Correct build failure for powerpc/pseries builds with CONFIG_SMP not defined.

The function cpu_sibling_mask has no meaning (or definition) when CONFIG_SMP
is not defined. Additionally, the updating of NUMA affinity for a CPU in a UP
system doesn't really make sense.

This patch ifdef's out the code making the affinity updates for PRRN events to
fix the following build break.

arch/powerpc/mm/numa.c: In function ‘stage_topology_update’:
arch/powerpc/mm/numa.c:1535: error: implicit declaration of function ‘cpu_sibling_mask’
arch/powerpc/mm/numa.c:1535: warning: passing argument 3 of ‘cpumask_or’ makes pointer from integer without a cast
make[1]: *** [arch/powerpc/mm/numa.o] Error 1
Signed-off-by: NNathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

This is stale and not used by anyone now.
Signed-off-by: NKevin Hao <haokexin@gmail.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

We use kmem_cache_alloc() to allocate memory to hold the new firmware
which will be flashed. kmem_cache_alloc() calls rtas_block_ctor() to
set memory to NULL. But these constructor is called only for newly
allocated slabs.

If we run below command multiple time without rebooting, allocator may
allocate memory from the area which was free'd by kmem_cache_free and
it will not call constructor. In this situation we may hit kernel oops.

dd if=<fw image> of=/proc/ppc64/rtas/firmware_flash bs=4096

oops message:
-------------
[ 1602.399755] Oops: Kernel access of bad area, sig: 11 [#1]
[ 1602.399772] SMP NR_CPUS=1024 NUMA pSeries
[ 1602.399779] Modules linked in: rtas_flash nfsd lockd auth_rpcgss nfs_acl sunrpc fuse loop dm_mod sg ipv6 ses enclosure ehea ehci_pci ohci_hcd ehci_hcd usbcore sd_mod usb_common crc_t10dif scsi_dh_alua scsi_dh_emc scsi_dh_hp_sw scsi_dh_rdac scsi_dh ipr libata scsi_mod
[ 1602.399817] NIP: d00000000a170b9c LR: d00000000a170b64 CTR: c00000000079cd58
[ 1602.399823] REGS: c0000003b9937930 TRAP: 0300   Not tainted  (3.9.0-rc4-0.27-ppc64)
[ 1602.399828] MSR: 8000000000009032 <SF,EE,ME,IR,DR,RI>  CR: 22000428  XER: 20000000
[ 1602.399841] SOFTE: 1
[ 1602.399844] CFAR: c000000000005f24
[ 1602.399848] DAR: 8c2625a820631fef, DSISR: 40000000
[ 1602.399852] TASK = c0000003b4520760[3655] 'dd' THREAD: c0000003b9934000 CPU: 3
GPR00: 8c2625a820631fe7 c0000003b9937bb0 d00000000a179f28 d00000000a171f08
GPR04: 0000000010040000 0000000000001000 c0000003b9937df0 c0000003b5fb2080
GPR08: c0000003b58f7200 d00000000a179f28 c0000003b40058d4 c00000000079cd58
GPR12: d00000000a171450 c000000007f40900 0000000000000005 0000000010178d20
GPR16: 00000000100cb9d8 000000000000001d 0000000000000000 000000001003ffff
GPR20: 0000000000000001 0000000000000000 00003fffa0b50d30 000000001001f010
GPR24: 0000000010020888 0000000010040000 d00000000a171f08 d00000000a172808
GPR28: 0000000000001000 0000000010040000 c0000003b4005880 8c2625a820631fe7
[ 1602.399924] NIP [d00000000a170b9c] .rtas_flash_write+0x7c/0x1e8 [rtas_flash]
[ 1602.399930] LR [d00000000a170b64] .rtas_flash_write+0x44/0x1e8 [rtas_flash]
[ 1602.399934] Call Trace:
[ 1602.399939] [c0000003b9937bb0] [d00000000a170b64] .rtas_flash_write+0x44/0x1e8 [rtas_flash] (unreliable)
[ 1602.399948] [c0000003b9937c60] [c000000000282830] .proc_reg_write+0x90/0xe0
[ 1602.399955] [c0000003b9937ce0] [c0000000001ff374] .vfs_write+0x114/0x238
[ 1602.399961] [c0000003b9937d80] [c0000000001ff5d8] .SyS_write+0x70/0xe8
[ 1602.399968] [c0000003b9937e30] [c000000000009cdc] syscall_exit+0x0/0xa0
[ 1602.399973] Instruction dump:
[ 1602.399977] eb698010 801b0028 2f80dcd6 419e00a4 2fbc0000 419e009c ebfb0030 2fbf0000
[ 1602.399989] 409e0010 480000d8 60000000 7c1f0378 <e81f0008> 2fa00000 409efff4 e81f0000
[ 1602.400012] ---[ end trace b4136d115dc31dac ]---
[ 1602.402178]
[ 1602.402185] Sending IPI to other CPUs
[ 1602.403329] IPI complete

This patch uses kmem_cache_zalloc() instead of kmem_cache_alloc() to
allocate memory, which makes sure memory is set to 0 before using.
Also removes rtas_block_ctor(), which is no longer required.
Signed-off-by: NVasant Hegde <hegdevasant@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

After merging the cgroup tree, today's linux-next build (powerpc
ppc64_defconfig) failed like this:

arch/powerpc/mm/numa.c: In function 'arch_update_cpu_topology':
arch/powerpc/mm/numa.c:1465:2: error: implicit declaration of function 'kzalloc' [-Werror=implicit-function-declaration]
arch/powerpc/mm/numa.c:1465:10: error: assignment makes pointer from integer without a cast [-Werror]
arch/powerpc/mm/numa.c:1497:2: error: implicit declaration of function 'kfree' [-Werror=implicit-function-declaration]

Caused by commit 30c05350 ("powerpc/pseries: Use stop machine to
update cpu maps") from the powerpc tree interacting with (probably)
commit ff794dea ("cpuset: remove include of cgroup.h from cpuset.h")
from the cgroup tree.  Removing includes from header files is fraught
with danger ...

The former should have added an include of linux/slab.h to
arch/powerpc/mm/numa.c.

I have added the following merge fix patch for today (but it should be
applied to the powerpc tree ASAP).

From: Stephen Rothwell <sfr@canb.auug.org.au>
Date: Mon, 29 Apr 2013 14:01:44 +1000
Subject: [PATCH] powerpc: numa.c: using kzalloc/kfree requires including
 slab.h

fixes these build errors:

arch/powerpc/mm/numa.c: In function 'arch_update_cpu_topology':
arch/powerpc/mm/numa.c:1465:2: error: implicit declaration of function 'kzalloc' [-Werror=implicit-function-declaration]
arch/powerpc/mm/numa.c:1465:10: error: assignment makes pointer from integer without a cast [-Werror]
arch/powerpc/mm/numa.c:1497:2: error: implicit declaration of function 'kfree' [-Werror=implicit-function-declaration]
Signed-off-by: NStephen Rothwell <sfr@canb.auug.org.au>
Acked-by: NTejun Heo <tj@kernel.org>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Linux next is currently failing to compile mpc85xx_defconfig with:
  arch/powerpc/sysdev/fsl_pci.c:944:2: error: too many arguments to function 'setup_pci_atmu'

This is caused by (from Kumar's next branch):
  commit 34642bbb
  Author: Kumar Gala <galak@kernel.crashing.org>
  powerpc/fsl-pci: Keep PCI SoC controller registers in pci_controller

Which changed definition of setup_pci_atmu() but didn't update one of
the callers.  Below fixes this.
Signed-off-by: NMichael Neuling <mikey@neuling.org>
Reviewed-by: NKim Phillips <kim.phillips@freescale.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Update the powerpc slice_get_unmapped_area function to make use of
vm_unmapped_area() instead of implementing a brute force search.
Signed-off-by: NMichel Lespinasse <walken@google.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Tested-by: N"Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>
Acked-by: NDavid Gibson <david@gibson.dropbear.id.au>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

As all other architectures have been converted to use vm_unmapped_area(),
we are about to retire the free_area_cache.

This change simply removes the use of that cache in
slice_get_unmapped_area(), which will most certainly have a
performance cost. Next one will convert that function to use the
vm_unmapped_area() infrastructure and regain the performance.
Signed-off-by: NMichel Lespinasse <walken@google.com>
Acked-by: NRik van Riel <riel@redhat.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

The reg property in the pci bridge device node is used to bind this
device node to the pci bridge device. Then all the pci devices under
this bridge could use the interrupt maps defined in this device node
to do the irq translation. So if this property is missed, the pci
traditional irq mechanism will not work.
Signed-off-by: NKevin Hao <haokexin@gmail.com>
Signed-off-by: NKumar Gala <galak@kernel.crashing.org>

In patch 34642bbb (powerpc/fsl-pci: Keep PCI SoC controller registers in
pci_controller) we choose to keep the map of the PCI SoC controller
registers. But we missed to delete the unmap in setup_pci_atmu
function. This will cause the following call trace once we access
the PCI SoC controller registers later.

Unable to handle kernel paging request for data at address 0x8000080080040f14
Faulting instruction address: 0xc00000000002ea58
Oops: Kernel access of bad area, sig: 11 [#1]
SMP NR_CPUS=24 T4240 QDS
Modules linked in:
NIP: c00000000002ea58 LR: c00000000002eaf4 CTR: c00000000002eac0
REGS: c00000017e10b4a0 TRAP: 0300   Not tainted  (3.9.0-rc1-00052-gfa3529f-dirty)
MSR: 0000000080029000 <CE,EE,ME>  CR: 28adbe22  XER: 00000000
SOFTE: 0
DEAR: 8000080080040f14, ESR: 0000000000000000
TASK = c00000017e100000[1] 'swapper/0' THREAD: c00000017e108000 CPU: 2
GPR00: 0000000000000000 c00000017e10b720 c0000000009928d8 c00000017e578e00
GPR04: 0000000000000000 000000000000000c 0000000000000001 c00000017e10bb40
GPR08: 0000000000000000 8000080080040000 0000000000000000 0000000000000016
GPR12: 0000000088adbe22 c00000000fffa800 c000000000001ba0 0000000000000000
GPR16: 0000000000000000 0000000000000000 0000000000000000 0000000000000000
GPR20: 0000000000000000 0000000000000000 0000000000000000 c0000000008a5b70
GPR24: c0000000008af938 c0000000009a28d8 c0000000009bb5dc c00000017e10bb40
GPR28: c00000017e32a400 c00000017e10bc00 c00000017e32a400 c00000017e578e00
NIP [c00000000002ea58] .fsl_pcie_check_link+0x88/0xf0
LR [c00000000002eaf4] .fsl_indirect_read_config+0x34/0xb0
Call Trace:
[c00000017e10b720] [c00000017e10b7a0] 0xc00000017e10b7a0 (unreliable)
[c00000017e10ba30] [c00000000002eaf4] .fsl_indirect_read_config+0x34/0xb0
[c00000017e10bad0] [c00000000033aa08] .pci_bus_read_config_byte+0x88/0xd0
[c00000017e10bb90] [c00000000088d708] .pci_apply_final_quirks+0x9c/0x18c
[c00000017e10bc40] [c0000000000013dc] .do_one_initcall+0x5c/0x1f0
[c00000017e10bcf0] [c00000000086ebac] .kernel_init_freeable+0x180/0x26c
[c00000017e10bdb0] [c000000000001bbc] .kernel_init+0x1c/0x460
[c00000017e10be30] [c000000000000880] .ret_from_kernel_thread+0x64/0xe4
Instruction dump:
38210310 2b800015 4fdde842 7c600026 5463fffe e8010010 7c0803a6 4e800020
60000000 60000000 e92301d0 7c0004ac <80690f14> 0c030000 4c00012c 38210310
---[ end trace 7a8fe0cbccb7d992 ]---

Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
Signed-off-by: NKevin Hao <haokexin@gmail.com>
Acked-by: NRoy Zang <tie-fei.zang@freescale.com>
Signed-off-by: NKumar Gala <galak@kernel.crashing.org>

Fix the following errors:

Error: p1025rdb.dtsi:326.2-3 label or path, 'qe', not found
Error: p1021si-post.dtsi:242.2-3 label or path, 'qe', not found
FATAL ERROR: Syntax error parsing input tree
Signed-off-by: NZhicheng Fan <B32736@freescale.com>
Signed-off-by: NKumar Gala <galak@kernel.crashing.org>

Provides basic enablement for perf branch stack sampling framework on
POWER8 processor based platforms. Adds new BHRB related elements into
cpu_hw_event structure to represent current BHRB config, BHRB filter
configuration, manage context and to hold output BHRB buffer during
PMU interrupt before passing to the user space. This also enables
processing of BHRB data and converts them into generic perf branch
stack data format.
Signed-off-by: NAnshuman Khandual <khandual@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

This patch populates BHRB specific data for power_pmu structure. It
also implements POWER8 specific BHRB filter and configuration functions.
Signed-off-by: NAnshuman Khandual <khandual@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

This patch adds couple of generic functions to power_pmu structure
which would configure the BHRB and it's filters. It also adds
representation of the number of BHRB entries present on the PMU.
A new PMU flag PPMU_BHRB would indicate presence of BHRB feature.
Signed-off-by: NAnshuman Khandual <khandual@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

This patch adds the basic assembly code to read BHRB buffer. BHRB entries
are valid only after a PMU interrupt has happened (when MMCR0[PMAO]=1)
and BHRB has been freezed. BHRB read should not be attempted when it is
still enabled (MMCR0[PMAE]=1) and getting updated, as this can produce
non-deterministic results.
Signed-off-by: NAnshuman Khandual <khandual@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

This patch adds new POWER8 instruction encoding for reading
and clearing Branch History Rolling Buffer entries. The new
instruction 'mfbhrbe' (move from branch history rolling buffer
entry) is used to read BHRB buffer entries and instruction
'clrbhrb' (clear branch history rolling buffer) is used to
clear the entire buffer. The instruction 'clrbhrb' has straight
forward encoding. But the instruction encoding format for
reading the BHRB entries is like 'mfbhrbe RT, BHRBE' where it
takes two arguments, i.e the index for the BHRB buffer entry to
read and a general purpose register to put the value which was
read from the buffer entry.
Signed-off-by: NAnshuman Khandual <khandual@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

This patch adds support for the power8 PMU to perf.

Work is ongoing to add generic cache events.
Signed-off-by: NMichael Ellerman <michael@ellerman.id.au>
Acked-by: NPaul Mackerras <paulus@samba.org>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

On power8 we have a new SIER (Sampled Instruction Event Register), which
captures information about instructions when we have random sampling
enabled.

Add support for loading the SIER into pt_regs, overloading regs->dar.
Also set the new NO_SIPR flag in regs->result if we don't have SIPR.

Update regs_sihv/sipr() to look for SIPR/SIHV in SIER.
Signed-off-by: NMichael Ellerman <michael@ellerman.id.au>
Acked-by: NPaul Mackerras <paulus@samba.org>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

On power8 the presence or absence of SIPR depends on settings at runtime,
so convert to using a dynamic flag for NO_SIPR. Existing backends that
set NO_SIPR unconditionally set the dynamic flag obviously.
Signed-off-by: NMichael Ellerman <michael@ellerman.id.au>
Acked-by: NPaul Mackerras <paulus@samba.org>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Add an accessor for regs->result so we can use it to store more flags in
future.
Signed-off-by: NMichael Ellerman <michael@ellerman.id.au>
Acked-by: NPaul Mackerras <paulus@samba.org>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

On power8 the SIPR and SIHV are not in MMCRA, so convert the routines
to take regs and change the names accordingly.
Signed-off-by: NMichael Ellerman <michael@ellerman.id.au>
Acked-by: NPaul Mackerras <paulus@samba.org>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

In perf_ip_adjust() we potentially use the MMCRA[SLOT] field to adjust
the reported IP of a sampled instruction.

Currently the logic is written so that if the backend does NOT have
the PPMU_ALT_SIPR flag set then we assume MMCRA[SLOT] exists.

However on power8 we do not want to set ALT_SIPR (it's in a third
location), and we also do not have MMCRA[SLOT].

So add a new flag which only indicates whether MMCRA[SLOT] exists.

Naively we'd set it on everything except power6/7, because they set
ALT_SIPR, and we've reversed the polarity of the flag. But it's more
complicated than that.

mpc7450 is 32-bit, and uses its own version of perf_ip_adjust()
which doesn't use MMCRA[SLOT], so it doesn't need the new flag set and
the behaviour is unchanged.

PPC970 (and I assume power4) don't have MMCRA[SLOT], so shouldn't have
the new flag set. This is a behaviour change on those cpus, though we
were probably getting lucky and the bits in question were 0.

power5 and power5+ set the new flag, behaviour unchanged.

power6 & power7 do not set the new flag, behaviour unchanged.
Signed-off-by: NMichael Ellerman <michael@ellerman.id.au>
Acked-by: NPaul Mackerras <paulus@samba.org>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

For both HV and guest kernels, intialise PMU regs to something sane.
Signed-off-by: NMichael Ellerman <michael@ellerman.id.au>
Acked-by: NPaul Mackerras <paulus@samba.org>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Ben found the root cause. Commit 37f02195
("powerpc/pci: fix PCI-e devices rescan issue on powerpc platform")
overwrites the IOMMU table of PCI device while enabling PCI device.
The patch intends to fix the IOMMU table after that point.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

The patch intends to build 32-bits DMA space for individual PEs on
PHB3. The TVE# is recognized by the combo of PE# and fixed bits
from DMA address, which is zero for 32-bits DMA space.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

The TCE should be invalidated while it's created or free'd. The
approach to do that for IODA1 and IODA2 compliant PHBs are different.
So the patch differentiate them with different functions called to
do that for IODA1 and IODA2 compliant PHBs. It's notable that the
PCI address is used to invalidate the corresponding TCE on IODA2
compliant PHB3.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

The EOI handler of MSI/MSI-X interrupts for P8 (PHB3) need additional
steps to handle the P/Q bits in IVE before EOIing the corresponding
interrupt. The patch changes the EOI handler to cover that. we have
individual IRQ chip in each PHB instance. During the MSI IRQ setup
time, the IRQ chip is copied over from the original one for that IRQ,
and the EOI handler is patched with the one that will handle the P/Q
bits (As Ben suggested).
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

As Michael Ellerman suggested, to add CONFIG_POWERNV_MSI for PowerNV
platform. That's similar to CONFIG_PSERIES_MSI for pSeries platform.
For now, we don't make it dependent on CONFIG_EEH since it's not ready
to enable that yet.

Apart from that, we also enable CONFIG_PPC_MSI_BITMAP on selecting
CONFIG_POWERNV_MSI.
Signed-off-by: NGavin Shan <shangw@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

The patch intends to initialize PHB3 during system boot stage. The
flag "PNV_PHB_MODEL_PHB3" is introduced to differentiate IODA2
compatible PHB3 from other types of PHBs.
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Building a 64-bit powerpc kernel with PR KVM enabled currently gives
this error:

  AS      arch/powerpc/kernel/head_64.o
arch/powerpc/kernel/exceptions-64s.S: Assembler messages:
arch/powerpc/kernel/exceptions-64s.S:258: Error: attempt to move .org backwards
make[2]: *** [arch/powerpc/kernel/head_64.o] Error 1

This happens because the MASKABLE_EXCEPTION_PSERIES macro turns into
33 instructions, but we only have space for 32 at the decrementer
interrupt vector (from 0x900 to 0x980).

In the code generated by the MASKABLE_EXCEPTION_PSERIES macro, we
currently have two instances of the HMT_MEDIUM macro, which has the
effect of setting the SMT thread priority to medium.  One is the
first instruction, and is overwritten by a no-op on processors where
we save the PPR (processor priority register), that is, POWER7 or
later.  The other is after we have saved the PPR.

In order to reduce the code at 0x900 by one instruction, we omit the
first HMT_MEDIUM.  On processors without SMT this will have no effect
since HMT_MEDIUM is a no-op there.  On POWER5 and RS64 machines this
will mean that the first few instructions take a little longer in the
case where a decrementer interrupt occurs when the hardware thread is
running at low SMT priority.  On POWER6 and later machines, the
hardware automatically boosts the thread priority when a decrementer
interrupt is taken if the thread priority was below medium, so this
change won't make any difference.

The alternative would be to branch out of line after saving the CFAR.
However, that would incur an extra overhead on all processors, whereas
the approach adopted here only adds overhead on older threaded processors.
Signed-off-by: NPaul Mackerras <paulus@samba.org>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

There are instances in which we do not want topology updates to occur.
In order to allow this a /proc interface (/proc/powerpc/topology_updates)
is introduced so that topology updates can be enabled and disabled.

This patch also adds a prrn_is_enabled() call so that PRRN events are
handled in the kernel only if topology updating is enabled.
Signed-off-by: NNathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

The Linux kernel and platform firmware negotiate their mutual support
of the PRRN option via the ibm,client-architecture-support interface.
This patch simply sets the appropriate fields in the client architecture
vector to indicate Linux support for PRRN and will allow the firmware to
report PRRN events via the RTAS event-scan mechanism.
Signed-off-by: NNathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

The new PRRN firmware feature provides a more convenient and event-driven
interface than VPHN for notifying Linux of changes to the NUMA affinity of
platform resources. However, for practical reasons, it may not be feasible
for some customers to update to the latest firmware. For these customers,
the VPHN feature supported on previous firmware versions may still be the
best option.

The VPHN feature was previously disabled due to races with the load
balancing code when accessing the NUMA cpu maps, but the new stop_machine()
approach protects the NUMA cpu maps from these concurrent accesses. It
should be safe to re-enable this feature now.
Signed-off-by: NNathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

The following patch adds vdso_getcpu_init(), which stores the NUMA node for
a cpu in SPRG3:

Commit 18ad51dd ("powerpc: Add VDSO version of getcpu") adds
vdso_getcpu_init(), which stores the NUMA node for a cpu in SPRG3.

This patch ensures that this information is also updated when the NUMA
affinity of a cpu changes.
Signed-off-by: NNathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

The new PRRN firmware feature allows CPU and memory resources to be
transparently reassigned across NUMA boundaries. When this happens, the
kernel must update the node maps to reflect the new affinity information.

Although the NUMA maps can be protected by locking primitives during the
update itself, this is insufficient to prevent concurrent accesses to these
structures. Since cpumask_of_node() hands out a pointer to these
structures, they can still be modified outside of the lock. Furthermore,
tracking down each usage of these pointers and adding locks would be quite
invasive and difficult to maintain.

The approach used is to make a list of affected cpus and call stop_machine
to have the update routine run on each of the affected cpus allowing them
to update themselves. Each cpu finds itself in the list of cpus and makes
the appropriate updates. We need to have each cpu do this for themselves to
handle calls to vdso_getcpu_init() added in a subsequent patch.

Situations like these are best handled using stop_machine(). Since the NUMA
affinity updates are exceptionally rare events, this approach has the
benefit of not adding any overhead while accessing the NUMA maps during
normal operation.
Signed-off-by: NNathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Platform events such as partition migration or the new PRRN firmware
feature can cause the NUMA characteristics of a CPU to change, and these
changes will be reflected in the device tree nodes for the affected
CPUs.

This patch registers a handler for Open Firmware device tree updates
and reconfigures the CPU and node maps whenever the associativity
changes. Currently, this is accomplished by marking the affected CPUs in
the cpu_associativity_changes_mask and allowing
arch_update_cpu_topology() to retrieve the new associativity information
using hcall_vphn().

Protecting the NUMA cpu maps from concurrent access during an update
operation will be addressed in a subsequent patch in this series.
Signed-off-by: NNathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Update the numa code to use the updated firmware_has_feature() when checking
for type 1 affinity.
Signed-off-by: NNathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

The firmware_has_feature() function makes it easy to check for supported
features of the hypervisor. This patch extends the capability of
firmware_has_feature() to include checking for specified bits
in vector 5 of the architecture vector as reported in the device tree.

As part of this the #defines used for the architecture vector are re-defined
such that each option has the index into vector 5 and the feature bit encoded
into it. This makes checking for architecture bits when initiating data
for firmware_has_feature much easier.
Signed-off-by: NNathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

When iterating over the entries in firmware_features_table we only need
to go over the actual number of entries in the array instead of declaring
it to be bigger and checking to make sure there is a valid entry in every
slot.

This patch removes the FIRMWARE_MAX_FEATURES #define and replaces the
array looping with the use of ARRAY_SIZE().
Signed-off-by: NNathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

As part of handling of PRRN events we need to check vector 5 of the
architecture vector bits reported in the device tree to ensure PRRN event
handling is enabled. To do this firmware_has_feature() is updated (in a
subsequent patch) to make this check vector 5 bits. To avoid having to
re-define bits in the architecture vector the bit definitions are moved
to prom.h.
Signed-off-by: NNathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

A PRRN event is signaled via the RTAS event-scan mechanism, which
returns a Hot Plug Event message "fixed part" indicating "Platform
Resource Reassignment". In response to the Hot Plug Event message,
we must call ibm,update-nodes to determine which resources were
reassigned and then ibm,update-properties to obtain the new affinity
information about those resources.

The PRRN event-scan RTAS message contains only the "fixed part" with
the "Type" field set to the value 160 and no Extended Event Log. The
four-byte Extended Event Log Length field is re-purposed (since no
Extended Event Log message is included) to pass the "scope" parameter
that causes the ibm,update-nodes to return the nodes affected by the
specific resource reassignment.

This patch adds a handler for RTAS events. The function
pseries_devicetree_update() (from mobility.c) is used to make the
ibm,update-nodes/ibm,update-properties RTAS calls. Updating the NUMA maps
(handled by a subsequent patch) will require significant processing,
so pseries_devicetree_update() is called from an asynchronous workqueue
to allow event processing to continue.

PRRN RTAS events on pseries systems are rare events that have to be
initiated from the HMC console for the system by an IBM tech. This allows
us to assume that these events are widely spaced. Additionally, all work
on the queue is flushed before handling any new work to ensure we only have
one event in flight being handled at a time.
Signed-off-by: NNathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>