The idea comes from Benjamin Herrenschmidt. The eeh cache helps
fetching the pci device according to the given I/O address. Since
the eeh cache is serving for eeh, it's reasonable for eeh cache
to trace eeh device except pci device.

The patch make eeh cache to trace eeh device. Also, the major
eeh entry function eeh_dn_check_failure has been renamed to
eeh_dev_check_failure since it will take eeh device as input
parameter.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

While EEH module is installed, PCI devices is checked one by one
to see if it supports eeh. On different platforms, the PCI devices
are referred through different ways when the EEH module is loaded.
For example, on pSeries platform, that is done by OF node. However,
we would do that by real PCI devices (struct pci_dev) on PowerNV
platform in future. So we needs some mechanism to differentiate
those cases by classifying them to probe modes, either from OF
nodes or real PCI devices.

The patch implements the support to eeh probe mode. Also, the
EEH on pSeries has set it into EEH_PROBE_MODE_DEVTREE. That means
the probe will be done based on OF nodes on pSeries platform.

In addition, On pSeries platform, it's done by OF nodes. The patch
moves the the probe function from EEH core to platform dependent
backend and some cleanup applied.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

The patch removes the eeh related statistics for eeh device since
they have been maintained by the corresponding eeh PE. Also, the
flags used to trace the state of eeh device and PE have been reworked
for a little bit.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

The patch reworks the current implementation so that the eeh errors
will be handled basing on PE instead of eeh device.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Once eeh error is found, eeh event will be created and put it into
the global linked list. At the mean while, kernel thread will be
started to process it. The handler for the kernel thread originally
was eeh device sensitive.

The patch reworks the handler of the kernel thread so that it's PE
sensitive.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

The patch implements reset based on PE instead of eeh device. Also,
The functions used to retrieve the reset type, either hot or fundamental
reset, have been reworked for a little bit. More specificly, it's
implemented based the the eeh device traverse function.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

The patch refactors the original implementation in order to enable
I/O and retrieve EEH log based on PE.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

The patch introduces the function to traverse the devices of the
specified PE and its child PEs. Also, the restore on device bars
is implemented based on the traverse function.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Originally, all the EEH operations were implemented based on OF node.

Actually, it explicitly breaks the rules that the operation target
is PE instead of device. Therefore, the patch makes all the operations
based on PE instead of device.

Unfortunately, the backend for config space has to be kept as original
because it doesn't depend on PE.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

There're 2 conditions to trigger EEH error detection: invalid value
returned from reading I/O or config space. On each case, the function
eeh_dn_check_failure will be called to initialize EEH event and put
it into the poll for further processing.

The patch changes the function for a little bit so that the EEH error
will be traced based on PE instead of EEH device any more. Also, the
function eeh_find_device_pe() has been removed since the eeh device
is tracing the PE by struct eeh_dev::pe.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Since we've introduced dedicated struct to trace individual PEs,
it's reasonable to trace its state through the dedicated struct
instead of using "eeh_dev" any more.

The patches implements the state tracing based on PE. It's notable
that the PE state will be applied to the specified PE as well as
its child PEs. That complies with the rule that problematic parent
PE will prevent those child PEs from working properly.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

The original implementation builds EEH event based on EEH device.
We already had dedicated struct to depict PE. It's reasonable to
build EEH event based on PE.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

During PCI hotplug and EEH recovery, the PE hierarchy tree might be
changed due to the PCI topology changes. At later point when the
PCI device is added, the PE will be created dynamically again.

The patch introduces new function to remove EEH devices from the
associated PE. That also can cause that the parent PE is removed
from the PE tree if the parent PE doesn't include valid EEH devices
and child PEs.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

The patch creates PEs and associated the newly created PEs with
it parent/silbing as well as EEH devices. It would become more
straight to trace EEH errors and recover them accordingly.

Once the EEH functionality on one PCI IOA has been enabled, we
tries to create PE against it. If there's existing PE, to which
the current PCI IOA should be attached, the existing PE will be
converted from "device" type to "bus" type accordingly.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

The patch implements searching PE based on the following
requirements:

 * Search PE according to PE address, which is traditional
   PE address that is composed of PCI bus/device/function
   number, or unified PE address assigned by firmware or
   platform.
 * Search parent PE according to the given EEH device. It's
   useful when creating new PE and put it into right position.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

For one particular PE, it's only meaningful in the ancestor PHB
domain. Therefore, each PHB should have its own PE hierarchy tree
to trace those PEs created against the PHB.

The patch creates PEs for the PHBs and put those PEs into the
global link list traced by "eeh_phb_pe". The link list of PEs
would be first level of overall PE hierarchy tree across the
system.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

The patch introduces global mutex for EEH so that the core data
structures can be protected by that. Also, 2 inline functions
are exported for that: eeh_lock() and eeh_unlock().
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

As defined in PAPR 2.4, Partitionable Endpoint (PE) is an I/O subtree
that can be treated as a unit for the purposes of partitioning and error
recovery. Therefore, eeh core should be aware of PE. With eeh_pe struct,
we can support PE explicitly. Further more, it makes all the stuff much
more data centralized. Another important reason is for eeh core to support
multiple platforms. Some of them like pSeries figures out PEs through
OF nodes while others like powernv have to do that through PCI bus/device
tree. With explicit PE support, eeh core will be implemented based on
the centrialized data and platform dependent implementations figure it
out by their feasible ways.

When the struct is designed, following factors are taken in account:
  * Reflecting the relationships of PEs. PE might have parent
    as well children.
  * Reflecting the association of PE and (eeh) devices.
  * PEs have PHB boundary.
  * PE should have unique address assigned in the corresponding
    PHB domain.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

The patch adds more logs to EEH initialization functions for
debugging purpose. Also, the machine type (pSeries) is checked
in the platform initialization to assure it's the correct platform
to invoke it.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

The EEH initialization functions have been postponed until slab/slub
are ready. So we use slab/slub to allocate the memory chunks for newly
creatd EEH devices. That would save lots of memory.

The patch also does cleanup to replace "kmalloc" with "kzalloc" so
that we needn't clear the allocated memory chunk explicitly.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Currently, we have 3 phases for EEH initialization on pSeries platform.
All of them are done through builtin functions: platform initialization,
EEH device creation, and EEH subsystem enablement. All of them are done
no later than ppc_md.setup_arch. That means that the slab/slub isn't ready
yet, so we have to allocate memory chunks on basis of PAGE_SIZE for those
dynamically created EEH devices. That's pretty expensive.

In order to utilize slab/slub for memory allocation, we have to move the EEH
initialization functions around, but all of them should be called after slab
is ready.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

It's possible for the cpu_possible_mask to change between the time we
initialise the pacas and the time we setup per_cpu areas.

Obviously impossible cpus shouldn't ever be running, but stranger things
have happened. So be paranoid and initialise data_offset with a poison
value in case we don't set it up later.

Based on a patch from Anton Blanchard.
Signed-off-by: NMichael Ellerman <michael@ellerman.id.au>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

We never use the XDABR hcall since we check for DABR hcall first.
XDABR syscall is better since it allows us to also set the DABRX.
Signed-off-by: NMichael Neuling <mikey@neuling.org>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Change bp_info to info to be consistent with the rest of this file.
Signed-off-by: NMichael Neuling <mikey@neuling.org>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

No functional change
Signed-off-by: NMichael Neuling <mikey@neuling.org>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

The powerpc kernel doesn't export the memory limit enforced by 'mem='
kernel parameter. This is required for building the ELF header in
kexec-tools to limit the vmcore to capture only the used memory. On
powerpc the kexec-tools depends on the device-tree for memory related
information, unlike /proc/iomem on the x86.

Without this information, the kexec-tools assumes the entire System
RAM and vmcore creates an unnecessarily larger dump.

This patch exports the memory limit, if present, via
chosen/linux,memory-limit
property, so that the vmcore can be limited to the memory limit.

The prom_init seems to export this value in the same node. But doesn't
really
appear there.  Also the memory_limit gets adjusted with the processing of
crashkernel= parameter. This patch makes sure we get the actual limit.

The kexec-tools will use the value to limit the 'end' of the memory
regions.

Tested this patch on ppc64 and ppc32(ppc440) with a kexec-tools
patch by Mahesh.
Signed-off-by: NSuzuki K. Poulose <suzuki@in.ibm.com>
Tested-by: NMahesh J. Salgaonkar <mahesh@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

There are some device-tree nodes, whose values are of type phys_addr_t.
The phys_addr_t is variable sized based on the CONFIG_PHSY_T_64BIT.

Change these to a fixed unsigned long long for consistency.

This patch does the change only for memory_limit.

The following is a list of such variables which need the change:

 1) kernel_end, crashk_size - in arch/powerpc/kernel/machine_kexec.c

 2) (struct resource *)crashk_res.start - We could export a local static
    variable from machine_kexec.c.

Changing the above values might break the kexec-tools. So, I will
fix kexec-tools first to handle the different sized values and then change
 the above.
Suggested-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: NSuzuki K. Poulose <suzuki@in.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Several files in obj-plat depend on libfdt header file. Sometimes
when building one can see the following issue. This patch adds
libfdt as dependency to those object files

| In file included from arch/powerpc/boot/treeboot-iss4xx.c:33:0:
| arch/powerpc/boot/libfdt.h:854:1: error: unterminated comment
| In file included from arch/powerpc/boot/treeboot-iss4xx.c:33:0:
| arch/powerpc/boot/libfdt.h:1:0: error: unterminated #ifndef
|   BOOTCC  arch/powerpc/boot/inffast.o
| make[1]: *** [arch/powerpc/boot/treeboot-iss4xx.o] Error 1
| make[1]: *** Waiting for unfinished jobs....
|   BOOTCC  arch/powerpc/boot/inflate.o
| make: *** [uImage] Error 2
| ERROR: oe_runmake failed
| ERROR: Function failed: do_compile (see /srv/home/pokybuild/yocto-autobuilder/yocto-slave/p1022ds/build/build/tmp/work/p1022ds-poky-linux-gnuspe/linux-qoriq-sdk-3.0.34-r5/temp/log.do_compile.2167 for further information)
NOTE: recipe linux-qoriq-sdk-3.0.34-r5: task do_compile: Failed
Signed-off-by: NMatthew McClintock <msm@freescale.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Starting with Power 7+ we need to check for marked events if the SIAR
register is valid, i.e. it contains the correct address of the instruction
at the time the performance counter overflowed.  The mmcra register on
Power 7+, contains a new bit to indicate that the contents of the SIAR
is valid. If the event is not marked, then the sample is recorded
independently of the SIAR valid bit setting.  For older processors, there
is no SIAR valid bit to check so the samples are always recorded.  This is
done by forcing the cntr_marked_events bit mask to zero.  The code will
always record the sample in this case since the bit mask says the event is
not a marked event even if it really is a marked event.
Signed-off-by: NCarl Love <cel@us.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

This definition will be used by subsequent perf and oprofile patches
Signed-off-by: NSukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

The pseries firmware currently refuses any non power of two MSI-X
request. Unfortunately most network drivers end up asking for that
because they want a power of two for RX queues and one or two extra
for everything else.

This patch rounds up the firmware request to the next power of two
if the quota allows it. If this fails we fall back to using the
original request size.
Signed-off-by: NAnton Blanchard <anton@samba.org>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

When PCI probe flag PCI_REASSIGN_ALL_RSRC has been passed into PCI
core, it's hoped that all resources to be reassigned by PCI core.
As to particular P2P (PCI-to-PCI) bridge, the size of the corresponding
BAR (I/O, MMIO, prefetchable MMIO) is calculated by the resources
required by the PCI devices behind the P2P bridge. That means that
the information like start/end address retrieved from the hardware
registers of the P2P bridge is meainingless in the case. However,
we still count that in and the BARs might have been configured by
firmware with non-zero size. That leads to space waste.

The patch explicitly sets the size of P2P bridge BARs to zero in
case that resource reassignment is expected with PCI probe flag
PCI_REASSIGN_ALL_RSRC. In the result, it will save overall resource
required by the system without waste.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Brings in various bug fixes from 3.6-rcX

commit: 8b7b80b9
[24/29] powerpc: Uprobes port to powerpc

Caused a clash with the fore200e driver:

In file included from drivers/atm/fore200e.c:70:0:
drivers/atm/fore200e.h:263:3: error: redefinition of typedef 'opcode_t' with different type
arch/powerpc/include/asm/probes.h:25:13: note: previous declaration of 'opcode_t' was here

Fix the namespace clash by making opcode_t in probes.h to ppc_opcode_t.
Signed-off-by: NAnanth N Mavinakayanahalli <ananth@in.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Critical exception on 64-bit booke uses user-visible SPRG3 as scratch.
Restore VDSO information in SPRG3 on exception prolog.

Use a common sprg3 field in PACA for all powerpc64 architectures.
Signed-off-by: NMihai Caraman <mihai.caraman@freescale.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

patch_instruction() can be called very early on ppc32, when the kernel
isn't yet running at it's linked address. That can cause the !
is_kernel_addr() test in __put_user() to trip and call might_sleep()
which is very bad at that point during boot.

Use a lower level function instead for now, at least until we get to
rework ppc32 boot process to do the code patching later, like ppc64
does.
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

We have been observing hangs, both of KVM guest vcpu tasks and more
generally, where a process that is woken doesn't properly wake up and
continue to run, but instead sticks in TASK_WAKING state.  This
happens because the update of rq->wake_list in ttwu_queue_remote()
is not ordered with the update of ipi_message in
smp_muxed_ipi_message_pass(), and the reading of rq->wake_list in
scheduler_ipi() is not ordered with the reading of ipi_message in
smp_ipi_demux().  Thus it is possible for the IPI receiver not to see
the updated rq->wake_list and therefore conclude that there is nothing
for it to do.

In order to make sure that anything done before smp_send_reschedule()
is ordered before anything done in the resulting call to scheduler_ipi(),
this adds barriers in smp_muxed_message_pass() and smp_ipi_demux().
The barrier in smp_muxed_message_pass() is a full barrier to ensure that
there is a full ordering between the smp_send_reschedule() caller and
scheduler_ipi().  In smp_ipi_demux(), we use xchg() rather than
xchg_local() because xchg() includes release and acquire barriers.
Using xchg() rather than xchg_local() makes sense given that
ipi_message is not just accessed locally.

This moves the barrier between setting the message and calling the
cause_ipi() function into the individual cause_ipi implementations.
Most of them -- those that used outb, out_8 or similar -- already had
a full barrier because out_8 etc. include a sync before the MMIO
store.  This adds an explicit barrier in the two remaining cases.

These changes made no measurable difference to the speed of IPIs as
measured using a simple ping-pong latency test across two CPUs on
different cores of a POWER7 machine.

The analysis of the reason why processes were not waking up properly
is due to Milton Miller.

Cc: stable@vger.kernel.org # v3.0+
Reported-by: NMilton Miller <miltonm@bga.com>
Signed-off-by: NPaul Mackerras <paulus@samba.org>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

During a context switch we always restore the per thread DSCR value.
If we aren't doing explicit DSCR management
(ie thread.dscr_inherit == 0) and the default DSCR changed while
the process has been sleeping we end up with the wrong value.

Check thread.dscr_inherit and select the default DSCR or per thread
DSCR as required.

This was found with the following test case, when running with
more threads than CPUs (ie forcing context switching):

http://ozlabs.org/~anton/junkcode/dscr_default_test.c

With the four patches applied I can run a combination of all
test cases successfully at the same time:

http://ozlabs.org/~anton/junkcode/dscr_default_test.c
http://ozlabs.org/~anton/junkcode/dscr_explicit_test.c
http://ozlabs.org/~anton/junkcode/dscr_inherit_test.cSigned-off-by: NAnton Blanchard <anton@samba.org>
Cc: <stable@kernel.org> # 3.0+
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

If the default DSCR is non zero we set thread.dscr_inherit in
copy_thread() meaning the new thread and all its children will ignore
future updates to the default DSCR. This is not intended and is
a change in behaviour that a number of our users have hit.

We just need to inherit thread.dscr and thread.dscr_inherit from
the parent which ends up being much simpler.

This was found with the following test case:

http://ozlabs.org/~anton/junkcode/dscr_default_test.cSigned-off-by: NAnton Blanchard <anton@samba.org>
Cc: <stable@kernel.org> # 3.0+
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

When we update the DSCR either via emulation of mtspr(DSCR) or via
a change to dscr_default in sysfs we don't update thread.dscr.
We will eventually update it at context switch time but there is
a period where thread.dscr is incorrect.

If we fork at this point we will copy the old value of thread.dscr
into the child. To avoid this, always keep thread.dscr in sync with
reality.

This issue was found with the following testcase:

http://ozlabs.org/~anton/junkcode/dscr_inherit_test.cSigned-off-by: NAnton Blanchard <anton@samba.org>
Cc: <stable@kernel.org> # 3.0+
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>