There are three ways pseries_suspend_begin() can be reached:
1. When "mem" is written to /sys/power/state:
kobj_attr_store()
-> state_store()
-> pm_suspend()
-> suspend_devices_and_enter()
-> pseries_suspend_begin()
This never works because there is no way to supply a valid stream id
using this interface, and H_VASI_STATE is called with a stream id of
zero. So this call path is useless at best.
2. When a stream id is written to /sys/devices/system/power/hibernate.
pseries_suspend_begin() is polled directly from store_hibernate()
until the stream is in the "Suspending" state (i.e. the platform is
ready for the OS to suspend execution):
dev_attr_store()
-> store_hibernate()
-> pseries_suspend_begin()
3. When a stream id is written to /sys/devices/system/power/hibernate
(continued). After #2, pseries_suspend_begin() is called once again
from the pm core:
dev_attr_store()
-> store_hibernate()
-> pm_suspend()
-> suspend_devices_and_enter()
-> pseries_suspend_begin()
This is redundant because the VASI suspend state is already known to
be Suspending.
The begin() callback of platform_suspend_ops is optional, so we can
simply remove that assignment with no loss of function.
Fixes: 32d8ad4e62 ("powerpc/pseries: Partition hibernation support")
Signed-off-by: Nathan Lynch <nathanl@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20201207215200.1785968-18-nathanl@linux.ibm.com
sys_rtas() cannot call ibm,suspend-me directly in the same way it
handles other inputs. Instead it must dispatch the request to code
that can first perform the H_JOIN sequence before any call to
ibm,suspend-me can succeed. Over time kernel/rtas.c has accreted a fair
amount of platform-specific code to implement this.
Since a different, more robust implementation of the suspend sequence
is now in the pseries platform code, we want to dispatch the request
there.
Note that invoking ibm,suspend-me via the RTAS syscall is all but
deprecated; this change preserves ABI compatibility for old programs
while providing to them the benefit of the new partition suspend
implementation. This is a behavior change in that the kernel performs
the device tree update and firmware activation before returning, but
experimentation indicates this is tolerated fine by legacy user space.
Signed-off-by: Nathan Lynch <nathanl@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20201207215200.1785968-16-nathanl@linux.ibm.com
This is a mitigation for the relatively rare occurrence where a
virtual IOA can be in a transient state that prevents the
suspend/migration from succeeding, resulting in an error from
ibm,suspend-me.
If the join/suspend sequence returns an error, it is acceptable to
retry as long as the VASI suspend session state is still
"Suspending" (i.e. the platform is still waiting for the OS to
suspend).
Retry a few times on suspend failure while this condition holds,
progressively increasing the delay between attempts. We don't want to
retry indefinitey because firmware emits an error log event on each
unsuccessful attempt.
Signed-off-by: Nathan Lynch <nathanl@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20201207215200.1785968-15-nathanl@linux.ibm.com
If we're returning an error to user space, use H_VASI_SIGNAL to send a
cancellation request to the platform. This isn't strictly required but
it communicates that Linux will not attempt to complete the suspend,
which allows the various entities involved to promptly end the
operation in progress.
Signed-off-by: Nathan Lynch <nathanl@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20201207215200.1785968-14-nathanl@linux.ibm.com
The partition suspend sequence as specified in the platform
architecture requires that all active processor threads call
H_JOIN, which:
- suspends the calling thread until it is the target of
an H_PROD; or
- immediately returns H_CONTINUE, if the calling thread is the last to
call H_JOIN. This thread is expected to call ibm,suspend-me to
completely suspend the partition.
Upon returning from ibm,suspend-me the calling thread must wake all
others using H_PROD.
rtas_ibm_suspend_me_unsafe() uses on_each_cpu() to implement this
protocol, but because of its synchronizing nature this is susceptible
to deadlock versus users of stop_machine() or other callers of
on_each_cpu().
Not only is stop_machine() intended for use cases like this, it
handles error propagation and allows us to keep the data shared
between CPUs minimal: a single atomic counter which ensures exactly
one CPU will wake the others from their joined states.
Switch the migration code to use stop_machine() and a less complex
local implementation of the H_JOIN/ibm,suspend-me logic, which
carries additional benefits:
- more informative error reporting, appropriately ratelimited
- resets the lockup detector / watchdog on resume to prevent lockup
warnings when the OS has been suspended for a time exceeding the
threshold.
Fixes: 91dc182ca6 ("[PATCH] powerpc: special-case ibm,suspend-me RTAS call")
Signed-off-by: Nathan Lynch <nathanl@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20201207215200.1785968-13-nathanl@linux.ibm.com
The behavior of rtas_ibm_suspend_me_unsafe() is to return -EAGAIN to
the caller until the specified VASI suspend session state makes the
transition from H_VASI_ENABLED to H_VASI_SUSPENDING. In the interest
of separating concerns to prepare for a new implementation of the
join/suspend sequence, extract VASI session polling logic into a
couple of local functions. Waiting for the session state to reach
H_VASI_SUSPENDING before calling rtas_ibm_suspend_me_unsafe() ensures
that we will never get an EAGAIN result necessitating a retry. No
user-visible change in behavior is intended.
Signed-off-by: Nathan Lynch <nathanl@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20201207215200.1785968-12-nathanl@linux.ibm.com
Provide a documented wrapper function for the ibm,activate-firmware
service, which must be called after a partition migration or
hibernation.
If the function is absent or the call fails, the OS will continue to
run normally with the current firmware, so there is no need to perform
any recovery. Just log it and continue.
Signed-off-by: Nathan Lynch <nathanl@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20201207215200.1785968-6-nathanl@linux.ibm.com
The pseries partition suspend sequence requires that all active CPUs
call H_JOIN, which suspends all but one of them with interrupts
disabled. The "chosen" CPU is then to call ibm,suspend-me to complete
the suspend. Upon returning from ibm,suspend-me, the chosen CPU is to
use H_PROD to wake the joined CPUs.
Using on_each_cpu() for this, as rtas_ibm_suspend_me() does to
implement partition migration, is susceptible to deadlock with other
users of on_each_cpu() and with users of stop_machine APIs. The
callback passed to on_each_cpu() is not allowed to synchronize with
other CPUs in the way it is used here.
Complicating the fix is the fact that rtas_ibm_suspend_me() also
occupies the function name that should be used to provide a more
conventional wrapper for ibm,suspend-me. Rename rtas_ibm_suspend_me()
to rtas_ibm_suspend_me_unsafe() to free up the name and indicate that
it should not gain users.
Signed-off-by: Nathan Lynch <nathanl@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20201207215200.1785968-4-nathanl@linux.ibm.com
While drmgr has had work in some areas to make its RTAS syscall
interactions endian-neutral, its code for performing partition
migration via the syscall has never worked on LE. While it is able to
complete ibm,suspend-me successfully, it crashes when attempting the
subsequent ibm,update-nodes call.
drmgr is the only known (or plausible) user of ibm,suspend-me,
ibm,update-nodes, and ibm,update-properties, so allow them only in
big-endian configurations.
Signed-off-by: Nathan Lynch <nathanl@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20201207215200.1785968-2-nathanl@linux.ibm.com
This partially reverts commit eb232b1624 ("powerpc/book3s64/kuap: Improve
error reporting with KUAP") and update the fault handler to print
[ 55.022514] Kernel attempted to access user page (7e6725b70000) - exploit attempt? (uid: 0)
[ 55.022528] BUG: Unable to handle kernel data access on read at 0x7e6725b70000
[ 55.022533] Faulting instruction address: 0xc000000000e8b9bc
[ 55.022540] Oops: Kernel access of bad area, sig: 11 [#1]
....
when the kernel access userspace address without unlocking AMR.
bad_kuap_fault() is added as part of commit 5e5be3aed2 ("powerpc/mm: Detect
bad KUAP faults") to catch userspace access incorrectly blocked by AMR. Hence
retain the full stack dump there even with hash translation. Also, add a comment
explaining the difference between hash and radix.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20201208031539.84878-1-aneesh.kumar@linux.ibm.com
The value in CIABR persists across kexec which can lead to unintended
results when the new kernel hits the old kernel's breakpoint. For
example:
0:mon> bi $loadavg_proc_show
0:mon> b
type address
1 inst c000000000519060 loadavg_proc_show+0x0/0x130
0:mon> x
$ kexec -l /mnt/vmlinux --initrd=/mnt/rootfs.cpio.gz --append='xmon=off'
$ kexec -e
$ cat /proc/loadavg
Trace/breakpoint trap
Make sure CIABR is cleared so this does not happen.
Signed-off-by: Jordan Niethe <jniethe5@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20201207010519.15597-1-jniethe5@gmail.com
Pull irq fixes from Thomas Gleixner:
"A set of updates for the interrupt subsystem:
- Make multiqueue devices which use the managed interrupt affinity
infrastructure work on PowerPC/Pseries. PowerPC does not use the
generic infrastructure for setting up PCI/MSI interrupts and the
multiqueue changes failed to update the legacy PCI/MSI
infrastructure. Make this work by passing the affinity setup
information down to the mapping and allocation functions.
- Move Jason Cooper from MAINTAINERS to CREDITS as his mail is
bouncing and he's not reachable. We hope all is well with him and
say thanks for his work over the years"
* tag 'irq-urgent-2020-12-06' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
powerpc/pseries: Pass MSI affinity to irq_create_mapping()
genirq/irqdomain: Add an irq_create_mapping_affinity() function
MAINTAINERS: Move Jason Cooper to CREDITS
Pull Kbuild fixes from Masahiro Yamada:
- Move -Wcast-align to W=3, which tends to be false-positive and there
is no tree-wide solution.
- Pass -fmacro-prefix-map to KBUILD_CPPFLAGS because it is a
preprocessor option and makes sense for .S files as well.
- Disable -gdwarf-2 for Clang's integrated assembler to avoid warnings.
- Disable --orphan-handling=warn for LLD 10.0.1 to avoid warnings.
- Fix undesirable line breaks in *.mod files.
* tag 'kbuild-fixes-v5.10-2' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:
kbuild: avoid split lines in .mod files
kbuild: Disable CONFIG_LD_ORPHAN_WARN for ld.lld 10.0.1
kbuild: Hoist '--orphan-handling' into Kconfig
Kbuild: do not emit debug info for assembly with LLVM_IAS=1
kbuild: use -fmacro-prefix-map for .S sources
Makefile.extrawarn: move -Wcast-align to W=3
Pull powerpc fixes from Michael Ellerman:
"Some more powerpc fixes for 5.10:
- Three commits fixing possible missed TLB invalidations for
multi-threaded processes when CPUs are hotplugged in and out.
- A fix for a host crash triggerable by host userspace (qemu) in KVM
on Power9.
- A fix for a host crash in machine check handling when running HPT
guests on a HPT host.
- One commit fixing potential missed TLB invalidations when using the
hash MMU on Power9 or later.
- A regression fix for machines with CPUs on node 0 but no memory.
Thanks to Aneesh Kumar K.V, Cédric Le Goater, Greg Kurz, Milan
Mohanty, Milton Miller, Nicholas Piggin, Paul Mackerras, and Srikar
Dronamraju"
* tag 'powerpc-5.10-5' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
powerpc/64s/powernv: Fix memory corruption when saving SLB entries on MCE
KVM: PPC: Book3S HV: XIVE: Fix vCPU id sanity check
powerpc/numa: Fix a regression on memoryless node 0
powerpc/64s: Trim offlined CPUs from mm_cpumasks
kernel/cpu: add arch override for clear_tasks_mm_cpumask() mm handling
powerpc/64s/pseries: Fix hash tlbiel_all_isa300 for guest kernels
powerpc/64s: Fix hash ISA v3.0 TLBIEL instruction generation
The placeholder for instruction selection should use the second
argument's operand, which is %1, not %0. This could generate incorrect
assembly code if the memory addressing of operand %0 is a different
form from that of operand %1.
Also remove the %Un placeholder because having %Un placeholders
for two operands which are based on the same local var (ptep) doesn't
make much sense. By the way, it doesn't change the current behaviour
because "<>" constraint is missing for the associated "=m".
[chleroy: revised commit log iaw segher's comments and removed %U0]
Fixes: 9bf2b5cdc5 ("powerpc: Fixes for CONFIG_PTE_64BIT for SMP support")
Cc: <stable@vger.kernel.org> # v2.6.28+
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Acked-by: Segher Boessenkool <segher@kernel.crashing.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/96354bd77977a6a933fe9020da57629007fdb920.1603358942.git.christophe.leroy@csgroup.eu
We execute certain NPU2 setup code (such as mapping an LPID to a device
in NPU2) unconditionally if an Nvlink bridge is detected. However this
cannot succeed on POWER8NVL machines and errors appear in dmesg. This is
harmless as skiboot returns an error and the only place we check it is
vfio-pci but that code does not get called on P8+ either.
This adds a check if pnv_npu2_xxx helpers are called on a machine with
NPU2 which initializes pnv_phb::npu in pnv_npu2_init();
pnv_phb::npu==NULL on POWER8/NVL (Naples).
While at this, fix NULL derefencing in pnv_npu_peers_take_ownership/
pnv_npu_peers_release_ownership which occurs when GPUs on mentioned P8s
cause EEH which happens if "vfio-pci" disables devices using
the D3 power state; the vfio-pci's disable_idle_d3 module parameter
controls this and must be set on Naples. The EEH handling clears
the entire pnv_ioda_pe struct in pnv_ioda_free_pe() hence
the NULL derefencing. We cannot recover from that but at least we stop
crashing.
Tested on
- POWER9 pvr=004e1201, Ubuntu 19.04 host, Ubuntu 18.04 vm,
NVIDIA GV100 10de:1db1 driver 418.39
- POWER8 pvr=004c0100, RHEL 7.6 host, Ubuntu 16.10 vm,
NVIDIA P100 10de:15f9 driver 396.47
Fixes: 1b785611e1 ("powerpc/powernv/npu: Add release_ownership hook")
Cc: stable@vger.kernel.org # 5.0
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20201122073828.15446-1-aik@ozlabs.ru
If the platform decides to block enabling the device nothing is printed
currently. This can lead to some confusion since the dmesg output will
usually print an error with no context e.g.
e1000e: probe of 0022:01:00.0 failed with error -22
This shouldn't be spammy since pci_enable_device() already prints a
messages when it succeeds.
Signed-off-by: Oliver O'Halloran <oohall@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20200409061337.9187-1-oohall@gmail.com
When a passthrough IO adapter is removed from a pseries machine using hash
MMU and the XIVE interrupt mode, the POWER hypervisor expects the guest OS
to clear all page table entries related to the adapter. If some are still
present, the RTAS call which isolates the PCI slot returns error 9001
"valid outstanding translations" and the removal of the IO adapter fails.
This is because when the PHBs are scanned, Linux maps automatically the
INTx interrupts in the Linux interrupt number space but these are never
removed.
This problem can be fixed by adding the corresponding unmap operation when
the device is removed. There's no pcibios_* hook for the remove case, but
the same effect can be achieved using a bus notifier.
Because INTx are shared among PHBs (and potentially across the system),
this adds tracking of virq to unmap them only when the last user is gone.
[aik: added refcounter]
Signed-off-by: Oliver O'Halloran <oohall@gmail.com>
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Tested-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Frederic Barrat <fbarrat@linux.ibm.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20201202005222.5477-1-aik@ozlabs.ru
It's sometimes handy to have a config that boots a bit like a system
under secure boot (forcing lockdown=integrity, without needing any
extra stuff like a command line option).
This config file allows that, and also turns on a few assorted security
and hardening options for good measure.
Suggested-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20201203042807.1293655-1-dja@axtens.net
Use SPRN_SPRG_SCRATCH2 as an alternative scratch register in
the early part of DSI prolog in order to avoid clobbering
SPRN_SPRG_SCRATCH0/1 used by other prologs.
The 603 doesn't like a jump from DataLoadTLBMiss to the 10 nops
that are now in the beginning of DSI exception as a result of
the feature section. To workaround this, add a jump as alternative.
It also avoids fetching 10 nops for nothing.
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/f9f8df2a2be93568768ef1ac793639f7914cf103.1606285014.git.christophe.leroy@csgroup.eu
