Book3e defines both _PAGE_USER and _PAGE_PRIVILEGED, so the nohash
default pte_mkprivileged() and pte_mkuser() are not usable.
This patch redefines them for book3e.
In theorie, only pte_mkprivileged() needs to be redefined because
_PAGE_USER includes _PAGE_PRIVILEGED, but it is less confusing
to redefine both.
Fixes: a0da4bc166 ("powerpc/mm: Allow platforms to redefine some helpers")
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Other archs do the same and instead of adding required pte bits (which
got masked out) in __ioremap_at(), make sure we filter only pfn bits
out.
Fixes: 26973fa5ac ("powerpc/mm: use pte helpers in generic code")
Reviewed-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Currently we limit the max addressable memory to 128TB. This patch increase the
limit to 2PB. We can have devices like nvdimm which adds memory above 512TB
limit.
We still don't support regular system ram above 512TB. One of the challenge with
that is the percpu allocator, that allocates per node memory and use the max
distance between them as the percpu offsets. This means with large gap in
address space ( system ram above 1PB) we will run out of vmalloc space to map
the percpu allocation.
In order to support addressable memory above 512TB, kernel should be able to
linear map this range. To do that with hash translation we now add 4 context
to kernel linear map region. Our per context addressable range is 512TB. We
still keep VMALLOC and VMEMMAP region to old size. SLB miss handlers is updated
to validate these limit.
We also limit this update to SPARSEMEM_VMEMMAP and SPARSEMEM_EXTREME
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
We will be adding get_kernel_context later. Update function name to indicate
this handle context allocation user space address.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
This adds CONFIG_DEBUG_VM checks to ensure:
- The kernel stack is in the SLB after it's flushed and bolted.
- We don't insert an SLB for an address that is aleady in the SLB.
- The kernel SLB miss handler does not take an SLB miss.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
slb_flush_and_rebolt() is misleading, it is called in virtual mode, so
it can not possibly change the stack, so it should not be touching the
shadow area. And since vmalloc is no longer bolted, it should not
change any bolted mappings at all.
Change the name to slb_flush_and_restore_bolted(), and have it just
load the kernel stack from what's currently in the shadow SLB area.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
When switching processes, currently all user SLBEs are cleared, and a
few (exec_base, pc, and stack) are preloaded. In trivial testing with
small apps, this tends to miss the heap and low 256MB segments, and it
will also miss commonly accessed segments on large memory workloads.
Add a simple round-robin preload cache that just inserts the last SLB
miss into the head of the cache and preloads those at context switch
time. Every 256 context switches, the oldest entry is removed from the
cache to shrink the cache and require fewer slbmte if they are unused.
Much more could go into this, including into the SLB entry reclaim
side to track some LRU information etc, which would require a study of
large memory workloads. But this is a simple thing we can do now that
is an obvious win for common workloads.
With the full series, process switching speed on the context_switch
benchmark on POWER9/hash (with kernel speculation security masures
disabled) increases from 140K/s to 178K/s (27%).
POWER8 does not change much (within 1%), it's unclear why it does not
see a big gain like POWER9.
Booting to busybox init with 256MB segments has SLB misses go down
from 945 to 69, and with 1T segments 900 to 21. These could almost all
be eliminated by preloading a bit more carefully with ELF binary
loading.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
This will be used by the SLB code in the next patch, but for now this
sets the slb_addr_limit to the correct size for 32-bit tasks.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Add 32-entry bitmaps to track the allocation status of the first 32
SLB entries, and whether they are user or kernel entries. These are
used to allocate free SLB entries first, before resorting to the round
robin allocator.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
This patch moves SLB miss handlers completely to C, using the standard
exception handler macros to set up the stack and branch to C.
This can be done because the segment containing the kernel stack is
always bolted, so accessing it with relocation on will not cause an
SLB exception.
Arbitrary kernel memory must not be accessed when handling kernel
space SLB misses, so care should be taken there. However user SLB
misses can access any kernel memory, which can be used to move some
fields out of the paca (in later patches).
User SLB misses could quite easily reconcile IRQs and set up a first
class kernel environment and exit via ret_from_except, however that
doesn't seem to be necessary at the moment, so we only do that if a
bad fault is encountered.
[ Credit to Aneesh for bug fixes, error checks, and improvements to
bad address handling, etc ]
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
[mpe: Disallow tracing for all of slb.c for now.]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
PPR is the odd register out when it comes to interrupt handling, it is
saved in current->thread.ppr while all others are saved on the stack.
The difficulty with this is that accessing thread.ppr can cause a SLB
fault, but the SLB fault handler implementation in C change had
assumed the normal exception entry handlers would not cause an SLB
fault.
Fix this by allocating room in the interrupt stack to save PPR.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Now that we've split the user & kernel versions of pt_regs we need to
be more careful in the ptrace code.
For now we've ensured the location of the fields in both structs is
the same, so most of the ptrace code doesn't need updating.
But there are a few places where we use sizeof(pt_regs), and these
will be wrong as soon as we increase the size of the kernel structure.
So flip them all to use sizeof(user_pt_regs).
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
We use a shared definition for struct pt_regs in uapi/asm/ptrace.h.
That means the layout of the structure is ABI, ie. we can't change it.
That would be fine if it was only used to describe the user-visible
register state of a process, but it's also the struct we use in the
kernel to describe the registers saved in an interrupt frame.
We'd like more flexibility in the content (and possibly layout) of the
kernel version of the struct, but currently that's not possible.
So split the definition into a user-visible definition which remains
unchanged, and a kernel internal one.
At the moment they're still identical, and we check that at build
time. That's because we have code (in ptrace etc.) that assumes that
they are the same. We will fix that code in future patches, and then
we can break the strict symmetry between the two structs.
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
In the same spirit as already done in pte query helpers,
this patch changes pte setting helpers to perform endian
conversions on the constants rather than on the pte value.
In the meantime, it changes pte_access_permitted() to use
pte helpers for the same reason.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
_PAGE_PRIVILEGED corresponds to the SH bit which doesn't protect
against user access but only disables ASID verification on kernel
accesses. User access is controlled with _PMD_USER flag.
Name it _PAGE_SH instead of _PAGE_PRIVILEGED
_PAGE_HUGE corresponds to the SPS bit which doesn't really tells
that's it is a huge page but only that it is not a 4k page.
Name it _PAGE_SPS instead of _PAGE_HUGE
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Do not include pte-common.h in nohash/32/pgtable.h
As that was the last includer, get rid of pte-common.h
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Cache related flags like _PAGE_COHERENT and _PAGE_WRITETHRU
are defined on most platforms. The platforms not defining
them don't define any alternative. So we can give them a NUL
value directly for those platforms directly.
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The 40xx defines _PAGE_HWWRITE while others don't.
The 8xx defines _PAGE_RO instead of _PAGE_RW.
The 8xx defines _PAGE_PRIVILEGED instead of _PAGE_USER.
The 8xx defines _PAGE_HUGE and _PAGE_NA while others don't.
Lets those platforms redefine pte_write(), pte_wrprotect() and
pte_mkwrite() and get _PAGE_RO and _PAGE_HWWRITE off the common
helpers.
Lets the 8xx redefine pte_user(), pte_mkprivileged() and pte_mkuser()
and get rid of _PAGE_PRIVILEGED and _PAGE_USER default values.
Lets the 8xx redefine pte_mkhuge() and get rid of
_PAGE_HUGE default value.
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
nohash/64 only uses book3e PTE flags, so it doesn't need pte-common.h
This also allows to drop PAGE_SAO and H_PAGE_4K_PFN from pte_common.h
as they are only used by PPC64
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The base kernel PAGE_XXXX definition sets are more or less platform
specific. Lets distribute them close to platform _PAGE_XXX flags
definition, and customise them to their exact platform flags.
Also defines _PAGE_PSIZE and _PTE_NONE_MASK for each platform
allthough they are defined as 0.
Do the same with _PMD flags like _PMD_USER and _PMD_PRESENT_MASK
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Now the pte-common.h is only for nohash platforms, lets
move pte_user() helper out of pte-common.h to put it
together with other helpers.
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
As done for book3s/64, add necessary flags/defines in
book3s/32/pgtable.h and do not include pte-common.h
It allows in the meantime to remove all related hash
definitions from pte-common.h and to also remove
_PAGE_EXEC default as _PAGE_EXEC is defined on all
platforms except book3s/32.
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
__P and __S flags are the same for all platform and should remain
as is in the future, so avoid duplication.
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The following page flags in pte-common.h can be dropped:
_PAGE_ENDIAN is only used in mm/fsl_booke_mmu.c and is defined in
asm/nohash/32/pte-fsl-booke.h
_PAGE_4K_PFN is nowhere defined nor used
_PAGE_READ, _PAGE_WRITE and _PAGE_PTE are only defined and used
in book3s/64
The following page flags in book3s/64/pgtable.h can be dropped as
they are not used on this platform nor by common code.
_PAGE_NA, _PAGE_RO, _PAGE_USER and _PAGE_PSIZE
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
To reduce the complexity of flag_array, and allow the removal of
default 0 value of non existing flags, lets have one flag_array
table for each platform family with only the really existing flags.
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Get rid of platform specific _PAGE_XXXX in powerpc common code and
use helpers instead.
mm/dump_linuxpagetables.c will be handled separately
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The 'access' parameter of hash_preload() is either 0 or _PAGE_EXEC.
Among the two versions of hash_preload(), only the PPC64 one is
doing something with this 'access' parameter.
In order to remove the use of _PAGE_EXEC outside platform code,
'access' parameter is replaced by 'is_exec' which will be either
true of false, and the PPC64 version of hash_preload() creates
the access flag based on 'is_exec'.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
In order to avoid using generic _PAGE_XXX flags in powerpc
core functions, define helpers for all needed flags:
- pte_mkuser() and pte_mkprivileged() to set/unset and/or
unset/set _PAGE_USER and/or _PAGE_PRIVILEGED
- pte_hashpte() to check if _PAGE_HASHPTE is set.
- pte_ci() check if cache is inhibited (already existing on book3s/64)
- pte_exprotect() to protect against execution
- pte_exec() and pte_mkexec() to query and set page execution
- pte_mkpte() to set _PAGE_PTE flag.
- pte_hw_valid() to check _PAGE_PRESENT since pte_present does
something different on book3s/64.
On book3s/32 there is no exec protection, so pte_mkexec() and
pte_exprotect() are nops and pte_exec() returns always true.
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
In order to allow their use in nohash/32/pgtable.h, we have to move the
following helpers in nohash/[32:64]/pgtable.h:
- pte_mkwrite()
- pte_mkdirty()
- pte_mkyoung()
- pte_wrprotect()
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
book3s/32 doesn't define _PAGE_EXEC, so no need to use it.
All other platforms define _PAGE_EXEC so no need to check
it is not NUL when not book3s/32.
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
In order to avoid multiple conversions, handover directly a
pgprot_t to map_kernel_page() as already done for radix.
Do the same for __ioremap_caller() and __ioremap_at().
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Set PAGE_KERNEL directly in the caller and do not rely on a
hack adding PAGE_KERNEL flags when _PAGE_PRESENT is not set.
As already done for PPC64, use pgprot_cache() helpers instead of
_PAGE_XXX flags in PPC32 ioremap() derived functions.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
In many places, ioremap_prot() and __ioremap() can be replaced with
higher level functions like ioremap(), ioremap_coherent(),
ioremap_cache(), ioremap_wc() ...
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Other arches have ioremap_wt() to map IO areas write-through.
Implement it on PPC as well in order to avoid drivers using
__ioremap(_PAGE_WRITETHRU)
Also implement ioremap_coherent() to avoid drivers using
__ioremap(_PAGE_COHERENT)
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Live Partition Migrations require all the present CPUs to execute the
H_JOIN call, and hence rtas_ibm_suspend_me() onlines any offline CPUs
before initiating the migration for this purpose.
The commit 85a88cabad
("powerpc/pseries: Disable CPU hotplug across migrations")
disables any CPU-hotplug operations once all the offline CPUs are
brought online to prevent any further state change. Once the
CPU-Hotplug operation is disabled, the code assumes that all the CPUs
are online.
However, there is a minor window in rtas_ibm_suspend_me() between
onlining the offline CPUs and disabling CPU-Hotplug when a concurrent
CPU-offline operations initiated by the userspace can succeed thereby
nullifying the the aformentioned assumption. In this unlikely case
these offlined CPUs will not call H_JOIN, resulting in a system hang.
Fix this by verifying that all the present CPUs are actually online
after CPU-Hotplug has been disabled, failing which we restore the
state of the offline CPUs in rtas_ibm_suspend_me() and return an
-EBUSY.
Cc: Nathan Fontenot <nfont@linux.vnet.ibm.com>
Cc: Tyrel Datwyler <tyreld@linux.vnet.ibm.com>
Suggested-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Gautham R. Shenoy <ego@linux.vnet.ibm.com>
Reviewed-by: Nathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Currently on POWER9 SMT8 cores systems, in sysfs, we report the
shared_cache_map for L1 caches (both data and instruction) to be the
cpu-ids of the threads in SMT8 cores. This is incorrect since on
POWER9 SMT8 cores there are two groups of threads, each of which
shares its own L1 cache.
This patch addresses this by reporting the shared_cpu_map correctly in
sysfs for L1 caches.
Before the patch
/sys/devices/system/cpu/cpu0/cache/index0/shared_cpu_map : 000000ff
/sys/devices/system/cpu/cpu0/cache/index1/shared_cpu_map : 000000ff
/sys/devices/system/cpu/cpu1/cache/index0/shared_cpu_map : 000000ff
/sys/devices/system/cpu/cpu1/cache/index1/shared_cpu_map : 000000ff
After the patch
/sys/devices/system/cpu/cpu0/cache/index0/shared_cpu_map : 00000055
/sys/devices/system/cpu/cpu0/cache/index1/shared_cpu_map : 00000055
/sys/devices/system/cpu/cpu1/cache/index0/shared_cpu_map : 000000aa
/sys/devices/system/cpu/cpu1/cache/index1/shared_cpu_map : 000000aa
Signed-off-by: Gautham R. Shenoy <ego@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
POWER9 SMT8 cores consist of two groups of threads, where threads in
each group shares L1-cache. The scheduler is not aware of this
distinction as the current sched-domain hierarchy has all the threads
of the core defined at the SMT domain.
SMT [Thread siblings of the SMT8 core]
DIE [CPUs in the same die]
NUMA [All the CPUs in the system]
Due to this, we can observe run-to-run variance when we run a
multi-threaded benchmark bound to a single core based on how the
scheduler spreads the software threads across the two groups in the
core.
We fix this in this patch by defining each group of threads which
share L1-cache to be the SMT level. The group of threads in the SMT8
core is defined to be the CACHE level. The sched-domain hierarchy
after this patch will be :
SMT [Thread siblings in the core that share L1 cache]
CACHE [Thread siblings that are in the SMT8 core]
DIE [CPUs in the same die]
NUMA [All the CPUs in the system]
Signed-off-by: Gautham R. Shenoy <ego@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
On IBM POWER9, the device tree exposes a property array identifed by
"ibm,thread-groups" which will indicate which groups of threads share
a particular set of resources.
As of today we only have one form of grouping identifying the group of
threads in the core that share the L1 cache, translation cache and
instruction data flow.
This patch adds helper functions to parse the contents of
"ibm,thread-groups" and populate a per-cpu variable to cache
information about siblings of each CPU that share the L1, traslation
cache and instruction data-flow.
It also defines a new global variable named "has_big_cores" which
indicates if the cores on this configuration have multiple groups of
threads that share L1 cache.
For each online CPU, it maintains a cpu_smallcore_mask, which
indicates the online siblings which share the L1-cache with it.
Signed-off-by: Gautham R. Shenoy <ego@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
If GCC is not built with glibc support then we must explicitly tell it
which register to use for TLS mode stack protector, otherwise it will
error out and the cc-option check will fail.
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Commit 6c1719942e ("powerpc/of: Remove useless register save/restore
when calling OF back") removed the saving of srr0 and srr1 when calling
into OpenFirmware. Commit e31aa453bb ("powerpc: Use LOAD_REG_IMMEDIATE
only for constants on 64-bit") did the same for rtas.
This means we don't need to save the extra stack space and can use
the common SWITCH_FRAME_SIZE.
There were already no users of _SRR0 and _SRR1 so we can remove them
too.
Link: https://github.com/linuxppc/linux/issues/83
Signed-off-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The powerpc mobility code may receive RTAS requests to perform PRRN
(Platform Resource Reassignment Notification) topology changes at any
time, including during LPAR migration operations.
In some configurations where the affinity of CPUs or memory is being
changed on that platform, the PRRN requests may apply or refer to
outdated information prior to the complete update of the device-tree.
This patch changes the duration for which topology updates are
suppressed during LPAR migrations from just the rtas_ibm_suspend_me()
/ 'ibm,suspend-me' call(s) to cover the entire migration_store()
operation to allow all changes to the device-tree to be applied prior
to accepting and applying any PRRN requests.
For tracking purposes, pr_info notices are added to the functions
start_topology_update() and stop_topology_update() of 'numa.c'.
Signed-off-by: Michael Bringmann <mwb@linux.vnet.ibm.com>
Reviewed-by: Nathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Ever since commit 15a3204d24 ("powerpc/64s: Set assembler machine type
to POWER4") we force -mpower4 to be passed to the assembler
irrespective of the CFLAGS used (for Book3s 64).
When building a powerpc64 kernel with clang, clang will not add -many
to the assembler flags, so any instructions that the compiler has
generated that are not available on power4 will cause an error:
/usr/bin/as -a64 -mppc64 -mlittle-endian -mpower8 \
-I ./arch/powerpc/include -I ./arch/powerpc/include/generated \
-I ./include -I ./arch/powerpc/include/uapi \
-I ./arch/powerpc/include/generated/uapi -I ./include/uapi \
-I ./include/generated/uapi -I arch/powerpc -I arch/powerpc \
-maltivec -mpower4 -o init/do_mounts.o /tmp/do_mounts-3b0a3d.s
/tmp/do_mounts-51ce54.s:748: Error: unrecognized opcode: `isel'
GCC does include -many, so the GCC driven gas call will succeed:
as -v -I ./arch/powerpc/include -I ./arch/powerpc/include/generated -I
./include -I ./arch/powerpc/include/uapi
-I ./arch/powerpc/include/generated/uapi -I ./include/uapi
-I ./include/generated/uapi -I arch/powerpc -I arch/powerpc
-a64 -mpower8 -many -mlittle -maltivec -mpower4 -o init/do_mounts.o
Note that isel is power7 and above for IBM CPUs. GCC only generates it
for Power9 and above, but the above test was run against the clang
generated assembly.
Peter Bergner explains:
When using -many -mpower4, gas will first try and find a matching
power4 mnemonic and failing that, it will then allow any valid
mnemonic that gas knows about. GCC's use of -many predates me
though.
IIRC, Alan looked at trying to remove it, but I forget why he
didn't. Could be either a gcc or gas issue at the time. I'm not sure
whether issue still exists or not. He and I have modified how gas
works internally a fair amount since he tried removing gcc use of
-many.
I will also note that when using -many, gas will choose the first
mnemonic that matches in the mnemonic table and we have (mostly)
sorted the table so that server mnemonics show up earlier in the
table than other mnemonics, so they'll be seen/chosen first.
By explicitly setting -many we can build with Clang and GCC while
retaining the -mpower4 option.
Signed-off-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The variable 'aa_index' is defined as an unsigned value in
update_lmb_associativity_index(), but find_aa_index() may return -1
when dlpar_clone_property() fails. So change find_aa_index() to return
a bool, which indicates whether 'aa_index' was found or not.
Fixes: c05a5a4096 ("powerpc/pseries: Dynamic add entires to associativity lookup array")
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Reviewed-by: Nathan Fontenot nfont@linux.vnet.ibm.com>
[mpe: Tweak changelog, rename is_found to just found]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Rather than mixing "if (state)" blocks and gotos, convert entirely to
"if (state)" blocks to make the state machine behaviour clearer.
Signed-off-by: Sam Bobroff <sbobroff@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The wait_state member of eeh_ops does not need to be platform
dependent; it's just logic around eeh_ops.get_state(). Therefore,
merge the two (slightly different!) platform versions into a new
function, eeh_wait_state() and remove the eeh_ops member.
While doing this, also correct:
* The wait logic, so that it never waits longer than max_wait.
* The wait logic, so that it never waits less than
EEH_STATE_MIN_WAIT_TIME.
* One call site where the result is treated like a bit field before
it's checked for negative error values.
* In pseries_eeh_get_state(), rename the "state" parameter to "delay"
because that's what it is.
Signed-off-by: Sam Bobroff <sbobroff@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Currently, eeh_pe_state_mark() marks a PE (and it's children) with a
state and then performs additional processing if that state included
EEH_PE_ISOLATED.
The state parameter is always a constant at the call site, so
rearrange eeh_pe_state_mark() into two functions and just call the
appropriate one at each site.
Signed-off-by: Sam Bobroff <sbobroff@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The function eeh_pe_state_mark_with_cfg() just performs the work of
eeh_pe_state_mark() and then, conditionally, the work of
eeh_pe_state_clear(). However it is only ever called with a constant
state such that the condition is always true, so replace it by direct
calls.
Signed-off-by: Sam Bobroff <sbobroff@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Move the call to eeh_dev_to_pe() up, so that later it's clear that
"pe" isn't NULL.
Signed-off-by: Sam Bobroff <sbobroff@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Change the name of the fields in eeh_rmv_data to clarify their usage.
Change "edev_list" to "removed_vf_list" because it does not contain
generic edevs, but rather only edevs that contain virtual functions
(which need to be removed during recovery).
Similarly, change "removed" to "removed_dev_count" because it is a
count of any removed devices, not just those in the above list.
Signed-off-by: Sam Bobroff <sbobroff@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Instances of struct eeh_pe are placed in a tree structure using the
fields "child_list" and "child", so place these next to each other
in the definition.
The field "child" is a list entry, so remove the unnecessary and
misleading use of the list initializer, LIST_HEAD(), on it.
The eeh_dev struct contains two list entry fields, called "list" and
"rmv_list". Rename them to "entry" and "rmv_entry" and, as above, stop
initializing them with LIST_HEAD().
Signed-off-by: Sam Bobroff <sbobroff@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Remove the unnecessary cast through void * on the first parameter and
remove the unused second parameter (always NULL).
Signed-off-by: Sam Bobroff <sbobroff@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The 'bus' member of struct eeh_dev is assigned to once but never used,
so remove it.
Signed-off-by: Sam Bobroff <sbobroff@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Currently a flag, EEH_POSTPONED_PROBE, is used to prevent an incorrect
message "EEH: No capable adapters found" from being displayed during
the boot of powernv systems.
It is necessary because, on powernv, the call to eeh_probe_devices()
made from eeh_init() is too early and EEH can't yet be enabled. A
second call is made later from eeh_pnv_post_init(), which succeeds.
(On pseries, the first call succeeds because PCI devices are set up
early enough and no second call is made.)
This can be simplified by moving the early call to eeh_probe_devices()
from eeh_init() (where it's seen by both platforms) to
pSeries_final_fixup(), so that each platform only calls
eeh_probe_devices() once, at a point where it can succeed.
This is slightly later in the boot sequence, but but still early
enough and it is now in the same place in the sequence for both
platforms (the pcibios_fixup hook).
The display of the message can be cleaned up as well, by moving it
into eeh_probe_devices().
Signed-off-by: Sam Bobroff <sbobroff@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
eeh_add_to_parent_pe() sometimes removes the EEH_PE_KEEP flag, but it
incorrectly removes it from pe->type, instead of pe->state.
However, rather than clearing it from the correct field, remove it.
Inspection of the code shows that it can't ever have had any effect
(even if it had been cleared from the correct field), because the
field is never tested after it is cleared by the statement in
question.
The clear statement was added by commit 807a827d4e ("powerpc/eeh:
Keep PE during hotplug"), but it didn't explain why it was necessary.
Signed-off-by: Sam Bobroff <sbobroff@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
If a device is removed during EEH processing (either by a driver's
handler or as part of recovery), it can lead to a null dereference
in eeh_pe_report_edev().
To handle this, skip devices that have been removed.
Signed-off-by: Sam Bobroff <sbobroff@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
If an error occurs during an unplug operation, it's possible for
eeh_dump_dev_log() to be called when edev->pdn is null, which
currently leads to dereferencing a null pointer.
Handle this by skipping the error log for those devices.
Signed-off-by: Sam Bobroff <sbobroff@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The boot wrapper is currently built with -Os. By building with O2 we
can meaningfully reduce the time decompressing the kernel.
I tested by comparing 10 runs of each option in Qemu and on hardware.
The kernel is compressed with KERNEL_XZ built with GCC 8.2.0-7ubuntu1.
The values are counts of the timebase.
Qemu TCG powernv Power8:
Os O2 O3
median 10221123889 6201518438 6568186825
stddev 1361267211 429090641 657930076
improvement 39.33% 35.74%
Palmetto Power8:
Os O2 O3
median 50279 50599 35790
stddev 992144533 627130655 623721078
improvement 36.79% 37.13%
Romulus Power9:
Os O2 O3
median 670312391 454733720 448881398
stddev 157569 107276 108760
improvement 32.16% 33.03%
TCG was quite noisy, with every few runs producing an outlier. Even so,
O2 is faster than O3. On hardware the numbers were less noisy and O3 is
slightly faster than O2.
The wrapper size increases when moving from Os. Comparing zImage.epapr
to the existing Os build using bloat-o-meter:
Before=43401, After=56837 (13KB), chg +30.96%
Before=43401, After=64305 (20KB), chg +48.16%
I chose O2 for a balance between Qemu and hardware speed up.
Signed-off-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
This will avoid auto-vectorisation when building with higher
optimisation levels.
We don't know if the machine can support VSX and even if it's present
it's probably not going to be enabled at this point in boot.
These flag were both added prior to GCC 4.6 which is the minimum
compiler version supported by upstream, thanks to Segher for the
details.
Signed-off-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
As of commit 10c77dba40 ("powerpc/boot: Fix build failure in 32-bit
boot wrapper") the opal code is hidden behind CONFIG_PPC64_BOOT_WRAPPER,
but the boot wrapper avoids include/linux, so it does not get the normal
Kconfig flags.
We can drop the guard entirely as in commit f8e8e69cea ("powerpc/boot:
Only build OPAL code when necessary") the makefile only includes opal.c
in the build if CONFIG_PPC64_BOOT_WRAPPER is set.
Fixes: 10c77dba40 ("powerpc/boot: Fix build failure in 32-bit boot wrapper")
Signed-off-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Currently the wrapper is built without including anything in
$(src)/include/, which means there are no CONFIG_ symbols defined.
This means the platform specific serial drivers were never enabled.
We now copy the definitions into the boot directory, so any C file can
now include autoconf.h to depend on configuration options.
Fixes: 866bfc75f4 ("powerpc: conditionally compile platform-specific serial drivers")
Signed-off-by: Joel Stanley <joel@jms.id.au>
[mpe: Fix to use $(objtree) to find autoconf.h]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
'default n' is the default value for any bool or tristate Kconfig
setting so there is no need to write it explicitly.
Also since commit f467c5640c ("kconfig: only write '# CONFIG_FOO
is not set' for visible symbols") the Kconfig behavior is the same
regardless of 'default n' being present or not:
...
One side effect of (and the main motivation for) this change is making
the following two definitions behave exactly the same:
config FOO
bool
config FOO
bool
default n
With this change, neither of these will generate a
'# CONFIG_FOO is not set' line (assuming FOO isn't selected/implied).
That might make it clearer to people that a bare 'default n' is
redundant.
...
Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Currently when we get an unknown RTAS event it prints the type as
"Unknown" and no other useful information. Add the raw type code to the
log message so that we have something to work off.
Signed-off-by: Oliver O'Halloran <oohall@gmail.com>
Reviewed-by: Vasant Hegde <hegdevasant@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The powerpc kernel uses setjmp which causes a warning when building
with clang:
In file included from arch/powerpc/xmon/xmon.c:51:
./arch/powerpc/include/asm/setjmp.h:15:13: error: declaration of
built-in function 'setjmp' requires inclusion of the header <setjmp.h>
[-Werror,-Wbuiltin-requires-header]
extern long setjmp(long *);
^
./arch/powerpc/include/asm/setjmp.h:16:13: error: declaration of
built-in function 'longjmp' requires inclusion of the header <setjmp.h>
[-Werror,-Wbuiltin-requires-header]
extern void longjmp(long *, long);
^
This *is* the header and we're not using the built-in setjump but
rather the one in arch/powerpc/kernel/misc.S. As the compiler warning
does not make sense, it for the files where setjmp is used.
Signed-off-by: Joel Stanley <joel@jms.id.au>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
[mpe: Move subdir-ccflags in xmon/Makefile to not clobber -Werror]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The "count < sizeof(struct os_area_db)" comparison is type promoted to
size_t so negative values of "count" are treated as very high values
and we accidentally return success instead of a negative error code.
This doesn't really change runtime much but it fixes a static checker
warning.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Geoff Levand <geoff@infradead.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
On a Power9 box we get a few screens full of these on boot. Drop
them to pr_debug.
[ 5.993645] nest_centaur6_imc performance monitor hardware support registered
[ 5.993728] nest_centaur7_imc performance monitor hardware support registered
[ 5.996510] core_imc performance monitor hardware support registered
[ 5.996569] nest_mba0_imc performance monitor hardware support registered
[ 5.996631] nest_mba1_imc performance monitor hardware support registered
[ 5.996685] nest_mba2_imc performance monitor hardware support registered
Signed-off-by: Joel Stanley <joel@jms.id.au>
Reviewed-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Reviewed-by: Stewart Smith <stewart@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
instructions_to_print var is assigned value 16 and there is no
way to change it.
This patch replaces it by a constant.
Reviewed-by: Murilo Opsfelder Araujo <muriloo@linux.ibm.com>
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
When two processes crash at the same time, we sometimes encounter
interleaving in the middle of a line:
init[1]: segfault (11) at 0 nip 0 lr 0 code 1
init[1]: code: XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX
init[74]: segfault (11) at 10a74 nip 1000c198 lr 100078c8 code 1 in sh[10000000+14000]
XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX
init[1]: code: XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX
init[74]: code: 90010024 bf61000c 91490a7c 3fa01002 3be00000 7d3e4b78 3bbd0c20 3b600000
init[74]: code: 3b9d0040 7c7fe02e 2f830000 419e0028 <89230000> 2f890000 41be001c 4b7f6e79
This patch fixes it by preparing complete lines in a buffer and
printing it at once.
Fixes: 88b0fe1757 ("powerpc: Add show_user_instructions()")
Reviewed-by: Murilo Opsfelder Araujo <muriloo@linux.ibm.com>
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
[mpe: Use seq_buf_printf() not seq_buf_puts() which doesn't NULL terminate]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
As spotted by sparse:
arch/powerpc/kernel/process.c:1302:6: warning: symbol 'show_user_instructions' was not declared. Should it be static?
Fixes: 88b0fe1757 ("powerpc: Add show_user_instructions()")
Reviewed-by: Murilo Opsfelder Araujo <muriloo@linux.ibm.com>
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
[mpe: Split out of larger patch]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
This patch fixes the following warnings, which are leftovers
from when __get_user() was replaced by probe_kernel_address().
arch/powerpc/kernel/process.c:1287:22: warning: incorrect type in argument 2 (different address spaces)
arch/powerpc/kernel/process.c:1287:22: expected void const *src
arch/powerpc/kernel/process.c:1287:22: got unsigned int [noderef] <asn:1>*<noident>
arch/powerpc/kernel/process.c:1319:21: warning: incorrect type in argument 2 (different address spaces)
arch/powerpc/kernel/process.c:1319:21: expected void const *src
arch/powerpc/kernel/process.c:1319:21: got unsigned int [noderef] <asn:1>*<noident>
Fixes: 7b051f665c ("powerpc: Use probe_kernel_address in show_instructions")
Reviewed-by: Murilo Opsfelder Araujo <muriloo@linux.ibm.com>
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
[mpe: Split out of larger patch]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Now that the 68k Mac port has adopted the via-pmu driver, the same RTC
code can be shared between m68k and powerpc. Replace duplicated code in
arch/powerpc and arch/m68k with common RTC accessors for Cuda and PMU.
Drop the problematic WARN_ON which was introduced in commit 22db552b50
("powerpc/powermac: Fix rtc read/write functions").
Tested-by: Stan Johnson <userm57@yahoo.com>
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Acked-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Recently we implemented show_user_instructions() which dumps the code
around the NIP when a user space process dies with an unhandled
signal. This was modelled on the x86 code, and we even went so far as
to implement the exact same bug, namely that if the user process
crashed with its NIP pointing into the kernel we will dump kernel text
to dmesg. eg:
bad-bctr[2996]: segfault (11) at c000000000010000 nip c000000000010000 lr 12d0b0894 code 1
bad-bctr[2996]: code: fbe10068 7cbe2b78 7c7f1b78 fb610048 38a10028 38810020 fb810050 7f8802a6
bad-bctr[2996]: code: 3860001c f8010080 48242371 60000000 <7c7b1b79> 4082002c e8010080 eb610048
This was discovered on x86 by Jann Horn and fixed in commit
342db04ae7 ("x86/dumpstack: Don't dump kernel memory based on usermode RIP").
Fix it by checking the adjusted NIP value (pc) and number of
instructions against USER_DS, and bail if we fail the check, eg:
bad-bctr[2969]: segfault (11) at c000000000010000 nip c000000000010000 lr 107930894 code 1
bad-bctr[2969]: Bad NIP, not dumping instructions.
Fixes: 88b0fe1757 ("powerpc: Add show_user_instructions()")
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Local radix TLB flush operations that operate on congruence classes
have explicit ERAT flushes for POWER9. The process scoped LPID flush
did not have a flush, so add it.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
PPC_INVALIDATE_ERAT is slbia IH=7 which is a new variant introduced
with POWER9, and the result is undefined on earlier CPUs.
Commits 7b9f71f974 ("powerpc/64s: POWER9 machine check handler") and
d4748276ae ("powerpc/64s: Improve local TLB flush for boot and MCE on
POWER9") caused POWER7/8 code to use this instruction. Remove it. An
ERAT flush can be made by invalidatig the SLB, but before POWER9 that
requires a flush and rebolt.
Fixes: 7b9f71f974 ("powerpc/64s: POWER9 machine check handler")
Fixes: d4748276ae ("powerpc/64s: Improve local TLB flush for boot and MCE on POWER9")
Cc: stable@vger.kernel.org # v4.11+
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
If CONFIG_PPC_WATCHDOG is enabled we always cap the decrementer to
0x7fffffff:
if (IS_ENABLED(CONFIG_PPC_WATCHDOG))
set_dec(0x7fffffff);
else
set_dec(decrementer_max);
If there are no future events, we don't reprogram the decrementer
after this and we end up with 0x7fffffff even on a large decrementer
capable system.
As suggested by Nick, add a set_state_oneshot_stopped callback
so we program the decrementer with decrementer_max if there are
no future events.
Signed-off-by: Anton Blanchard <anton@ozlabs.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
We currently cap the decrementer clockevent at 4 seconds, even on systems
with large decrementer support. Fix this by converting the code to use
clockevents_register_device() which calculates the upper bound based on
the max_delta passed in.
Signed-off-by: Anton Blanchard <anton@ozlabs.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
This threshold is no longer used now that all invalidates issue a single
ATSD to each active NPU.
Signed-off-by: Mark Hairgrove <mhairgrove@nvidia.com>
Reviewed-by: Alistair Popple <alistair@popple.id.au>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Prior to this change only two types of ATSDs were issued to the NPU:
invalidates targeting a single page and invalidates targeting the whole
address space. The crossover point happened at the configurable
atsd_threshold which defaulted to 2M. Invalidates that size or smaller
would issue per-page invalidates for the whole range.
The NPU supports more invalidation sizes however: 64K, 2M, 1G, and all.
These invalidates target addresses aligned to their size. 2M is a common
invalidation size for GPU-enabled applications because that is a GPU
page size, so reducing the number of invalidates by 32x in that case is a
clear improvement.
ATSD latency is high in general so now we always issue a single invalidate
rather than multiple. This will over-invalidate in some cases, but for any
invalidation size over 2M it matches or improves the prior behavior.
There's also an improvement for single-page invalidates since the prior
version issued two invalidates for that case instead of one.
With this change all issued ATSDs now perform a flush, so the flush
parameter has been removed from all the helpers.
To show the benefit here are some performance numbers from a
microbenchmark which creates a 1G allocation then uses mprotect with
PROT_NONE to trigger invalidates in strides across the allocation.
One NPU (1 GPU):
mprotect rate (GB/s)
Stride Before After Speedup
64K 5.3 5.6 5%
1M 39.3 57.4 46%
2M 49.7 82.6 66%
4M 286.6 285.7 0%
Two NPUs (6 GPUs):
mprotect rate (GB/s)
Stride Before After Speedup
64K 6.5 7.4 13%
1M 33.4 67.9 103%
2M 38.7 93.1 141%
4M 356.7 354.6 -1%
Anything over 2M is roughly the same as before since both cases issue a
single ATSD.
Signed-off-by: Mark Hairgrove <mhairgrove@nvidia.com>
Reviewed-By: Alistair Popple <alistair@popple.id.au>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
There are two types of ATSDs issued to the NPU: invalidates targeting a
specific virtual address and invalidates targeting the whole address
space. In both cases prior to this change, the sequence was:
for each NPU
- Write the target address to the XTS_ATSD_AVA register
- EIEIO
- Write the launch value to issue the ATSD
First, a target address is not required when invalidating the whole
address space, so that write and the EIEIO have been removed. The AP
(size) field in the launch is not needed either.
Second, for per-address invalidates the above sequence is inefficient in
the common case of multiple NPUs because an EIEIO is issued per NPU. This
unnecessarily forces the launches of later ATSDs to be ordered with the
launches of earlier ones. The new sequence only issues a single EIEIO:
for each NPU
- Write the target address to the XTS_ATSD_AVA register
EIEIO
for each NPU
- Write the launch value to issue the ATSD
Performance results were gathered using a microbenchmark which creates a
1G allocation then uses mprotect with PROT_NONE to trigger invalidates in
strides across the allocation.
With only a single NPU active (one GPU) the difference is in the noise for
both types of invalidates (+/-1%).
With two NPUs active (on a 6-GPU system) the effect is more noticeable:
mprotect rate (GB/s)
Stride Before After Speedup
64K 5.9 6.5 10%
1M 31.2 33.4 7%
2M 36.3 38.7 7%
4M 322.6 356.7 11%
Signed-off-by: Mark Hairgrove <mhairgrove@nvidia.com>
Reviewed-by: Alistair Popple <alistair@popple.id.au>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Clean up the leftover of commit f2910f0e68 ("powerpc: remove old
GCC version checks").
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Acked-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
When enumerating page size definitions to check hardware support,
we construct a constant which is (1U << (def->shift - 10)).
However, the array of page size definitions is only initalised for
various MMU_PAGE_* constants, so it contains a number of 0-initialised
elements with def->shift == 0. This means we end up shifting by a
very large number, which gives the following UBSan splat:
================================================================================
UBSAN: Undefined behaviour in /home/dja/dev/linux/linux/arch/powerpc/mm/tlb_nohash.c:506:21
shift exponent 4294967286 is too large for 32-bit type 'unsigned int'
CPU: 0 PID: 0 Comm: swapper Not tainted 4.19.0-rc3-00045-ga604f927b012-dirty #6
Call Trace:
[c00000000101bc20] [c000000000a13d54] .dump_stack+0xa8/0xec (unreliable)
[c00000000101bcb0] [c0000000004f20a8] .ubsan_epilogue+0x18/0x64
[c00000000101bd30] [c0000000004f2b10] .__ubsan_handle_shift_out_of_bounds+0x110/0x1a4
[c00000000101be20] [c000000000d21760] .early_init_mmu+0x1b4/0x5a0
[c00000000101bf10] [c000000000d1ba28] .early_setup+0x100/0x130
[c00000000101bf90] [c000000000000528] start_here_multiplatform+0x68/0x80
================================================================================
Fix this by first checking if the element exists (shift != 0) before
constructing the constant.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Add call to early_memtest() so that kernel compiled with
CONFIG_MEMTEST really perform memtest at startup when requested
via 'memtest' boot parameter.
Tested-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
When a process allocates a hugepage, the following leak is
reported by kmemleak. This is a false positive which is
due to the pointer to the table being stored in the PGD
as physical memory address and not virtual memory pointer.
unreferenced object 0xc30f8200 (size 512):
comm "mmap", pid 374, jiffies 4872494 (age 627.630s)
hex dump (first 32 bytes):
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
backtrace:
[<e32b68da>] huge_pte_alloc+0xdc/0x1f8
[<9e0df1e1>] hugetlb_fault+0x560/0x8f8
[<7938ec6c>] follow_hugetlb_page+0x14c/0x44c
[<afbdb405>] __get_user_pages+0x1c4/0x3dc
[<b8fd7cd9>] __mm_populate+0xac/0x140
[<3215421e>] vm_mmap_pgoff+0xb4/0xb8
[<c148db69>] ksys_mmap_pgoff+0xcc/0x1fc
[<4fcd760f>] ret_from_syscall+0x0/0x38
See commit a984506c54 ("powerpc/mm: Don't report PUDs as
memory leaks when using kmemleak") for detailed explanation.
To fix that, this patch tells kmemleak to ignore the allocated
hugepage table.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The comments in this file don't conform to the coding style so take
them to "Comment Formatting Re-Education Camp".
Suggested-by: Michael "Camp Drill Sergeant" Ellerman <mpe@ellerman.id.au>
Signed-off-by: Michael Neuling <mikey@neuling.org>
[mpe: Reflow some comments and add full stops, fix spelling of Sergeant.]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
for 64bit configs which use for CONFIG_LOG_BUF_SHIFT the same
or higher value than the default (currently 17).
Signed-off-by: Petr Vorel <pvorel@suse.cz>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The code in machine_check_exception excludes 64s hvmode when
incrementing the MCE counter only to call opal_machine_check to
increment it specifically for this case.
Remove the exclusion and special case.
Fixes: a43c159042 ("powerpc/pseries: Flush SLB contents on SLB MCE
errors.")
Signed-off-by: Michal Suchanek <msuchanek@suse.de>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
On a kernel TM Bad thing program exception, the Machine State Register
(MSR) is not being properly displayed. The exception code dumps a 32-bits
value but MSR is a 64 bits register for all platforms that have HTM
enabled.
This patch dumps the MSR value as a 64-bits value instead of 32 bits. In
order to do so, the 'reason' variable could not be used, since it trimmed
MSR to 32-bits (int).
Signed-off-by: Breno Leitao <leitao@debian.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Currently msr_tm_active() is a wrapper around MSR_TM_ACTIVE() if
CONFIG_PPC_TRANSACTIONAL_MEM is set, or it is just a function that
returns false if CONFIG_PPC_TRANSACTIONAL_MEM is not set.
This function is not necessary, since MSR_TM_ACTIVE() just do the same and
could be used, removing the dualism and simplifying the code.
This patchset remove every instance of msr_tm_active() and replaced it
by MSR_TM_ACTIVE().
Signed-off-by: Breno Leitao <leitao@debian.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
There is a mismatch between function pnv_platform_error_reboot() definition
and declaration regarding function modifiers. In the declaration part, it
contains the function attribute __noreturn, while function definition
itself lacks it.
This was reported by sparse tool as an error:
arch/powerpc/platforms/powernv/opal.c:538:6: error: symbol 'pnv_platform_error_reboot' redeclared with different type (originally declared at arch/powerpc/platforms/powernv/powernv.h:11) - different modifiers
I checked and the function is already being considered as being 'noreturn'
by the compiler, thus, I understand this patch does not change any code
being generated.
Signed-off-by: Breno Leitao <leitao@debian.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
This is a patch that adds support for PTRACE_SYSEMU ptrace request in
PowerPC architecture.
When ptrace(PTRACE_SYSEMU, ...) request is called, it will be handled by
the arch independent function ptrace_resume(), which will tag the task with
the TIF_SYSCALL_EMU flag. This flag needs to be handled from a platform
dependent point of view, which is what this patch does.
This patch adds this task's flag as part of the _TIF_SYSCALL_DOTRACE, which
is the MACRO that is used to trace syscalls at entrance/exit.
Since TIF_SYSCALL_EMU is now part of _TIF_SYSCALL_DOTRACE, if the task has
_TIF_SYSCALL_DOTRACE set, it will hit do_syscall_trace_enter() at syscall
entrance and do_syscall_trace_leave() at syscall leave.
do_syscall_trace_enter() needs to handle the TIF_SYSCALL_EMU flag properly,
which will interrupt the syscall executing if TIF_SYSCALL_EMU is set. The
output values should not be changed, i.e. the return value (r3) should
contain the original syscall argument on exit.
With this flag set, the syscall is not executed fundamentally, because
do_syscall_trace_enter() is returning -1 which is bigger than NR_syscall,
thus, skipping the syscall execution and exiting userspace.
Signed-off-by: Breno Leitao <leitao@debian.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Moving TIF_32BIT to use bit 20 instead of 4 in the task flag field.
This change is making room for an upcoming new task macro
(_TIF_SYSCALL_EMU) which is preferred to set a bit in the lower 16-bits
part of the word.
This upcoming flag macro will take part in a composed macro
(_TIF_SYSCALL_DOTRACE) which will contain other flags as well, and it is
preferred that the whole _TIF_SYSCALL_DOTRACE macro only sets the lower 16
bits of a word, so, it could be handled using immediate operations (as load
immediate, add immediate, ...) where the immediate operand (SI) is limited
to 16-bits.
Another possible solution would be using the LOAD_REG_IMMEDIATE() macro
to load a full 64-bits word immediate, but it takes 5 operations instead of
one.
Having TIF_32BITS being redefined to use an upper bit is not a problem
since there is only one place in the assembly code where TIF_32BIT is being
used, and it could be replaced with an operation with right shift (addis),
since it is used alone, i.e. not being part of a composed macro, which has
different bits set, and would require LOAD_REG_IMMEDIATE().
Tested on a 64 bits Big Endian machine running a 32 bits task.
Signed-off-by: Breno Leitao <leitao@debian.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
On PPC64, as register r13 points to the paca_struct at all time,
this patch adds a copy of the canary there, which is copied at
task_switch.
That new canary is then used by using the following GCC options:
-mstack-protector-guard=tls
-mstack-protector-guard-reg=r13
-mstack-protector-guard-offset=offsetof(struct paca_struct, canary))
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
