commit 11276d5306 upstream.
There are various problems and short-comings with the current
static_key interface:
- static_key_{true,false}() read like a branch depending on the key
value, instead of the actual likely/unlikely branch depending on
init value.
- static_key_{true,false}() are, as stated above, tied to the
static_key init values STATIC_KEY_INIT_{TRUE,FALSE}.
- we're limited to the 2 (out of 4) possible options that compile to
a default NOP because that's what our arch_static_branch() assembly
emits.
So provide a new static_key interface:
DEFINE_STATIC_KEY_TRUE(name);
DEFINE_STATIC_KEY_FALSE(name);
Which define a key of different types with an initial true/false
value.
Then allow:
static_branch_likely()
static_branch_unlikely()
to take a key of either type and emit the right instruction for the
case.
This means adding a second arch_static_branch_jump() assembly helper
which emits a JMP per default.
In order to determine the right instruction for the right state,
encode the branch type in the LSB of jump_entry::key.
This is the final step in removing the naming confusion that has led to
a stream of avoidable bugs such as:
a833581e37 ("x86, perf: Fix static_key bug in load_mm_cr4()")
... but it also allows new static key combinations that will give us
performance enhancements in the subsequent patches.
Tested-by: Rabin Vincent <rabin@rab.in> # arm
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Michael Ellerman <mpe@ellerman.id.au> # ppc
Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com> # s390
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
[bwh: Backported to 3.16:
- For s390, use the 31-bit-compatible macros in arch_static_branch_jump()
-
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit 1bd6a1c4b8 upstream.
Crash memory ranges is an array of memory ranges of the crashing kernel
to be exported as a dump via /proc/vmcore file. The size of the array
is set based on INIT_MEMBLOCK_REGIONS, which works alright in most cases
where memblock memory regions count is less than INIT_MEMBLOCK_REGIONS
value. But this count can grow beyond INIT_MEMBLOCK_REGIONS value since
commit 142b45a72e ("memblock: Add array resizing support").
On large memory systems with a few DLPAR operations, the memblock memory
regions count could be larger than INIT_MEMBLOCK_REGIONS value. On such
systems, registering fadump results in crash or other system failures
like below:
task: c00007f39a290010 ti: c00000000b738000 task.ti: c00000000b738000
NIP: c000000000047df4 LR: c0000000000f9e58 CTR: c00000000010f180
REGS: c00000000b73b570 TRAP: 0300 Tainted: G L X (4.4.140+)
MSR: 8000000000009033 <SF,EE,ME,IR,DR,RI,LE> CR: 22004484 XER: 20000000
CFAR: c000000000008500 DAR: 000007a450000000 DSISR: 40000000 SOFTE: 0
...
NIP [c000000000047df4] smp_send_reschedule+0x24/0x80
LR [c0000000000f9e58] resched_curr+0x138/0x160
Call Trace:
resched_curr+0x138/0x160 (unreliable)
check_preempt_curr+0xc8/0xf0
ttwu_do_wakeup+0x38/0x150
try_to_wake_up+0x224/0x4d0
__wake_up_common+0x94/0x100
ep_poll_callback+0xac/0x1c0
__wake_up_common+0x94/0x100
__wake_up_sync_key+0x70/0xa0
sock_def_readable+0x58/0xa0
unix_stream_sendmsg+0x2dc/0x4c0
sock_sendmsg+0x68/0xa0
___sys_sendmsg+0x2cc/0x2e0
__sys_sendmsg+0x5c/0xc0
SyS_socketcall+0x36c/0x3f0
system_call+0x3c/0x100
as array index overflow is not checked for while setting up crash memory
ranges causing memory corruption. To resolve this issue, dynamically
allocate memory for crash memory ranges and resize it incrementally,
in units of pagesize, on hitting array size limit.
Fixes: 2df173d9e8 ("fadump: Initialize elfcore header and add PT_LOAD program headers.")
Signed-off-by: Hari Bathini <hbathini@linux.ibm.com>
Reviewed-by: Mahesh Salgaonkar <mahesh@linux.vnet.ibm.com>
[mpe: Just use PAGE_SIZE directly, fixup variable placement]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
[bwh: Backported to 3.16:
- register_fadump() returns void
- Include <linux/slab.h> for kfree(), krealloc()]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit 19c22f3a29 upstream.
arch/tile added word-at-a-time.h after the patch that added generic-y
entries; the generic-y entry is now stale.
arch/h8300 is newer than the generic-y patch for word-at-a-time.h,
and needs a generic-y entry.
arch/powerpc seems to have gotten a generic-y entry by mistake in
the first patch; this change removes it.
Signed-off-by: Chris Metcalf <cmetcalf@ezchip.com>
[bwh: Backported to 3.16:
- Drop change in arch/h8300, which doesn't exist here
- Drop change in arch/tile, which is still using the generic implementation]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit a6e2f029ae upstream.
Added the x86 implementation of word-at-a-time to the
generic version, which previously only supported big-endian.
Omitted the x86-specific load_unaligned_zeropad(), which in
any case is also not present for the existing BE-only
implementation of a word-at-a-time, and is only used under
CONFIG_DCACHE_WORD_ACCESS.
Added as a "generic-y" to the Kbuilds of all architectures
that didn't previously have it.
Signed-off-by: Chris Metcalf <cmetcalf@ezchip.com>
[bwh: Backported to 3.16:
- Drop change in arch/nios2
- Adjust context]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit 34dd25de9f upstream.
This is the start of an effort to tidy up and standardise all the
delays. Existing loops have a range of delay/sleep periods from 1ms
to 20ms, and some have no delay. They all loop forever except rtc,
which times out after 10 retries, and that uses 10ms delays. So use
10ms as our standard delay. The OPAL maintainer agrees 10ms is a
reasonable starting point.
The idea is to use the same recipe everywhere, once this is proven to
work then it will be documented as an OPAL API standard. Then both
firmware and OS can agree, and if a particular call needs something
else, then that can be documented with reasoning.
This is not the end-all of this effort, it's just a relatively easy
change that fixes some existing high latency delays. There should be
provision for standardising timeouts and/or interruptible loops where
possible, so non-fatal firmware errors don't cause hangs.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit 0bfdf59890 upstream.
asm/barrier.h is not always included after asm/synch.h, which meant
it was missing __SUBARCH_HAS_LWSYNC, so in some files smp_wmb() would
be eieio when it should be lwsync. kernel/time/hrtimer.c is one case.
__SUBARCH_HAS_LWSYNC is only used in one place, so just fold it in
to where it's used. Previously with my small simulator config, 377
instances of eieio in the tree. After this patch there are 55.
Fixes: 46d075be58 ("powerpc: Optimise smp_wmb")
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit c1e150ceb6 upstream.
When CONFIG_NUMA is not set, the build fails with:
arch/powerpc/platforms/pseries/hotplug-cpu.c:335:4:
error: déclaration implicite de la fonction « update_numa_cpu_lookup_table »
So we have to add update_numa_cpu_lookup_table() as an empty function
when CONFIG_NUMA is not set.
Fixes: 1d9a090783 ("powerpc/numa: Invalidate numa_cpu_lookup_table on cpu remove")
Signed-off-by: Corentin Labbe <clabbe@baylibre.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit 1d9a090783 upstream.
When DLPAR removing a CPU, the unmapping of the cpu from a node in
unmap_cpu_from_node() should also invalidate the CPUs entry in the
numa_cpu_lookup_table. There is not a guarantee that on a subsequent
DLPAR add of the CPU the associativity will be the same and thus
could be in a different node. Invalidating the entry in the
numa_cpu_lookup_table causes the associativity to be read from the
device tree at the time of the add.
The current behavior of not invalidating the CPUs entry in the
numa_cpu_lookup_table can result in scenarios where the the topology
layout of CPUs in the partition does not match the device tree
or the topology reported by the HMC.
This bug looks like it was introduced in 2004 in the commit titled
"ppc64: cpu hotplug notifier for numa", which is 6b15e4e87e32 in the
linux-fullhist tree. Hence tag it for all stable releases.
Signed-off-by: Nathan Fontenot <nfont@linux.vnet.ibm.com>
Reviewed-by: Tyrel Datwyler <tyreld@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit 07ae5389e9 upstream.
When copying between the vcpu and svcpu, we may get scheduled away onto
a different host CPU which in turn means our svcpu pointer may change.
That means we need to atomically copy to and from the svcpu with preemption
disabled, so that all code around it always sees a coherent state.
Reported-by: Simon Guo <wei.guo.simon@gmail.com>
Fixes: 3d3319b45e ("KVM: PPC: Book3S: PR: Enable interrupts earlier")
Signed-off-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit 93b2d3cf37 upstream.
The instruction code for xxlor that commit 0016a4cf55 ("powerpc:
Emulate most Book I instructions in emulate_step()", 2010-06-15)
added is actually the code for xxlnor. It is used in get_vsr()
and put_vsr() and the effect of the error is that if emulate_step
is used to emulate a VSX load or store from any register other
than vsr0, the bitwise complement of the correct value will be
loaded or stored. This corrects the error.
Fixes: 0016a4cf55 ("powerpc: Emulate most Book I instructions in emulate_step()")
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
[bwh: Backported to 3.16: adjust context]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit 01e6a61ace upstream.
Although it's not documented anywhere, there is an expectation that
atomic64_inc_not_zero() returns a result which fits in an int. This is
the behaviour implemented on all arches except powerpc.
This has caused at least one bug in practice, in the percpu-refcount
code, where the long result from our atomic64_inc_not_zero() was
truncated to an int leading to lost references and stuck systems. That
was worked around in that code in commit 966d2b04e0 ("percpu-refcount:
fix reference leak during percpu-atomic transition").
To the best of my grepping abilities there are no other callers
in-tree which truncate the value, but we should fix it anyway. Because
the breakage is subtle and potentially very harmful I'm also tagging
it for stable.
Code generation is largely unaffected because in most cases the
callers are just using the result for a test anyway. In particular the
case of fget() that was mentioned in commit a6cf7ed511
("powerpc/atomic: Implement atomic*_inc_not_zero") generates exactly
the same code.
Fixes: a6cf7ed511 ("powerpc/atomic: Implement atomic*_inc_not_zero")
Noticed-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit 2400fd822f upstream.
The workaround for the CELL timebase bug does not correctly mark cr0 as
being clobbered. This means GCC doesn't know that the asm block changes cr0 and
might leave the result of an unrelated comparison in cr0 across the block, which
we then trash, leading to basically random behaviour.
Fixes: 859deea949 ("[POWERPC] Cell timebase bug workaround")
Signed-off-by: Oliver O'Halloran <oohall@gmail.com>
[mpe: Tweak change log and flag for stable]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit ba4a648f12 upstream.
In commit 8c27226119 ("powerpc/numa: Enable USE_PERCPU_NUMA_NODE_ID"), we
switched to the generic implementation of cpu_to_node(), which uses a percpu
variable to hold the NUMA node for each CPU.
Unfortunately we neglected to notice that we use cpu_to_node() in the allocation
of our percpu areas, leading to a chicken and egg problem. In practice what
happens is when we are setting up the percpu areas, cpu_to_node() reports that
all CPUs are on node 0, so we allocate all percpu areas on node 0.
This is visible in the dmesg output, as all pcpu allocs being in group 0:
pcpu-alloc: [0] 00 01 02 03 [0] 04 05 06 07
pcpu-alloc: [0] 08 09 10 11 [0] 12 13 14 15
pcpu-alloc: [0] 16 17 18 19 [0] 20 21 22 23
pcpu-alloc: [0] 24 25 26 27 [0] 28 29 30 31
pcpu-alloc: [0] 32 33 34 35 [0] 36 37 38 39
pcpu-alloc: [0] 40 41 42 43 [0] 44 45 46 47
To fix it we need an early_cpu_to_node() which can run prior to percpu being
setup. We already have the numa_cpu_lookup_table we can use, so just plumb it
in. With the patch dmesg output shows two groups, 0 and 1:
pcpu-alloc: [0] 00 01 02 03 [0] 04 05 06 07
pcpu-alloc: [0] 08 09 10 11 [0] 12 13 14 15
pcpu-alloc: [0] 16 17 18 19 [0] 20 21 22 23
pcpu-alloc: [1] 24 25 26 27 [1] 28 29 30 31
pcpu-alloc: [1] 32 33 34 35 [1] 36 37 38 39
pcpu-alloc: [1] 40 41 42 43 [1] 44 45 46 47
We can also check the data_offset in the paca of various CPUs, with the fix we
see:
CPU 0: data_offset = 0x0ffe8b0000
CPU 24: data_offset = 0x1ffe5b0000
And we can see from dmesg that CPU 24 has an allocation on node 1:
node 0: [mem 0x0000000000000000-0x0000000fffffffff]
node 1: [mem 0x0000001000000000-0x0000001fffffffff]
Fixes: 8c27226119 ("powerpc/numa: Enable USE_PERCPU_NUMA_NODE_ID")
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Reviewed-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit e41e53cd4f upstream.
virt_addr_valid() is supposed to tell you if it's OK to call virt_to_page() on
an address. What this means in practice is that it should only return true for
addresses in the linear mapping which are backed by a valid PFN.
We are failing to properly check that the address is in the linear mapping,
because virt_to_pfn() will return a valid looking PFN for more or less any
address. That bug is actually caused by __pa(), used in virt_to_pfn().
eg: __pa(0xc000000000010000) = 0x10000 # Good
__pa(0xd000000000010000) = 0x10000 # Bad!
__pa(0x0000000000010000) = 0x10000 # Bad!
This started happening after commit bdbc29c19b ("powerpc: Work around gcc
miscompilation of __pa() on 64-bit") (Aug 2013), where we changed the definition
of __pa() to work around a GCC bug. Prior to that we subtracted PAGE_OFFSET from
the value passed to __pa(), meaning __pa() of a 0xd or 0x0 address would give
you something bogus back.
Until we can verify if that GCC bug is no longer an issue, or come up with
another solution, this commit does the minimal fix to make virt_addr_valid()
work, by explicitly checking that the address is in the linear mapping region.
Fixes: bdbc29c19b ("powerpc: Work around gcc miscompilation of __pa() on 64-bit")
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Reviewed-by: Paul Mackerras <paulus@ozlabs.org>
Reviewed-by: Balbir Singh <bsingharora@gmail.com>
Tested-by: Breno Leitao <breno.leitao@gmail.com>
[bwh: Backported to 3.16: open-code virt_to_pfn()]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit 8b8642af15 upstream.
Since commit 5093bb965a ("powerpc/QE: switch to the cpm_muram
implementation"), muram area is not part of immrbar mapping anymore
so immrbar_virt_to_phys() is not usable anymore.
Fixes: 5093bb965a ("powerpc/QE: switch to the cpm_muram implementation")
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Acked-by: David S. Miller <davem@davemloft.net>
Acked-by: Li Yang <pku.leo@gmail.com>
Signed-off-by: Scott Wood <oss@buserror.net>
[bwh: Backported to 3.16: adjust filename]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit 6f698df10c upstream.
Rather than forcing the whole function into the ".kprobes.text" section,
just add the symbol's address to the kprobe blacklist.
This also lets us drop the three versions of the_KPROBE macro, in
exchange for just one version of _ASM_NOKPROBE_SYMBOL - which is a good
cleanup.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
[bwh: Backported to 3.16: adjust context]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit 178f358208 upstream.
IBM bit 31 (for the rest of us - bit 0) is a reserved field in the
instruction definition of mtspr and mfspr. Hardware is encouraged to
(and does) ignore it.
As a result, if userspace executes an mtspr DSCR with the reserved bit
set, we get a DSCR facility unavailable exception. The kernel fails to
match against the expected value/mask, and we silently return to
userspace to try and re-execute the same mtspr DSCR instruction. We
loop forever until the process is killed.
We should do something here, and it seems mirroring what hardware does
is the better option vs killing the process. While here, relax the
matching of mfspr PVR too.
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit 0d808df06a upstream.
When switching from/to a guest that has a transaction in progress,
we need to save/restore the checkpointed register state. Although
XER is part of the CPU state that gets checkpointed, the code that
does this saving and restoring doesn't save/restore XER.
This fixes it by saving and restoring the XER. To allow userspace
to read/write the checkpointed XER value, we also add a new ONE_REG
specifier.
The visible effect of this bug is that the guest may see its XER
value being corrupted when it uses transactions.
Fixes: e4e3812150 ("KVM: PPC: Book3S HV: Add transactional memory support")
Fixes: 0a8eccefcb ("KVM: PPC: Book3S HV: Add missing code for transaction reclaim on guest exit")
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
[bwh: Backported to 3.16: adjust context, spacing]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit fa73c3b25b upstream.
The MMCR2 register is available twice, one time with number 785
(privileged access), and one time with number 769 (unprivileged,
but it can be disabled completely). In former times, the Linux
kernel was using the unprivileged register 769 only, but since
commit 8dd75ccb57 ("powerpc: Use privileged SPR number
for MMCR2"), it uses the privileged register 785 instead.
The KVM-PR code then of course also switched to use the SPR 785,
but this is causing older guest kernels to crash, since these
kernels still access 769 instead. So to support older kernels
with KVM-PR again, we have to support register 769 in KVM-PR, too.
Fixes: 8dd75ccb57
Signed-off-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit 88b02cf97b upstream.
POWER8 has one virtual timebase (VTB) register per subcore, not one
per CPU thread. The HV KVM code currently treats VTB as a per-thread
register, which can lead to spurious soft lockup messages from guests
which use the VTB as the time source for the soft lockup detector.
(CPUs before POWER8 did not have the VTB register.)
For HV KVM, this fixes the problem by making only the primary thread
in each virtual core save and restore the VTB value. With this,
the VTB state becomes part of the kvmppc_vcore structure. This
also means that "piggybacking" of multiple virtual cores onto one
subcore is not possible on POWER8, because then the virtual cores
would share a single VTB register.
PR KVM emulates a VTB register, which is per-vcpu because PR KVM
has no notion of CPU threads or SMT. For PR KVM we move the VTB
state into the kvmppc_vcpu_book3s struct.
Reported-by: Thomas Huth <thuth@redhat.com>
Tested-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
[bwh: Backported to 3.16:
- Adjust filenames, context
- Drop changes to kvmppc_core_emulate_mfspr_pr(), can_piggyback_subcore(),
kvmppc_copy_from_svcpu()]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit 224264657b upstream.
should clear on access_ok() failures. Also remove the useless
range truncation logics.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
[bwh: Backported to 3.16: no calls to check_object_size()]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit 2dd9c11b9d upstream.
This patch leverages 'struct pci_host_bridge' from the PCI subsystem
in order to free the pci_controller only after the last reference to
its devices is dropped (avoiding an oops in pcibios_release_device()
if the last reference is dropped after pcibios_free_controller()).
The patch relies on pci_host_bridge.release_fn() (and .release_data),
which is called automatically by the PCI subsystem when the root bus
is released (i.e., the last reference is dropped). Those fields are
set via pci_set_host_bridge_release() (e.g. in the platform-specific
implementation of pcibios_root_bridge_prepare()).
It introduces the 'pcibios_free_controller_deferred()' .release_fn()
and it expects .release_data to hold a pointer to the pci_controller.
The function implictly calls 'pcibios_free_controller()', so an user
must *NOT* explicitly call it if using the new _deferred() callback.
The functionality is enabled for pseries (although it isn't platform
specific, and may be used by cxl).
Details on not-so-elegant design choices:
- Use 'pci_host_bridge.release_data' field as pointer to associated
'struct pci_controller' so *not* to 'pci_bus_to_host(bridge->bus)'
in pcibios_free_controller_deferred().
That's because pci_remove_root_bus() sets 'host_bridge->bus = NULL'
(so, if the last reference is released after pci_remove_root_bus()
runs, which eventually reaches pcibios_free_controller_deferred(),
that would hit a null pointer dereference).
The cxl/vphb.c code calls pci_remove_root_bus(), and the cxl folks
are interested in this fix.
Test-case #1 (hold references)
# ls -ld /sys/block/sd* | grep -m1 0021:01:00.0
<...> /sys/block/sdaa -> ../devices/pci0021:01/0021:01:00.0/<...>
# ls -ld /sys/block/sd* | grep -m1 0021:01:00.1
<...> /sys/block/sdab -> ../devices/pci0021:01/0021:01:00.1/<...>
# cat >/dev/sdaa & pid1=$!
# cat >/dev/sdab & pid2=$!
# drmgr -w 5 -d 1 -c phb -s 'PHB 33' -r
Validating PHB DLPAR capability...yes.
[ 594.306719] pci_hp_remove_devices: PCI: Removing devices on bus 0021:01
[ 594.306738] pci_hp_remove_devices: Removing 0021:01:00.0...
...
[ 598.236381] pci_hp_remove_devices: Removing 0021:01:00.1...
...
[ 611.972077] pci_bus 0021:01: busn_res: [bus 01-ff] is released
[ 611.972140] rpadlpar_io: slot PHB 33 removed
# kill -9 $pid1
# kill -9 $pid2
[ 632.918088] pcibios_free_controller_deferred: domain 33, dynamic 1
Test-case #2 (don't hold references)
# drmgr -w 5 -d 1 -c phb -s 'PHB 33' -r
Validating PHB DLPAR capability...yes.
[ 916.357363] pci_hp_remove_devices: PCI: Removing devices on bus 0021:01
[ 916.357386] pci_hp_remove_devices: Removing 0021:01:00.0...
...
[ 920.566527] pci_hp_remove_devices: Removing 0021:01:00.1...
...
[ 933.955873] pci_bus 0021:01: busn_res: [bus 01-ff] is released
[ 933.955977] pcibios_free_controller_deferred: domain 33, dynamic 1
[ 933.955999] rpadlpar_io: slot PHB 33 removed
Suggested-By: Gavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: Mauricio Faria de Oliveira <mauricfo@linux.vnet.ibm.com>
Reviewed-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Reviewed-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com>
Tested-by: Andrew Donnellan <andrew.donnellan@au1.ibm.com> # cxl
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit 8dd75ccb57 upstream.
We are already using the privileged versions of MMCR0, MMCR1
and MMCRA in the kernel, so for MMCR2, we should better use
the privileged versions, too, to be consistent.
Fixes: 240686c136 ("powerpc: Initialise PMU related regs on Power8")
Suggested-by: Paul Mackerras <paulus@ozlabs.org>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Acked-by: Paul Mackerras <paulus@ozlabs.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit d23fac2b27 upstream.
The SIAR and SDAR registers are available twice, one time as SPRs
780 / 781 (unprivileged, but read-only), and one time as the SPRs
796 / 797 (privileged, but read and write). The Linux kernel code
currently uses the unprivileged SPRs - while this is OK for reading,
writing to that register of course does not work.
Since the KVM code tries to write to this register, too (see the mtspr
in book3s_hv_rmhandlers.S), the contents of this register sometimes get
lost for the guests, e.g. during migration of a VM.
To fix this issue, simply switch to the privileged SPR numbers instead.
Signed-off-by: Thomas Huth <thuth@redhat.com>
Acked-by: Paul Mackerras <paulus@ozlabs.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit b4c112114a upstream.
In create_zero_mask() we have:
addi %1,%2,-1
andc %1,%1,%2
popcntd %0,%1
using the "r" constraint for %2. r0 is a valid register in the "r" set,
but addi X,r0,X turns it into an li:
li r7,-1
andc r7,r7,r0
popcntd r4,r7
Fix this by using the "b" constraint, for which r0 is not a valid
register.
This was found with a kernel build using gcc trunk, narrowed down to
when -frename-registers was enabled at -O2. It is just luck however
that we aren't seeing this on older toolchains.
Thanks to Segher for working with me to find this issue.
Fixes: d0cebfa650 ("powerpc: word-at-a-time optimization for 64-bit Little Endian")
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
[bwh: Backported to 3.16: same issue exists with a different variable in
find_zero()]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit 6997e57d69 upstream.
The REAL_LE feature entry in the ibm_pa_feature struct is missing an MMU
feature value, meaning all the remaining elements initialise the wrong
values.
This means instead of checking for byte 5, bit 0, we check for byte 0,
bit 0, and then we incorrectly set the CPU feature bit as well as MMU
feature bit 1 and CPU user feature bits 0 and 2 (5).
Checking byte 0 bit 0 (IBM numbering), means we're looking at the
"Memory Management Unit (MMU)" feature - ie. does the CPU have an MMU.
In practice that bit is set on all platforms which have the property.
This means we set CPU_FTR_REAL_LE always. In practice that seems not to
matter because all the modern cpus which have this property also
implement REAL_LE, and we've never needed to disable it.
We're also incorrectly setting MMU feature bit 1, which is:
#define MMU_FTR_TYPE_8xx 0x00000002
Luckily the only place that looks for MMU_FTR_TYPE_8xx is in Book3E
code, which can't run on the same cpus as scan_features(). So this also
doesn't matter in practice.
Finally in the CPU user feature mask, we're setting bits 0 and 2. Bit 2
is not currently used, and bit 0 is:
#define PPC_FEATURE_PPC_LE 0x00000001
Which says the CPU supports the old style "PPC Little Endian" mode.
Again this should be harmless in practice as no 64-bit CPUs implement
that mode.
Fix the code by adding the missing initialisation of the MMU feature.
Also add a comment marking CPU user feature bit 2 (0x4) as reserved. It
would be unsafe to start using it as old kernels incorrectly set it.
Fixes: 44ae3ab335 ("powerpc: Free up some CPU feature bits by moving out MMU-related features")
Signed-off-by: Anton Blanchard <anton@samba.org>
[mpe: Flesh out changelog, add comment reserving 0x4]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
[bwh: Backported to 3.16: adjust context]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
commit a61674bdfc upstream.
GCC 6 will include changes to generated code with -mcmodel=large,
which is used to build kernel modules on powerpc64le. This was
necessary because the large model is supposed to allow arbitrary
sizes and locations of the code and data sections, but the ELFv2
global entry point prolog still made the unconditional assumption
that the TOC associated with any particular function can be found
within 2 GB of the function entry point:
func:
addis r2,r12,(.TOC.-func)@ha
addi r2,r2,(.TOC.-func)@l
.localentry func, .-func
To remove this assumption, GCC will now generate instead this global
entry point prolog sequence when using -mcmodel=large:
.quad .TOC.-func
func:
.reloc ., R_PPC64_ENTRY
ld r2, -8(r12)
add r2, r2, r12
.localentry func, .-func
The new .reloc triggers an optimization in the linker that will
replace this new prolog with the original code (see above) if the
linker determines that the distance between .TOC. and func is in
range after all.
Since this new relocation is now present in module object files,
the kernel module loader is required to handle them too. This
patch adds support for the new relocation and implements the
same optimization done by the GNU linker.
Signed-off-by: Ulrich Weigand <ulrich.weigand@de.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Luis Henriques <luis.henriques@canonical.com>
commit 81d7a3294d upstream.
According to memory-barriers.txt, xchg*, cmpxchg* and their atomic_
versions all need to be fully ordered, however they are now just
RELEASE+ACQUIRE, which are not fully ordered.
So also replace PPC_RELEASE_BARRIER and PPC_ACQUIRE_BARRIER with
PPC_ATOMIC_ENTRY_BARRIER and PPC_ATOMIC_EXIT_BARRIER in
__{cmp,}xchg_{u32,u64} respectively to guarantee fully ordered semantics
of atomic{,64}_{cmp,}xchg() and {cmp,}xchg(), as a complement of commit
b97021f855 ("powerpc: Fix atomic_xxx_return barrier semantics")
This patch depends on patch "powerpc: Make value-returning atomics fully
ordered" for PPC_ATOMIC_ENTRY_BARRIER definition.
Signed-off-by: Boqun Feng <boqun.feng@gmail.com>
Reviewed-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Luis Henriques <luis.henriques@canonical.com>
commit 49e9cf3f0c upstream.
According to memory-barriers.txt:
> Any atomic operation that modifies some state in memory and returns
> information about the state (old or new) implies an SMP-conditional
> general memory barrier (smp_mb()) on each side of the actual
> operation ...
Which mean these operations should be fully ordered. However on PPC,
PPC_ATOMIC_ENTRY_BARRIER is the barrier before the actual operation,
which is currently "lwsync" if SMP=y. The leading "lwsync" can not
guarantee fully ordered atomics, according to Paul Mckenney:
https://lkml.org/lkml/2015/10/14/970
To fix this, we define PPC_ATOMIC_ENTRY_BARRIER as "sync" to guarantee
the fully-ordered semantics.
This also makes futex atomics fully ordered, which can avoid possible
memory ordering problems if userspace code relies on futex system call
for fully ordered semantics.
Fixes: b97021f855 ("powerpc: Fix atomic_xxx_return barrier semantics")
Signed-off-by: Boqun Feng <boqun.feng@gmail.com>
Reviewed-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Luis Henriques <luis.henriques@canonical.com>
commit d2b9d2a5ad upstream.
Currently we allow both the MSR T and S bits to be set by userspace on
a signal return. Unfortunately this is a reserved configuration and
will cause a TM Bad Thing exception if attempted (via rfid).
This patch checks for this case in both the 32 and 64 bit signals
code. If both T and S are set, we mark the context as invalid.
Found using a syscall fuzzer.
Fixes: 2b0a576d15 ("powerpc: Add new transactional memory state to the signal context")
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Luis Henriques <luis.henriques@canonical.com>
commit 74b5037baa upstream.
The powerpc kernel can be built to have either a 4K PAGE_SIZE or a 64K
PAGE_SIZE.
However when built with a 4K PAGE_SIZE there is an additional config
option which can be enabled, PPC_HAS_HASH_64K, which means the kernel
also knows how to hash a 64K page even though the base PAGE_SIZE is 4K.
This is used in one obscure configuration, to support 64K pages for SPU
local store on the Cell processor when the rest of the kernel is using
4K pages.
In this configuration, pte_pagesize_index() is defined to just pass
through its arguments to get_slice_psize(). However pte_pagesize_index()
is called for both user and kernel addresses, whereas get_slice_psize()
only knows how to handle user addresses.
This has been broken forever, however until recently it happened to
work. That was because in get_slice_psize() the large kernel address
would cause the right shift of the slice mask to return zero.
However in commit 7aa0727f33 ("powerpc/mm: Increase the slice range to
64TB"), the get_slice_psize() code was changed so that instead of a
right shift we do an array lookup based on the address. When passed a
kernel address this means we index way off the end of the slice array
and return random junk.
That is only fatal if we happen to hit something non-zero, but when we
do return a non-zero value we confuse the MMU code and eventually cause
a check stop.
This fix is ugly, but simple. When we're called for a kernel address we
return 4K, which is always correct in this configuration, otherwise we
use the slice mask.
Fixes: 7aa0727f33 ("powerpc/mm: Increase the slice range to 64TB")
Reported-by: Cyril Bur <cyrilbur@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Luis Henriques <luis.henriques@canonical.com>
commit 1c2cb59444 upstream.
The EPOW interrupt handler uses rtas_get_sensor(), which in turn
uses rtas_busy_delay() to wait for RTAS becoming ready in case it
is necessary. But rtas_busy_delay() is annotated with might_sleep()
and thus may not be used by interrupts handlers like the EPOW handler!
This leads to the following BUG when CONFIG_DEBUG_ATOMIC_SLEEP is
enabled:
BUG: sleeping function called from invalid context at arch/powerpc/kernel/rtas.c:496
in_atomic(): 1, irqs_disabled(): 1, pid: 0, name: swapper/1
CPU: 1 PID: 0 Comm: swapper/1 Not tainted 4.2.0-rc2-thuth #6
Call Trace:
[c00000007ffe7b90] [c000000000807670] dump_stack+0xa0/0xdc (unreliable)
[c00000007ffe7bc0] [c0000000000e1f14] ___might_sleep+0x134/0x180
[c00000007ffe7c20] [c00000000002aec0] rtas_busy_delay+0x30/0xd0
[c00000007ffe7c50] [c00000000002bde4] rtas_get_sensor+0x74/0xe0
[c00000007ffe7ce0] [c000000000083264] ras_epow_interrupt+0x44/0x450
[c00000007ffe7d90] [c000000000120260] handle_irq_event_percpu+0xa0/0x300
[c00000007ffe7e70] [c000000000120524] handle_irq_event+0x64/0xc0
[c00000007ffe7eb0] [c000000000124dbc] handle_fasteoi_irq+0xec/0x260
[c00000007ffe7ef0] [c00000000011f4f0] generic_handle_irq+0x50/0x80
[c00000007ffe7f20] [c000000000010f3c] __do_irq+0x8c/0x200
[c00000007ffe7f90] [c0000000000236cc] call_do_irq+0x14/0x24
[c00000007e6f39e0] [c000000000011144] do_IRQ+0x94/0x110
[c00000007e6f3a30] [c000000000002594] hardware_interrupt_common+0x114/0x180
Fix this issue by introducing a new rtas_get_sensor_fast() function
that does not use rtas_busy_delay() - and thus can only be used for
sensors that do not cause a BUSY condition - known as "fast" sensors.
The EPOW sensor is defined to be "fast" in sPAPR - mpe.
Fixes: 587f83e8dd ("powerpc/pseries: Use rtas_get_sensor in RAS code")
Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Nathan Fontenot <nfont@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Luis Henriques <luis.henriques@canonical.com>
commit 4ad04e5987 upstream.
After d905c5df9a ("PPC: POWERNV: move iommu_add_device earlier"), the
refcnt on the kobject backing the IOMMU group for a PCI device is
elevated by each call to pci_dma_dev_setup_pSeriesLP() (via
set_iommu_table_base_and_group). When we go to dlpar a multi-function
PCI device out:
iommu_reconfig_notifier ->
iommu_free_table ->
iommu_group_put
BUG_ON(tbl->it_group)
We trip this BUG_ON, because there are still references on the table, so
it is not freed. Fix this by moving the powernv bus notifier to common
code and calling it for both powernv and pseries.
Fixes: d905c5df9a ("PPC: POWERNV: move iommu_add_device earlier")
Signed-off-by: Nishanth Aravamudan <nacc@linux.vnet.ibm.com>
Tested-by: Nishanth Aravamudan <nacc@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
[ luis: backported to 3.16: adjusted context ]
Signed-off-by: Luis Henriques <luis.henriques@canonical.com>
commit 8117ac6a6c upstream.
Currently, when going idle, we set the flag indicating that we are in
nap mode (paca->kvm_hstate.hwthread_state) and then execute the nap
(or sleep or rvwinkle) instruction, all with the MMU on. This is bad
for two reasons: (a) the architecture specifies that those instructions
must be executed with the MMU off, and in fact with only the SF, HV, ME
and possibly RI bits set, and (b) this introduces a race, because as
soon as we set the flag, another thread can switch the MMU to a guest
context. If the race is lost, this thread will typically start looping
on relocation-on ISIs at 0xc...4400.
This fixes it by setting the MSR as required by the architecture before
setting the flag or executing the nap/sleep/rvwinkle instruction.
[ shreyas@linux.vnet.ibm.com: Edited to handle LE ]
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Shreyas B. Prabhu <shreyas@linux.vnet.ibm.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: linuxppc-dev@lists.ozlabs.org
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Luis Henriques <luis.henriques@canonical.com>
commit 51d7d5205d upstream.
The kernel defines the function spin_is_locked(), which can be used to
check if a spinlock is currently locked.
Using spin_is_locked() on a lock you don't hold is obviously racy. That
is, even though you may observe that the lock is unlocked, it may become
locked at any time.
There is (at least) one exception to that, which is if two locks are
used as a pair, and the holder of each checks the status of the other
before doing any update.
Assuming *A and *B are two locks, and *COUNTER is a shared non-atomic
value:
The first CPU does:
spin_lock(*A)
if spin_is_locked(*B)
# nothing
else
smp_mb()
LOAD r = *COUNTER
r++
STORE *COUNTER = r
spin_unlock(*A)
And the second CPU does:
spin_lock(*B)
if spin_is_locked(*A)
# nothing
else
smp_mb()
LOAD r = *COUNTER
r++
STORE *COUNTER = r
spin_unlock(*B)
Although this is a strange locking construct, it should work.
It seems to be understood, but not documented, that spin_is_locked() is
not a memory barrier, so in the examples above and below the caller
inserts its own memory barrier before acting on the result of
spin_is_locked().
For now we assume spin_is_locked() is implemented as below, and we break
it out in our examples:
bool spin_is_locked(*LOCK) {
LOAD l = *LOCK
return l.locked
}
Our intuition is that there should be no problem even if the two code
sequences run simultaneously such as:
CPU 0 CPU 1
==================================================
spin_lock(*A) spin_lock(*B)
LOAD b = *B LOAD a = *A
if b.locked # true if a.locked # true
# nothing # nothing
spin_unlock(*A) spin_unlock(*B)
If one CPU gets the lock before the other then it will do the update and
the other CPU will back off:
CPU 0 CPU 1
==================================================
spin_lock(*A)
LOAD b = *B
spin_lock(*B)
if b.locked # false LOAD a = *A
else if a.locked # true
smp_mb() # nothing
LOAD r1 = *COUNTER spin_unlock(*B)
r1++
STORE *COUNTER = r1
spin_unlock(*A)
However in reality spin_lock() itself is not indivisible. On powerpc we
implement it as a load-and-reserve and store-conditional.
Ignoring the retry logic for the lost reservation case, it boils down to:
spin_lock(*LOCK) {
LOAD l = *LOCK
l.locked = true
STORE *LOCK = l
ACQUIRE_BARRIER
}
The ACQUIRE_BARRIER is required to give spin_lock() ACQUIRE semantics as
defined in memory-barriers.txt:
This acts as a one-way permeable barrier. It guarantees that all
memory operations after the ACQUIRE operation will appear to happen
after the ACQUIRE operation with respect to the other components of
the system.
On modern powerpc systems we use lwsync for ACQUIRE_BARRIER. lwsync is
also know as "lightweight sync", or "sync 1".
As described in Power ISA v2.07 section B.2.1.1, in this scenario the
lwsync is not the barrier itself. It instead causes the LOAD of *LOCK to
act as the barrier, preventing any loads or stores in the locked region
from occurring prior to the load of *LOCK.
Whether this behaviour is in accordance with the definition of ACQUIRE
semantics in memory-barriers.txt is open to discussion, we may switch to
a different barrier in future.
What this means in practice is that the following can occur:
CPU 0 CPU 1
==================================================
LOAD a = *A LOAD b = *B
a.locked = true b.locked = true
LOAD b = *B LOAD a = *A
STORE *A = a STORE *B = b
if b.locked # false if a.locked # false
else else
smp_mb() smp_mb()
LOAD r1 = *COUNTER LOAD r2 = *COUNTER
r1++ r2++
STORE *COUNTER = r1
STORE *COUNTER = r2 # Lost update
spin_unlock(*A) spin_unlock(*B)
That is, the load of *B can occur prior to the store that makes *A
visibly locked. And similarly for CPU 1. The result is both CPUs hold
their lock and believe the other lock is unlocked.
The easiest fix for this is to add a full memory barrier to the start of
spin_is_locked(), so adding to our previous definition would give us:
bool spin_is_locked(*LOCK) {
smp_mb()
LOAD l = *LOCK
return l.locked
}
The new barrier orders the store to the lock we are locking vs the load
of the other lock:
CPU 0 CPU 1
==================================================
LOAD a = *A LOAD b = *B
a.locked = true b.locked = true
STORE *A = a STORE *B = b
smp_mb() smp_mb()
LOAD b = *B LOAD a = *A
if b.locked # true if a.locked # true
# nothing # nothing
spin_unlock(*A) spin_unlock(*B)
Although the above example is theoretical, there is code similar to this
example in sem_lock() in ipc/sem.c. This commit in addition to the next
commit appears to be a fix for crashes we are seeing in that code where
we believe this race happens in practice.
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 85101af13b upstream.
ABIv2 kernels are failing to backtrace through the kernel. An example:
39.30% readseek2_proce [kernel.kallsyms] [k] find_get_entry
|
--- find_get_entry
__GI___libc_read
The problem is in valid_next_sp() where we check that the new stack
pointer is at least STACK_FRAME_OVERHEAD below the previous one.
ABIv1 has a minimum stack frame size of 112 bytes consisting of 48 bytes
and 64 bytes of parameter save area. ABIv2 changes that to 32 bytes
with no paramter save area.
STACK_FRAME_OVERHEAD is in theory the minimum stack frame size,
but we over 240 uses of it, some of which assume that it includes
space for the parameter area.
We need to work through all our stack defines and rationalise them
but let's fix perf now by creating STACK_FRAME_MIN_SIZE and using
in valid_next_sp(). This fixes the issue:
30.64% readseek2_proce [kernel.kallsyms] [k] find_get_entry
|
--- find_get_entry
pagecache_get_page
generic_file_read_iter
new_sync_read
vfs_read
sys_read
syscall_exit
__GI___libc_read
Reported-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit fc04795575 upstream.
If we changed base page size of the segment, either via sub_page_protect
or via remap_4k_pfn, we do a demote_segment which doesn't flush the hash
table entries. We do a lazy hash page table flush for all mapped pages
in the demoted segment. This happens when we handle hash page fault for
these pages.
We use _PAGE_COMBO bit along with _PAGE_HASHPTE to indicate whether a
pte is backed by 4K hash pte. If we find _PAGE_COMBO not set on the pte,
that implies that we could possibly have older 64K hash pte entries in
the hash page table and we need to invalidate those entries.
Use _PAGE_COMBO to determine the page size with which we should
invalidate the hash table entries on unmap.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit fa1f8ae80f upstream.
The segment identifier and segment size will remain the same in
the loop, So we can compute it outside. We also change the
hugepage_invalidate interface so that we can use it the later patch
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 85c1fafd72 upstream.
On ppc64 we support 4K hash pte with 64K page size. That requires
us to track the hash pte slot information on a per 4k basis. We do that
by storing the slot details in the second half of pte page. The pte bit
_PAGE_COMBO is used to indicate whether the second half need to be
looked while building real_pte. We need to use read memory barrier while
doing that so that load of hidx is not reordered w.r.t _PAGE_COMBO
check. On the store side we already do a lwsync in __hash_page_4K
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit a0840240c0 upstream.
Unfortunately, the LPCR got defined as a 32-bit register in the
one_reg interface. This is unfortunate because KVM allows userspace
to control the DPFD (default prefetch depth) field, which is in the
upper 32 bits. The result is that DPFD always get set to 0, which
reduces performance in the guest.
We can't just change KVM_REG_PPC_LPCR to be a 64-bit register ID,
since that would break existing userspace binaries. Instead we define
a new KVM_REG_PPC_LPCR_64 id which is 64-bit. Userspace can still use
the old KVM_REG_PPC_LPCR id, but it now only modifies those fields in
the bottom 32 bits that userspace can modify (ILE, TC and AIL).
If userspace uses the new KVM_REG_PPC_LPCR_64 id, it can modify DPFD
as well.
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Pull powerpc fixes from Ben Herrenschmidt:
"Here is a handful of powerpc fixes for 3.16. They are all pretty
simple and self contained and should still make this release"
* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc:
powerpc: use _GLOBAL_TOC for memmove
powerpc/pseries: dynamically added OF nodes need to call of_node_init
powerpc: subpage_protect: Increase the array size to take care of 64TB
powerpc: Fix bugs in emulate_step()
powerpc: Disable doorbells on Power8 DD1.x
We now support TASK_SIZE of 16TB, hence the array should be 8.
Fixes the below crash:
Unable to handle kernel paging request for data at address 0x000100bd
Faulting instruction address: 0xc00000000004f914
cpu 0x13: Vector: 300 (Data Access) at [c000000fea75fa90]
pc: c00000000004f914: .sys_subpage_prot+0x2d4/0x5c0
lr: c00000000004fb5c: .sys_subpage_prot+0x51c/0x5c0
sp: c000000fea75fd10
msr: 9000000000009032
dar: 100bd
dsisr: 40000000
current = 0xc000000fea6ae490
paca = 0xc00000000fb8ab00 softe: 0 irq_happened: 0x00
pid = 8237, comm = a.out
enter ? for help
[c000000fea75fe30] c00000000000a164 syscall_exit+0x0/0x98
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
These processors do not currently support doorbell IPIs, so remove them
from the feature list if we are at DD 1.xx for the 0x004d part.
This fixes a regression caused by d4e58e5928 (powerpc/powernv: Enable
POWER8 doorbell IPIs). With that patch the kernel would hang at boot
when calling smp_call_function_many, as the doorbell would not be
received by the target CPUs:
.smp_call_function_many+0x2bc/0x3c0 (unreliable)
.on_each_cpu_mask+0x30/0x100
.cpuidle_register_driver+0x158/0x1a0
.cpuidle_register+0x2c/0x110
.powernv_processor_idle_init+0x23c/0x2c0
.do_one_initcall+0xd4/0x260
.kernel_init_freeable+0x25c/0x33c
.kernel_init+0x1c/0x120
.ret_from_kernel_thread+0x58/0x7c
Fixes: d4e58e5928 (powerpc/powernv: Enable POWER8 doorbell IPIs)
Signed-off-by: Joel Stanley <joel@jms.id.au>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
an x86 change too and it is a regression from 3.14. As it only affects
nested virtualization and there were other changes in this area in 3.16,
I am not nominating it for 3.15-stable.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (GNU/Linux)
iQIcBAABAgAGBQJTyTTfAAoJEBvWZb6bTYbyytIQAJare/EWQmNBDK57EcJBIlJS
6MW2XnASEW+KCoUw0+u3sm9eaRXQdmJRb1Aw5zxTiUIR3ZSI8MDSQr1XxEgTAOtE
vFZjonPwlbnE8edLMhH3v/6/v9oO7bwNTDYeOE2pKPRfgPRjFmj1QUOJkvzRnRwj
kS5M4RtI+VqhdyJW8f4HaWqoRaOAISp3ZjQUJQdab3DWsf9ZpNjwLNjKzGZKNvIN
Klcpi7JH32zawUfqnAvph/7NsrBGrpFRE+j+JU9LLnD9PehuXwqZbWh01g2Anbq2
TKVrmXW+YnoD1IZsDw7r/14FaeRweV7yALA/eA9F4KfSyF2Qm9RbjVVdrUYz0CHV
aIl0cZeZM8xRCLy/ZWj+dOQ23RWelZaslHSpshKOznoRsuuvVwpx93zVtRwlw2dx
4WJ2A5gYA+ZUQ7eWjk83381JXkbRDUb3cO+NL8t9GnFctCJzT/gQHjqu15f7TJ2Q
gKhmeciKOS3xY4sQ+ti6gv8CwIFYqgdTzkxedxSgS9xpiAmw9v57V7WukXoXB6zl
AyjEAk9FFOeBZ5nXs0ObK5LKjI+MJoZ3X0bin7PCuT6dFrIA2yHvo5EgMvOcUua9
8Tu9L8sWv/JsKjuqebkKxekAKvv0CV35Q8OsLpEF6RI0eXyiXy2extk1LzUuK9cx
ZVYbN263++En/tgH2AJM
=Vdqn
-----END PGP SIGNATURE-----
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Pull kvm fixes from Paolo Bonzini:
"These are mostly PPC changes for 3.16-new things. However, there is
an x86 change too and it is a regression from 3.14. As it only
affects nested virtualization and there were other changes in this
area in 3.16, I am not nominating it for 3.15-stable"
* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
KVM: x86: Check for nested events if there is an injectable interrupt
KVM: PPC: RTAS: Do byte swaps explicitly
KVM: PPC: Book3S PR: Fix ABIv2 on LE
KVM: PPC: Assembly functions exported to modules need _GLOBAL_TOC()
PPC: Add _GLOBAL_TOC for 32bit
KVM: PPC: BOOK3S: HV: Use base page size when comparing against slb value
KVM: PPC: Book3E: Unlock mmu_lock when setting caching atttribute
Instead of separate bits for every POWER8 PMU feature, have a single one
for v2.07 of the architecture.
This saves us adding a MMCR2 define for a future patch.
Cc: stable@vger.kernel.org
Signed-off-by: Joel Stanley <joel@jms.id.au>
Acked-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
In fb5a515704 "powerpc: Remove platforms/wsp and associated pieces",
we removed the last user of MMU_FTRS_A2. So remove it.
MMU_FTRS_A2 was the last user of MMU_FTR_TYPE_3E, so remove it also.
This leaves some unreachable code in mmu_context_nohash.c, so remove
that also.
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
A few bug fixes to make 3.16 work well with KVM on PowerPC:
- Fix ppc32 module builds
- Fix Little Endian hosts
- Fix Book3S HV HPTE lookup with huge pages in guest
- Fix BookE lock leak
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19 (GNU/Linux)
iQIcBAABAgAGBQJTu8HqAAoJECszeR4D/txg7u0QAMk3SC2+yBVsOAusB0YvERyU
Y9X4Lz9fALdeZf2Fd2Qk9BD5y283LDppZqyoy+dmef9DXopfCv0Kh4rl/GrlG9ny
aHeiBfJGpIpjqZnvkZP0Ln9zpyg7gMLRVNfNJvZWji8RcHly6m9/bxEkG0HnX6Hn
/2UkUzOdk2aymjzMqFXdHODdC0JsGtWtGBiVC+HOtIf1D3TX42R4KI+ieOSKjGDp
OYgN2XskOMgiXvPtEx2yMyHAAw5OTCVNdFt6Co1x0qUsz560Wy3Hy6QCwiroLrPH
rjxkHhcQN0GJJLXs/jajdDJoEp5wYLRomReZbdrKgBj+zGvQQgGRD+RO9iyfedlm
4hTw98tgmHcPgFTIXQlG5U8Cn0/oPr/k7FWBZJDpiUCTNRI/rsL6eHX7Wu/ylUfm
uvcwdl5tXdM2OMHE2wEB4pEwSAK4TNGjx237txNgaeLu4ZT8yk4TQnOXlxyMJQe7
/Bfh8oUKBqRlWAymwut8y/cazZCRDFAx88ovwqAW9GXxgB+tiCeIDLNnLYEkjEmV
8l+viAjZz3LbzLeFxCxHnNha9WhK7A7kNGhYaWn1+N2Zlz1F3u3mQm5QoZ1UJgIH
TtbwWsfM7jYrlUsJB1xTeL5Hs8JhOTp+kgLpMbRXe1sNX1xqh+OQZHsJ16VB6zU9
RiOjHnv2D9/icH0B2DsW
=+sQF
-----END PGP SIGNATURE-----
Merge tag 'signed-for-3.16' of git://github.com/agraf/linux-2.6 into kvm-master
Patch queue for 3.16 - 2014-07-08
A few bug fixes to make 3.16 work well with KVM on PowerPC:
- Fix ppc32 module builds
- Fix Little Endian hosts
- Fix Book3S HV HPTE lookup with huge pages in guest
- Fix BookE lock leak
Commit ac5a8ee8 started using _GLOBAL_TOC on ppc32 code. Unfortunately it's only
defined for 64bit targets though. Define it for ppc32 as well, fixing the build
breakage that commit introduced.
Signed-off-by: Alexander Graf <agraf@suse.de>
With guests supporting Multiple page size per segment (MPSS),
hpte_page_size returns the actual page size used. Add a new function to
return base page size and use that to compare against the the page size
calculated from SLB. Without this patch a hpte lookup can fail since
we are comparing wrong page size in kvmppc_hv_find_lock_hpte.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
