Introduce yet another X86_FEATURE flag framework to manage and cache KVM
governed features (for lack of a better name). "Governed" in this case
means that KVM has some level of involvement and/or vested interest in
whether or not an X86_FEATURE can be used by the guest. The intent of the
framework is twofold: to simplify caching of guest CPUID flags that KVM
needs to frequently query, and to add clarity to such caching, e.g. it
isn't immediately obvious that SVM's bundle of flags for "optional nested
SVM features" track whether or not a flag is exposed to L1.
Begrudgingly define KVM_MAX_NR_GOVERNED_FEATURES for the size of the
bitmap to avoid exposing governed_features.h in arch/x86/include/asm/, but
add a FIXME to call out that it can and should be cleaned up once
"struct kvm_vcpu_arch" is no longer expose to the kernel at large.
Cc: Zeng Guang <guang.zeng@intel.com>
Reviewed-by: Binbin Wu <binbin.wu@linux.intel.com>
Reviewed-by: Kai Huang <kai.huang@intel.com>
Reviewed-by: Yuan Yao <yuan.yao@intel.com>
Link: https://lore.kernel.org/r/20230815203653.519297-2-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Correct the spec_ctrl field in the VMCB save area based on the AMD
Programmer's manual.
Originally, the spec_ctrl was listed as u32 with 4 bytes of reserved
area. The AMD Programmer's Manual now lists the spec_ctrl as 8 bytes
in VMCB save area.
The Public Processor Programming reference for Genoa, shows SPEC_CTRL
as 64b register, but the AMD Programmer's Manual lists SPEC_CTRL as
32b register. This discrepancy will be cleaned up in next revision of
the AMD Programmer's Manual.
Since remaining bits above bit 7 are reserved bits in SPEC_CTRL MSR
and thus, not being used, the spec_ctrl added as u32 in the VMCB save
area is currently not an issue.
Fixes: 3dd2775b74 ("KVM: SVM: Create a separate mapping for the SEV-ES save area")
Suggested-by: Tom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: Manali Shukla <manali.shukla@amd.com>
Link: https://lore.kernel.org/r/20230717041903.85480-1-manali.shukla@amd.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
The status of global socket memory pressure is updated when:
a) __sk_mem_raise_allocated():
enter: sk_memory_allocated(sk) > sysctl_mem[1]
leave: sk_memory_allocated(sk) <= sysctl_mem[0]
b) __sk_mem_reduce_allocated():
leave: sk_under_memory_pressure(sk) &&
sk_memory_allocated(sk) < sysctl_mem[0]
So the conditions of leaving global pressure are inconstant, which
may lead to the situation that one pressured net-memcg prevents the
global pressure from being cleared when there is indeed no global
pressure, thus the global constrains are still in effect unexpectedly
on the other sockets.
This patch fixes this by ignoring the net-memcg's pressure when
deciding whether should leave global memory pressure.
Fixes: e1aab161e0 ("socket: initial cgroup code.")
Signed-off-by: Abel Wu <wuyun.abel@bytedance.com>
Acked-by: Shakeel Butt <shakeelb@google.com>
Link: https://lore.kernel.org/r/20230816091226.1542-1-wuyun.abel@bytedance.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
In VMX, ept_level looks better than tdp_level and is consistent with
SVM's get_npt_level().
Signed-off-by: Shiyuan Gao <gaoshiyuan@baidu.com>
Link: https://lore.kernel.org/r/20230810113853.98114-1-gaoshiyuan@baidu.com
[sean: massage changelog]
Signed-off-by: Sean Christopherson <seanjc@google.com>
Update the target pCPU for IOMMU doorbells when updating IRTE routing if
KVM is actively running the associated vCPU. KVM currently only updates
the pCPU when loading the vCPU (via avic_vcpu_load()), and so doorbell
events will be delayed until the vCPU goes through a put+load cycle (which
might very well "never" happen for the lifetime of the VM).
To avoid inserting a stale pCPU, e.g. due to racing between updating IRTE
routing and vCPU load/put, get the pCPU information from the vCPU's
Physical APIC ID table entry (a.k.a. avic_physical_id_cache in KVM) and
update the IRTE while holding ir_list_lock. Add comments with --verbose
enabled to explain exactly what is and isn't protected by ir_list_lock.
Fixes: 411b44ba80 ("svm: Implements update_pi_irte hook to setup posted interrupt")
Reported-by: dengqiao.joey <dengqiao.joey@bytedance.com>
Cc: stable@vger.kernel.org
Cc: Alejandro Jimenez <alejandro.j.jimenez@oracle.com>
Cc: Joao Martins <joao.m.martins@oracle.com>
Cc: Maxim Levitsky <mlevitsk@redhat.com>
Cc: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
Tested-by: Alejandro Jimenez <alejandro.j.jimenez@oracle.com>
Reviewed-by: Joao Martins <joao.m.martins@oracle.com>
Link: https://lore.kernel.org/r/20230808233132.2499764-3-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Hoist the acquisition of ir_list_lock from avic_update_iommu_vcpu_affinity()
to its two callers, avic_vcpu_load() and avic_vcpu_put(), specifically to
encapsulate the write to the vCPU's entry in the AVIC Physical ID table.
This will allow a future fix to pull information from the Physical ID entry
when updating the IRTE, without potentially consuming stale information,
i.e. without racing with the vCPU being (un)loaded.
Add a comment to call out that ir_list_lock does NOT protect against
multiple writers, specifically that reading the Physical ID entry in
avic_vcpu_put() outside of the lock is safe.
To preserve some semblance of independence from ir_list_lock, keep the
READ_ONCE() in avic_vcpu_load() even though acuiring the spinlock
effectively ensures the load(s) will be generated after acquiring the
lock.
Cc: stable@vger.kernel.org
Tested-by: Alejandro Jimenez <alejandro.j.jimenez@oracle.com>
Reviewed-by: Joao Martins <joao.m.martins@oracle.com>
Link: https://lore.kernel.org/r/20230808233132.2499764-2-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Existing comment in the source explains why we don't want efx_init_tc()
failure to be fatal. Cited commit erroneously consolidated failure
paths causing the probe to be failed in this case.
Fixes: 7e056e2360 ("sfc: obtain device mac address based on firmware handle for ef100")
Reviewed-by: Martin Habets <habetsm.xilinx@gmail.com>
Signed-off-by: Edward Cree <ecree.xilinx@gmail.com>
Link: https://lore.kernel.org/r/aa7f589dd6028bd1ad49f0a85f37ab33c09b2b45.1692114888.git.ecree.xilinx@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
In efx_init_tc(), move the setting of efx->tc->up after the
flow_indr_dev_register() call, so that if it fails, efx_fini_tc()
won't call flow_indr_dev_unregister().
Fixes: 5b2e12d51b ("sfc: bind indirect blocks for TC offload on EF100")
Suggested-by: Pieter Jansen van Vuuren <pieter.jansen-van-vuuren@amd.com>
Reviewed-by: Martin Habets <habetsm.xilinx@gmail.com>
Signed-off-by: Edward Cree <ecree.xilinx@gmail.com>
Link: https://lore.kernel.org/r/a81284d7013aba74005277bd81104e4cfbea3f6f.1692114888.git.ecree.xilinx@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Drop the WARN in KVM_RUN that asserts that KVM isn't using the hypervisor
timer, a.k.a. the VMX preemption timer, for a vCPU that is in the
UNINITIALIZIED activity state. The intent of the WARN is to sanity check
that KVM won't drop a timer interrupt due to an unexpected transition to
UNINITIALIZED, but unfortunately userspace can use various ioctl()s to
force the unexpected state.
Drop the sanity check instead of switching from the hypervisor timer to a
software based timer, as the only reason to switch to a software timer
when a vCPU is blocking is to ensure the timer interrupt wakes the vCPU,
but said interrupt isn't a valid wake event for vCPUs in UNINITIALIZED
state *and* the interrupt will be dropped in the end.
Reported-by: Yikebaer Aizezi <yikebaer61@gmail.com>
Closes: https://lore.kernel.org/all/CALcu4rbFrU4go8sBHk3FreP+qjgtZCGcYNpSiEXOLm==qFv7iQ@mail.gmail.com
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Link: https://lore.kernel.org/r/20230808232057.2498287-1-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Wrap kvm_{gfn,hva}_range.pte in a union so that future notifier events can
pass event specific information up and down the stack without needing to
constantly expand and churn the APIs. Lockless aging of SPTEs will pass
around a bitmap, and support for memory attributes will pass around the
new attributes for the range.
Add a "KVM_NO_ARG" placeholder to simplify handling events without an
argument (creating a dummy union variable is midly annoying).
Opportunstically drop explicit zero-initialization of the "pte" field, as
omitting the field (now a union) has the same effect.
Cc: Yu Zhao <yuzhao@google.com>
Link: https://lore.kernel.org/all/CAOUHufagkd2Jk3_HrVoFFptRXM=hX2CV8f+M-dka-hJU4bP8kw@mail.gmail.com
Reviewed-by: Oliver Upton <oliver.upton@linux.dev>
Acked-by: Yu Zhao <yuzhao@google.com>
Link: https://lore.kernel.org/r/20230729004144.1054885-1-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
The VFIO_DEVICE_GET_INFO, VFIO_DEVICE_GET_REGION_INFO, and
VFIO_IOMMU_GET_INFO ioctls fill in an info struct followed by capability
structs:
+------+---------+---------+-----+
| info | caps[0] | caps[1] | ... |
+------+---------+---------+-----+
Both the info and capability struct sizes are not always multiples of
sizeof(u64), leaving u64 fields in later capability structs misaligned.
Userspace applications currently need to handle misalignment manually in
order to support CPU architectures and programming languages with strict
alignment requirements.
Make life easier for userspace by ensuring alignment in the kernel. This
is done by padding info struct definitions and by copying out zeroes
after capability structs that are not aligned.
The new layout is as follows:
+------+---------+---+---------+-----+
| info | caps[0] | 0 | caps[1] | ... |
+------+---------+---+---------+-----+
In this example caps[0] has a size that is not multiples of sizeof(u64),
so zero padding is added to align the subsequent structure.
Adding zero padding between structs does not break the uapi. The memory
layout is specified by the info.cap_offset and caps[i].next fields
filled in by the kernel. Applications use these field values to locate
structs and are therefore unaffected by the addition of zero padding.
Note that code that copies out info structs with padding is updated to
always zero the struct and copy out as many bytes as userspace
requested. This makes the code shorter and avoids potential information
leaks by ensuring padding is initialized.
Originally-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Acked-by: Jason Gunthorpe <jgg@nvidia.com>
Link: https://lore.kernel.org/r/20230809203144.2880050-1-stefanha@redhat.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
The regex search for declarations of struct of_device_id was missing
cases that had a compiler annotation such as "__maybe_unused". Improve
the regex to allow for these. Use '\S' instead of specific characters to
shorten the regex. That also finds some more compatibles using '.'
characters.
Unfortunately, these changes add ~400 more compatibles without a
schema.
Reviewed-by: Conor Dooley <conor.dooley@microchip.com>
Link: https://lore.kernel.org/r/20230804190130.1936566-1-robh@kernel.org
Signed-off-by: Rob Herring <robh@kernel.org>
Conversion from ftgmac100.txt to yaml format version.
Signed-off-by: Ivan Mikhaylov <fr0st61te@gmail.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20230805135318.6102-1-fr0st61te@gmail.com
Signed-off-by: Rob Herring <robh@kernel.org>
Currently, the SystemCMOS address space handler is installed for the
ACPI RTC devices (PNP0B00/PNP0B01/PNP0B02) only. But there are platforms
with SystemCMOS Operetion Region defined under the ACPI Time and Alarm
Device (ACPI000E), which is used by the ACPI pre-defined control methods
like _GRT (Get the Real time) and _SRT (Set the Real time).
When accessing these control methods via the acpi_tad sysfs interface,
missing SystemCMOS address space handler causes errors like below
[ 478.255453] ACPI Error: No handler for Region [RTCM] (00000000a8d2dd39) [SystemCMOS] (20230331/evregion-130)
[ 478.255458] ACPI Error: Region SystemCMOS (ID=5) has no handler (20230331/exfldio-261)
[ 478.255461] Initialized Local Variables for Method [_GRT]:
[ 478.255461] Local1: 00000000f182542c <Obj> Integer 0000000000000000
[ 478.255464] No Arguments are initialized for method [_GRT]
[ 478.255465] ACPI Error: Aborting method \_SB.AWAC._GRT due to previous error (AE_NOT_EXIST) (20230331/psparse-529)
Export two APIs for SystemCMOS address space handler from acpi_cmos_rtc
scan handler and install the handler for the ACPI Time and Alarm Device
from the ACPI TAD driver.
Link: https://bugzilla.kernel.org/show_bug.cgi?id=217714
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
[ rjw: Subject and changelog edits, whitespace adjustment ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
When the value of ZT is set via ptrace we don't disable traps for SME.
This means that when a the task has never used SME before then the value
set via ptrace will never be seen by the target task since it will
trigger a SME access trap which will flush the register state.
Disable SME traps when setting ZT, this means we also need to allocate
storage for SVE if it is not already allocated, for the benefit of
streaming SVE.
Fixes: f90b529bcb ("arm64/sme: Implement ZT0 ptrace support")
Signed-off-by: Mark Brown <broonie@kernel.org>
Cc: <stable@vger.kernel.org> # 6.3.x
Link: https://lore.kernel.org/r/20230816-arm64-zt-ptrace-first-use-v2-1-00aa82847e28@kernel.org
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
When we use NT_ARM_SSVE to either enable streaming mode or change the
vector length for a process we do not currently do anything to ensure that
there is storage allocated for the SME specific register state. If the
task had not previously used SME or we changed the vector length then
the task will not have had TIF_SME set or backing storage for ZA/ZT
allocated, resulting in inconsistent register sizes when saving state
and spurious traps which flush the newly set register state.
We should set TIF_SME to disable traps and ensure that storage is
allocated for ZA and ZT if it is not already allocated. This requires
modifying sme_alloc() to make the flush of any existing register state
optional so we don't disturb existing state for ZA and ZT.
Fixes: e12310a0d3 ("arm64/sme: Implement ptrace support for streaming mode SVE registers")
Reported-by: David Spickett <David.Spickett@arm.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Cc: <stable@vger.kernel.org> # 5.19.x
Link: https://lore.kernel.org/r/20230810-arm64-fix-ptrace-race-v1-1-a5361fad2bd6@kernel.org
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
acpi_create_dir()/acpi_remove_dir() are never implemented since
the beginning of git history.
Commit f8d3148962 ("ACPICA: Debugger: Convert some mechanisms to OSPM specific")
declared but never implemented acpi_run_debugger().
Commit 781d737c74 ("ACPI: Drop power resources driver")
removed acpi_power_init() but not its declaration.
Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
[ rjw: Subject edits ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Kmemleak report a leak in graph_trace_open():
unreferenced object 0xffff0040b95f4a00 (size 128):
comm "cat", pid 204981, jiffies 4301155872 (age 99771.964s)
hex dump (first 32 bytes):
e0 05 e7 b4 ab 7d 00 00 0b 00 01 00 00 00 00 00 .....}..........
f4 00 01 10 00 a0 ff ff 00 00 00 00 65 00 10 00 ............e...
backtrace:
[<000000005db27c8b>] kmem_cache_alloc_trace+0x348/0x5f0
[<000000007df90faa>] graph_trace_open+0xb0/0x344
[<00000000737524cd>] __tracing_open+0x450/0xb10
[<0000000098043327>] tracing_open+0x1a0/0x2a0
[<00000000291c3876>] do_dentry_open+0x3c0/0xdc0
[<000000004015bcd6>] vfs_open+0x98/0xd0
[<000000002b5f60c9>] do_open+0x520/0x8d0
[<00000000376c7820>] path_openat+0x1c0/0x3e0
[<00000000336a54b5>] do_filp_open+0x14c/0x324
[<000000002802df13>] do_sys_openat2+0x2c4/0x530
[<0000000094eea458>] __arm64_sys_openat+0x130/0x1c4
[<00000000a71d7881>] el0_svc_common.constprop.0+0xfc/0x394
[<00000000313647bf>] do_el0_svc+0xac/0xec
[<000000002ef1c651>] el0_svc+0x20/0x30
[<000000002fd4692a>] el0_sync_handler+0xb0/0xb4
[<000000000c309c35>] el0_sync+0x160/0x180
The root cause is descripted as follows:
__tracing_open() { // 1. File 'trace' is being opened;
...
*iter->trace = *tr->current_trace; // 2. Tracer 'function_graph' is
// currently set;
...
iter->trace->open(iter); // 3. Call graph_trace_open() here,
// and memory are allocated in it;
...
}
s_start() { // 4. The opened file is being read;
...
*iter->trace = *tr->current_trace; // 5. If tracer is switched to
// 'nop' or others, then memory
// in step 3 are leaked!!!
...
}
To fix it, in s_start(), close tracer before switching then reopen the
new tracer after switching. And some tracers like 'wakeup' may not update
'iter->private' in some cases when reopen, then it should be cleared
to avoid being mistakenly closed again.
Link: https://lore.kernel.org/linux-trace-kernel/20230817125539.1646321-1-zhengyejian1@huawei.com
Fixes: d7350c3f45 ("tracing/core: make the read callbacks reentrants")
Signed-off-by: Zheng Yejian <zhengyejian1@huawei.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Fix by using acpi_hest_get_payload() to find out the correct
generic error data for v3 structure.
The revision v300 generic error data is different from the old one, so
for compatibility with old and new version, change to a new interface to
locate the right memory error section that was defined in CPER.
Signed-off-by: Xiaochun Lee <lixc17@lenovo.com>
Reviewed-by: Tony Luck <tony.luck@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
As following methods are not used outside ice_vf_lib,
they can be made static:
ice_vf_rebuild_host_vlan_cfg
ice_vf_rebuild_host_tx_rate_cfg
ice_vf_set_host_trust_cfg
ice_vf_rebuild_host_mac_cfg
ice_vf_rebuild_aggregator_node_cfg
ice_vf_rebuild_host_cfg
ice_set_vf_state_qs_dis
ice_vf_set_initialized
In order to achieve that, the order in which these
were defined was reorganized.
Signed-off-by: Jan Sokolowski <jan.sokolowski@intel.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
As following methods are not used outside of ice_lib,
they can be made static:
ice_vsi_is_vlan_pruning_ena
ice_vsi_cfg_frame_size
Signed-off-by: Jan Sokolowski <jan.sokolowski@intel.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
As following methods are not used outside of ice_ddp,
they can be made static:
ice_verify_pgk
ice_pkg_val_buf
ice_aq_download_pkg
ice_aq_update_pkg
ice_find_seg_in_pkg
Signed-off-by: Jan Sokolowski <jan.sokolowski@intel.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Following methods were found to no longer be in use:
ice_is_pca9575_present
ice_mac_fltr_exist
ice_napi_del
Remove them.
Signed-off-by: Jan Sokolowski <jan.sokolowski@intel.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
LKP reports below warning when building for RISC-V.
drivers/pnp/pnpacpi/core.c:253:17:
warning: 'strncpy' specified bound 50 equals destination
size [-Wstringop-truncation]
This appears to be a valid issue since the destination string may not be
null-terminated.
To fix this, append the NUL explicitly after the strncpy().
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202307241942.Rff2Nri5-lkp@intel.com/
Signed-off-by: Sunil V L <sunilvl@ventanamicro.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Merge series from Linus Walleij <linus.walleij@linaro.org>:
After dropping unused headers a few Realtek devices
actually using the GPIO descriptors remain.
Converting them to use optional GPIO descriptors is
pretty straight-forward.
Merge series from Richard Fitzgerald <rf@opensource.cirrus.com>:
These two patches add an ACPI HID and update the way the platform-
specific firmware identifier is extracted from the ACPI.
This is never used since commit 1e3590e2e4 ("[PATCH] pgdat allocation
for new node add (get node id by acpi)").
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Screen brightness can only be changed once on HP ZBook Fury 16 G10.
The vendor reports that the issue is related to the fact that Linux doesn't
invoke _PS0 at boot for all ACPI devices, as expected by the platform firmware:
Scope (\_SB.PC00.GFX0)
{
Scope (DD1F)
{
Method (_PS0, 0, Serialized) // _PS0: Power State 0
{
If (CondRefOf (\_SB.PC00.LPCB.EC0.SSBC))
{
\_SB.PC00.LPCB.EC0.SSBC ()
}
}
...
}
...
}
The \_SB.PC00.GFX0.DD1F is the panel device, and its _PS0 needs to be
executed at the initialization time to make the brightness control work
properly.
_PS0 is not evaluated for this device, because _PSC is missing,
which violates the ACPI specification (ACPI 6.5, section 7.3.6).
Commit b378549226 ("ACPI / PM: Do not power manage devices in unknown
initial states") tried to work around missing _PSC on platforms with
defective firmware, but got reverted due to a regression.
So the safest approach is to use acpi_device_fix_up_power_extended() to
put ACPI video and its child devices to D0 to address the issue at hand.
Link: https://bugzilla.kernel.org/show_bug.cgi?id=217683
Signed-off-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
[ rjw: Subject and changelog edits ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
While debugging a segfault on 'perf lock contention' without an
available perf.data file I noticed that it was basically calling:
perf_session__delete(ERR_PTR(-1))
Resulting in:
(gdb) run lock contention
Starting program: /root/bin/perf lock contention
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib64/libthread_db.so.1".
failed to open perf.data: No such file or directory (try 'perf record' first)
Initializing perf session failed
Program received signal SIGSEGV, Segmentation fault.
0x00000000005e7515 in auxtrace__free (session=0xffffffffffffffff) at util/auxtrace.c:2858
2858 if (!session->auxtrace)
(gdb) p session
$1 = (struct perf_session *) 0xffffffffffffffff
(gdb) bt
#0 0x00000000005e7515 in auxtrace__free (session=0xffffffffffffffff) at util/auxtrace.c:2858
#1 0x000000000057bb4d in perf_session__delete (session=0xffffffffffffffff) at util/session.c:300
#2 0x000000000047c421 in __cmd_contention (argc=0, argv=0x7fffffffe200) at builtin-lock.c:2161
#3 0x000000000047dc95 in cmd_lock (argc=0, argv=0x7fffffffe200) at builtin-lock.c:2604
#4 0x0000000000501466 in run_builtin (p=0xe597a8 <commands+552>, argc=2, argv=0x7fffffffe200) at perf.c:322
#5 0x00000000005016d5 in handle_internal_command (argc=2, argv=0x7fffffffe200) at perf.c:375
#6 0x0000000000501824 in run_argv (argcp=0x7fffffffe02c, argv=0x7fffffffe020) at perf.c:419
#7 0x0000000000501b11 in main (argc=2, argv=0x7fffffffe200) at perf.c:535
(gdb)
So just set it to NULL after using PTR_ERR(session) to decode the error
as perf_session__delete(NULL) is supported.
Fixes: eef4fee5e5 ("perf lock: Dynamically allocate lockhash_table")
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: K Prateek Nayak <kprateek.nayak@amd.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Mamatha Inamdar <mamatha4@linux.vnet.ibm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ravi Bangoria <ravi.bangoria@amd.com>
Cc: Ross Zwisler <zwisler@chromium.org>
Cc: Sean Christopherson <seanjc@google.com>
Cc: Steven Rostedt (VMware) <rostedt@goodmis.org>
Cc: Tiezhu Yang <yangtiezhu@loongson.cn>
Cc: Yang Jihong <yangjihong1@huawei.com>
Link: https://lore.kernel.org/lkml/ZN4R1AYfsD2J8lRs@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
While debugging a segfault on 'perf lock contention' without an
available perf.data file I noticed that it was basically calling:
perf_session__delete(ERR_PTR(-1))
Resulting in:
(gdb) run lock contention
Starting program: /root/bin/perf lock contention
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib64/libthread_db.so.1".
failed to open perf.data: No such file or directory (try 'perf record' first)
Initializing perf session failed
Program received signal SIGSEGV, Segmentation fault.
0x00000000005e7515 in auxtrace__free (session=0xffffffffffffffff) at util/auxtrace.c:2858
2858 if (!session->auxtrace)
(gdb) p session
$1 = (struct perf_session *) 0xffffffffffffffff
(gdb) bt
#0 0x00000000005e7515 in auxtrace__free (session=0xffffffffffffffff) at util/auxtrace.c:2858
#1 0x000000000057bb4d in perf_session__delete (session=0xffffffffffffffff) at util/session.c:300
#2 0x000000000047c421 in __cmd_contention (argc=0, argv=0x7fffffffe200) at builtin-lock.c:2161
#3 0x000000000047dc95 in cmd_lock (argc=0, argv=0x7fffffffe200) at builtin-lock.c:2604
#4 0x0000000000501466 in run_builtin (p=0xe597a8 <commands+552>, argc=2, argv=0x7fffffffe200) at perf.c:322
#5 0x00000000005016d5 in handle_internal_command (argc=2, argv=0x7fffffffe200) at perf.c:375
#6 0x0000000000501824 in run_argv (argcp=0x7fffffffe02c, argv=0x7fffffffe020) at perf.c:419
#7 0x0000000000501b11 in main (argc=2, argv=0x7fffffffe200) at perf.c:535
(gdb)
So just set it to NULL after using PTR_ERR(session) to decode the error
as perf_session__delete(NULL) is supported.
The same problem was found in 'perf top' after an audit of all
perf_session__new() failure handling.
Fixes: 6ef81c55a2 ("perf session: Return error code for perf_session__new() function on failure")
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexey Budankov <alexey.budankov@linux.intel.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Jeremie Galarneau <jeremie.galarneau@efficios.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kate Stewart <kstewart@linuxfoundation.org>
Cc: Mamatha Inamdar <mamatha4@linux.vnet.ibm.com>
Cc: Mukesh Ojha <mojha@codeaurora.org>
Cc: Nageswara R Sastry <rnsastry@linux.vnet.ibm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ravi Bangoria <ravi.bangoria@linux.ibm.com>
Cc: Shawn Landden <shawn@git.icu>
Cc: Song Liu <songliubraving@fb.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Tzvetomir Stoyanov <tstoyanov@vmware.com>
Link: https://lore.kernel.org/lkml/ZN4Q2rxxsL08A8rd@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Apart from some slight naming and grouping differences, the new metrics
are functionally the same as the existing ones. Any missing metrics were
manually appended to the end of the auto generated file.
For the events, the new data includes descriptions that may have product
specific details and new groupings that will be consistent with other
products.
After generating the metrics from the telemetry repo [1], the following
manual steps were performed:
* Change the topdown expressions to compare on CPUID and use
#slots so that the same data can be shared between N2 and V2. Apart
from these modifications, the expressions now match more closely with
the Arm telemetry data which will hopefully make future updates
easier.
* Append some metrics from the old N2/V2 data that aren't present in
the telemetry data. These will possibly be added to the
telemetry-solution repo at a later time:
l3d_cache_mpki, l3d_cache_miss_rate, branch_pki, ipc_rate, spec_ipc,
retired_rate, wasted_rate, branch_immed_spec_rate,
branch_return_spec_rate, branch_indirect_spec_rate
[1]: https://gitlab.arm.com/telemetry-solution/telemetry-solution/-/blob/main/data/pmu/cpu/neoverse/neoverse-n2.json
Signed-off-by: James Clark <james.clark@arm.com>
Reviewed-by: John Garry <john.g.garry@oracle.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andrii Nakryiko <andrii@kernel.org>
Cc: Eduard Zingerman <eddyz87@gmail.com>
Cc: Haixin Yu <yuhaixin.yhx@linux.alibaba.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jing Zhang <renyu.zj@linux.alibaba.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mike Leach <mike.leach@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Nick Forrington <nick.forrington@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rob Herring <robh@kernel.org>
Cc: Sohom Datta <sohomdatta1@gmail.com>
Cc: Will Deacon <will@kernel.org>
Cc: linux-arm-kernel@lists.infradead.org
Link: https://lore.kernel.org/r/20230816114841.1679234-7-james.clark@arm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
N2 r0p3 doesn't require the workaround [1], so gating on (#slots - 5) no
longer works because all N2s have 5 slots. Use the new expression
builtin that allows calling strcmp_cpuid_str() and comparing CPUIDs in
metric formulas.
In this case, the commented formula looks like this:
strcmp_cpuid_str(0x410fd493) # greater than or equal to N2 r0p3
| strcmp_cpuid_str(0x410fd490) ^ 1 # OR NOT any version of N2
[1]: https://gitlab.arm.com/telemetry-solution/telemetry-solution/-/blob/main/data/pmu/cpu/neoverse/neoverse-n2-r0p3.json
Signed-off-by: James Clark <james.clark@arm.com>
Reviewed-by: John Garry <john.g.garry@oracle.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Eduard Zingerman <eddyz87@gmail.com>
Cc: Haixin Yu <yuhaixin.yhx@linux.alibaba.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jing Zhang <renyu.zj@linux.alibaba.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mike Leach <mike.leach@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Nick Forrington <nick.forrington@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rob Herring <robh@kernel.org>
Cc: Sohom Datta <sohomdatta1@gmail.com>
Cc: Will Deacon <will@kernel.org>
Cc: linux-arm-kernel@lists.infradead.org
Link: https://lore.kernel.org/r/20230816114841.1679234-6-james.clark@arm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
This will allow writing formulas that are conditional on a specific
CPU type or CPU version. It calls through to the existing
strcmp_cpuid_str() function in Perf which has a default weak version,
and an arch specific version for x86 and arm64.
The function takes an 'ID' type value, which is a string. But in this
case Arm CPU IDs are hex numbers prefixed with '0x'. metric.py
assumes strings are only used by event names, and that they can't start
with a number ('0'), so an additional change has to be made to the
regex to convert hex numbers back to 'ID' types.
Signed-off-by: James Clark <james.clark@arm.com>
Reviewed-by: John Garry <john.g.garry@oracle.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Eduard Zingerman <eddyz87@gmail.com>
Cc: Haixin Yu <yuhaixin.yhx@linux.alibaba.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jing Zhang <renyu.zj@linux.alibaba.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mike Leach <mike.leach@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Nick Forrington <nick.forrington@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rob Herring <robh@kernel.org>
Cc: Sohom Datta <sohomdatta1@gmail.com>
Cc: Will Deacon <will@kernel.org>
Cc: linux-arm-kernel@lists.infradead.org
Link: https://lore.kernel.org/r/20230816114841.1679234-5-james.clark@arm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Now that variant and revision fields are taken into account the behavior
is slightly more complicated so add a test to ensure that this behaves
as expected.
Reviewed-by: John Garry <john.g.garry@oracle.com>
Signed-off-by: James Clark <james.clark@arm.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Eduard Zingerman <eddyz87@gmail.com>
Cc: Haixin Yu <yuhaixin.yhx@linux.alibaba.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jing Zhang <renyu.zj@linux.alibaba.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mike Leach <mike.leach@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Nick Forrington <nick.forrington@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rob Herring <robh@kernel.org>
Cc: Sohom Datta <sohomdatta1@gmail.com>
Cc: Will Deacon <will@kernel.org>
Cc: linux-arm-kernel@lists.infradead.org
Link: https://lore.kernel.org/r/20230816114841.1679234-3-james.clark@arm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Currently variant and revision fields are masked out of the MIDR so
it's not possible to compare different versions of the same CPU.
In a later commit a workaround will be removed just for N2 r0p3, so
enable comparisons on version.
This has the side effect of changing the MIDR stored in the header of
the perf.data file to no longer have masked version fields. It also
affects the lookups in mapfile.csv, but as that currently only has
zeroed version fields, it has no actual effect. The mapfile.csv
documentation also states to zero the version fields, so unless this
isn't done it will continue to have no effect.
There is an existing weak default strcmp_cpuid_str() function, and an
x86 version. This adds another version for arm64.
Signed-off-by: James Clark <james.clark@arm.com>
Reviewed-by: John Garry <john.g.garry@oracle.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andrii Nakryiko <andrii@kernel.org>
Cc: Eduard Zingerman <eddyz87@gmail.com>
Cc: Haixin Yu <yuhaixin.yhx@linux.alibaba.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jing Zhang <renyu.zj@linux.alibaba.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kajol Jain <kjain@linux.ibm.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mike Leach <mike.leach@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Nick Forrington <nick.forrington@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rob Herring <robh@kernel.org>
Cc: Sohom Datta <sohomdatta1@gmail.com>
Cc: Will Deacon <will@kernel.org>
Cc: linux-arm-kernel@lists.infradead.org
Link: https://lore.kernel.org/r/20230816114841.1679234-2-james.clark@arm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
If the GMU can't guarantee the required resources are up, trying to
bring up the GPU is a lost cause. Return early if setting GPU OOB
fails.
Tested-by: Neil Armstrong <neil.armstrong@linaro.org> # on SM8550-QRD
Tested-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> # sm8450
Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/551830/
Signed-off-by: Rob Clark <robdclark@chromium.org>
Move these wrappers in preparation for use in a6xx_gmu.c
Tested-by: Neil Armstrong <neil.armstrong@linaro.org> # on SM8550-QRD
Tested-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> # sm8450
Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/551824/
Signed-off-by: Rob Clark <robdclark@chromium.org>
Add a helper that does exactly what it says on the can, it'll be
required for A7xx.
Tested-by: Neil Armstrong <neil.armstrong@linaro.org> # on SM8550-QRD
Tested-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org> # sm8450
Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org>
Patchwork: https://patchwork.freedesktop.org/patch/551828/
Signed-off-by: Rob Clark <robdclark@chromium.org>
With ACPI Spec 6.5 chapter 5.2.12.20, each processor in LoongArch
systems has a Core Programmable Interrupt Controller in MADT. The
value of its type is 0x11 in the spec and defined as enum variable
ACPI_MADT_TYPE_CORE_PIC in the Linux kernel.
Physical IDs can be retrieved from MADT for LoongArch systems during
initialization and they can be retrieved from the _MAT output for
hotplug CPUs.
Add physical CPU ID enumeration for LoongArch systems.
Signed-off-by: Bibo Mao <maobibo@loongson.cn>
[ rjw: Subject and changelog edits ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
'nd_pmu->pmu.attr_groups' is dereferenced in function
'nvdimm_pmu_free_hotplug_memory' call after it has been freed. Because in
function 'nvdimm_pmu_free_hotplug_memory' memory pointed by the fields of
'nd_pmu->pmu.attr_groups' is deallocated it is necessary to call 'kfree'
after 'nvdimm_pmu_free_hotplug_memory'.
Fixes: 0fab1ba6ad ("drivers/nvdimm: Add perf interface to expose nvdimm performance stats")
Co-developed-by: Ivanov Mikhail <ivanov.mikhail1@huawei-partners.com>
Signed-off-by: Konstantin Meskhidze <konstantin.meskhidze@huawei.com>
Reviewed-by: Jeff Moyer <jmoyer@redhat.com>
Link: https://lore.kernel.org/r/20230817114103.754977-1-konstantin.meskhidze@huawei.com
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Memory pointed by 'nd_pmu->pmu.attr_groups' is allocated in function
'register_nvdimm_pmu' and is lost after 'kfree(nd_pmu)' call in function
'unregister_nvdimm_pmu'.
Fixes: 0fab1ba6ad ("drivers/nvdimm: Add perf interface to expose nvdimm performance stats")
Co-developed-by: Ivanov Mikhail <ivanov.mikhail1@huawei-partners.com>
Signed-off-by: Konstantin Meskhidze <konstantin.meskhidze@huawei.com>
Reviewed-by: Jeff Moyer <jmoyer@redhat.com>
Link: https://lore.kernel.org/r/20230817115945.771826-1-konstantin.meskhidze@huawei.com
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Inside IVSC, switching ownership requires an interface with two
different hardware modules, ACE and CSI. The software interface
to these modules is based on Intel MEI framework. Usually mei
client devices are dynamically created, so the info of consumers
depending on mei client devices is not present in the firmware
tables.
This causes problems with the probe ordering with respect to
drivers for consumers of these MEI client devices. But on these
camera sensor devices, the ACPI nodes describing the sensors all
have a _DEP dependency on the matching MEI bus ACPI device, so
adding IVSC MEI bus ACPI device to acpi_honor_dep_ids allows
solving the probe-ordering problem by deferring the enumeration of
ACPI-devices which have a _DEP dependency on an IVSC mei bus ACPI
device.
On TGL platform, the HID of IVSC MEI bus ACPI device is INTC1059,
and on ADL platform, the HID is INTC1095. So add both of them to
acpi_honor_dep_ids.
Signed-off-by: Wentong Wu <wentong.wu@intel.com>
Reviewed-by: Sakari Ailus <sakari.ailus@linux.intel.com>
[ rjw: Subject and changelog edits ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Since for MMIO driver using FIFO registers, also known as tpm_tis, the
default (and tbh recommended) behaviour is now the polling mode, the
"tristate" workaround is no longer for benefit.
If someone wants to explicitly enable IRQs for a TPM chip that should be
without question allowed. It could very well be a piece hardware in the
existing deny list because of e.g. firmware update or something similar.
While at it, document the module parameter, as this was not done in 2006
when it first appeared in the mainline.
Link: https://lore.kernel.org/linux-integrity/20201015214430.17937-1-jsnitsel@redhat.com/
Link: https://lore.kernel.org/all/1145393776.4829.19.camel@localhost.localdomain/
Reviewed-by: Jerry Snitselaar <jsnitsel@redhat.com>
Signed-off-by: Jarkko Sakkinen <jarkko@kernel.org>
The following warning is reported when frame pointers and kernel IBT are
enabled:
vmlinux.o: warning: objtool: ibt_selftest+0x11: sibling call from callable instruction with modified stack frame
The problem is that objtool interprets the indirect branch in
ibt_selftest() as a sibling call, and GCC inserts a (partial) frame
pointer prologue before it:
0000 000000000003f550 <ibt_selftest>:
0000 3f550: f3 0f 1e fa endbr64
0004 3f554: e8 00 00 00 00 call 3f559 <ibt_selftest+0x9> 3f555: R_X86_64_PLT32 __fentry__-0x4
0009 3f559: 55 push %rbp
000a 3f55a: 48 8d 05 02 00 00 00 lea 0x2(%rip),%rax # 3f563 <ibt_selftest_ip>
0011 3f561: ff e0 jmp *%rax
Note the inline asm is missing ASM_CALL_CONSTRAINT, so the 'push %rbp'
happens before the indirect branch and the 'mov %rsp, %rbp' happens
afterwards.
Simplify the generated code and make it easier to understand for both
tools and humans by moving the selftest to proper asm.
Signed-off-by: Josh Poimboeuf <jpoimboe@kernel.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/99a7e16b97bda97bf0a04aa141d6241cd8a839a2.1680912949.git.jpoimboe@kernel.org
Mike and others noticed that EEVDF does like to over-schedule quite a
bit -- which does hurt performance of a number of benchmarks /
workloads.
In particular, what seems to cause over-scheduling is that when lag is
of the same order (or larger) than the request / slice then placement
will not only cause the task to be placed left of current, but also
with a smaller deadline than current, which causes immediate
preemption.
[ notably, lag bounds are relative to HZ ]
Mike suggested we stick to picking 'current' for as long as it's
eligible to run, giving it uninterrupted runtime until it reaches
parity with the pack.
Augment Mike's suggestion by only allowing it to exhaust it's initial
request.
One random data point:
echo NO_RUN_TO_PARITY > /debug/sched/features
perf stat -a -e context-switches --repeat 10 -- perf bench sched messaging -g 20 -t -l 5000
3,723,554 context-switches ( +- 0.56% )
9.5136 +- 0.0394 seconds time elapsed ( +- 0.41% )
echo RUN_TO_PARITY > /debug/sched/features
perf stat -a -e context-switches --repeat 10 -- perf bench sched messaging -g 20 -t -l 5000
2,556,535 context-switches ( +- 0.51% )
9.2427 +- 0.0302 seconds time elapsed ( +- 0.33% )
Suggested-by: Mike Galbraith <umgwanakikbuti@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20230816134059.GC982867@hirez.programming.kicks-ass.net
- Fix new MSG_SPLICE_PAGES support in server's TCP sendmsg helper
-----BEGIN PGP SIGNATURE-----
iQIzBAABCAAdFiEEKLLlsBKG3yQ88j7+M2qzM29mf5cFAmTeKW4ACgkQM2qzM29m
f5dqfRAAjLXcxSixOnFqp49echmMlG1DI493WbiXG05hJ43TspnoThIGI+GpeTBG
LxADUeZuxbMor0In8qA/GzRGQjJpdL22E1EWfLUlFN7JtWjVyyeuYLSeECGskByP
VxAiQ/uo50dYPcRI3YN391ec2GZ9wnrNkHbfj3X0pEZoZVDea3GPRXnUQbnb4GVS
Xj433F1pmPfl7EFNWCSuMcb3OPjG1uzcbYWF6+1RjaLZ02QrgzvwmU4bMePGHRga
PsFGBaBdVJi6Dkv41NhsAflsg6DjZnzyuHnH8N6EygMz4py3kess9TRux1J9g22f
j+o/GM+U1r66YXdhMhg1kiYApemRCybrmzFgn0B9dtYC7r4MiHVVRQFgAoa95teo
g4iDQ/jxJOJMFm8ipXM/bOgMcznJfcHBX2pG2cN+qXZyPQnCn+luh81JI2s5pvqf
yAJ/uThL6cWdyLoM+ZBvozpfPn9QTwwqDWMtZlbVkIz/nM6kiJjdUrl2NDBpaDhq
WGd3CP6GWgxoCbURpfBJPLytjhb00Bep7xRm6BZavt/XSOHRWJdhEachVyWt/zFI
6zLOMHv9A7rPTBigFAIgdfZzIL9c+StdpW3oEafWovwTBKW537W+m680tdLQK44q
yl2EmeG2Z3hpqbRnDPiV5XCJ7ijArFslBBuZB6mwDCPKcbMs3FQ=
=jfUp
-----END PGP SIGNATURE-----
Merge tag 'nfsd-6.5-4' of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux
Pull nfsd fix from Chuck Lever:
- Fix new MSG_SPLICE_PAGES support in server's TCP sendmsg helper
* tag 'nfsd-6.5-4' of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux:
sunrpc: set the bv_offset of first bvec in svc_tcp_sendmsg
