] PS3: Add os-area rtc_diff set/get routines
Geoff Levand [Sat, 6 Oct 2007 21:35:46 +0000]
[POWERPC] PS3: Add os-area rtc_diff set/get routines

Updates for PS3 os-area rtc_diff set/get routines
o Add a new routine ps3_os_area_set_rtc_diff().
o Rename ps3_os_area_rtc_diff() to ps3_os_area_get_rtc_diff().
o Remove static variable rtc_shift with calls to ps3_os_area_get_rtc_diff().

Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] PS3: os-area workqueue processing
Geoff Levand [Sat, 6 Oct 2007 21:35:45 +0000]
[POWERPC] PS3: os-area workqueue processing

Add a workqueue to the PS3 os-area support.  This is needed to
support writing updates to flash memory and to update the /proc
device tree entries from the timer tick interrupt context.

Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] PS3: Remove unused os-area params
Geoff Levand [Sat, 6 Oct 2007 21:35:44 +0000]
[POWERPC] PS3: Remove unused os-area params

Updates for PS3 os-area startup params
o Remove some unused PS3 os-area startup params from struct saved_params.
o Rename ps3_os_area_init() to ps3_os_area_save_params().
o Zero mirrored header after saving params.

Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] PS3: Cleanup of os-area.c
Geoff Levand [Sat, 6 Oct 2007 21:35:43 +0000]
[POWERPC] PS3: Cleanup of os-area.c

Minor cleanup of the PS3 file os-area.c:
 o Correct file text header.
 o Add type names enum os_area_ldr_format, enum os_area_boot_flag,
   enum os_area_ctrl_button.
 o Change struct os_area_header.magic_num type to u8.
 o Add preprocessor macro SECONDS_FROM_1970_TO_2000.

Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] cell: Add Cell memory controller register defs and expose it
Benjamin Herrenschmidt [Thu, 4 Oct 2007 05:40:43 +0000]
[POWERPC] cell: Add Cell memory controller register defs and expose it

This adds definitions for the Cell memory controller registers (at
least some of them) for use by the EDAC driver for ECC error reporting.

It also expose the said MIC as a platform device that can be used
by the EDAC driver to match on.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] cell: Move cbe_regs.h to include/asm-powerpc/cell-regs.h
Benjamin Herrenschmidt [Thu, 4 Oct 2007 05:40:42 +0000]
[POWERPC] cell: Move cbe_regs.h to include/asm-powerpc/cell-regs.h

The new Cell EDAC driver needs that file, oprofile also does ugly
path tricks to get to it, it's time to move it to asm-powerpc. While
at it, rename it to be consistent with cell-pmu.h (and dashes look
nicer than underscores anyway).

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] Enable debug info on boot wrapper
Grant Likely [Thu, 4 Oct 2007 04:05:01 +0000]
[POWERPC] Enable debug info on boot wrapper

Add '-g' to BOOTCFLAGS if CONFIG_DEBUG_INFO is set.

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] Remove unused old code from powermac setup code
Paul Mackerras [Thu, 4 Oct 2007 03:47:06 +0000]
[POWERPC] Remove unused old code from powermac setup code

Since bootdevice never gets initialized, it's always NULL, and hence a
whole pile of code in arch/powerpc/platforms/setup.c never gets used.
(This was the code that originally was there so that the automatic
root partition selection mechanism would prefer a rootish-looking
partition on the device that OF loaded the kernel from over a similar
partition on other devices.)

This removes the unused code.

Signed-off-by: Paul Mackerras <paulus@samba.org>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

12 years ago[POWERPC] Remove some more section mismatch warnings
Stephen Rothwell [Thu, 4 Oct 2007 02:00:28 +0000]
[POWERPC] Remove some more section mismatch warnings

WARNING: vmlinux.o(.text+0x2ff5c): Section mismatch: reference to .init.text:.pmac_find_ide_boot (between '.note_bootable_part' and '.note_scsi_host')

>From holly_defconfig:
WARNING: vmlinux.o(.text+0x164fe): Section mismatch: reference to .init.data:boot_command_line (between 'note_bootable_part' and 'find_via_pmu')
WARNING: vmlinux.o(.text+0x16506): Section mismatch: reference to .init.data:boot_command_line (between 'note_bootable_part' and 'find_via_pmu')

>From linkstation_defconfig:
WARNING: vmlinux.o(.text+0x158fe): Section mismatch: reference to .init.data:boot_command_line (between 'note_bootable_part' and 'find_via_pmu')
WARNING: vmlinux.o(.text+0x15906): Section mismatch: reference to .init.data:boot_command_line (between 'note_bootable_part' and 'find_via_pmu')

>From mpc7448_hpc2_defconfig:
WARNING: vmlinux.o(.text+0x1583e): Section mismatch: reference to .init.data:boot_command_line (between 'note_bootable_part' and 'find_via_pmu')
WARNING: vmlinux.o(.text+0x15846): Section mismatch: reference to .init.data:boot_command_line (between 'note_bootable_part' and 'find_via_pmu')

>From pmac32_defconfig:
WARNING: vmlinux.o(.text+0x154ca): Section mismatch: reference to .init.data:boot_command_line (between 'note_bootable_part' and 'note_scsi_host')
WARNING: vmlinux.o(.text+0x154d2): Section mismatch: reference to .init.data:boot_command_line (between 'note_bootable_part' and 'note_scsi_host')
WARNING: vmlinux.o(.text+0x1553c): Section mismatch: reference to .init.text:pmac_find_ide_boot (between 'note_bootable_part' and 'note_scsi_host')

>From ppc64_defconfig:
WARNING: vmlinux.o(.text+0x3acdc): Section mismatch: reference to .init.text:.pmac_find_ide_boot (between '.note_bootable_part' and '.note_scsi_host')

>From prpmc2800_defconfig:
WARNING: vmlinux.o(.text+0x1611e): Section mismatch: reference to .init.data:boot_command_line (between 'note_bootable_part' and 'find_via_pmu')
WARNING: vmlinux.o(.text+0x16126): Section mismatch: reference to .init.data:boot_command_line (between 'note_bootable_part' and 'find_via_pmu')

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] Use cache-inhibited large page bit from firmware
Paul Mackerras [Wed, 3 Oct 2007 04:41:15 +0000]
[POWERPC] Use cache-inhibited large page bit from firmware

Discussions with firmware architects have confirmed that the bit in
the ibm,pa-features property that indicates support for
cache-inhibited large (>= 64kB) page mappings does in fact mean that
the hypervisor allows 64kB mappings to I/O devices.

Thus we can now enable the code that tests that bit and sets our
CPU_FTR_CI_LARGE_PAGE feature bit.

Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] Prepare to remove of_platform_driver name
Stephen Rothwell [Fri, 21 Sep 2007 08:08:17 +0000]
[POWERPC] Prepare to remove of_platform_driver name

The name field of of_platform_driver is just copied into the included
device_driver.  By not overriding an already initialised device_driver
name, we can convert the drivers over time to stop using the
of_platform_driver name.

Also we were not copying the owner field from of_platform_driver, so do
the same with it.

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] iSeries: Correct missing newline in printk
Stephen Rothwell [Fri, 21 Sep 2007 04:36:47 +0000]
[POWERPC] iSeries: Correct missing newline in printk

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Paul Mackerras <paulus@samba.org>

12 years ago[POWERPC] Lite5200: Use comma delimiter format for lists in device tree
Grant Likely [Mon, 8 Oct 2007 07:24:22 +0000]
[POWERPC] Lite5200: Use comma delimiter format for lists in device tree

DTC now supports "foo","bar" format for lists of strings; use the new
format on the lite5200 device trees.

Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] spi: Support non-QE processors
Peter Korsgaard [Sat, 6 Oct 2007 20:06:40 +0000]
[POWERPC] spi: Support non-QE processors

On non-QE processors (mpc831x/mpc834x) the SPI clock is the SoC clock.

Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] 85xx: mpc85xx_mds - reset UCC ethernet properly
Anton Vorontsov [Fri, 5 Oct 2007 17:46:47 +0000]
[POWERPC] 85xx: mpc85xx_mds - reset UCC ethernet properly

Apart from that the current code doesn't compile it's also
meaningless with regard to the MPC8568E-MDS' BCSR.

This patch used to reset UCCs properly.

Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] 85xx: mpc8568mds - update dts to be able to use UCCs
Anton Vorontsov [Fri, 5 Oct 2007 17:46:53 +0000]
[POWERPC] 85xx: mpc8568mds - update dts to be able to use UCCs

1. UCC1's RX_DV pin is 16, not 15;
2. UCC1's phy is at 0x7, not 0x1. Schematics says 0x7, and recent
   u-boot also using 0x7.
3. Use gianfar's (eTSEC) mdio bus. This is hardware default setup.
4. tx-clock should be CLK16 (GE125, PB31);
5. phy-connection-type is RGMII-ID;

Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] QE: pario - support for MPC85xx layout
Anton Vorontsov [Fri, 5 Oct 2007 17:47:09 +0000]
[POWERPC] QE: pario - support for MPC85xx layout

8 bytes padding required to match MPC85xx registers layout.

Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Reviewed-by: Kim Phillips <kim.phillips@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] QEIC: Implement pluggable handlers, fix MPIC cascading
Anton Vorontsov [Fri, 5 Oct 2007 17:47:29 +0000]
[POWERPC] QEIC: Implement pluggable handlers, fix MPIC cascading

set_irq_chained_handler overwrites MPIC's handle_irq function
(handle_fasteoi_irq) thus MPIC never gets eoi event from the
cascaded IRQ. This situation hangs MPIC on MPC8568E.

To solve this problem efficiently, QEIC needs pluggable handlers,
specific to the underlaying interrupt controller.

Patch extends qe_ic_init() function to accept low and high interrupt
handlers. To avoid #ifdefs, stack of interrupt handlers specified in
the header file and functions are marked 'static inline', thus
handlers are compiled-in only if actually used (in the board file).
Another option would be to lookup for parent controller and
automatically detect handlers (will waste text size because of
never used handlers, so this option abolished).

qe_ic_init() also changed in regard to support multiplexed high/low
lines as found in MPC8568E-MDS, plus qe_ic_cascade_muxed_mpic()
handler implemented appropriately.

Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] mpc85xx_mds: select QUICC_ENGINE
Anton Vorontsov [Fri, 5 Oct 2007 17:47:38 +0000]
[POWERPC] mpc85xx_mds: select QUICC_ENGINE

Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] 86xx: update immap_86xx.h for the 8610
Timur Tabi [Tue, 2 Oct 2007 21:27:13 +0000]
[POWERPC] 86xx: update immap_86xx.h for the 8610

Update the definition of the global utilities structure (ccsr_guts) in
immap_86xx.h and add some related macros for the Freescale 8610 SOC.

Signed-off-by: Timur Tabi <timur@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] 85xx/86xx: refactor RSTCR reset code
Kumar Gala [Thu, 4 Oct 2007 06:04:57 +0000]
[POWERPC] 85xx/86xx: refactor RSTCR reset code

On the majority of 85xx & 86xx we have a register that's ability to
assert HRESET_REQ to reset the board.  We refactored that code so it
can be shared between both platforms into fsl_soc.c and removed all
the duplication in each platform directory.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] Use for_each_ matching routinues for pci PHBs
Kumar Gala [Thu, 4 Oct 2007 05:28:43 +0000]
[POWERPC] Use for_each_ matching routinues for pci PHBs

On the Freescale embedded (83xx, 85xx, 86xx) and a few of the discrete
bridges (mpc10x, tsi108) use the new for_each_compatible_node() or
for_each_node_by_type() to provide more exact matching when looking for
PHBs in the device tree.

With the previous code it was possible to match on pci bridges since
we were only matching on device_type.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] FSL: Access PCIe LTSSM register with correct size
Kumar Gala [Thu, 4 Oct 2007 04:37:33 +0000]
[POWERPC] FSL: Access PCIe LTSSM register with correct size

The LTSSM register is actual 32-bits wide so we should be doing a
dword access.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] 85xx: Failure with odd memory sizes and CONFIG_HIGHMEM
Dale Farnsworth [Wed, 3 Oct 2007 19:01:40 +0000]
[POWERPC] 85xx: Failure with odd memory sizes and CONFIG_HIGHMEM

The CONFIG_FSL_BOOKE mmu setup code fails when CONFIG_HIGHMEM=y
and the 3 fixed TLB entries cannot exactly map the lowmem size.
Each TLB entry can map 4MB, 16MB, 64MB or 256MB, so the failure
is observed when the kernel lowmem size is not equal to the
sum of up to 3 of those values.

Normally, memory is sized in nice numbers, but I observed this
problem while testing a crash dump kernel.  The failure can
also be observed by artificially reducing the kernel's main
memory via the mem= kernel command line parameter.

This commit fixes the problem by setting __initial_memory_limit
in adjust_total_lowmem().

Signed-off-by: Dale Farnsworth <dale@farnsworth.org>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] Add initial MPC8610 HPCD Platform files.
Xianghua Xiao [Wed, 3 Oct 2007 20:09:33 +0000]
[POWERPC] Add initial MPC8610 HPCD Platform files.

Add basic board support for the MPC8610 HPCD.  This does
not include any support the SoC Display or Audio controllers.

Signed-off-by: Xianghua Xiao <x.xiao@freescale.com>
Signed-off-by: Jason Jin <Jason.jin@freescale.com>
Signed-off-by: Jon Loelier <jdl@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] Treat 8610 PCIe host bridge as transparent
Jason Jin [Wed, 3 Oct 2007 20:09:50 +0000]
[POWERPC] Treat 8610 PCIe host bridge as transparent

Signed-off-by: Jason Jin <Jason.jin@freescale.com>
Signed-off-by: Jon Loeliger <jdl@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] Add initial MPC8610 HPCD Device Tree Source file.
Xianghua Xiao [Wed, 3 Oct 2007 20:09:15 +0000]
[POWERPC] Add initial MPC8610 HPCD Device Tree Source file.

Signed-off-by: Xianghua Xiao <x.xiao@freescale.com>
Signed-off-by: Jon Loeliger <jdl@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] spi: mode should be "cpu-qe" instead of "qe"
Peter Korsgaard [Wed, 3 Oct 2007 16:29:09 +0000]
[POWERPC] spi: mode should be "cpu-qe" instead of "qe"

Mode should be "cpu-qe" for QE in CPU mode. "qe" should be reserved
for native QE mode.

Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] spi: Use fsl_spi instead of mpc83xx_spi
Peter Korsgaard [Wed, 3 Oct 2007 15:44:58 +0000]
[POWERPC] spi: Use fsl_spi instead of mpc83xx_spi

According to booting-without-of.txt, compatible should be "fsl_spi" and
mode "cpu" or "qe" for the fsl SPI controllers.

Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] qe: miscellaneous code improvements and fixes to the QE library
Timur Tabi [Wed, 3 Oct 2007 16:34:59 +0000]
[POWERPC] qe: miscellaneous code improvements and fixes to the QE library

This patch makes numerous miscellaneous code improvements to the QE library.

1. Remove struct ucc_common and merge ucc_init_guemr() into ucc_set_type()
   (every caller of ucc_init_guemr() also calls ucc_set_type()).  Modify all
   callers of ucc_set_type() accordingly.

2. Remove the unused enum ucc_pram_initial_offset.

3. Refactor qe_setbrg(), also implement work-around for errata QE_General4.

4. Several printk() calls were missing the terminating \n.

5. Add __iomem where needed, and change u16 to __be16 and u32 to __be32 where
   appropriate.

6. In ucc_slow_init() the RBASE and TBASE registers in the PRAM were programmed
   with the wrong value.

7. Add the protocol type to struct us_info and updated ucc_slow_init() to
   use it, instead of always programming QE_CR_PROTOCOL_UNSPECIFIED.

8. Rename ucc_slow_restart_x() to ucc_slow_restart_tx()

9. Add several macros in qe.h (mostly for slow UCC support, but also to
   standardize some naming convention) and remove several unused macros.

10. Update ucc_geth.c to use the new macros.

11. Add ucc_slow_info.protocol to specify which QE_CR_PROTOCOL_xxx protcol
    to use when initializing the UCC in ucc_slow_init().

12. Rename ucc_slow_pram.rfcr to rbmr and ucc_slow_pram.tfcr to tbmr, since
    these are the real names of the registers.

13. Use the setbits, clrbits, and clrsetbits where appropriate.

14. Refactor ucc_set_qe_mux_rxtx().

15. Remove all instances of 'volatile'.

16. Simplify get_cmxucr_reg();

17. Replace qe_mux.cmxucrX with qe_mux.cmxucr[].

18. Updated struct ucc_geth because struct ucc_fast is not padded any more.

Signed-off-by: Timur Tabi <timur@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] Update .gitignore for new vdso generated files
Kumar Gala [Wed, 3 Oct 2007 15:43:10 +0000]
[POWERPC] Update .gitignore for new vdso generated files

We now generate vdso[32,64].so.dbg as part of the build so
add them to .gitignore

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] Fixup MPC8568 dts
Kumar Gala [Tue, 2 Oct 2007 14:51:32 +0000]
[POWERPC] Fixup MPC8568 dts

The PCI nodes on the MPC8568 dts didn't get moved up to be sibilings of the
SOC node when we did that clean up for some reason.  Fix that up and some
minor whitespace and adjusting the size of the soc reg property.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] mpc8568mds.dts: fix PCIe I/O address space location and size
Anton Vorontsov [Tue, 2 Oct 2007 13:48:07 +0000]
[POWERPC] mpc8568mds.dts: fix PCIe I/O address space location and size

According to u-boot/board/mpc8568mds/init.S:

LAW(Local Access Window) configuration:
2)   0xa000_0000   0xbfff_ffff     PCIe MEM                512MB
4)   0xe280_0000   0xe2ff_ffff     PCIe I/O                8M

Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] MPC8568E-MDS: add support for ds1374 rtc
Anton Vorontsov [Tue, 2 Oct 2007 13:47:43 +0000]
[POWERPC] MPC8568E-MDS: add support for ds1374 rtc

MPC8568E-MDS have DS1374 chip on the I2C bus, thus let's use it.
This patch also adds #address-cells and #size-cells to the I2C
controllers nodes.

p.s. DS1374 rtc class driver is in the -mm tree, its name is
rtc-rtc-class-driver-for-the-ds1374.patch.

Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] fsl_soc: fix uninitialized i2c_board_info structure
Anton Vorontsov [Tue, 2 Oct 2007 13:47:40 +0000]
[POWERPC] fsl_soc: fix uninitialized i2c_board_info structure

i2c_board_info used semi-initialized, causing garbage in the
info->flags, and that, in turn, causes various symptoms of i2c
malfunctioning, like PEC mismatches.

Signed-off-by: Anton Vorontsov <avorontsov@ru.mvista.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] cpm: Describe multi-user ram in its own device node.
Scott Wood [Fri, 28 Sep 2007 19:06:16 +0000]
[POWERPC] cpm: Describe multi-user ram in its own device node.

The way the current CPM binding describes available multi-user (a.k.a.
dual-ported) RAM doesn't work well when there are multiple free regions,
and it doesn't work at all if the region doesn't begin at the start of
the muram area (as the hardware needs to be programmed with offsets into
this area).  The latter situation can happen with SMC UARTs on CPM2, as its
parameter RAM is relocatable, u-boot puts it at zero, and the kernel doesn't
support moving it.

It is now described with a muram node, similar to QE.  The current CPM
binding is sufficiently recent (i.e. never appeared in an official release)
that compatibility with existing device trees is not an issue.

The code supporting the new binding is shared between cpm1 and cpm2, rather
than remain separated.  QE should be able to use this code as well, once
minor fixes are made to its device trees.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] Cleaned up whitespace in head_fsl_booke.S
Kumar Gala [Thu, 27 Sep 2007 13:43:35 +0000]
[POWERPC] Cleaned up whitespace in head_fsl_booke.S

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] QE: Added missing CEURNR register
Emil Medve [Wed, 26 Sep 2007 17:03:40 +0000]
[POWERPC] QE: Added missing CEURNR register

According to the publicly available MPC8360E RM (rev. 1 from 09/2006 and rev. 2
from 05/2007) and MPC8323E RM (rev. 1 from 09/2006), CEURNR is the QE microcode
revision number register and is located at offset 0x1b8 within the QE internal
register space

Signed-off-by: Emil Medve <Emilian.Medve@Freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] fsl_soc: rtc-ds1307 support
Peter Korsgaard [Thu, 20 Sep 2007 10:42:13 +0000]
[POWERPC] fsl_soc: rtc-ds1307 support

Add support for the I2C devices handled by the rtc-ds1307 driver to
of_register_i2c_devices.

Cc: G. Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] fsl_soc: Fix trivial printk typo.
Peter Korsgaard [Thu, 20 Sep 2007 10:42:12 +0000]
[POWERPC] fsl_soc: Fix trivial printk typo.

Fix a trivial printk typo in fsl_soc.

Cc: G. Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] 8xx: Move softemu8xx.c from arch/ppc
Scott Wood [Tue, 18 Sep 2007 20:29:35 +0000]
[POWERPC] 8xx: Move softemu8xx.c from arch/ppc

Previously, Soft_emulate_8xx was called with no implementation, resulting in
build failures whenever building 8xx without math emulation.  The
implementation is copied from arch/ppc to resolve this issue.

However, this sort of minimal emulation is not a very good idea other than
for compatibility with existing userspaces, as it's less efficient than
soft-float and can mislead users into believing they have soft-float.  Thus,
it is made a configurable option, off by default.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] bootwrapper: adds cuboot for MPC7448HPC2 platform
Roy Zang [Mon, 24 Sep 2007 10:31:55 +0000]
[POWERPC] bootwrapper: adds cuboot for MPC7448HPC2 platform

This patch adds cuboot support for MPC7448HPC2 platform.
The cuImage can be used with legacy u-boot without FDT support.

Signed-off-by: Roy Zang <tie-fei.zang@freescale.com>
Acked-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] mpc82xx: Add pq2fads board support.
Scott Wood [Wed, 5 Sep 2007 19:00:54 +0000]
[POWERPC] mpc82xx: Add pq2fads board support.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] mpc82xx: Update mpc8272ads, and factor out PCI and reset.
Scott Wood [Fri, 14 Sep 2007 20:41:56 +0000]
[POWERPC] mpc82xx: Update mpc8272ads, and factor out PCI and reset.

1. PCI and reset are factored out into pq2.c.  I renamed them from m82xx
to pq2 because they won't work on the Integrated Host Processor line of
82xx chips (i.e. 8240, 8245, and such).

2. The PCI PIC, which is nominally board-specific, is used on multiple
boards, and thus is used into pq2ads-pci-pic.c.

3. The new CPM binding is used.

4. General cleanup.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] 8xx/wrapper: Embedded Planet EP88xC support
Scott Wood [Fri, 14 Sep 2007 19:58:25 +0000]
[POWERPC] 8xx/wrapper: Embedded Planet EP88xC support

This board is also resold by Freescale under the names
"QUICCStart MPC885 Evaluation System" and "CWH-PPC-885XN-VE".

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] 8xx: mpc885ads cleanup
Scott Wood [Fri, 14 Sep 2007 19:38:16 +0000]
[POWERPC] 8xx: mpc885ads cleanup

It now uses the new CPM binding and the generic pin/clock functions, and
has assorted fixes and cleanup.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] Document local bus nodes in the device tree, and update cuboot-pq2.
Scott Wood [Fri, 14 Sep 2007 18:24:02 +0000]
[POWERPC] Document local bus nodes in the device tree, and update cuboot-pq2.

The localbus node is used to describe devices that are connected via a chip
select or similar mechanism.  The advantages over placing the devices under
the root node are that it can be probed without probing other random things
under the root, and that the description of which chip select a given device
uses can be used to set up mappings if the firmware failed to do so in a
useful manner.

cuboot-pq2 is updated to match the binding; previously, it called itself
chipselect rather than localbus, and used phandle linkage between the
actual bus node and the control node (the current agreement is to simply use
the fully-qualified address of the control registers, and ignore the overlap
with the IMMR node).

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] mpc8272ads: Change references from 82xx_ADS to 8272_ADS.
Scott Wood [Thu, 26 Jul 2007 18:52:28 +0000]
[POWERPC] mpc8272ads: Change references from 82xx_ADS to 8272_ADS.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] mpc82xx: Rename mpc82xx_ads to mpc8272_ads.
Scott Wood [Thu, 26 Jul 2007 18:51:42 +0000]
[POWERPC] mpc82xx: Rename mpc82xx_ads to mpc8272_ads.

This is just a rename patch; internal references to mpc82xx_ads will be
changed in the next one.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] mpc82xx: Remove a bunch of cruft that duplicates generic functionality.
Scott Wood [Mon, 27 Aug 2007 21:56:43 +0000]
[POWERPC] mpc82xx: Remove a bunch of cruft that duplicates generic functionality.

m82xx_calibrate_decr(), mpc82xx_ads_show_cpuinfo(), and mpc82xx_halt() do
anything useful beyond what the generic code does.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] mpc82xx: Define CPU_FTR_NEED_COHERENT
Scott Wood [Fri, 14 Sep 2007 20:32:14 +0000]
[POWERPC] mpc82xx: Define CPU_FTR_NEED_COHERENT

The 8272 (and presumably other PCI PQ2 chips) appear to have the
same issue as the 83xx regarding PCI streaming DMA.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] cpm2: Add cpm2_set_pin().
Scott Wood [Mon, 16 Jul 2007 18:32:24 +0000]
[POWERPC] cpm2: Add cpm2_set_pin().

This provides a generic way for board code to set up CPM pins, rather
than directly poking magic values into registers.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] cpm2: Add SCCs to cpm2_clk_setup(), and cpm2_smc_clk_setup().
Scott Wood [Mon, 16 Jul 2007 18:26:35 +0000]
[POWERPC] cpm2: Add SCCs to cpm2_clk_setup(), and cpm2_smc_clk_setup().

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] cpm2: Infrastructure code cleanup.
Scott Wood [Fri, 14 Sep 2007 20:30:44 +0000]
[POWERPC] cpm2: Infrastructure code cleanup.

Mostly sparse fixes (__iomem annotations, etc); also, cpm2_immr
is used rather than creating many temporary mappings.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] 8xx: Set initial memory limit.
John Traill [Tue, 17 Jul 2007 01:17:23 +0000]
[POWERPC] 8xx: Set initial memory limit.

The 8xx can only support a max of 8M during early boot (it seems a lot of
8xx boards only have 8M so the bug was never triggered), but the early
allocator isn't aware of this.  The following change makes it able to run
with larger memory.

Signed-off-by: John Traill <john.traill@freescale.com>
Signed-off-by: Vitaly Bordug <vitb@kernel.crashing.org>
Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] 8xx: Work around CPU15 erratum.
Scott Wood [Mon, 25 Jun 2007 19:50:41 +0000]
[POWERPC] 8xx: Work around CPU15 erratum.

The CPU15 erratum on MPC8xx chips can cause incorrect code execution
under certain circumstances, where there is a conditional or indirect
branch in the last word of a page, with a target in the last cache line
of the next page.  This patch implements one of the suggested
workarounds, by forcing a TLB miss whenever execution crosses a page
boundary.  This is done by invalidating the pages before and after the
one being loaded into the TLB in the ITLB miss handler.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] 8xx: Add pin and clock setting functions.
Scott Wood [Mon, 16 Jul 2007 22:22:01 +0000]
[POWERPC] 8xx: Add pin and clock setting functions.

These let board code set up pins and clocks without having to
put magic numbers directly into the registers.

The clock function is mostly duplicated from the cpm2 version;
hopefully this stuff can be merged at some point.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] 8xx: Infrastructure code cleanup.
Scott Wood [Fri, 14 Sep 2007 19:22:36 +0000]
[POWERPC] 8xx: Infrastructure code cleanup.

1. Keep a global mpc8xx_immr mapping, rather than constantly
creating temporary mappings.
2. Look for new fsl,cpm1 and fsl,cpm1-pic names.
3. Always reset the CPM when not using the udbg console;
this is required in case the firmware initialized a device
that is incompatible with one that the kernel is about to
use.
4. Remove some superfluous casts and header includes.
5. Change a usage of IMAP_ADDR to get_immrbase().
6. Use phys_addr_t, not uint, for dpram_pbase.
7. Various sparse-related fixes, such as __iomem annotations.
8. Remove mpc8xx_show_cpuinfo, which doesn't provide anything
useful beyond the generic cpuinfo handler.
9. Move prototypes for 8xx support functions from board files
to sysdev/commproc.h.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] 8xx: Fix CONFIG_PIN_TLB.
Scott Wood [Mon, 16 Jul 2007 16:28:18 +0000]
[POWERPC] 8xx: Fix CONFIG_PIN_TLB.

1. Move CONSISTENT_START on 8xx so that it doesn't overlap the IMMR mapping.
2. The wrong register was being loaded into SPRN_MD_RPN.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] cpm_uart: Issue STOP_TX command before initializing console.
Scott Wood [Tue, 17 Jul 2007 23:09:33 +0000]
[POWERPC] cpm_uart: Issue STOP_TX command before initializing console.

This prevents some bootloader/bootwrapper characters from being lost.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] cpm_uart: sparse fixes
Scott Wood [Tue, 24 Jul 2007 20:53:07 +0000]
[POWERPC] cpm_uart: sparse fixes

Mostly a bunch of direct access to in/out conversions, plus a few
cast removals, __iomem annotations, and miscellaneous cleanup.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] cpm_uart: Be an of_platform device when CONFIG_PPC_CPM_NEW_BINDING is set.
Scott Wood [Tue, 17 Jul 2007 22:59:06 +0000]
[POWERPC] cpm_uart: Be an of_platform device when CONFIG_PPC_CPM_NEW_BINDING is set.

The existing OF glue code was crufty and broken.  Rather than fix it,
it has been removed, and the serial driver now talks to the device tree
directly.

The non-CONFIG_PPC_CPM_NEW_BINDING code can go away once CPM platforms
are dropped from arch/ppc (which will hopefully be soon), and existing
arch/powerpc boards that I wasn't able to test on for this patchset get
converted (which should be even sooner).

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] bootwrapper: Use fsl_get_immr() in cuboot-pq2.c.
Scott Wood [Fri, 31 Aug 2007 22:18:28 +0000]
[POWERPC] bootwrapper: Use fsl_get_immr() in cuboot-pq2.c.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] bootwrapper: Add fsl_get_immr() and 8xx/pq2 clock functions.
Scott Wood [Mon, 27 Aug 2007 18:46:38 +0000]
[POWERPC] bootwrapper: Add fsl_get_immr() and 8xx/pq2 clock functions.

fsl_get_immr() uses /soc/ranges to determine the immr.

mpc885_get_clock() transforms a crystal frequency into a system frequency
according to the PLL register settings.

pq2_get_clocks() does the same as the above for the PowerQUICC II,
except that it produces several different clocks.

The mpc8xx/pq2 set_clocks() functions modify common properties in
the device tree based on the given clock data.

The mpc885/pq2 fixup_clocks() functions call get_clocks(), and
pass the results to set_clocks().

Signed-off-by: Scott Wood <scottwood@freescale.com>
Acked-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] bootwrapper: Support all-in-one PCI nodes in cuboot-pq2.
Scott Wood [Thu, 30 Aug 2007 17:06:21 +0000]
[POWERPC] bootwrapper: Support all-in-one PCI nodes in cuboot-pq2.

Consensus was reached to put PCI nodes at the root of the tree (and not
under /soc), but the phandle to a control node was rejected in favor of
simply not worrying about /pci/reg overlapping /soc/ranges.

This updates cuboot-82xx to not look for the phandle.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Acked-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>

12 years ago[POWERPC] Add early debug console for CPM serial ports.
Scott Wood [Mon, 16 Jul 2007 16:43:43 +0000]
[POWERPC] Add early debug console for CPM serial ports.

This code assumes that the ports have been previously set up, with
buffers in DPRAM.

Signed-off-by: Scott Wood <scottwood@freescale.com>
Acked-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>