When CONFIG_PPC_IRQ_SOFT_MASK_DEBUG is enabled (uncommon), we have a
series of WARN_ON's in arch_local_irq_restore().

These are "should never happen" conditions, but if they do happen they
can flood the console and render the system unusable. So switch them
to WARN_ON_ONCE().

Fixes: e2b36d59 ("powerpc/64: Don't trace code that runs with the soft irq mask unreconciled")
Fixes: 9b81c021 ("powerpc/64s: make PACA_IRQ_HARD_DIS track MSR[EE] closely")
Fixes: 7c0482e3 ("powerpc/irq: Fix another case of lazy IRQ state getting out of sync")
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20190708061046.7075-1-mpe@ellerman.id.au

To increase readability/maintainability, replace hard coded
instructions values by symbolic names.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
[mpe: Fix R_PPC64_ENTRY case, the addi reads from r2 not r12]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

To increase readability/maintainability, replace hard coded
instructions values by symbolic names.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

PPC_HA() PPC_HI() and PPC_LO() macros are nice macros. Move them
from module64.c to ppc-opcode.h in order to use them in other places.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
[mpe: Clean up formatting in new code, drop duplicates in ftrace.c]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

The comment here is wrong, the addi reads from r2 not r12. The code is
correct, 0x38420000 = addi r2,r2,0.

Fixes: a61674bd ("powerpc/module: Handle R_PPC64_ENTRY relocations")
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

This patch allows to generate lzo compressed uImage
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

This patch allows to generate lzma compressed uImage
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

This patch modifies the generation of uImage by handing over
the selected compression type instead of forcing gzip
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

Some SCC functions like the QMC requires an extended parameter RAM.
On modern 8xx (ie 866 and 885), SPI area can already be relocated,
allowing the use of those functions on SCC2. But SCC3 and SCC4
parameter RAM collide with SMC1 and SMC2 parameter RAMs.

This patch adds microcode to allow the relocation of both SMC1 and
SMC2, and relocate them at offsets 0x1ec0 and 0x1fc0.
Those offsets are by default for the CPM1 DSP1 and DSP2, but there
is no kernel driver using them at the moment so this area can be
reused.

This microcode is provided by Freescale/NXP in Engineering Bulletin
EB662 ("MPC8xx I2C/SPI and SMC Relocation Microcode Packages")
dated 2006. The binary code is public. The source is not available.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

Change microcode functions to use IO accessors and get rid
of volatile attributes.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

Reduce #ifdef mess by using IS_ENABLED() instead.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

The CPM registers RCCR and CPMCR1..4 registers has to be set in
accordance with the microcode patch beeing programmed. Lets
define them as part of the patch set and refactor their
programming from that definition.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

Define patch name together with the patch code, and refactor
the associated printk() while replacing it by a pr_info()
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

Add empty microcode tables so that all tables are defined
all the time. Regroup the writing of the 3 tables regardless
of the selected microcode.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

Create a function to refactor the writing of CPM microcode arrays.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

Compact obscure microcode arrays by putting 4 values per line
in order to reduce number of lines in the file to increase
readability.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

verify_patch() has been opted out since many years, and
the comment suggests it doesn't work. So drop it.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

Only 8xx selects CPM1 and related CONFIG options are already
in platforms/8xx/Kconfig

Move the related C files to platforms/8xx/.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
[mpe: Minor formatting fixes]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

This patch drops the assembly PPC64 version of flush_dcache_range()
and re-uses the PPC32 static inline version.

With GCC 8.1, the following code is generated:

void flush_test(unsigned long start, unsigned long stop)
{
	flush_dcache_range(start, stop);
}

0000000000000130 <.flush_test>:
 130:	3d 22 00 00 	addis   r9,r2,0
			132: R_PPC64_TOC16_HA	.data+0x8
 134:	81 09 00 00 	lwz     r8,0(r9)
			136: R_PPC64_TOC16_LO	.data+0x8
 138:	3d 22 00 00 	addis   r9,r2,0
			13a: R_PPC64_TOC16_HA	.data+0xc
 13c:	80 e9 00 00 	lwz     r7,0(r9)
			13e: R_PPC64_TOC16_LO	.data+0xc
 140:	7d 48 00 d0 	neg     r10,r8
 144:	7d 43 18 38 	and     r3,r10,r3
 148:	7c 00 04 ac 	hwsync
 14c:	4c 00 01 2c 	isync
 150:	39 28 ff ff 	addi    r9,r8,-1
 154:	7c 89 22 14 	add     r4,r9,r4
 158:	7c 83 20 50 	subf    r4,r3,r4
 15c:	7c 89 3c 37 	srd.    r9,r4,r7
 160:	41 82 00 1c 	beq     17c <.flush_test+0x4c>
 164:	7d 29 03 a6 	mtctr   r9
 168:	60 00 00 00 	nop
 16c:	60 00 00 00 	nop
 170:	7c 00 18 ac 	dcbf    0,r3
 174:	7c 63 42 14 	add     r3,r3,r8
 178:	42 00 ff f8 	bdnz    170 <.flush_test+0x40>
 17c:	7c 00 04 ac 	hwsync
 180:	4c 00 01 2c 	isync
 184:	4e 80 00 20 	blr
 188:	60 00 00 00 	nop
 18c:	60 00 00 00 	nop
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

This patch defines C helpers to retrieve the size of
cache blocks and uses them in the cacheflush functions.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

On most arches having function flush_dcache_range(), including PPC32,
this function does a writeback and invalidation of the cache bloc.

On PPC64, flush_dcache_range() only does a writeback while
flush_inval_dcache_range() does the invalidation in addition.

In addition it looks like within arch/powerpc/, there are no PPC64
platforms using flush_dcache_range()

This patch drops the existing 64 bits version of flush_dcache_range()
and renames flush_inval_dcache_range() into flush_dcache_range().
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

Cache instructions (dcbz, dcbi, dcbf and dcbst) take two registers
that are summed to obtain the target address. Using 'Z' constraint
and '%y0' argument gives GCC the opportunity to use both registers
instead of only one with the second being forced to 0.
Suggested-by: Segher Boessenkool <segher@kernel.crashing.org>
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

This makes sure we don't enable HugeTLB if the cache is not configured.
I am still not sure about this. IMHO hugetlb support should be a hardware
support derivative and any cache allocation failure should be handled as I did
in the earlier patch. But then if we were not able to create hugetlb page table
cache, we can as well declare hugetlb support disabled thereby avoiding calling
into allocation routines.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

We only check for hugetlb allocations, because with hugetlb we do conditional
registration. For PGD/PUD/PMD levels we register them always in
pgtable_cache_init.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

This fixes kernel crash that arises due to not handling page table allocation
failures while allocating hugetlb page table.

Fixes: e2b3d202 ("powerpc: Switch 16GB and 16MB explicit hugepages to a different page table format")
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

Now that we have switched the page table walk to use pmd_is_leaf we can now
revert commit 8adddf34 ("powerpc/mm/radix: Make Radix require HUGETLB_PAGE")
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

large devmap usage is dependent on THP. Hence once check is sufficient.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

Even when we have HugeTLB and THP disabled, kernel linear map can still be
mapped with hugepages. This is only an issue with radix translation because hash
MMU doesn't map kernel linear range in linux page table and other kernel
map areas are not mapped using hugepage.

Add config independent helpers and put WARN_ON() when we don't expect things
to be mapped via hugepages.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

We used uuid_parse to convert uuid string from device tree to two u64
components. We want to make sure we look at the uuid read from device
tree in an endian-neutral fashion. For now, I am picking little-endian
to be format so that we don't end up doing an additional conversion.

The reason to store in a specific endian format is to enable reading
the namespace created with a little-endian kernel config on a
big-endian kernel. We do store the device tree uuid string as a 64-bit
little-endian cookie in the label area. When booting the kernel we
also compare this cookie against what is read from the device tree.
For this, to work we have to store and compare these values in a CPU
endian config independent fashion.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

SCM_READ/WRITE_MEATADATA hcall supports multibyte read/write. This patch
updates the metadata read/write to use 1, 2, 4 or 8 byte read/write as
mentioned in PAPR document.

READ/WRITE_METADATA hcall supports the 1, 2, 4, or 8 bytes read/write.
For other values hcall results H_P3.

Hypervisor stores the metadata contents in big-endian format and in-order
to enable read/write in different granularity, we need to switch the contents
to big-endian before calling HCALL.

Based on an patch from Oliver O'Halloran <oohall@gmail.com>
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

The device tree node is documented as below:

  “ibm,cache-flush-required”:
  property name indicates Cache Flush Required for this Persistent Memory Segment to persist memory
  prop-encoded-array: None, this is a name only property.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

Allocation from altmap area can fail based on vmemmap page size used.
Add kernel info message to indicate the failure. That allows the user
to identify whether they are really using persistent memory reserved
space for per-page metadata.

The message looks like:
  [  136.587212] altmap block allocation failed, falling back to system memory
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Reviewed-by: Oliver O'Halloran <oohall@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

If we fail to parse min_common_depth from device tree we boot with
numa disabled. Reflect the same by updating numa_enabled variable
to false. Also, switch all min_common_depth failure check to
if (!numa_enabled) check.

This helps us to avoid checking for both in different code paths.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

If we boot with numa=off, we need to make sure we return NUMA_NO_NODE when
looking up associativity details of resources. Without this, we hit crash
like below

BUG: Unable to handle kernel data access at 0x40000000008
Faulting instruction address: 0xc000000008f31704
cpu 0x1b: Vector: 380 (Data SLB Access) at [c00000000b9bb320]
    pc: c000000008f31704: _raw_spin_lock+0x14/0x100
    lr: c0000000083f41fc: ____cache_alloc_node+0x5c/0x290
    sp: c00000000b9bb5b0
   msr: 800000010280b033
   dar: 40000000008
  current = 0xc00000000b9a2700
  paca    = 0xc00000000a740c00   irqmask: 0x03   irq_happened: 0x01
    pid   = 1, comm = swapper/27
Linux version 5.2.0-rc4-00925-g74e188c620b1 (root@linux-d8ip) (gcc version 7.4.1 20190424 [gcc-7-branch revision 270538] (SUSE Linux)) #34 SMP Sat Jun 29 00:41:02 EDT 2019
enter ? for help
[link register   ] c0000000083f41fc ____cache_alloc_node+0x5c/0x290
[c00000000b9bb5b0] 0000000000000dc0 (unreliable)
[c00000000b9bb5f0] c0000000083f48c8 kmem_cache_alloc_node_trace+0x138/0x360
[c00000000b9bb670] c000000008aa789c devres_alloc_node+0x4c/0xa0
[c00000000b9bb6a0] c000000008337218 devm_memremap+0x58/0x130
[c00000000b9bb6f0] c000000008aed00c devm_nsio_enable+0xdc/0x170
[c00000000b9bb780] c000000008af3b6c nd_pmem_probe+0x4c/0x180
[c00000000b9bb7b0] c000000008ad84cc nvdimm_bus_probe+0xac/0x260
[c00000000b9bb840] c000000008aa0628 really_probe+0x148/0x500
[c00000000b9bb8d0] c000000008aa0d7c driver_probe_device+0x19c/0x1d0
[c00000000b9bb950] c000000008aa11bc device_driver_attach+0xcc/0x100
[c00000000b9bb990] c000000008aa12ec __driver_attach+0xfc/0x1e0
[c00000000b9bba10] c000000008a9d0a4 bus_for_each_dev+0xb4/0x130
[c00000000b9bba70] c000000008a9fc04 driver_attach+0x34/0x50
[c00000000b9bba90] c000000008a9f118 bus_add_driver+0x1d8/0x300
[c00000000b9bbb20] c000000008aa2358 driver_register+0x98/0x1a0
[c00000000b9bbb90] c000000008ad7e6c __nd_driver_register+0x5c/0x100
[c00000000b9bbbf0] c0000000093efbac nd_pmem_driver_init+0x34/0x48
[c00000000b9bbc10] c0000000080106c0 do_one_initcall+0x60/0x2d0
[c00000000b9bbce0] c00000000938463c kernel_init_freeable+0x384/0x48c
[c00000000b9bbdb0] c000000008010a5c kernel_init+0x2c/0x160
[c00000000b9bbe20] c00000000800ba54 ret_from_kernel_thread+0x5c/0x68
Reported-and-debugged-by: Vaibhav Jain <vaibhav@linux.ibm.com>
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

If we fail to parse the associativity array we should default to
NUMA_NO_NODE instead of NODE 0. Rest of the code fallback to the
right default if we find the numa node value NUMA_NO_NODE.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

We use mmu_vmemmap_psize to find the page size for mapping the vmmemap area.
With radix translation, we are suboptimally setting this value to PAGE_SIZE.

We do check for 2M page size support and update mmu_vmemap_psize to use
hugepage size but we suboptimally reset the value to PAGE_SIZE in
radix__early_init_mmu(). This resulted in always mapping vmemmap area with
64K page size.

Fixes: 2bfd65e4 ("powerpc/mm/radix: Add radix callbacks for early init routines")
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

With hash translation and 4K PAGE_SIZE config, we need to make sure we don't
use 64K page size for vmemmap.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

Since commit 0034d395 ("powerpc/mm/hash64: Map all the kernel
regions in the same 0xc range") __kernel_virt_size is not used
anymore.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

Add a document describing the fields provided by
/proc/powerpc/vcpudispatch_stats.
Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>

When enabling or disabling the vcpu dispatch statistics, we do a lot of
work including allocating/deallocating memory across all possible cpus
for the DTL buffer. In order to guard against hogging the cpu for too
long, track the time we're taking and yield the processor if necessary.
Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>