This adds the low level irq tracing hooks to the powerpc architecture
needed to enable full lockdep functionality.

This is partly based on Johannes Berg's initial version.  I removed
the asm trampoline that isn't needed (thus improving performance) and
modified all sorts of bits and pieces, reworking most of the assembly,
etc...
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This adds stacktrace support for powerpc, which will be needed for
lockdep.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This moves various definitions used all over the place to parse stack
frames to ptrace.h so only one definition is needed.
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Lockdep found out that we can occasionally take the device-tree
lock for reading from softirq time (from rtas_token called
by the rtas real time clock code called by the NTP code),
while we take it occasionally for writing without masking
interrupts. The combination of those two can thus deadlock.

While some of those cases of interrupt read lock could be fixed
(such as caching the RTAS tokens) I figured that taking the
lock for writing is so rare (device-tree modification) that we
may as well penalize that case and allow reading from interrupts.

Thus, this turns all the writers to take the lock with irqs
masked to avoid the situation.
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

global_dbcr0 needs to be a per cpu set of save areas instead of a single
global on all processors.

Also, we switch to using DBCR0_IDM to determine if the user space app is
being debugged as its a more consistent way.  In the future we should
support features like hardware breakpoint and watchpoints which will
have DBCR0_IDM set but not necessarily DBCR0_IC (single step).
Signed-off-by: NKumar Gala <galak@kernel.crashing.org>

The architecture allows for "Book-E" style debug interrupts to either go
to critial interrupts of their own debug interrupt level.  To allow for
a dynamic kernel to support machines of either type we want to be able to
compile in the interrupt handling code for both exception levels.

Towards this goal we renamed the debug handling macros to specify the
interrupt level in their name (DEBUG_CRIT_EXCEPTION/DebugCrit and
DEBUG_DEBUG_EXCEPTION/DebugDebug).

Additionally, on the Freescale Book-e parts we expanded the exception
stacks to cover the maximum case of needing three exception stacks (normal,
machine check and debug).

There is some kernel text space optimization to be gained if a kernel is
configured for a specific Freescale implementation but we aren't handling
that now to allow for the single kernel image support.
Signed-off-by: NKumar Gala <galak@kernel.crashing.org>

This changes the way we calculate how much space to reserve for the
pHyp dump.  Currently we reserve 256MB only.  With this change, the
code first checks to see if an amount has been specified on the boot
command line with the "phyp_dump_reserve_size" option, and if so, uses
that much.

Otherwise it computes 5% of total ram and rounds it down to a multiple
of 256MB, and uses the larger of that or 256MB.

This is for large systems with a lot of memory (10GB or more).  The
aim is to have more space available for the kernel on reboot on
machines with more resources.  Although the dump will be collected
pretty fast and the memory released really early on allowing the
machine to have the full memory available, this alleviates any issues
that can be caused by having way too little memory on very very large
systems during those few minutes.
Signed-off-by: NManish Ahuja <mahuja@us.ibm.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

We can set LOAD_OFFSET and use the AT attribute on sections and the
linker will properly set the physical address of the LOAD program
header for us.

This allows us to know how the PHYSICAL_START the user configured a
kernel with by just looking at the resulting vmlinux ELF.

This is pretty much stolen from how x86 does things in their linker
scripts.
Signed-off-by: NKumar Gala <galak@kernel.crashing.org>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

* PAGE_OFFSET is not always the start of code, use _stext instead.
* grab PAGE_SIZE and KERNELBASE from asm/page.h like ppc64 does.  Makes the
  code a bit more common and provide a single place to manipulate the
  defines for things like kdump.
Signed-off-by: NKumar Gala <galak@kernel.crashing.org>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

* Determine the RPN we are running the kernel at runtime rather
  than using compile time constant for initial TLB

* Cleanup adjust_total_lowmem() to respect memstart_addr and
  be a bit more clear on variables that are sizes vs addresses.
Signed-off-by: NKumar Gala <galak@kernel.crashing.org>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Fedora 9 works on Efika without the separate 'device-tree supplement',
thanks to the kernel's own fixups. With one exception -- because 'CHRP'
still appears on the 'machine:' line in /proc/cpuinfo, the installer
misdetects the platform and misconfigures yaboot, putting it into a PReP
boot partition instead of in the /boot filesystem where the Efika's
firmware could find it.

The kernel's fixups for Efika already correct one instance of 'chrp', in
the 'device_type' property. This fixes it in the 'CODEGEN,description'
property too, since that's what's exposed to userspace in /proc/cpuinfo.
Signed-off-by: NDavid Woodhouse <dwmw2@infradead.org>
Acked-by: NGrant Likely <grant.likely@secretlab.ca>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This fixes the handling of the preempt count when switching
interrupt stacks so that HW interrupt properly get the softirq
mask copied over from the previous stack.

It also initializes the softirq stack preempt_count to 0 instead
of SOFTIRQ_OFFSET, like x86, as __do_softirq() does the increment,
and we hit some lockdep checks if we have it twice.

That means we do run for a little while off the softirq stack
with the preempt-count set to 0, which could be deadly if we
try to take a softirq at that point, however we do so with
interrupts disabled, so I think we are ok.
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Currently, we initialize the "current" pointer in the PACA (which
is used by the "current" macro in the kernel) before calling
setup_system(). That means that early_setup() is called with
current still "NULL" which is -not- a good idea. It happens to
work so far but breaks with lockdep when early code calls printk.

This changes it so that all PACAs are statically initialized with
__current pointing to the init task. For non-0 CPUs, this is fixed
up before use.
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Now that we have the alpaca, the reg_save_ptr is no longer needed in the
paca.  Eradicate all global uses of it and make it static in the iSeries
lpardata.c
Signed-off-by: NStephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

The iSeries HV only needs the first two fields of the paca statically
initialised, so create an alternate paca that contains only those and
switch to our real paca immediately after boot.

This is in order to make the 1024 cpu patches easier since they will no
longer have to statically initialise the pacas for iSeries.
Signed-off-by: NStephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

If an SLB miss interrupt happens while the RI bit of MSR is zero, we
can't just return, because RI being zero indicates that SRR0/SRR1
potentially had live values in them, and the process of taking an
interrupt overwrites them.

This should never happen, but if it does, we try to print a nice oops
message.  That doesn't work, however, because the code at unrecov_slb
assumes that the MMU has been turned on, but we call it with the MMU
off (and have done so since the SLB miss handler was rewritten to run
without turning the MMU on) -- except on iSeries, where everything runs
with the MMU on.

This fixes it by adding the necessary code to turn the MMU on if
necessary.
Signed-off-by: NPaul Mackerras <paulus@samba.org>

A couple of places are duplicating the function of
of_device_is_available; convert them to use it.
Signed-off-by: NNathan Lynch <ntl@pobox.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

We have an assembly version of strncmp for the bootwrapper, but not
for the kernel, so we end up using the C version in the kernel.  This
takes the strncmp code from the bootup and copies it to the kernel
proper, adding two instructions so it copes correctly with len==0.
Signed-off-by: NSteven Rostedt <srostedt@redhat.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Replace logical "&&" by bit "&" for ISA_SPACE_MASK.
Signed-off-by: NRoel Kluin <12o3l@tiscali.nl>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Handling of the proc_dir_entry->count was changed in 2.6.24-rc5.
After this change, the default value for pde->count is 1 and not 0 as
before.  Therefore, if we want to check whether our procfs file is
already opened (already in use), we have to check if pde->count is
greater than 2 rather than 1.
Signed-off-by: NMaxim Shchetynin <maxim@de.ibm.com>
Signed-off-by: NJens Osterkamp <jens@de.ibm.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

A subtle bug sneaked into iSeries recently.  On this platform, we must
not normally clear MSR:EE (the hardware external interrupt enable)
except for short periods of time.  Taking an interrupt while
soft-disabled doesn't cause us to clear it for example.

The iSeries kernel expects to mostly run with MSR:EE enabled at all
times except in a few exception entry/exit code paths.  Thus
local_irq_enable() doesn't check if it needs to hard-enable as it
expects this to be unnecessary on iSeries.

However, hard_irq_disable() _does_ cause MSR:EE to be cleared,
including on iSeries.  A call to it was recently added to the
context switch code, thus causing interrupts to become disabled
for a long periods of time, causing the iSeries watchdog to kick
in under some circumstances and other nasty things.

This patch fixes it by making local_irq_enable() properly re-enable
MSR:EE on iSeries.  It basically removes a return statement here
to make iSeries use the same code path as everybody else.  That does
mean that we might occasionally get spurious decrementer interrupts
but I don't think that matters.

Another option would have been to make hard_irq_disable() a nop
on iSeries but I didn't like it much, in case we have good reasons
to hard-disable.

Part of the patch is fixes to make sure the hard_enabled PACA field
is properly set on iSeries as it used not to be before, since it
was mostly unused.
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

__FUNCTION__ is gcc-specific, use __func__
Signed-off-by: NHarvey Harrison <harvey.harrison@gmail.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Old-world powermacs don't set L2CR or L3CR on processor upgrade cards.
This simple patch allows the setting of L3CR via a kernel parameter
(like the existing kernel parameter to set L2CR).
Signed-off-by: NRobert Brose <bob@qbjnet.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

WARNING: vmlinux.o(.text+0x1e4c0): Section mismatch in reference from the
function .move_device_tree() to the function .init.text:.lmb_alloc_base()
The function .move_device_tree() references
the function __init .lmb_alloc_base().
This is often because .move_device_tree lacks a __init
annotation or the annotation of .lmb_alloc_base is wrong.

move_device_tree() is called from early_init_devtree() only, which is __init
Signed-off-by: NGeert Uytterhoeven <Geert.Uytterhoeven@sonycom.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

There is a bug in the powerpc DABR (data access breakpoint) handling,
which can result in us missing breakpoints if several threads are trying
to break on the same address.

The circumstances are that do_page_fault() calls do_dabr(), this clears
the DABR (sets it to 0) and sets up the signal which will report to
userspace that the DABR was hit. The do_signal() code will restore the DABR
value on the way out to userspace.

If we reschedule before calling do_signal(), __switch_to() will check the
cached DABR value and compare it to the new thread's value, if they match
we don't set the DABR in hardware.

So if two threads have the same DABR value, and we schedule from one to
the other after taking the interrupt for the first thread hitting the DABR,
the second thread will run without the DABR set in hardware.

The cleanest fix is to move the cache update into set_dabr(), that way we
can't forget to do it.
Reported-by: NJan Kratochvil <jan.kratochvil@redhat.com>
Signed-off-by: NMichael Ellerman <michael@ellerman.id.au>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This patch adds basic support for the AMCC 460EX/460GT PPC's to arch/powerpc.
Currently those PPC's are still based on a 440 core and *not* a 460 core.

Here some basic features of those SoC's:

460EX:
- Up to 1.2GHz, 32kB L1 I-cache and D-cache, 256kB L2-cache, FPU
- 1 * PCI (max 66MHz), 2 * PCIe (one 4-lane, one 1-lane)
- 2 * GBit Ethernet with TCP/IP acceleration
- USB 2.0 Host/Device OTG and Host interface
- SATA controller
- Optional security feature

460GT (only changes to 460EX):
- 4 * GBit Ethernet with TCP/IP acceleration
- RapidIO
- No SATA
- No USB
Signed-off-by: NStefan Roese <sr@denx.de>
Signed-off-by: NJosh Boyer <jwboyer@linux.vnet.ibm.com>

This adds a kernel command line option "phyp_dump", which takes a 0/1
value for disabling/ enabling phyp_dump at boot time.  Kdump can use
this on cmdline (phyp_dump=0) to disable phyp-dump during boot when
enabling itself.  This will ensure only one dumping mechanism is active
at any given time.
Signed-off-by: NManish Ahuja <mahuja@us.ibm.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Initial patch for reserving memory in early boot, and freeing it
later.  If the previous boot had ended with a crash, the reserved
memory would contain a copy of the crashed kernel data.
Signed-off-by: NManish Ahuja <mahuja@us.ibm.com>
Signed-off-by: NLinas Vepstas <linasvepstas@gmail.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

These items in asm-offsets.c are not used anywhere.  This removes them.
Signed-off-by: NRoland McGrath <roland@redhat.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

The PT_DTRACE flag is meaningless and obsolete.
Don't touch it.
Signed-off-by: NRoland McGrath <roland@redhat.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

It was being protected by CONFIG_PPC32, but we want to export it on
64-bit also.  This moves it out of the ifdef.
Signed-off-by: NStephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Print out 'model' property of '/' node as a machine name
in generic show_cpuinfo() routine.
Signed-off-by: NMarian Balakowicz <m8@semihalf.com>
Acked-by: NDavid Gibson <david@gibson.dropbear.id.au>
Acked-by: NOlof Johansson <olof@lixom.net>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Since the PMU is an NMI now, it can come at any time we are only soft
disabled.  We must hard disable around the two places we allow the kernel
stack SLB and r1 to go out of sync.  Otherwise the PMU exception can
force a kernel stack SLB into another slot, which can lead to it
getting evicted, which can lead to a nasty unrecoverable SLB miss
in the exception entry code.
Signed-off-by: NAnton Blanchard <anton@samba.org>
Acked-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>
Acked-by: NOlof Johansson <olof@lixom.net>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

The PTRACE_SETREGS request was only recently added on powerpc,
and gdb does not use it.  So it slipped through without getting
all the testing it should have had.

The user_regset changes had a simple bug in storing to all of
the 32-bit general registers block on 64-bit kernels.  This bug
only comes up with PTRACE_SETREGS, not PPC_PTRACE_SETREGS.
It causes a BUG_ON to hit, so this fix needs to go in ASAP.
Signed-off-by: NRoland McGrath <roland@redhat.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Once again, this time with feeling....

						- Ted

>From c91cfaabc17f8a53807a2f31f067a732e34a1550 Mon Sep 17 00:00:00 2001
From: Theodore Ts'o <tytso@mit.edu>
Date: Wed, 12 Mar 2008 11:50:39 -0400
Subject: Export empty_zero_page

The empty_zero_page symbol is exported by most other architectures
(s390, ia64, x86, um), and an upcoming ext4 patch needs it because
ZERO_PAGE() references empty_zero_page, and we need it to zero out an
unitialized extents in ext4 files.
Signed-off-by: N"Theodore Ts'o" <tytso@mit.edu>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

A bogus test for unassigned resources that came from our 32-bit
PCI code ended up being "merged" by my previous patch series,
breaking some 64-bit setups where devices have legal resources
ending at 0xffffffff.

This fixes it by completely changing the test.  We now test for
res->start == 0, as the generic code expects, and we also only
do so on platforms that don't have the PPC_PCI_PROBE_ONLY flag
set, as there are cases of pSeries and iSeries where it could
be a valid value and those can't reassign devices.
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

Some drivers (such as V4L2) have code that causes gcc to generate
calls to __ucmpdi2 when compiling for 32-bit powerpc, which results
in either a link-time error or a module that can't be loaded, as
we don't currently have a __ucmpdi2.  This adds one so these drivers
can be used.
Signed-off-by: NPaul Mackerras <paulus@samba.org>

This makes swap routines operate correctly on the ppc_8xx based machines.
Code has been revalidated on mpc885ads (8M sdram) with recent kernel. Based
on patch from Yuri Tikhonov <yur@emcraft.com> to do the same on arch/ppc
instance.

Recent kernel's size makes swap feature very important on low-memory platforms,
those are actually non-operable without it.
Signed-off-by: NYuri Tikhonov <yur@emcraft.com>
Signed-off-by: NVitaly Bordug <vitb@kernel.crashing.org>
Signed-off-by: NKumar Gala <galak@kernel.crashing.org>

This code isn't referenced anywhere, so remove it.
Signed-off-by: NDale Farnsworth <dale@farnsworth.org>
Signed-off-by: NPaul Mackerras <paulus@samba.org>

dt_mem_next_cell() currently does of_read_ulong().  This does not
allow for the case where #size-cells and/or #address-cells = 2 on a
32-bit system, as it will end up reading 32 bits instead of the
expected 64.  Change it to use of_read_number instead and always
return a u64.

Signed-off-by: Becky Bruce <becky.bruce at freescale.com>
Acked-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: NPaul Mackerras <paulus@samba.org>