Sandipan Das [Tue, 4 Aug 2020 17:31:37 +0000 (23:01 +0530)]
selftests/powerpc: Fix pkey syscall redefinitions
On distros using older glibc versions, the pkey tests encounter build
failures due to redefinition of the pkey syscall numbers.
For compatibility, commit d28c7004dc6d added a wrapper for the
gettid() syscall and included syscall.h if the version of glibc used
is older than 2.30. This leads to different definitions of SYS_pkey_*
as the ones in the pkey test header set numeric constants where as the
ones from syscall.h reuse __NR_pkey_*. The compiler complains about
redefinitions since they are different.
This replaces SYS_pkey_* definitions with __NR_pkey_* such that the
definitions in both syscall.h and pkeys.h are alike. This way, if
syscall.h has to be included for compatibility reasons, builds will
still succeed.
Fixes: d28c7004dc6d ("selftests/powerpc: Add wrapper for gettid") Reported-by: Sachin Sant <sachinp@linux.vnet.ibm.com> Suggested-by: David Laight <david.laight@aculab.com> Suggested-by: Michael Ellerman <mpe@ellerman.id.au> Signed-off-by: Sandipan Das <sandipan@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/a4956d838bf59b0a71a2553c5ca81131ea8b49b9.1596561758.git.sandipan@linux.ibm.com
powerpc: Fix circular dependency between percpu.h and mmu.h
Recently random.h started including percpu.h (see commit d1b61c3d6f0c ("random32: update the net random state on interrupt and
activity")), which broke corenet64_smp_defconfig:
In file included from /linux/arch/powerpc/include/asm/paca.h:18,
from /linux/arch/powerpc/include/asm/percpu.h:13,
from /linux/include/linux/random.h:14,
from /linux/lib/uuid.c:14:
/linux/arch/powerpc/include/asm/mmu.h:139:22: error: unknown type name 'next_tlbcam_idx'
139 | DECLARE_PER_CPU(int, next_tlbcam_idx);
This is due to a circular header dependency:
asm/mmu.h includes asm/percpu.h, which includes asm/paca.h, which
includes asm/mmu.h
Which means DECLARE_PER_CPU() isn't defined when mmu.h needs it.
We can fix it by moving the include of paca.h below the include of
asm-generic/percpu.h.
This moves the include of paca.h out of the #ifdef __powerpc64__, but
that is OK because paca.h is almost entirely inside #ifdef
CONFIG_PPC64 anyway.
It also moves the include of paca.h out of the #ifdef CONFIG_SMP,
which could possibly break something, but seems to have no ill
effects.
Fixes: d1b61c3d6f0c ("random32: update the net random state on interrupt and activity") Cc: stable@vger.kernel.org # v5.8 Reported-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200804130558.292328-1-mpe@ellerman.id.au
powerpc/powernv/sriov: Fix use of uninitialised variable
Initialising the value before using it is generally regarded as a good
idea so do that.
Fixes: d117cb334531 ("powerpc/powernv/sriov: Make single PE mode a per-BAR setting") Reported-by: Nathan Chancellor <natechancellor@gmail.com> Signed-off-by: Oliver O'Halloran <oohall@gmail.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200803075408.132601-1-oohall@gmail.com
Michael Ellerman [Wed, 22 Jul 2020 02:24:22 +0000 (12:24 +1000)]
powerpc/40x: Fix assembler warning about r0
The assembler says:
arch/powerpc/kernel/head_40x.S:623: Warning: invalid register expression
It's objecting to the use of r0 as the RA argument. That's because
when RA = 0 the literal value 0 is used, rather than the content of
r0, making the use of r0 in the source potentially confusing.
Fix it to use a literal 0, the generated code is identical.
powerpc/papr_scm: Add support for fetching nvdimm 'fuel-gauge' metric
We add support for reporting 'fuel-gauge' NVDIMM metric via
PAPR_PDSM_HEALTH pdsm payload. 'fuel-gauge' metric indicates the usage
life remaining of a papr-scm compatible NVDIMM. PHYP exposes this
metric via the H_SCM_PERFORMANCE_STATS.
The metric value is returned from the pdsm by extending the return
payload 'struct nd_papr_pdsm_health' without breaking the ABI. A new
field 'dimm_fuel_gauge' to hold the metric value is introduced at the
end of the payload struct and its presence is indicated by by
extension flag PDSM_DIMM_HEALTH_RUN_GAUGE_VALID.
The patch introduces a new function papr_pdsm_fuel_gauge() that is
called from papr_pdsm_health(). If fetching NVDIMM performance stats
is supported then 'papr_pdsm_fuel_gauge()' allocated an output buffer
large enough to hold the performance stat and passes it to
drc_pmem_query_stats() that issues the HCALL to PHYP. The return value
of the stat is then populated in the 'struct
nd_papr_pdsm_health.dimm_fuel_gauge' field with extension flag
'PDSM_DIMM_HEALTH_RUN_GAUGE_VALID' set in 'struct
nd_papr_pdsm_health.extension_flags'
powerpc/papr_scm: Fetch nvdimm performance stats from PHYP
Update papr_scm.c to query dimm performance statistics from PHYP via
H_SCM_PERFORMANCE_STATS hcall and export them to user-space as PAPR
specific NVDIMM attribute 'perf_stats' in sysfs. The patch also
provide a sysfs ABI documentation for the stats being reported and
their meanings.
During NVDIMM probe time in papr_scm_nvdimm_init() a special variant
of H_SCM_PERFORMANCE_STATS hcall is issued to check if collection of
performance statistics is supported or not. If successful then a PHYP
returns a maximum possible buffer length needed to read all
performance stats. This returned value is stored in a per-nvdimm
attribute 'stat_buffer_len'.
The layout of request buffer for reading NVDIMM performance stats from
PHYP is defined in 'struct papr_scm_perf_stats' and 'struct
papr_scm_perf_stat'. These structs are used in newly introduced
drc_pmem_query_stats() that issues the H_SCM_PERFORMANCE_STATS hcall.
The sysfs access function perf_stats_show() uses value
'stat_buffer_len' to allocate a buffer large enough to hold all
possible NVDIMM performance stats and passes it to
drc_pmem_query_stats() to populate. Finally statistics reported in the
buffer are formatted into the sysfs access function output buffer.
We are currently assuming that CEDE(0) has exit latency 10us, since
there is no way for us to query from the platform. However, if the
wakeup latency of an Extended CEDE state is smaller than 10us, then we
can be sure that the exit latency of CEDE(0) cannot be more than that.
In this patch, we fix the exit latency of CEDE(0) if we discover an
Extended CEDE state with wakeup latency smaller than 10us.
Benchmark results:
On POWER8, this patch does not have any impact since the advertized
latency of Extended CEDE (1) is 30us which is higher than the default
latency of CEDE (0) which is 10us.
On POWER9 we see improvement the single-threaded performance of
ebizzy, and no regression in the wakeup latency or the number of
context-switches.
ebizzy:
2 ebizzy threads bound to the same big-core. 25% improvement in the
avg records/s with patch.
context_switch2:
There is no major regression observed with this patch as seen from the
context_switch2 benchmark.
context_switch2 across CPU0 CPU1 (Both belong to same big-core, but
different small cores). We observe a minor 0.14% regression in the
number of context-switches (higher is better).
x without_patch
* with_patch
N Min Max Median Avg Stddev
x 500 348872 362236 354712 354745.69 2711.827
* 500 349422 361452 353942 354215.4 2576.9258
Difference at 99.0% confidence
-530.288 +/- 430.963
-0.149484% +/- 0.121485%
(Student's t, pooled s = 2645.24)
context_switch2 across CPU0 CPU8 (Different big-cores). We observe a
0.37% improvement in the number of context-switches (higher is
better).
x without_patch
* with_patch
N Min Max Median Avg Stddev
x 500 287956 294940 288896 288977.23 646.59295
* 500 288300 294646 289582 290064.76 1161.9992
Difference at 99.0% confidence
1087.53 +/- 153.194
0.376337% +/- 0.0530125%
(Student's t, pooled s = 940.299)
schbench:
No major difference could be seen until the 99.9th percentile.
cpuidle: pseries: Add function to parse extended CEDE records
Currently we use CEDE with latency-hint 0 as the only other idle state
on a dedicated LPAR apart from the polling "snooze" state.
The platform might support additional extended CEDE idle states, which
can be discovered through the "ibm,get-system-parameter" rtas-call
made with CEDE_LATENCY_TOKEN.
This patch adds a function to obtain information about the extended
CEDE idle states from the platform and parse the contents to populate
an array of extended CEDE states. These idle states thus discovered
will be added to the cpuidle framework in the next patch.
dmesg on a POWER8 and POWER9 LPAR, demonstrating the output of parsing
the extended CEDE latency parameters are as follows
POWER8
[ 10.093279] xcede : xcede_record_size = 10
[ 10.093285] xcede : Record 0 : hint = 1, latency = 0x3c00 tb ticks, Wake-on-irq = 1
[ 10.093291] xcede : Record 1 : hint = 2, latency = 0x4e2000 tb ticks, Wake-on-irq = 0
[ 10.093297] cpuidle : Skipping the 2 Extended CEDE idle states
POWER9
[ 5.913180] xcede : xcede_record_size = 10
[ 5.913183] xcede : Record 0 : hint = 1, latency = 0x400 tb ticks, Wake-on-irq = 1
[ 5.913188] xcede : Record 1 : hint = 2, latency = 0x3e8000 tb ticks, Wake-on-irq = 0
[ 5.913193] cpuidle : Skipping the 2 Extended CEDE idle states
cpuidle: pseries: Set the latency-hint before entering CEDE
As per the PAPR, each H_CEDE call is associated with a latency-hint to
be passed in the VPA field "cede_latency_hint". The CEDE states that
we were implicitly entering so far is CEDE with latency-hint = 0.
This patch explicitly sets the latency hint corresponding to the CEDE
state that we are currently entering. While at it, we save the
previous hint, to be restored once we wakeup from CEDE. This will be
required in the future when we expose extended-cede states through the
cpuidle framework, where each of them will have a different
cede-latency hint.
Sandipan Das [Thu, 30 Jul 2020 05:08:46 +0000 (10:38 +0530)]
selftests/powerpc: Fix online CPU selection
The size of the CPU affinity mask must be large enough for
systems with a very large number of CPUs. Otherwise, tests
which try to determine the first online CPU by calling
sched_getaffinity() will fail. This makes sure that the size
of the allocated affinity mask is dependent on the number of
CPUs as reported by get_nprocs_conf().
perf_callchain_user_64() and perf_callchain_user_32() are nearly
identical. Consolidate into one function with thin wrappers.
Suggested-by: Nicholas Piggin <npiggin@gmail.com> Signed-off-by: Michal Suchanek <msuchanek@suse.de>
[mpe: Adapt to copy_from_user_nofault(), minor formatting] Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200406210022.32265-1-msuchanek@suse.de
powerpc/pseries/hotplug-cpu: Remove double free in error path
In the unlikely event that the device tree lacks a /cpus node,
find_dlpar_cpus_to_add() oddly frees the cpu_drcs buffer it has been
passed before returning an error. Its only caller also frees the
buffer on error.
Remove the less conventional kfree() of a caller-supplied buffer from
find_dlpar_cpus_to_add().
Nathan Lynch [Thu, 27 Jun 2019 05:30:44 +0000 (00:30 -0500)]
powerpc/pseries/mobility: Add pr_debug() for device tree changes
When investigating issues with partition migration or resource
reassignments it is helpful to have a log of which nodes and
properties in the device tree have changed. Use pr_debug() so it's
easy to enable these at runtime with the dynamic debug facility.
Nathan Lynch [Thu, 27 Jun 2019 05:30:43 +0000 (00:30 -0500)]
powerpc/pseries/mobility: Set pr_fmt()
The pr_err() callsites in mobility.c already manually include a
"mobility:" prefix, let's make it official for the benefit of messages
to be added later.
Nathan Lynch [Thu, 27 Jun 2019 05:15:37 +0000 (00:15 -0500)]
powerpc/cacheinfo: Warn if cache object chain becomes unordered
This can catch cases where the device tree has gotten mishandled into
an inconsistent state at runtime, e.g. the cache nodes for both the
source and the destination are present after a migration.
Nathan Lynch [Thu, 27 Jun 2019 05:15:36 +0000 (00:15 -0500)]
powerpc/cacheinfo: Improve diagnostics about malformed cache lists
If we have a bug which causes us to start with the wrong kind of OF
node when linking up the cache tree, it's helpful for debugging to
print information about what we found vs what we expected. So replace
uses of WARN_ON_ONCE with WARN_ONCE, which lets us include an
informative message instead of a contentless backtrace.
Nathan Lynch [Thu, 27 Jun 2019 05:15:35 +0000 (00:15 -0500)]
powerpc/cacheinfo: Use name@unit instead of full DT path in debug messages
We know that every OF node we deal with in this code is under /cpus,
so we can make the debug messages a little less verbose without losing
information.
E.g.
cacheinfo: creating L1 dcache and icache for /cpus/PowerPC,POWER8@0
cacheinfo: creating L2 ucache for /cpus/l2-cache@2006
cacheinfo: creating L3 ucache for /cpus/l3-cache@3106
becomes
cacheinfo: creating L1 dcache and icache for PowerPC,POWER8@0
cacheinfo: creating L2 ucache for l2-cache@2006
cacheinfo: creating L3 ucache for l3-cache@3106
powerpc: fix function annotations to avoid section mismatch warnings with gcc-10
Certain warnings are emitted for powerpc code when building with a gcc-10
toolset:
WARNING: modpost: vmlinux.o(.text.unlikely+0x377c): Section mismatch in
reference from the function remove_pmd_table() to the function
.meminit.text:split_kernel_mapping()
The function remove_pmd_table() references
the function __meminit split_kernel_mapping().
This is often because remove_pmd_table lacks a __meminit
annotation or the annotation of split_kernel_mapping is wrong.
Add the appropriate __init and __meminit annotations to make modpost not
complain. In all the cases there are just a single callsite from another
__init or __meminit function:
Hari Bathini [Wed, 29 Jul 2020 11:44:06 +0000 (17:14 +0530)]
powerpc/kexec_file: Enable early kernel OPAL calls
Kernels built with CONFIG_PPC_EARLY_DEBUG_OPAL enabled expects r8 & r9
to be filled with OPAL base & entry addresses respectively. Setting
these registers allows the kernel to perform OPAL calls before the
device tree is parsed.
Hari Bathini [Wed, 29 Jul 2020 11:43:49 +0000 (17:13 +0530)]
powerpc/kexec_file: Fix kexec load failure with lack of memory hole
The kexec purgatory has to run in real mode. Only the first memory
block maybe accessible in real mode. And, unlike the case with panic
kernel, no memory is set aside for regular kexec load. Another thing
to note is, the memory for crashkernel is reserved at an offset of
128MB. So, when crashkernel memory is reserved, the memory ranges to
load kexec segments shrink further as the generic code only looks for
memblock free memory ranges and in all likelihood only a tiny bit of
memory from 0 to 128MB would be available to load kexec segments.
With kdump being used by default in general, kexec file load is likely
to fail almost always. This can be fixed by changing the memory hole
lookup logic for regular kexec to use the same method as kdump. This
would mean that most kexec segments will overlap with crashkernel
memory region. That should still be ok as the pages, whose destination
address isn't available while loading, are placed in an intermediate
location till a flush to the actual destination address happens during
kexec boot sequence.
Hari Bathini [Wed, 29 Jul 2020 11:43:33 +0000 (17:13 +0530)]
powerpc/kexec_file: Add appropriate regions for memory reserve map
While initrd, elfcorehdr and backup regions are already added to the
reserve map, there are a few missing regions that need to be added to
the memory reserve map. Add them here. And now that all the changes to
load panic kernel are in place, claim likewise.
Hari Bathini [Wed, 29 Jul 2020 11:43:14 +0000 (17:13 +0530)]
powerpc/kexec_file: Prepare elfcore header for crashing kernel
Prepare elf headers for the crashing kernel's core file using
crash_prepare_elf64_headers() and pass on this info to kdump kernel by
updating its command line with elfcorehdr parameter. Also, add
elfcorehdr location to reserve map to avoid it from being stomped on
while booting.
Signed-off-by: Hari Bathini <hbathini@linux.ibm.com> Tested-by: Pingfan Liu <piliu@redhat.com> Reviewed-by: Thiago Jung Bauermann <bauerman@linux.ibm.com>
[mpe: Ensure cmdline is nul terminated] Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/159602298855.575379.15819225623219909517.stgit@hbathini
Hari Bathini [Wed, 29 Jul 2020 11:42:58 +0000 (17:12 +0530)]
powerpc/kexec_file: Setup backup region for kdump kernel
Though kdump kernel boots from loaded address, the first 64KB of it is
copied down to real 0. So, setup a backup region and let purgatory
copy the first 64KB of crashed kernel into this backup region before
booting into kdump kernel. Update reserve map with backup region and
crashed kernel's memory to avoid kdump kernel from accidentially using
that memory.
Hari Bathini [Wed, 29 Jul 2020 11:42:16 +0000 (17:12 +0530)]
powerpc/kexec_file: Restrict memory usage of kdump kernel
Kdump kernel, used for capturing the kernel core image, is supposed
to use only specific memory regions to avoid corrupting the image to
be captured. The regions are crashkernel range - the memory reserved
explicitly for kdump kernel, memory used for the tce-table, the OPAL
region and RTAS region as applicable. Restrict kdump kernel memory
to use only these regions by setting up usable-memory DT property.
Also, tell the kdump kernel to run at the loaded address by setting
the magic word at 0x5c.
Hari Bathini [Wed, 29 Jul 2020 11:40:32 +0000 (17:10 +0530)]
powerpc/drmem: Make LMB walk a bit more flexible
Currently, numa & prom are the only users of drmem LMB walk code.
Loading kdump with kexec_file also needs to walk the drmem LMBs to
setup the usable memory ranges for kdump kernel. But there are couple
of issues in using the code as is. One, walk_drmem_lmb() code is built
into the .init section currently, while kexec_file needs it later.
Two, there is no scope to pass data to the callback function for
processing and/or erroring out on certain conditions.
Fix that by, moving drmem LMB walk code out of .init section, adding
scope to pass data to the callback function and bailing out when an
error is encountered in the callback function.
Hari Bathini [Wed, 29 Jul 2020 11:40:16 +0000 (17:10 +0530)]
powerpc/kexec_file: Avoid stomping memory used by special regions
crashkernel region could have an overlap with special memory regions
like OPAL, RTAS, TCE table & such. These regions are referred to as
excluded memory ranges. Setup these ranges during image probe in order
to avoid them while finding the buffer for different kdump segments.
Override arch_kexec_locate_mem_hole() to locate a memory hole taking
these ranges into account.
Hari Bathini [Wed, 29 Jul 2020 11:40:00 +0000 (17:10 +0530)]
powerpc/kexec_file: Add helper functions for getting memory ranges
In kexec case, the kernel to be loaded uses the same memory layout as
the running kernel. So, passing on the DT of the running kernel would
be good enough.
But in case of kdump, different memory ranges are needed to manage
loading the kdump kernel, booting into it and exporting the elfcore of
the crashing kernel. The ranges are exclude memory ranges, usable
memory ranges, reserved memory ranges and crash memory ranges.
Exclude memory ranges specify the list of memory ranges to avoid while
loading kdump segments. Usable memory ranges list the memory ranges
that could be used for booting kdump kernel. Reserved memory ranges
list the memory regions for the loading kernel's reserve map. Crash
memory ranges list the memory ranges to be exported as the crashing
kernel's elfcore.
Add helper functions for setting up the above mentioned memory ranges.
This helpers facilitate in understanding the subsequent changes better
and make it easy to setup the different memory ranges listed above, as
and when appropriate.
Hari Bathini [Wed, 29 Jul 2020 11:39:41 +0000 (17:09 +0530)]
powerpc/kexec_file: Mark PPC64 specific code
Some of the kexec_file_load code isn't PPC64 specific. Move PPC64
specific code from kexec/file_load.c to kexec/file_load_64.c. Also,
rename purgatory/trampoline.S to purgatory/trampoline_64.S in the same
spirit. No functional changes.
Signed-off-by: Hari Bathini <hbathini@linux.ibm.com> Tested-by: Pingfan Liu <piliu@redhat.com> Reviewed-by: Laurent Dufour <ldufour@linux.ibm.com> Reviewed-by: Thiago Jung Bauermann <bauerman@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/159602276920.575379.10390965946438306388.stgit@hbathini
Hari Bathini [Wed, 29 Jul 2020 11:39:19 +0000 (17:09 +0530)]
kexec_file: Allow archs to handle special regions while locating memory hole
Some architectures may have special memory regions, within the given
memory range, which can't be used for the buffer in a kexec segment.
Implement weak arch_kexec_locate_mem_hole() definition which arch code
may override, to take care of special regions, while trying to locate
a memory hole.
Also, add the missing declarations for arch overridable functions and
and drop the __weak descriptors in the declarations to avoid non-weak
definitions from becoming weak.
Signed-off-by: Hari Bathini <hbathini@linux.ibm.com> Tested-by: Pingfan Liu <piliu@redhat.com> Reviewed-by: Thiago Jung Bauermann <bauerman@linux.ibm.com> Acked-by: Dave Young <dyoung@redhat.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/159602273603.575379.17665852963340380839.stgit@hbathini
Anton Blanchard [Wed, 29 Jul 2020 04:08:28 +0000 (14:08 +1000)]
powerpc/configs: Add BLK_DEV_NVME to pseries_defconfig
I've forgotten to manually enable NVME when building pseries kernels
for machines with NVME adapters. Since it's a reasonably common
configuration, enable it by default.
powerpc/64s: Move HMI IRQ stat from percpu variable to paca.
With the proposed change in percpu bootmem allocator to use page
mapping [1], the percpu first chunk memory area can come from vmalloc
ranges. This makes the HMI (Hypervisor Maintenance Interrupt) handler
crash the kernel whenever percpu variable is accessed in real mode.
This patch fixes this issue by moving the HMI IRQ stat inside paca for
safe access in realmode.
David Lamparter [Thu, 20 Sep 2018 23:04:22 +0000 (01:04 +0200)]
powerpc/fsl/dts: add missing P4080DS I2C devices
This just adds the zl2006 voltage regulators / power monitors and the
onboard I2C eeproms. The ICS9FG108 clock chip doesn't seem to have a
driver, so it is left in the DTS as a comment. And for good measure,
the SPD eeproms are tagged as such.
selftests/powerpc: Return skip code for spectre_v2
When running under older versions of qemu of under newer versions with
old machine types, some security features will not be reported to the
guest. This will lead the guest OS to consider itself Vulnerable to
spectre_v2.
So, spectre_v2 test fails in such cases when the host is mitigated and
miss predictions cannot be detected as expected by the test.
Make it return the skip code instead, for this particular case. We
don't want to miss the case when the test fails and the system reports
as mitigated or not affected. But it is not a problem to miss failures
when the system reports as Vulnerable.
Balamuruhan S [Tue, 28 Jul 2020 13:03:08 +0000 (18:33 +0530)]
powerpc/test_emulate_step: Add testcases for divde[.] and divdeu[.] instructions
Add testcases for divde, divde., divdeu, divdeu. emulated instructions
to cover few scenarios,
- with same dividend and divisor to have undefine RT
for divdeu[.]
- with divide by zero to have undefine RT for both
divde[.] and divdeu[.]
- with negative dividend to cover -|divisor| < r <= 0 if
the dividend is negative for divde[.]
- normal case with proper dividend and divisor for both
divde[.] and divdeu[.]
Signed-off-by: Balamuruhan S <bala24@linux.ibm.com> Reviewed-by: Sandipan Das <sandipan@linux.ibm.com> Acked-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200728130308.1790982-4-bala24@linux.ibm.com
Balamuruhan S [Tue, 28 Jul 2020 13:03:07 +0000 (18:33 +0530)]
powerpc/sstep: Add support for divde[.] and divdeu[.] instructions
This patch adds emulation support for divde, divdeu instructions,
- Divide Doubleword Extended (divde[.])
- Divide Doubleword Extended Unsigned (divdeu[.])
Signed-off-by: Balamuruhan S <bala24@linux.ibm.com> Reviewed-by: Sandipan Das <sandipan@linux.ibm.com> Acked-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200728130308.1790982-3-bala24@linux.ibm.com
Balamuruhan S [Tue, 28 Jul 2020 13:03:06 +0000 (18:33 +0530)]
powerpc/ppc-opcode: Add divde and divdeu opcodes
Include instruction opcodes for divde and divdeu as macros.
Signed-off-by: Balamuruhan S <bala24@linux.ibm.com> Reviewed-by: Sandipan Das <sandipan@linux.ibm.com> Acked-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200728130308.1790982-2-bala24@linux.ibm.com
Harish [Tue, 9 Jun 2020 08:14:23 +0000 (13:44 +0530)]
selftests/powerpc: Fix CPU affinity for child process
On systems with large number of cpus, test fails trying to set
affinity by calling sched_setaffinity() with smaller size for affinity
mask. This patch fixes it by making sure that the size of allocated
affinity mask is dependent on the number of CPUs as reported by
get_nprocs().
powerpc/powernv/sriov: Remove unused but set variable 'phb'
Gcc report warning as follows:
arch/powerpc/platforms/powernv/pci-sriov.c:602:25: warning:
variable 'phb' set but not used [-Wunused-but-set-variable]
602 | struct pnv_phb *phb;
| ^~~
This variable is not used, so this commit removing it.
Replace the existing /* fall through */ comments and its variants with
the new pseudo-keyword macro fallthrough[1]. Also, remove unnecessary
fall-through markings when it is the case.
powerpc/book3s64/radix: Add kernel command line option to disable radix GTSE
This adds a kernel command line option that can be used to disable GTSE support.
Disabling GTSE implies kernel will make hcalls to invalidate TLB entries.
This was done so that we can do VM migration between configs that enable/disable
GTSE support via hypervisor. To migrate a VM from a system that supports
GTSE to a system that doesn't, we can boot the guest with
radix_hcall_invalidate=on, thereby forcing the guest to use hcalls for TLB
invalidates.
The check for hcall availability is done in pSeries_setup_arch so that
the panic message appears on the console. This should only happen on
a hypervisor that doesn't force the guest to hash translation even
though it can't handle the radix GTSE=0 request via CAS. With
radix_hcall_invalidate=on if the hypervisor doesn't support hcall_rpt_invalidate
hcall it should force the LPAR to hash translation.
[ 0.000000] cma: Reserved 26224 MiB at 0x0000007959000000
[ 0.000000] hugetlb_cma: reserve 65536 MiB, up to 16384 MiB per node
[ 0.000000] cma: Reserved 16384 MiB at 0x0000001800000000
With the fix
[ 0.000000] kvm_cma_reserve: reserving 26214 MiB for global area
[ 0.000000] cma: Reserved 26224 MiB at 0x0000007959000000
[ 0.000000] hugetlb_cma: reserve 65536 MiB, up to 16384 MiB per node
[ 0.000000] cma: Reserved 16384 MiB at 0x0000001800000000
powerpc/hugetlb/cma: Allocate gigantic hugetlb pages using CMA
commit: e80bf5d82e21 ("mm: hugetlb: optionally allocate gigantic hugepages using cma")
added support for allocating gigantic hugepages using CMA. This patch
enables the same for powerpc
Balamuruhan S [Mon, 30 Mar 2020 07:59:54 +0000 (13:29 +0530)]
powerpc/xmon: Use `dcbf` inplace of `dcbi` instruction for 64bit Book3S
Data Cache Block Invalidate (dcbi) instruction implemented back in
PowerPC architecture version 2.03. But as per Power Processor Users Manual
it is obsolete and not supported by POWER8/POWER9 core. Attempt to use of
this illegal instruction results in a hypervisor emulation assistance
interrupt. So, ifdef it out the option `i` in xmon for 64bit Book3S.
Michael Ellerman [Fri, 24 Jul 2020 13:17:28 +0000 (23:17 +1000)]
powerpc: Drop old comment about CONFIG_POWER
There's a comment in time.h referring to CONFIG_POWER, which doesn't
exist. That confuses scripts/checkkconfigsymbols.py.
Presumably the comment was referring to a CONFIG_POWER vs CONFIG_PPC,
in which case for CONFIG_POWER we would #define __USE_RTC to 1. But
instead we have CONFIG_PPC_BOOK3S_601, and these days we have
IS_ENABLED().
Michael Ellerman [Fri, 24 Jul 2020 13:17:26 +0000 (23:17 +1000)]
powerpc/boot: Fix CONFIG_PPC_MPC52XX references
Commit 9cdfb4daed17 ("powerpc: conditionally compile platform-specific
serial drivers") made some code depend on CONFIG_PPC_MPC52XX, which
doesn't exist.
Michael Ellerman [Fri, 24 Jul 2020 13:17:25 +0000 (23:17 +1000)]
powerpc/32s: Remove TAUException wart in traps.c
All 32 and 64-bit builds that don't have CONFIG_TAU_INT enabled (all
of them), get a definition of TAUException() in traps.c.
On 64-bit it's completely useless, and just wastes ~120 bytes of text.
On 32-bit it allows the kernel to link because head_32.S calls it
unconditionally.
Instead follow the example of altivec_assist_exception(), and if
CONFIG_TAU_INT is not enabled just point it at unknown_exception using
the preprocessor.
Michael Ellerman [Fri, 24 Jul 2020 13:17:23 +0000 (23:17 +1000)]
powerpc/64e: Drop dead BOOK3E_MMU_TLB_STATS code
This code was merged 11 years ago in commit e285d82a2539 ("powerpc:
Add definitions used by exception handling on 64-bit Book3E") but was
never able to be built because CONFIG_BOOK3E_MMU_TLB_STATS never
existed. Remove it.
powerpc/mm: Limit resize_hpt_for_hotplug() call to hash guests only
During memory hotplug and unplug, resize_hpt_for_hotplug() gets called
for both hash and radix guests but it should be called only for hash
guests. Though the call does nothing in the radix guest case, it is
cleaner to push this call into hash specific memory hotplug routines.
We have powerpc specific logic in our page fault handling to decide if
an access to an unmapped address below the stack pointer should expand
the stack VMA.
The logic aims to prevent userspace from doing bad accesses below the
stack pointer. However as long as the stack is < 1MB in size, we allow
all accesses without further checks. Adding some debug I see that I
can do a full kernel build and LTP run, and not a single process has
used more than 1MB of stack. So for the majority of processes the
logic never even fires.
We also recently found a nasty bug in this code which could cause
userspace programs to be killed during signal delivery. It went
unnoticed presumably because most processes use < 1MB of stack.
The generic mm code has also grown support for stack guard pages since
this code was originally written, so the most heinous case of the
stack expanding into other mappings is now handled for us.
Finally although some other arches have special logic in this path,
from what I can tell none of x86, arm64, arm and s390 impose any extra
checks other than those in expand_stack().
So drop our complicated logic and like other architectures just let
the stack expand as long as its within the rlimit.
Michael Ellerman [Fri, 24 Jul 2020 09:25:25 +0000 (19:25 +1000)]
powerpc: Allow 4224 bytes of stack expansion for the signal frame
We have powerpc specific logic in our page fault handling to decide if
an access to an unmapped address below the stack pointer should expand
the stack VMA.
The code was originally added in 2004 "ported from 2.4". The rough
logic is that the stack is allowed to grow to 1MB with no extra
checking. Over 1MB the access must be within 2048 bytes of the stack
pointer, or be from a user instruction that updates the stack pointer.
The 2048 byte allowance below the stack pointer is there to cover the
288 byte "red zone" as well as the "about 1.5kB" needed by the signal
delivery code.
Unfortunately since then the signal frame has expanded, and is now
4224 bytes on 64-bit kernels with transactional memory enabled. This
means if a process has consumed more than 1MB of stack, and its stack
pointer lies less than 4224 bytes from the next page boundary, signal
delivery will fault when trying to expand the stack and the process
will see a SEGV.
The total size of the signal frame is the size of struct rt_sigframe
(which includes the red zone) plus __SIGNAL_FRAMESIZE (128 bytes on
64-bit).
The 2048 byte allowance was correct until 2008 as the signal frame
was:
At this point we should have been exposed to the bug, though as far as
I know it was never reported. I no longer have a system old enough to
easily test on.
Then in 2010 commit 121be24a1dca ("mm: keep a guard page below a
grow-down stack segment") caused our stack expansion code to never
trigger, as there was always a VMA found for a write up to PAGE_SIZE
below r1.
That meant the bug was hidden as we continued to expand the signal
frame in commit f68633cf0b04 ("powerpc: Add new transactional memory
state to the signal context") (Feb 2013):
Then finally in 2017, commit b0e13d6f9c6a ("mm: larger stack guard
gap, between vmas") exposed us to the existing bug, because it changed
the stack VMA to be the correct/real size, meaning our stack expansion
code is now triggered.
Fix it by increasing the allowance to 4224 bytes.
Hard-coding 4224 is obviously unsafe against future expansions of the
signal frame in the same way as the existing code. We can't easily use
sizeof() because the signal frame structure is not in a header. We
will either fix that, or rip out all the custom stack expansion
checking logic entirely.
Fixes: 6bd8af432365 ("powerpc: Add VSX context save/restore, ptrace and signal support") Cc: stable@vger.kernel.org # v2.6.27+ Reported-by: Tom Lane <tgl@sss.pgh.pa.us> Tested-by: Daniel Axtens <dja@axtens.net> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200724092528.1578671-2-mpe@ellerman.id.au
Michael Ellerman [Fri, 24 Jul 2020 09:25:24 +0000 (19:25 +1000)]
selftests/powerpc: Add test of stack expansion logic
We have custom stack expansion checks that it turns out are extremely
badly tested and contain bugs, surprise. So add some tests that
exercise the code and capture the current boundary conditions.
The signal test currently fails on 64-bit kernels because the 2048
byte allowance for the signal frame is too small, we will fix that in
a subsequent patch.
selftests/powerpc: Squash spurious errors due to device removal
For drivers that don't have the error handling callbacks we implement
recovery by removing the device and re-probing it. This causes the sysfs
directory for the PCI device to be removed which causes the following
spurious error to be printed when checking the PE state:
Breaking 0005:03:00.0...
./eeh-basic.sh: line 13: can't open /sys/bus/pci/devices/0005:03:00.0/eeh_pe_state: no such file
0005:03:00.0, waited 0/60
0005:03:00.0, waited 1/60
0005:03:00.0, waited 2/60
0005:03:00.0, waited 3/60
0005:03:00.0, waited 4/60
0005:03:00.0, waited 5/60
0005:03:00.0, waited 6/60
0005:03:00.0, waited 7/60
0005:03:00.0, Recovered after 8 seconds
We currently try to avoid this by checking if the PE state file exists
before reading from it. This is however inherently racy so re-work the
state checking so that we only read from the file once, and we squash any
errors that occur while reading.
Sandipan Das [Mon, 27 Jul 2020 04:00:40 +0000 (09:30 +0530)]
selftests/powerpc: Add test for pkey siginfo verification
Commit 5203b4e34bc56 ("powerpc/pkeys: Check vma before
returning key fault error to the user") fixes a bug which
causes the kernel to set the wrong pkey in siginfo when a
pkey fault occurs after two competing threads that have
allocated different pkeys, one fully permissive and the
other restrictive, attempt to protect a common page at the
same time. This adds a test to detect the bug.
Sandipan Das [Mon, 27 Jul 2020 04:00:37 +0000 (09:30 +0530)]
selftests/powerpc: Harden test for execute-disabled pkeys
Commit 631a894de8cfd ("powerpc/book3s64/pkeys: Fix
pkey_access_permitted() for execute disable pkey") fixed a
bug that caused repetitive faults for pkeys with no execute
rights alongside some combination of read and write rights.
This removes the last two cases of the test, which check
the behaviour of pkeys with read, write but no execute
rights and all the rights, in favour of checking all the
possible combinations of read, write and execute rights
to be able to detect bugs like the one mentioned above.
Sandipan Das [Mon, 27 Jul 2020 04:00:36 +0000 (09:30 +0530)]
selftests/powerpc: Add pkey helpers for rights
This adds some new pkey-related helper to print
access rights of a pkey in the "rwx" format and
to generate different valid combinations of pkey
rights starting from a given combination.
Sandipan Das [Mon, 27 Jul 2020 04:00:35 +0000 (09:30 +0530)]
selftests/powerpc: Move pkey helpers to headers
This moves all the pkey-related helpers to a new header
file and also a helper to print error messages in signal
handlers to the existing utils header file.
KVM guests have certain restrictions and performance quirks when using
doorbells. This patch moves the EPAPR KVM guest test so it can be shared
with PSERIES, and uses that in doorbell setup code to apply the KVM
guest quirks and improves IPI performance for two cases:
- PowerVM guests may now use doorbells even if they are secure.
- KVM guests no longer use doorbells if XIVE is available.
There is a valid complaint that "KVM guest" is not a very reasonable
thing to test for, it's preferable for the hypervisor to advertise
particular behaviours to the guest so they could change if the
hypervisor implementation or configuration changes. However in this case
we were already assuming a KVM guest worst case, so this patch is about
containing those quirks. If KVM later advertises fast doorbells, we
should test for that and override the quirks.
Nicholas Piggin [Sun, 26 Jul 2020 03:51:54 +0000 (13:51 +1000)]
powerpc/pseries: Use doorbells even if XIVE is available
KVM supports msgsndp in guests by trapping and emulating the
instruction, so it was decided to always use XIVE for IPIs if it is
available. However on PowerVM systems, msgsndp can be used and gives
better performance. On large systems, high XIVE interrupt rates can
have sub-linear scaling, and using msgsndp can reduce the load on
the interrupt controller.
So switch to using core local doorbells even if XIVE is available.
This reduces performance for KVM guests with an SMT topology by
about 50% for ping-pong context switching between SMT vCPUs. An
option vector (or dt-cpu-ftrs) could be defined to disable msgsndp
to get KVM performance back.
powerpc/perf: Fix MMCRA_BHRB_DISABLE define for binutils < 2.28
Commit 9daaee1c83e5 ("powerpc/perf: Add Power10 PMU feature to DT CPU
features") defines MMCRA_BHRB_DISABLE as `0x2000000000UL`. Binutils
version less than 2.28 doesn't support UL suffix.
arch/powerpc/kernel/cpu_setup_power.S: Assembler messages:
arch/powerpc/kernel/cpu_setup_power.S:250: Error: found 'L', expected: ')'
arch/powerpc/kernel/cpu_setup_power.S:250: Error: junk at end of line, first unrecognized character is `L'
arch/powerpc/kernel/cpu_setup_power.S:250: Error: found 'L', expected: ')'
arch/powerpc/kernel/cpu_setup_power.S:250: Error: found 'L', expected: ')'
arch/powerpc/kernel/cpu_setup_power.S:250: Error: junk at end of line, first unrecognized character is `L'
arch/powerpc/kernel/cpu_setup_power.S:250: Error: found 'L', expected: ')'
arch/powerpc/kernel/cpu_setup_power.S:250: Error: found 'L', expected: ')'
arch/powerpc/kernel/cpu_setup_power.S:250: Error: operand out of range (0x0000002000000000 is not between 0xffffffffffff8000 and 0x000000000000ffff)
Fix this by wrapping it with the `_UL` macro.
Fixes: 9daaee1c83e5 ("Add Power10 PMU feature to DT CPU features") Suggested-by: Michael Ellerman <mpe@ellerman.id.au> Signed-off-by: Athira Rajeev <atrajeev@linux.vnet.ibm.com> Reviewed-by: Madhavan Srinivasan <maddy@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/1595996214-5833-1-git-send-email-atrajeev@linux.vnet.ibm.com
Christophe Leroy [Mon, 29 Jun 2020 11:17:19 +0000 (11:17 +0000)]
powerpc/ptdump: Refactor update of pg_state
In note_page(), the pg_state is updated the same way in two places.
Add note_page_update_state() to do it.
Also include the display of boundary markers there as it is missing
"no level" leg, leading to a mismatch when the first two markers
are at the same address and the first displayed area uses that
address.
Christophe Leroy [Mon, 29 Jun 2020 11:15:26 +0000 (11:15 +0000)]
powerpc/32s: Use dedicated segment for modules with STRICT_KERNEL_RWX
When STRICT_KERNEL_RWX is set, we want to set NX bit on vmalloc
segments. But modules require exec.
Use a dedicated segment for modules. There is not much space
above kernel, and we don't waste vmalloc space to do alignment.
Therefore, we take the segment before PAGE_OFFSET for modules.
arch/powerpc/platforms/pseries/papr_scm.c:97:1: warning:
symbol 'papr_nd_regions' was not declared. Should it be static?
arch/powerpc/platforms/pseries/papr_scm.c:98:1: warning:
symbol 'papr_ndr_lock' was not declared. Should it be static?
Those variables are not used outside of papr_scm.c, so this
commit marks them static.
Fixes: 4b3300d7bb88 ("powerpc/papr/scm: Add bad memory ranges to nvdimm bad ranges") Reported-by: Hulk Robot <hulkci@huawei.com> Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200725091949.75234-1-weiyongjun1@huawei.com
Bill Wendling [Fri, 24 Jul 2020 22:49:01 +0000 (15:49 -0700)]
powerpc/64s: allow for clang's objdump differences
Clang's objdump emits slightly different output from GNU's objdump,
causing a list of warnings to be emitted during relocatable builds.
E.g., clang's objdump emits this:
Adjust llvm-objdump's output to remove the extraneous '0x' and convert
'bf' and 'bt' to 'bne' and 'beq' resp. to more closely match GNU
objdump's output.
Note that clang's objdump doesn't yet output the relocation symbols on
PPC.
Nicholas Piggin [Fri, 24 Jul 2020 13:14:23 +0000 (23:14 +1000)]
powerpc: Implement smp_cond_load_relaxed()
This implements smp_cond_load_relaxed() with the slowpath busy loop
using the preferred SMT priority pattern.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Acked-by: Waiman Long <longman@redhat.com>
[mpe: Make it 64-bit only to fix build errors on 32-bit] Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200724131423.1362108-7-npiggin@gmail.com
Nicholas Piggin [Fri, 24 Jul 2020 13:14:21 +0000 (23:14 +1000)]
powerpc/pseries: Implement paravirt qspinlocks for SPLPAR
This implements the generic paravirt qspinlocks using H_PROD and
H_CONFER to kick and wait.
This uses an un-directed yield to any CPU rather than the directed
yield to a pre-empted lock holder that paravirtualised simple
spinlocks use, that requires no kick hcall. This is something that
could be investigated and improved in future.
Performance results can be found in the commit which added queued
spinlocks.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Acked-by: Waiman Long <longman@redhat.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200724131423.1362108-5-npiggin@gmail.com
Nicholas Piggin [Fri, 24 Jul 2020 13:14:20 +0000 (23:14 +1000)]
powerpc/64s: Implement queued spinlocks and rwlocks
These have shown significantly improved performance and fairness when
spinlock contention is moderate to high on very large systems.
With this series including subsequent patches, on a 16 socket 1536
thread POWER9, a stress test such as same-file open/close from all
CPUs gets big speedups, 11620op/s aggregate with simple spinlocks vs
384158op/s (33x faster), where the difference in throughput between
the fastest and slowest thread goes from 7x to 1.4x.
Thanks to the fast path being identical in terms of atomics and
barriers (after a subsequent optimisation patch), single threaded
performance is not changed (no measurable difference).
On smaller systems, performance and fairness seems to be generally
improved. Using dbench on tmpfs as a test (that starts to run into
kernel spinlock contention), a 2-socket OpenPOWER POWER9 system was
tested with bare metal and KVM guest configurations. Results can be
found here:
- Queued spinlocks are equal when contention is insignificant, as
expected and as measured with microbenchmarks.
- When there is contention, on bare metal queued spinlocks have better
throughput and max latency at all points.
- When virtualised, queued spinlocks are slightly worse approaching
peak throughput, but significantly better throughput and max latency
at all points beyond peak, until queued spinlock maximum latency
rises when clients are 2x vCPUs.
The regressions haven't been analysed very well yet, there are a lot
of things that can be tuned, particularly the paravirtualised locking,
but the numbers already look like a good net win even on relatively
small systems.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Acked-by: Waiman Long <longman@redhat.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://lore.kernel.org/r/20200724131423.1362108-4-npiggin@gmail.com
projects / kernel.git / log