Adds the core platform support for the PS3 game console and other devices
using the PS3 hypervisor.
Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
This fixes the xmon support for the cell spu to be compatable with the split
spu platform code.
Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
This adds a platform specific spu management abstraction and the coresponding
routines to support the IBM Cell Blade. It also removes the hypervisor only
resources that were included in struct spu.
Three new platform specific routines are introduced, spu_enumerate_spus(),
spu_create_spu() and spu_destroy_spu(). The underlying design uses a new
type, struct spu_management_ops, to hold function pointers that the platform
setup code is expected to initialize to instances appropriate to that platform.
For the IBM Cell Blade support, I put the hypervisor only resources that were
in struct spu into a platform specific data structure struct spu_pdata.
Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
This adds an accessor routine virq_to_hw() to the
virq routines which hides the implementation details
of the virq to hwirq map.
Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
This includes:
* version 1.24 of ppc-dis.c
* version 1.88 of ppc-opc.c
* version 1.23 of ppc.h
I can't vouch for the accuracy etc. of these changes, but it brings
us into line with binutils - and from a cursory test appears to work
fine.
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
It saves #ifdef'ing in callers if we at least define the 64-bit cpu
features for 32-bit also.
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
While adding spu disassembly support it struck me that we're actually
carrying quite a lot of code around, just to do disassembly in the case
of a crash.
While on large systems it's not an issue, on smaller ones it might be
nice to have xmon - but without the weight of the disassembly support.
For a Cell build this saves ~230KB (!), and for pSeries ~195KB.
We still support the 'di' and 'sdi' commands, however they just dump
the instruction in hex.
Move the definitions into a header to clean xmon.c just a tiny bit.
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
This patch adds a "sdi" command to xmon, to disassemble the contents
of an spu's local store.
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
This patch imports and munges the spu disassembly code from binutils.
All files originated from version 1.1 in binutils cvs.
* spu.h, spu-insns.h and spu-opc.c are unchanged except for pathnames.
* spu-dis.c has been edited heavily:
* use printf instead of info->fprintf_func and similar.
* pass the instruction in rather than reading it.
* we have no equivalent to symbol_at_address_func, so we just assume
there is never a symbol at the address given.
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
In order to do disassembly of spu binaries in xmon, we need to abstract
the disassembly function from ppc_inst_dump.
We do this by making the actual disassembly function a function pointer
that we pass to ppc_inst_dump(). To save updating all the callers, we
turn ppc_inst_dump() into generic_inst_dump() and make ppc_inst_dump()
a wrapper which always uses print_insn_powerpc().
Currently we pass the dialect into print_insn_powerpc(), but we always
pass 0 - so just make it a local.
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Add a command to xmon to dump the memory of a spu's local store.
This mimics the 'd' command which dumps regular memory, but does
a little hand holding by taking the user supplied address and
finding that offset in the local store for the specified spu.
This makes it easy for example to look at what was executing on a spu:
1:mon> ss
...
Stopped spu 04 (was running)
...
1:mon> sf 4
Dumping spu fields at address c0000000019e0a00:
...
problem->spu_npc_RW = 0x228
...
1:mon> sd 4 0x228
d000080080318228 01a00c021cffc408 4020007f217ff488 |........@ ..!...|
Aha, 01a00c02, which is of course rdch $2,$ch24 !
--
Updated to only do the setjmp goo around the spu access, and not
around prdump because it does its own (via mread).
Also the num variable is now common between sf and sd, so you don't
have to keep typing the spu number in if you're repeating commands
on the same spu.
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
After stopping spus in xmon I often find myself trawling through the
field dumps to find out which spus were running. The spu stopping
code actually knows what's running, so let's print it out to save
the user some futzing.
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
My patch to add spu helpers to xmon (a898497088)
introduced a few sparse warnings, because I was dereferencing an __iomem
pointer.
I think the best way to handle it is to actually use the appropriate in_beXX
functions. Need to rejigger the DUMP macro a little to accomodate that.
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
With soft-disabled interrupts in power_save, we can
still get external exceptions on Cell, even if we are
in pause(0) a.k.a. sleep state.
When the CPU really wakes up through the 0x100 (system reset)
vector, while we have already started processing the 0x500
(external) exception, we get a panic in unrecoverable_exception()
because of the lost state.
This occurred in Systemsim for Cell, but as far as I can see,
it can theoretically occur on any machine that uses the
system reset exception to get out of sleep state.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
This patch adds SPU elf notes to the coredump. It creates a separate note
for each of /regs, /fpcr, /lslr, /decr, /decr_status, /mem, /signal1,
/signal1_type, /signal2, /signal2_type, /event_mask, /event_status,
/mbox_info, /ibox_info, /wbox_info, /dma_info, /proxydma_info, /object-id.
A new macro, ARCH_HAVE_EXTRA_NOTES, was created for architectures to
specify they have extra elf core notes.
A new macro, ELF_CORE_EXTRA_NOTES_SIZE, was created so the size of the
additional notes could be calculated and added to the notes phdr entry.
A new macro, ELF_CORE_WRITE_EXTRA_NOTES, was created so the new notes
would be written after the existing notes.
The SPU coredump code resides in spufs. Stub functions are provided in the
kernel which are hooked into the spufs code which does the actual work via
register_arch_coredump_calls().
A new set of __spufs_<file>_read/get() functions was provided to allow the
coredump code to read from the spufs files without having to lock the
SPU context for each file read from.
Cc: <linux-arch@vger.kernel.org>
Signed-off-by: Dwayne Grant McConnell <decimal@us.ibm.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Devices with no "reg" nor "dcr-reg" property are given a bus_id which
is the node name alone. This means that if more than one such device
with the same names are present in the system, sysfs will have
collisions when creating the symlinks and will fail registering the
devices.
This works around that problem by assigning successive numbers to such
devices.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
This adds code to look at the properties firmware puts in the device
tree to determine what compatibility mode the partition is in on
POWER6 machines, and set the ELF aux vector AT_HWCAP and AT_PLATFORM
entries appropriately.
Specifically, we look at the cpu-version property in the cpu node(s).
If that contains a "logical" PVR value (of the form 0x0f00000x), we
call identify_cpu again with this PVR value. A value of 0x0f000001
indicates the partition is in POWER5+ compatibility mode, and a value
of 0x0f000002 indicates "POWER6 architected" mode, with various
extensions disabled. We also look for various other properties:
ibm,dfp, ibm,purr and ibm,spurr.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Add PPU event-based and cycle-based profiling support to Oprofile for Cell.
Oprofile is expected to collect data on all CPUs simultaneously.
However, there is one set of performance counters per node. There are
two hardware threads or virtual CPUs on each node. Hence, OProfile must
multiplex in time the performance counter collection on the two virtual
CPUs.
The multiplexing of the performance counters is done by a virtual
counter routine. Initially, the counters are configured to collect data
on the even CPUs in the system, one CPU per node. In order to capture
the PC for the virtual CPU when the performance counter interrupt occurs
(the specified number of events between samples has occurred), the even
processors are configured to handle the performance counter interrupts
for their node. The virtual counter routine is called via a kernel
timer after the virtual sample time. The routine stops the counters,
saves the current counts, loads the last counts for the other virtual
CPU on the node, sets interrupts to be handled by the other virtual CPU
and restarts the counters, the virtual timer routine is scheduled to run
again. The virtual sample time is kept relatively small to make sure
sampling occurs on both CPUs on the node with a relatively small
granularity. Whenever the counters overflow, the performance counter
interrupt is called to collect the PC for the CPU where data is being
collected.
The oprofile driver relies on a firmware RTAS call to setup the debug bus
to route the desired signals to the performance counter hardware to be
counted. The RTAS call must set the routing registers appropriately in
each of the islands to pass the signals down the debug bus as well as
routing the signals from a particular island onto the bus. There is a
second firmware RTAS call to reset the debug bus to the non pass thru
state when the counters are not in use.
Signed-off-by: Carl Love <carll@us.ibm.com>
Signed-off-by: Maynard Johnson <mpjohn@us.ibm.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
The following routines are added to arch/powerpc/platforms/cell/pmu.c:
cbe_clear_pm_interrupts()
cbe_enable_pm_interrupts()
cbe_disable_pm_interrupts()
cbe_query_pm_interrupts()
cbe_pm_irq()
cbe_init_pm_irq()
This also adds a routine in arch/powerpc/platforms/cell/interrupt.c and
some macros in cbe_regs.h to manipulate the IIC_IR register:
iic_set_interrupt_routing()
Signed-off-by: Kevin Corry <kevcorry@us.ibm.com>
Signed-off-by: Carl Love <carll@us.ibm.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Move some PMU-related macros and function prototypes from cbe_regs.h
and pmu.h in arch/powerpc/platforms/cell/ to a new header at
include/asm-powerpc/cell-pmu.h
This is cleaner to use from the oprofile code, since that sits in
arch/powerpc/oprofile, not in the cell platform directory.
Signed-off-by: Kevin Corry <kevcorry@us.ibm.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
More macros for manipulating bits in the Cell PMU control registers.
Signed-off-by: Kevin Corry <kevcorry@us.ibm.com>
Signed-off-by: Carl Love <carll@us.ibm.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Add symbol-exports for the new routines in arch/powerpc/platforms/cell/pmu.c.
They are needed for Oprofile, which can be built as a module.
Signed-off-by: Kevin Corry <kevcorry@us.ibm.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
In order to fit with the "don't-run-spus-outside-of-spu_run" model, this
patch starts the isolated-mode loader in spu_run, rather than
spu_create. If spu_run is passed an isolated-mode context that isn't in
isolated mode state, it will run the loader.
This fixes potential races with the isolated SPE app doing a
stop-and-signal before the PPE has called spu_run: bugzilla #29111.
Also (in conjunction with a mambo patch), this addresses #28565, as we
always set the runcntrl register when entering spu_run.
It is up to libspe to ensure that isolated-mode apps are cleaned up
after running to completion - ie, put the app through the "ISOLATE EXIT"
state (see Ch11 of the CBEA).
Signed-off-by: Jeremy Kerr <jk@ozlabs.org>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
This change adds a read accessor for the SPE problem-state run control
register.
This is required for for applying (userspace) changes made to the run
control register while the SPE is stopped - simply asserting the master
run control bit is not sufficient. My next patch for isolated-mode
setup requires this.
Signed-off-by: Jeremy Kerr <jk@ozlabs.org>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
When the user changes the runcontrol register, an SPU might be
running without a process being attached to it and waiting for
events. In order to prevent this, make sure we always disable
the priv1 master control when we're not inside of spu_run.
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
When fixing spufs to map the 'mem' file backing store cacheable,
I incorrectly set the physical mapping to use both cache-inhibited
and guarded mapping, which resulted in a serious performance
degradation.
Debugged-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
When one of the spufs files is mapped into a process address
space, regular users can use ptrace to attempt accessing
them with access_process_vm(). With the way that the
mappings currently work, this likely causes an oops.
Setting the vm_flags to VM_IO makes sure that ptrace can
not access them but returns an error code. This is not
the perfect solution in case of the local store mapping,
but it fixes the oops in a well-defined way.
Also remove leftover VM_RESERVED flags in spufs. The
VM_RESERVED flag is on it's way out and not checked by
the memory managment code anymore.
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Christoph Hellwig <chellwig@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
When there is pending signals, current spufs_run_spu() always returns
-ERESTARTSYS and it is called again automatically.
But, if spe already stopped by stop-and-signal or halt instruction,
returning -ERESTARTSYS makes stop-and-signal/halt lost and
spu run over the end-point.
For your convenience, I attached a sample code to restage this bug.
If there is no bug, printed NPC will be 0x4000.
Signed-off-by: Masato Noguchi <Masato.Noguchi@jp.sony.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
When we attempt an MFC DMA to an unmapped address, the event
returned from spu_run should be SPE_EVENT_SPE_DATA_STORAGE,
not SPE_EVENT_INVALID_DMA.
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Replace the use of the platform specific variable spu.nid with the
platform independednt variable spu.node.
Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
We need to check the channel count of the signal notification registers
before reading them, because it can be undefined when the count is
zero. In order to read count and data atomically, we read from the
saved context.
This patch uses spu_acquire_saved() to force a context save before a
/signal1 or /signal2 read. Because of this it is no longer necessary to
have backing_ops and hw_ops versions of this function so they have been
removed.
Regular applications should not rely on reading this register
to be fast, as it's conceptually a write-only file from the PPE
perspective.
Signed-off-by: Dwayne Grant McConnell <decimal@us.ibm.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
This patch implements read only access to
/mbox_info - SPU Write Outbound Mailbox
/ibox_info - SPU Write Outbound Interrupt Mailbox
/wbox_info - SPU Read Inbound Mailbox
These files are used by gdb in order to look into the current mailbox
queues without changing the contents at the same time. They are
not meant for general programming use, since the access requires
a context save and is therefore rather slow.
It would be good to complement this patch with one that adds
write support as well.
Signed-off-by: Dwayne Grant McConnell <decimal@us.ibm.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
This patch removes the /spu_tag_mask file from spufs. The data provided by
this file is also available from the /dma_info file in the dma_info_mask
of the spu_dma_info struct.
The file was intended to be used by gdb, but that never used it, and
now it has been replaced with the more verbose dma_info file.
Signed-off-by: Dwayne Grant McConnell <decimal@us.ibm.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
The /lslr file gives read access to the SPU_LSLR register in hex; 0x3fff
for example The /dma_info file provides read access to the SPU Command
Queue in a binary format. The /proxydma_info files provides read access
access to the Proxy Command Queue in a binary format. The spu_info.h
file provides data structures for interpreting the binary format of
/dma_info and /proxydma_info.
Signed-off-by: Dwayne Grant McConnell <decimal@us.ibm.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
This patches changes /npc, /decr, /decr_status, /spu_tag_mask,
/event_mask, /event_status, and /srr0 files to provide output according to
the format string "0x%llx" instead of "%llx".
Before this patch some files used "0x%llx" and other used "%llx" which is
inconsistent and potentially confusing. A user might assume "%llx" numbers
were decimal if they happened to not contain any a-f digits. This change
will break any code cannot tolerate a leading 0x in the file contents. The
only known users of these files are the libspe but there might also be
some scripts which access these files. This risk is deemed acceptable for
future consistency.
Signed-off-by: Dwayne Grant McConnell <decimal@us.ibm.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Change the definition of powerpc's cond_syscall() to use the standard gcc
weak attribute specifier which provides proper support for C linkage as
needed by spu_syscall_table[].
Fixes this powerpc build error with CONFIG_SPU_FS=y, CONFIG_PPC_RTAS=n:
arch/powerpc/platforms/built-in.o: undefined reference to `ppc_rtas'
Signed-off-by: Geoff Levand <geoffrey.levand@am.sony.com>
Signed-off-by: Arnd Bergmann <arnd.bergmann@de.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
The SCC parameter RAM areas are mapped wrong in MPC8xx device descriptions. All
memory areas overlap with the next one, so that I2C, SPI, SMC1 and SMC2 cannot
be enabled if the four SCCs are.
Signed-off-by: Kalle Pokki <kalle.pokki@iki.fi>
Acked-by: Kumar Gala <galak@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
arch/powerpc/sysdev/mpc52xx_pic.c breaks the ppc build
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Signed-off-by: Paul Mackerras <paulus@samba.org>
The patch below fixes an arithmetic wrap-around issue on 32bit machines
using smp-tbsync. Without this patch a timebase value over
0x000000007fffffff will hang the boot process while bringing up
secondary CPUs.
Signed-off-by: Adrian Cox <adrian@humboldt.co.uk>
Signed-off-by: Paul Mackerras <paulus@samba.org>
This makes 2 changes to clean up the flat device tree handling
logic in the zImage wrapper.
First, there were two callbacks from the dt_ops structure used for
producing a final flat tree to pass to the kerne: dt_ops.ft_pack()
which packed the flat tree (possibly a no-op) and dt_ops.ft_addr()
which retreived the address of the final blob. Since they were only
ever called together, this patch combines the two into a single new
callback, dt_ops.finalize(). This new callback does whatever
platform-dependent things are necessary to produce a final flat device
tree blob, and returns the blob's addres.
Second, the current logic calls the kernel with a flat device tree if
one is build into the zImage wrapper, otherwise it boots the kernel
with a PROM pointer, expecting the kernel to copy the OF device tree
itself. This approach precludes the possibility of the platform
wrapper code building a flat device tree from whatever
platform-specific information firmware provides. Thus, this patch
takes the more sensible approach of invoking the kernel with a flat
tree if the dt_ops.finalize callback provides one (by whatever means).
So, the dt_ops.finalize callback can be NULL, or can be a function
which returns NULL. In either case, the zImage wrapper logic assumes
that this is a platform with OF and invokes the kernel accordingly.
Signed-off-by: David Gibson <dwg@au1.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
This patch makes the handling of the initrd (or initramfs) in the
zImage wrapper a little easier to follow. Instead of passing the
initrd addresses out from prep_kernel() via the cryptic a1 and a2
parameters, use the global struct add_range, 'initrd'. prep_kernel()
already passes information through the 'vmlinux' addr_range struct, so
this seems like a reasonable extension.
Some comments also clarify the logic with prep_kernel(): we use an
initrd included in the zImage if present, otherwise we use an initrd
passed in by the bootloader in the a1 and a2 parameters (yaboot, at
least, uses this mechanism to pass an initrd).
Signed-off-by: David Gibson <dwg@au1.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
This is a small driver for the Xserve G5 CPU-meter blue LEDs on the
front-panel. It might work on the Xserve G4 as well though that was
not tested. It's pretty basic and could use some improvements if
somebody cares doing them. :)
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Per email discussion, it appears that rtas_stop_self()
and pSeries_mach_cpu_die() should not be compiled if
CONFIG_HOTPLUG_CPU is not defined. This patch adds
#ifdefs around these bits of code.
Signed-off-by: Linas Vepstas <linas@austin.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
The following patch adds a tsi108/9 pci interrupt controller host.
On mpc7448hpc2 board, pci_irq_fixup function is removed, which makes the
pci_read_irq_line be the default pci irq fixup.
Signed-off-by: Roy Zang <tie-fei.zang@freescale.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
The below hunk was missed from the recent patch, and now, there are somewhat
inconsistent definitions:
in cpm_uart.h:
int __init cpm_uart_init_portdesc(void);
in cpm_uart_cpm1.c:
int __init cpm_uart_init_portdesc(void)
{
}
in cpm_uart_cpm2.c:
int cpm_uart_init_portdesc(void)
{
}
Signed-off-by: Kalle Pokki <kalle.pokki@iki.fi>
Signed-off-by: Vitaly Bordug <vbordug@ru.mvista.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
