That was a workaround in the XICS domain because of the lack of MSI
domain. This is now handled.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-24-clg@kaod.org
The PowerNV and pSeries platforms now have support for both the XICS
and XIVE IRQ domains.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-23-clg@kaod.org
PHB3s need an extra OPAL call to EOI the interrupt. The call takes an
OPAL HW IRQ number but it is translated into a vector number in OPAL.
Here, we directly use the vector number of the in-the-middle "PNV-MSI"
domain instead of grabbing the OPAL HW IRQ number in the XICS parent
domain.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-22-clg@kaod.org
XICS doesn't have any state associated with the IRQ. The support is
straightforward and simpler than for XIVE.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-21-clg@kaod.org
This moves the IRQ initialization done under the different ICS backends
in the common part of XICS. The 'map' handler becomes a simple 'check'
on the HW IRQ at the FW level.
As we don't need an ICS anymore in xics_migrate_irqs_away(), the XICS
domain does not set a chip data for the IRQ.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-18-clg@kaod.org
We always had only one ICS per machine. Simplify the XICS driver by
removing the ICS list.
The ICS stored in the chip data of the XICS domain becomes useless and
we don't need it anymore to migrate away IRQs from a CPU. This will be
removed in a subsequent patch.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-17-clg@kaod.org
PCI MSI interrupt numbers are now mapped in a PCI-MSI domain but the
underlying calls handling the passthrough of the interrupt in the
guest need a number in the XIVE IRQ domain.
Use the IRQ data mapped in the XIVE IRQ domain and not the one in the
PCI-MSI domain.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-16-clg@kaod.org
The routine kvmppc_set_passthru_irq() calls kvmppc_xive_set_mapped()
and kvmppc_xive_clr_mapped() with an IRQ descriptor. Use directly the
host IRQ number to remove a useless conversion.
Add some debug.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-15-clg@kaod.org
Passthrough PCI MSI interrupts are detected in KVM with a check on a
specific EOI handler (P8) or on XIVE (P9). We can now check the
PCI-MSI IRQ chip which is cleaner.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-14-clg@kaod.org
This is very similar to the MSI domains of the pSeries platform. The
MSI allocator is directly handled under the Linux PHB in the
in-the-middle "PNV-MSI" domain.
Only the XIVE (P9/P10) parent domain is supported for now. Support for
XICS will come later.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-13-clg@kaod.org
Simply allocate or release the MSI domains when a PHB is inserted in
or removed from the machine.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-11-clg@kaod.org
The MSI domain clears the IRQ with msi_domain_free(), which calls
irq_domain_free_irqs_top(), which clears the handler data. This is a
problem for the XIVE controller since we need to unmap MMIO pages and
free a specific XIVE structure.
The 'msi_free()' handler is called before irq_domain_free_irqs_top()
when the handler data is still available. Use that to clear the XIVE
controller data.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-10-clg@kaod.org
The RTAS firmware can not disable one MSI at a time. It's all or
nothing. We need a custom free IRQ handler for that.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-9-clg@kaod.org
In the early days of XIVE support, commit cffb717ceb ("powerpc/xive:
Ensure active irqd when setting affinity") tried to fix an issue
related to interrupt migration. If the root cause was related to CPU
unplug, it should have been fixed and there is no reason to keep the
irqd_is_started() check. This test is also breaking affinity setting
of MSIs which can set before starting the associated IRQ.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-8-clg@kaod.org
That was a workaround in the XIVE domain because of the lack of MSI
domain. This is now handled.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-7-clg@kaod.org
Two IRQ domains are added on top of default machine IRQ domain.
First, the top level "pSeries-PCI-MSI" domain deals with the MSI
specificities. In this domain, the HW IRQ numbers are generated by the
PCI MSI layer, they compose a unique ID for an MSI source with the PCI
device identifier and the MSI vector number.
These numbers can be quite large on a pSeries machine running under
the IBM Hypervisor and /sys/kernel/irq/ and /proc/interrupts will
require small fixes to show them correctly.
Second domain is the in-the-middle "pSeries-MSI" domain which acts as
a proxy between the PCI MSI subsystem and the machine IRQ subsystem.
It usually allocate the MSI vector numbers but, on pSeries machines,
this is done by the RTAS FW and RTAS returns IRQ numbers in the IRQ
number space of the machine. This is why the in-the-middle "pSeries-MSI"
domain has the same HW IRQ numbers as its parent domain.
Only the XIVE (P9/P10) parent domain is supported for now. We still
need to add support for IRQ domain hierarchy under XICS.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-6-clg@kaod.org
pr_debug() is easier to activate and it helps to know how the kernel
configures the HW when tweaking the IRQ subsystem.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-5-clg@kaod.org
This adds handlers to allocate/free IRQs in a domain hierarchy. We
could try to use xive_irq_domain_map() in xive_irq_domain_alloc() but
we rely on xive_irq_alloc_data() to set the IRQ handler data and
duplicating the code is simpler.
xive_irq_free_data() needs to be called when IRQ are freed to clear
the MMIO mappings and free the XIVE handler data, xive_irq_data
structure. This is going to be a problem with MSI domains which we
will address later.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-4-clg@kaod.org
This splits the routine setting the MSIs in two parts: allocation of
MSIs for the PCI device at the FW level (RTAS) and the actual mapping
and activation of the IRQs.
rtas_prepare_msi_irqs() will serve as a handler for the PCI MSI domain.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210701132750.1475580-3-clg@kaod.org
The powernv_get_random_long() does not work in nested KVM (which is
pseries) and produces a crash when accessing in_be64(rng->regs) in
powernv_get_random_long().
This replaces powernv_get_random_long with the ppc_md machine hook
wrapper.
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Reviewed-by: Fabiano Rosas <farosas@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210805075649.2086567-1-aik@ozlabs.ru
We shouldn't need legacy ptys, and disabling the option improves boot
time by about 0.5 seconds.
Signed-off-by: Anton Blanchard <anton@ozlabs.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210805112005.3cb1f412@kryten.localdomain
This is the same as commit acdad8fb4a ("powerpc: Force inlining of
mmu_has_feature to fix build failure") but for radix_enabled(). The
config in the linked bugzilla causes the following build failure:
LD .tmp_vmlinux.kallsyms1
powerpc64-linux-ld: arch/powerpc/mm/pgtable.o: in function `.__ptep_set_access_flags':
pgtable.c:(.text+0x17c): undefined reference to `.radix__ptep_set_access_flags'
powerpc64-linux-ld: arch/powerpc/mm/pageattr.o: in function `.change_page_attr':
pageattr.c:(.text+0xc0): undefined reference to `.radix__flush_tlb_kernel_range'
etc.
This is due to radix_enabled() not being inlined. See extract from
building with -Winline:
In file included from arch/powerpc/include/asm/lppaca.h:46,
from arch/powerpc/include/asm/paca.h:17,
from arch/powerpc/include/asm/current.h:13,
from include/linux/thread_info.h:23,
from include/asm-generic/preempt.h:5,
from ./arch/powerpc/include/generated/asm/preempt.h:1,
from include/linux/preempt.h:78,
from include/linux/spinlock.h:51,
from include/linux/mmzone.h:8,
from include/linux/gfp.h:6,
from arch/powerpc/mm/pgtable.c:21:
arch/powerpc/include/asm/book3s/64/pgtable.h: In function '__ptep_set_access_flags':
arch/powerpc/include/asm/mmu.h:327:20: error: inlining failed in call to 'radix_enabled': call is unlikely and code size would grow [-Werror=inline]
The code relies on constant folding of MMU_FTRS_POSSIBLE at buildtime
and elimination of non possible parts of code at compile time. For this
to work radix_enabled() must be inlined so make it __always_inline.
Reported-by: Erhard F. <erhard_f@mailbox.org>
Suggested-by: Michael Ellerman <mpe@ellerman.id.au>
Tested-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Jordan Niethe <jniethe5@gmail.com>
[mpe: Trimmed error messages in change log]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://bugzilla.kernel.org/show_bug.cgi?id=213803
Link: https://lore.kernel.org/r/20210804013724.514468-1-jniethe5@gmail.com
The functions get_online_cpus() and put_online_cpus() have been
deprecated during the CPU hotplug rework. They map directly to
cpus_read_lock() and cpus_read_unlock().
Replace deprecated CPU-hotplug functions with the official version.
The behavior remains unchanged.
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210803141621.780504-4-bigeasy@linutronix.de
for_each_node_by_type should have of_node_put() before return.
Generated by: scripts/coccinelle/iterators/for_each_child.cocci
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: kernel test robot <lkp@intel.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/alpine.DEB.2.22.394.2108031654080.17639@hadrien
When a CPU is hot added, the CPU ids are taken from the available mask
from the lower possible set. If that set of values was previously used
for a CPU attached to a different node, it appears to an application as
if these CPUs have migrated from one node to another node which is not
expected.
To prevent this, it is needed to record the CPU ids used for each node
and to not reuse them on another node. However, to prevent CPU hot plug
to fail, in the case the CPU ids is starved on a node, the capability to
reuse other nodes’ free CPU ids is kept. A warning is displayed in such
a case to warn the user.
A new CPU bit mask (node_recorded_ids_map) is introduced for each
possible node. It is populated with the CPU onlined at boot time, and
then when a CPU is hot plugged to a node. The bits in that mask remain
when the CPU is hot unplugged, to remind this CPU ids have been used for
this node.
If no id set was found, a retry is made without removing the ids used on
the other nodes to try reusing them. This is the way ids have been
allocated prior to this patch.
The effect of this patch can be seen by removing and adding CPUs using
the Qemu monitor. In the following case, the first CPU from the node 2
is removed, then the first one from the node 1 is removed too. Later,
the first CPU of the node 2 is added back. Without that patch, the
kernel will number these CPUs using the first CPU ids available which
are the ones freed when removing the second CPU of the node 0. This
leads to the CPU ids 16-23 to move from the node 1 to the node 2. With
the patch applied, the CPU ids 32-39 are used since they are the lowest
free ones which have not been used on another node.
At boot time:
[root@vm40 ~]# numactl -H | grep cpus
node 0 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
node 1 cpus: 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
node 2 cpus: 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47
Vanilla kernel, after the CPU hot unplug/plug operations:
[root@vm40 ~]# numactl -H | grep cpus
node 0 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
node 1 cpus: 24 25 26 27 28 29 30 31
node 2 cpus: 16 17 18 19 20 21 22 23 40 41 42 43 44 45 46 47
Patched kernel, after the CPU hot unplug/plug operations:
[root@vm40 ~]# numactl -H | grep cpus
node 0 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
node 1 cpus: 24 25 26 27 28 29 30 31
node 2 cpus: 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47
Signed-off-by: Laurent Dufour <ldufour@linux.ibm.com>
Reviewed-by: Nathan Lynch <nathanl@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210429174908.16613-1-ldufour@linux.ibm.com
After a LPM, the device tree node ibm,dynamic-reconfiguration-memory may be
updated by the hypervisor in the case the NUMA topology of the LPAR's
memory is updated.
This is handled by the kernel, but the memory's node is not updated because
there is no way to move a memory block between nodes from the Linux kernel
point of view.
If later a memory block is added or removed, drmem_update_dt() is called
and it is overwriting the DT node ibm,dynamic-reconfiguration-memory to
match the added or removed LMB. But the LMB's associativity node has not
been updated after the DT node update and thus the node is overwritten by
the Linux's topology instead of the hypervisor one.
Introduce a hook called when the ibm,dynamic-reconfiguration-memory node is
updated to force an update of the LMB's associativity. However, ignore the
call to that hook when the update has been triggered by drmem_update_dt().
Because, in that case, the LMB tree has been used to set the DT property
and thus it doesn't need to be updated back. Since drmem_update_dt() is
called under the protection of the device_hotplug_lock and the hook is
called in the same context, use a simple boolean variable to detect that
call.
Signed-off-by: Laurent Dufour <ldufour@linux.ibm.com>
Reviewed-by: Nathan Lynch <nathanl@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210517090606.56930-1-ldufour@linux.ibm.com
When a LPAR is migratable, we should consider the maximum possible NUMA
node instead of the number of NUMA nodes from the actual system.
The DT property 'ibm,current-associativity-domains' defines the maximum
number of nodes the LPAR can see when running on that box. But if the
LPAR is being migrated on another box, it may see up to the nodes
defined by 'ibm,max-associativity-domains'. So if a LPAR is migratable,
that value should be used.
Unfortunately, there is no easy way to know if an LPAR is migratable or
not. The hypervisor exports the property 'ibm,migratable-partition' in
the case it set to migrate partition, but that would not mean that the
current partition is migratable.
Without this patch, when a LPAR is started on a 2 node box and then
migrated to a 3 node box, the hypervisor may spread the LPAR's CPUs on
the 3rd node. In that case if a CPU from that 3rd node is added to the
LPAR, it will be wrongly assigned to the node because the kernel has
been set to use up to 2 nodes (the configuration of the departure node).
With this patch applies, the CPU is correctly added to the 3rd node.
Fixes: f9f130ff2e ("powerpc/numa: Detect support for coregroup")
Signed-off-by: Laurent Dufour <ldufour@linux.ibm.com>
Reviewed-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210511073136.17795-1-ldufour@linux.ibm.com
If an interrupt is taken in kernel mode, always use SIAR for it rather than
looking at regs_sipr. This prevents samples piling up around interrupt
enable (hard enable or interrupt replay via soft enable) in PMUs / modes
where the PR sample indication is not in synch with SIAR.
This results in better sampling of interrupt entry and exit in particular.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Tested-by: Athira Rajeev <atrajeev@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210720141504.420110-1-npiggin@gmail.com
On POWER10 systems, the "ibm,thread-groups" property "2" indicates the cpus
in thread-group share both L2 and L3 caches. Hence, use cache_property = 2
itself to find both the L2 and L3 cache siblings.
Hence, create a new thread_group_l3_cache_map to keep list of L3 siblings,
but fill the mask using same property "2" array.
Signed-off-by: Parth Shah <parth@linux.ibm.com>
Reviewed-by: Gautham R. Shenoy <ego@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210728175607.591679-4-parth@linux.ibm.com
The helper function get_shared_cpu_map() was added in
'commit 500fe5f550 ("powerpc/cacheinfo: Report the correct
shared_cpu_map on big-cores")'
and subsequently expanded upon in
'commit 0be47634db ("powerpc/cacheinfo: Print correct cache-sibling
map/list for L2 cache")'
in order to help report the correct groups of threads sharing these caches
on big-core systems where groups of threads within a core can share
different sets of caches.
Now that powerpc/cacheinfo is aware of "ibm,thread-groups" property,
cache->shared_cpu_map contains the correct set of thread-siblings
sharing the cache. Hence we no longer need the functions
get_shared_cpu_map(). This patch removes this function. We also remove
the helper function index_dir_to_cpu() which was only called by
get_shared_cpu_map().
With these functions removed, we can still see the correct
cache-sibling map/list for L1 and L2 caches on systems with L1 and L2
caches distributed among groups of threads in a core.
With this patch, on a SMT8 POWER10 system where the L1 and L2 caches
are split between the two groups of threads in a core, for CPUs 8,9,
the L1-Data, L1-Instruction, L2, L3 cache CPU sibling list is as
follows:
$ grep . /sys/devices/system/cpu/cpu[89]/cache/index[0123]/shared_cpu_list
/sys/devices/system/cpu/cpu8/cache/index0/shared_cpu_list:8,10,12,14
/sys/devices/system/cpu/cpu8/cache/index1/shared_cpu_list:8,10,12,14
/sys/devices/system/cpu/cpu8/cache/index2/shared_cpu_list:8,10,12,14
/sys/devices/system/cpu/cpu8/cache/index3/shared_cpu_list:8-15
/sys/devices/system/cpu/cpu9/cache/index0/shared_cpu_list:9,11,13,15
/sys/devices/system/cpu/cpu9/cache/index1/shared_cpu_list:9,11,13,15
/sys/devices/system/cpu/cpu9/cache/index2/shared_cpu_list:9,11,13,15
/sys/devices/system/cpu/cpu9/cache/index3/shared_cpu_list:8-15
$ ppc64_cpu --smt=4
$ grep . /sys/devices/system/cpu/cpu[89]/cache/index[0123]/shared_cpu_list
/sys/devices/system/cpu/cpu8/cache/index0/shared_cpu_list:8,10
/sys/devices/system/cpu/cpu8/cache/index1/shared_cpu_list:8,10
/sys/devices/system/cpu/cpu8/cache/index2/shared_cpu_list:8,10
/sys/devices/system/cpu/cpu8/cache/index3/shared_cpu_list:8-11
/sys/devices/system/cpu/cpu9/cache/index0/shared_cpu_list:9,11
/sys/devices/system/cpu/cpu9/cache/index1/shared_cpu_list:9,11
/sys/devices/system/cpu/cpu9/cache/index2/shared_cpu_list:9,11
/sys/devices/system/cpu/cpu9/cache/index3/shared_cpu_list:8-11
$ ppc64_cpu --smt=2
$ grep . /sys/devices/system/cpu/cpu[89]/cache/index[0123]/shared_cpu_list
/sys/devices/system/cpu/cpu8/cache/index0/shared_cpu_list:8
/sys/devices/system/cpu/cpu8/cache/index1/shared_cpu_list:8
/sys/devices/system/cpu/cpu8/cache/index2/shared_cpu_list:8
/sys/devices/system/cpu/cpu8/cache/index3/shared_cpu_list:8-9
/sys/devices/system/cpu/cpu9/cache/index0/shared_cpu_list:9
/sys/devices/system/cpu/cpu9/cache/index1/shared_cpu_list:9
/sys/devices/system/cpu/cpu9/cache/index2/shared_cpu_list:9
/sys/devices/system/cpu/cpu9/cache/index3/shared_cpu_list:8-9
$ ppc64_cpu --smt=1
$ grep . /sys/devices/system/cpu/cpu[89]/cache/index[0123]/shared_cpu_list
/sys/devices/system/cpu/cpu8/cache/index0/shared_cpu_list:8
/sys/devices/system/cpu/cpu8/cache/index1/shared_cpu_list:8
/sys/devices/system/cpu/cpu8/cache/index2/shared_cpu_list:8
/sys/devices/system/cpu/cpu8/cache/index3/shared_cpu_list:8
Signed-off-by: Gautham R. Shenoy <ego@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210728175607.591679-3-parth@linux.ibm.com
Currently the cacheinfo code on powerpc indexes the "cache" objects
(modelling the L1/L2/L3 caches) where the key is device-tree node
corresponding to that cache. On some of the POWER server platforms
thread-groups within the core share different sets of caches (Eg: On
SMT8 POWER9 systems, threads 0,2,4,6 of a core share L1 cache and
threads 1,3,5,7 of the same core share another L1 cache). On such
platforms, there is a single device-tree node corresponding to that
cache and the cache-configuration within the threads of the core is
indicated via "ibm,thread-groups" device-tree property.
Since the current code is not aware of the "ibm,thread-groups"
property, on the aforementoined systems, cacheinfo code still treats
all the threads in the core to be sharing the cache because of the
single device-tree node (In the earlier example, the cacheinfo code
would says CPUs 0-7 share L1 cache).
In this patch, we make the powerpc cacheinfo code aware of the
"ibm,thread-groups" property. We indexe the "cache" objects by the
key-pair (device-tree node, thread-group id). For any CPUX, for a
given level of cache, the thread-group id is defined to be the first
CPU in the "ibm,thread-groups" cache-group containing CPUX. For levels
of cache which are not represented in "ibm,thread-groups" property,
the thread-group id is -1.
[parth: Remove "static" keyword for the definition of "thread_group_l1_cache_map"
and "thread_group_l2_cache_map" to get rid of the compile error.]
Signed-off-by: Gautham R. Shenoy <ego@linux.vnet.ibm.com>
Signed-off-by: Parth Shah <parth@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210728175607.591679-2-parth@linux.ibm.com
Currently, the install target in arch/powerpc/Makefile descends into
arch/powerpc/boot/Makefile to invoke the shell script, but there is no
good reason to do so.
arch/powerpc/Makefile can run the shell script directly.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210729141937.445051-3-masahiroy@kernel.org
The install target should not depend on any build artifact.
The reason is explained in commit 19514fc665 ("arm, kbuild: make
"make install" not depend on vmlinux").
Change the PowerPC installation code in a similar way.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210729141937.445051-2-masahiroy@kernel.org
Commit c913e5f95e ("powerpc/boot: Don't install zImage.* from make
install") added the zInstall target to arch/powerpc/boot/Makefile,
but you cannot use it since the corresponding hook is missing in
arch/powerpc/Makefile.
It has never worked since its addition. Nobody has complained about
it for 7 years, which means this code was unneeded.
With this removal, the install.sh will be passed in with 4 parameters.
Simplify the shell script.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210729141937.445051-1-masahiroy@kernel.org
After commit 7cbd631d4dec ("cpuidle: pseries: Fixup CEDE0 latency only
for POWER10 onwards"), pseries_idle_probe() is no longer inlined when
compiling with clang, which causes a modpost warning:
WARNING: modpost: vmlinux.o(.text+0xc86a54): Section mismatch in
reference from the function pseries_idle_probe() to the function
.init.text:fixup_cede0_latency()
The function pseries_idle_probe() references
the function __init fixup_cede0_latency().
This is often because pseries_idle_probe lacks a __init
annotation or the annotation of fixup_cede0_latency is wrong.
pseries_idle_probe() is a non-init function, which calls
fixup_cede0_latency(), which is an init function, explaining the
mismatch. pseries_idle_probe() is only called from
pseries_processor_idle_init(), which is an init function, so mark
pseries_idle_probe() as __init so there is no more warning.
Fixes: 054e44ba99 ("cpuidle: pseries: Add function to parse extended CEDE records")
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210803211547.1093820-1-nathan@kernel.org
commit 7c6986ade6 ("powerpc/stacktrace: Fix spurious "stale" traces in raise_backtrace_ipi()")
introduces udelay() call without including the linux/delay.h header.
This may happen to work on master but the header that declares the
functionshould be included nonetheless.
Fixes: 7c6986ade6 ("powerpc/stacktrace: Fix spurious "stale" traces in raise_backtrace_ipi()")
Signed-off-by: Michal Suchanek <msuchanek@suse.de>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20210729180103.15578-1-msuchanek@suse.de
Currently in fixup_cede0_latency() code, we perform the fixup the
CEDE(0) exit latency value only if minimum advertized extended CEDE
latency values are less than 10us. This was done so as to not break
the expected behaviour on POWER8 platforms where the advertised
latency was higher than the default 10us, which would delay the SMT
folding on the core.
However, after the earlier patch "cpuidle/pseries: Fixup CEDE0 latency
only for POWER10 onwards", we can be sure that the fixup of CEDE0
latency is going to happen only from POWER10 onwards. Hence
unconditionally use the minimum exit latency provided by the platform.
Signed-off-by: Gautham R. Shenoy <ego@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/1626676399-15975-3-git-send-email-ego@linux.vnet.ibm.com
Commit d947fb4c96 ("cpuidle: pseries: Fixup exit latency for
CEDE(0)") sets the exit latency of CEDE(0) based on the latency values
of the Extended CEDE states advertised by the platform
On POWER9 LPARs, the firmwares advertise a very low value of 2us for
CEDE1 exit latency on a Dedicated LPAR. The latency advertized by the
PHYP hypervisor corresponds to the latency required to wakeup from the
underlying hardware idle state. However the wakeup latency from the
LPAR perspective should include
1. The time taken to transition the CPU from the Hypervisor into the
LPAR post wakeup from platform idle state
2. Time taken to send the IPI from the source CPU (waker) to the idle
target CPU (wakee).
1. can be measured via timer idle test, where we queue a timer, say
for 1ms, and enter the CEDE state. When the timer fires, in the timer
handler we compute how much extra timer over the expected 1ms have we
consumed. On a a POWER9 LPAR the numbers are
CEDE latency measured using a timer (numbers in ns)
N Min Median Avg 90%ile 99%ile Max Stddev
400 2601 5677 5668.74 5917 6413 9299 455.01
1. and 2. combined can be determined by an IPI latency test where we
send an IPI to an idle CPU and in the handler compute the time
difference between when the IPI was sent and when the handler ran. We
see the following numbers on POWER9 LPAR.
CEDE latency measured using an IPI (numbers in ns)
N Min Median Avg 90%ile 99%ile Max Stddev
400 711 7564 7369.43 8559 9514 9698 1200.01
Suppose, we consider the 99th percentile latency value measured using
the IPI to be the wakeup latency, the value would be 9.5us This is in
the ballpark of the default value of 10us.
Hence, use the exit latency of CEDE(0) based on the latency values
advertized by platform only from POWER10 onwards. The values
advertized on POWER10 platforms is more realistic and informed by the
latency measurements. For earlier platforms stick to the default value
of 10us. The fix was suggested by Michael Ellerman.
Fixes: d947fb4c96 ("cpuidle: pseries: Fixup exit latency for CEDE(0)")
Reported-by: Enrico Joedecke <joedecke@de.ibm.com>
Signed-off-by: Gautham R. Shenoy <ego@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/1626676399-15975-2-git-send-email-ego@linux.vnet.ibm.com
The H_ENTER_NESTED hypercall is handled by the L0, and it is a request
by the L1 to switch the context of the vCPU over to that of its L2
guest, and return with an interrupt indication. The L1 is responsible
for switching some registers to guest context, and the L0 switches
others (including all the hypervisor privileged state).
If the L2 MSR has TM active, then the L1 is responsible for
recheckpointing the L2 TM state. Then the L1 exits to L0 via the
H_ENTER_NESTED hcall, and the L0 saves the TM state as part of the exit,
and then it recheckpoints the TM state as part of the nested entry and
finally HRFIDs into the L2 with TM active MSR. Not efficient, but about
the simplest approach for something that's horrendously complicated.
Problems arise if the L1 exits to the L0 with a TM state which does not
match the L2 TM state being requested. For example if the L1 is
transactional but the L2 MSR is non-transactional, or vice versa. The
L0's HRFID can take a TM Bad Thing interrupt and crash.
Fix this by disallowing H_ENTER_NESTED in TM[T] state entirely, and then
ensuring that if the L1 is suspended then the L2 must have TM active,
and if the L1 is not suspended then the L2 must not have TM active.
Fixes: 360cae3137 ("KVM: PPC: Book3S HV: Nested guest entry via hypercall")
Cc: stable@vger.kernel.org # v4.20+
Reported-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Acked-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The kvmppc_rtas_hcall() sets the host rtas_args.rets pointer based on
the rtas_args.nargs that was provided by the guest. That guest nargs
value is not range checked, so the guest can cause the host rets pointer
to be pointed outside the args array. The individual rtas function
handlers check the nargs and nrets values to ensure they are correct,
but if they are not, the handlers store a -3 (0xfffffffd) failure
indication in rets[0] which corrupts host memory.
Fix this by testing up front whether the guest supplied nargs and nret
would exceed the array size, and fail the hcall directly without storing
a failure indication to rets[0].
Also expand on a comment about why we kill the guest and try not to
return errors directly if we have a valid rets[0] pointer.
Fixes: 8e591cb720 ("KVM: PPC: Book3S: Add infrastructure to implement kernel-side RTAS calls")
Cc: stable@vger.kernel.org # v3.10+
Reported-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
