Now we add a case where we skip dma unmap, the vq->premapped is true.
We can't just rely on use_dma_api to determine whether to skip the dma
operation. For convenience, I introduced the "do_unmap". By default, it
is the same as use_dma_api. If the driver is configured with premapped,
then do_unmap is false.
So as long as do_unmap is false, for addr of desc, we should skip dma
unmap operation.
Signed-off-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
Message-Id: <20230810123057.43407-7-xuanzhuo@linux.alibaba.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This helper allows the driver change the dma mode to premapped mode.
Under the premapped mode, the virtio core do not do dma mapping
internally.
This just work when the use_dma_api is true. If the use_dma_api is false,
the dma options is not through the DMA APIs, that is not the standard
way of the linux kernel.
Signed-off-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
Message-Id: <20230810123057.43407-4-xuanzhuo@linux.alibaba.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This patch put the dma addr error check in vring_map_one_sg().
The benefits of doing this:
1. reduce one judgment of vq->use_dma_api.
2. make vring_map_one_sg more simple, without calling
vring_mapping_error to check the return value. simplifies subsequent
code
Signed-off-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20230810123057.43407-3-xuanzhuo@linux.alibaba.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This operation allow vdpa parent to expose its own backend feature bits.
Next patches introduce a feature not compatible with all parent drivers:
the ability to enable vq after driver_ok. Each parent must declare if
it allows it or not.
Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Acked-by: Shannon Nelson <shannon.nelson@amd.com>
Message-Id: <20230609092127.170673-4-eperezma@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This feature flag allows the driver enabling virtqueues both before and
after DRIVER_OK.
This is needed for software assisted live migration, so userland can
restore the device status in devices with control virtqueue before the
dataplane is enabled.
Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Acked-by: Shannon Nelson <shannon.nelson@amd.com>
Message-Id: <20230609092127.170673-2-eperezma@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Pull dmaengine updates from Vinod Koul:
"New controller support and updates to drivers.
New support:
- Qualcomm SM6115 and QCM2290 dmaengine support
- at_xdma support for microchip,sam9x7 controller
Updates:
- idxd updates for wq simplification and ats knob updates
- fsl edma updates for v3 support
- Xilinx AXI4-Stream control support
- Yaml conversion for bcm dma binding"
* tag 'dmaengine-6.6-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/dmaengine: (53 commits)
dmaengine: fsl-edma: integrate v3 support
dt-bindings: fsl-dma: fsl-edma: add edma3 compatible string
dmaengine: fsl-edma: move tcd into struct fsl_dma_chan
dmaengine: fsl-edma: refactor chan_name setup and safety
dmaengine: fsl-edma: move clearing of register interrupt into setup_irq function
dmaengine: fsl-edma: refactor using devm_clk_get_enabled
dmaengine: fsl-edma: simply ATTR_DSIZE and ATTR_SSIZE by using ffs()
dmaengine: fsl-edma: move common IRQ handler to common.c
dmaengine: fsl-edma: Remove enum edma_version
dmaengine: fsl-edma: transition from bool fields to bitmask flags in drvdata
dmaengine: fsl-edma: clean up EXPORT_SYMBOL_GPL in fsl-edma-common.c
dmaengine: fsl-edma: fix build error when arch is s390
dmaengine: idxd: Fix issues with PRS disable sysfs knob
dmaengine: idxd: Allow ATS disable update only for configurable devices
dmaengine: xilinx_dma: Program interrupt delay timeout
dmaengine: xilinx_dma: Use tasklet_hi_schedule for timing critical usecase
dmaengine: xilinx_dma: Freeup active list based on descriptor completion bit
dmaengine: xilinx_dma: Increase AXI DMA transaction segment count
dmaengine: xilinx_dma: Pass AXI4-Stream control words to dma client
dt-bindings: dmaengine: xilinx_dma: Add xlnx,irq-delay property
...
Pull soundwire updates from Vinod Koul:
"Device numbering and intel driver changes are main features:
- Core support for soundwire device number allocation
- intel driver updates for adding hw_params for DAI ops, hybrid
number allocation and power managemnt callback updates
- DT header include changes for subsystem"
* tag 'soundwire-6.6-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/vkoul/soundwire:
soundwire: intel_ace2x: add DAI hw_params/prepare/hw_free callbacks
soundwire: intel_auxdevice: add hybrid IDA-based device_number allocation
soundwire: bus: add callbacks for device_number allocation
soundwire: extend parameters of new_peripheral_assigned() callback
soundWire: intel_auxdevice: resume 'sdw-master' on startup and system resume
soundwire: intel_auxdevice: enable pm_runtime earlier on startup
soundwire: Explicitly include correct DT includes
Currently the Kconfig fragments in kernel/configs and arch/*/configs
that aren't used internally aren't discoverable through "make help",
which consists of hard-coded lists of config fragments. Instead, list
all the fragment targets that have a "# Help: " comment prefix so the
targets can be generated dynamically.
Add logic to the Makefile to search for and display the fragment and
comment. Add comments to fragments that are intended to be direct targets.
Signed-off-by: Kees Cook <keescook@chromium.org>
Co-developed-by: Masahiro Yamada <masahiroy@kernel.org>
Acked-by: Michael Ellerman <mpe@ellerman.id.au> (powerpc)
Reviewed-by: Nicolas Schier <nicolas@fjasle.eu>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Pull MTD updates from Miquel Raynal:
"Core MTD changes:
- Use refcount to prevent corruption
- Call external _get and _put in right order
- Fix use-after-free in mtd release
- Explicitly include correct DT includes
- Clean refcounting with MTD_PARTITIONED_MASTER
- mtdblock: make warning messages ratelimited
- dt-bindings: Add SEAMA partition bindings
Device driver changes:
- Use devm helper functions
- Fix questionable cast, remove pointless ones.
- error handling fixes
- add support for new chip versions
- update DT bindings
- misc cleanups - fix typos, whitespace, indentation"
* tag 'mtd/for-6.6' of git://git.kernel.org/pub/scm/linux/kernel/git/mtd/linux: (105 commits)
dt-bindings: mtd: amlogic,meson-nand: drop unneeded quotes
mtd: spear_smi: Use helper function devm_clk_get_enabled()
mtd: rawnand: orion: Use helper function devm_clk_get_optional_enabled()
mtd: rawnand: vf610_nfc: Use helper function devm_clk_get_enabled()
mtd: rawnand: sunxi: Use helper function devm_clk_get_enabled()
mtd: rawnand: stm32_fmc2: Use helper function devm_clk_get_enabled()
mtd: rawnand: mtk: Use helper function devm_clk_get_enabled()
mtd: rawnand: mpc5121: Use helper function devm_clk_get_enabled()
mtd: rawnand: lpc32xx_slc: Use helper function devm_clk_get_enabled()
mtd: rawnand: intel: Use helper function devm_clk_get_enabled()
mtd: rawnand: fsmc: Use helper function devm_clk_get_enabled()
mtd: rawnand: arasan: Use helper function devm_clk_get_enabled()
mtd: rawnand: qcom: Add read/read_start ops in exec_op path
mtd: rawnand: qcom: Clear buf_count and buf_start in raw read
mtd: maps: fix -Wvoid-pointer-to-enum-cast warning
mtd: rawnand: fix -Wvoid-pointer-to-enum-cast warning
mtd: rawnand: fsmc: handle clk prepare error in fsmc_nand_resume()
mtd: rawnand: Propagate error and simplify ternary operators for brcmstb_nand_wait_for_completion()
mtd: rawnand: qcom: Sort includes alphabetically
mtd: rawnand: qcom: Do not override the error no of submit_descs()
...
Disable virtualization features on 82580 just as on i210/i211.
This avoids that virt functions are acidentally called on 82850.
Fixes: 55cac248ca ("igb: Add full support for 82580 devices")
Signed-off-by: Corinna Vinschen <vinschen@redhat.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Pull f2fs updates from Jaegeuk Kim:
"In this cycle, we don't have a highlighted feature enhancement, but
mostly have fixed issues mainly in two parts: 1) zoned block device,
and 2) compression support.
For zoned block device, we've tried to improve the power-off recovery
flow as much as possible. For compression, we found some corner cases
caused by wrong compression policy and logics. Other than them, there
were some reverts and stat corrections.
Bug fixes:
- use finish zone command when closing a zone
- check zone type before sending async reset zone command
- fix to assign compress_level for lz4 correctly
- fix error path of f2fs_submit_page_read()
- don't {,de}compress non-full cluster
- send small discard commands during checkpoint back
- flush inode if atomic file is aborted
- correct to account gc/cp stats
And, there are minor bug fixes, avoiding false lockdep warning, and
clean-ups"
* tag 'f2fs-for-6-6-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs: (25 commits)
f2fs: use finish zone command when closing a zone
f2fs: compress: fix to assign compress_level for lz4 correctly
f2fs: fix error path of f2fs_submit_page_read()
f2fs: clean up error handling in sanity_check_{compress_,}inode()
f2fs: avoid false alarm of circular locking
Revert "f2fs: do not issue small discard commands during checkpoint"
f2fs: doc: fix description of max_small_discards
f2fs: should update REQ_TIME for direct write
f2fs: fix to account cp stats correctly
f2fs: fix to account gc stats correctly
f2fs: remove unneeded check condition in __f2fs_setxattr()
f2fs: fix to update i_ctime in __f2fs_setxattr()
Revert "f2fs: fix to do sanity check on extent cache correctly"
f2fs: increase usage of folio_next_index() helper
f2fs: Only lfs mode is allowed with zoned block device feature
f2fs: check zone type before sending async reset zone command
f2fs: compress: don't {,de}compress non-full cluster
f2fs: allow f2fs_ioc_{,de}compress_file to be interrupted
f2fs: don't reopen the main block device in f2fs_scan_devices
f2fs: fix to avoid mmap vs set_compress_option case
...
Commit bde5f6bc68 ("kmemleak: add scheduling point to kmemleak_scan()")
added a cond_resched() call to the struct page scanning loop to prevent
soft lockup from happening. However, soft lockup can still happen in that
loop in some corner cases when the pages that satisfy the "!(pfn & 63)"
check are skipped for some reasons.
Fix this corner case by moving up the cond_resched() check so that it will
be called every 64 pages unconditionally.
Link: https://lkml.kernel.org/r/20230825164947.1317981-1-longman@redhat.com
Fixes: bde5f6bc68 ("kmemleak: add scheduling point to kmemleak_scan()")
Signed-off-by: Waiman Long <longman@redhat.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Yisheng Xie <xieyisheng1@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
In the past, movable allocations could be disallowed from CMA through
PF_MEMALLOC_PIN. As CMA pages are funneled through the MOVABLE pcplist,
this required filtering that cornercase during allocations, such that
pinnable allocations wouldn't accidentally get a CMA page.
However, since 8e3560d963 ("mm: honor PF_MEMALLOC_PIN for all movable
pages"), PF_MEMALLOC_PIN automatically excludes __GFP_MOVABLE. Once
again, MOVABLE implies CMA is allowed.
Remove the stale filtering code. Also remove a stale comment that was
introduced as part of the filtering code, because the filtering let
order-0 pages fall through to the buddy allocator. See 1d91df85f3
("mm/page_alloc: handle a missing case for memalloc_nocma_{save/restore}
APIs") for context. The comment's been obsolete since the introduction of
the explicit ALLOC_HIGHATOMIC flag in eb2e2b425c ("mm/page_alloc:
explicitly record high-order atomic allocations in alloc_flags").
Link: https://lkml.kernel.org/r/20230824153821.243148-1-hannes@cmpxchg.org
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Cc: David Hildenbrand <david@redhat.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Miaohe Lin <linmiaohe@huawei.com>
Cc: Pasha Tatashin <pasha.tatashin@soleen.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
With madvise and prctl KSM can be enabled for different VMA's. Once it is
enabled we can query how effective KSM is overall. However we cannot
easily query if an individual VMA benefits from KSM.
This commit adds a KSM section to the /prod/<pid>/smaps file. It reports
how many of the pages are KSM pages. Note that KSM-placed zeropages are
not included, only actual KSM pages.
Here is a typical output:
7f420a000000-7f421a000000 rw-p 00000000 00:00 0
Size: 262144 kB
KernelPageSize: 4 kB
MMUPageSize: 4 kB
Rss: 51212 kB
Pss: 8276 kB
Shared_Clean: 172 kB
Shared_Dirty: 42996 kB
Private_Clean: 196 kB
Private_Dirty: 7848 kB
Referenced: 15388 kB
Anonymous: 51212 kB
KSM: 41376 kB
LazyFree: 0 kB
AnonHugePages: 0 kB
ShmemPmdMapped: 0 kB
FilePmdMapped: 0 kB
Shared_Hugetlb: 0 kB
Private_Hugetlb: 0 kB
Swap: 202016 kB
SwapPss: 3882 kB
Locked: 0 kB
THPeligible: 0
ProtectionKey: 0
ksm_state: 0
ksm_skip_base: 0
ksm_skip_count: 0
VmFlags: rd wr mr mw me nr mg anon
This information also helps with the following workflow:
- First enable KSM for all the VMA's of a process with prctl.
- Then analyze with the above smaps report which VMA's benefit the most
- Change the application (if possible) to add the corresponding madvise
calls for the VMA's that benefit the most
[shr@devkernel.io: v5]
Link: https://lkml.kernel.org/r/20230823170107.1457915-1-shr@devkernel.io
Link: https://lkml.kernel.org/r/20230822180539.1424843-1-shr@devkernel.io
Signed-off-by: Stefan Roesch <shr@devkernel.io>
Reviewed-by: David Hildenbrand <david@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Rik van Riel <riel@surriel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Pull probes updates from Masami Hiramatsu:
- kprobes: use struct_size() for variable size kretprobe_instance data
structure.
- eprobe: Simplify trace_eprobe list iteration.
- probe events: Data structure field access support on BTF argument.
- Update BTF argument support on the functions in the kernel
loadable modules (only loaded modules are supported).
- Move generic BTF access function (search function prototype and
get function parameters) to a separated file.
- Add a function to search a member of data structure in BTF.
- Support accessing BTF data structure member from probe args by
C-like arrow('->') and dot('.') operators. e.g.
't sched_switch next=next->pid vruntime=next->se.vruntime'
- Support accessing BTF data structure member from $retval. e.g.
'f getname_flags%return +0($retval->name):string'
- Add string type checking if BTF type info is available. This will
reject if user specify ":string" type for non "char pointer"
type.
- Automatically assume the fprobe event as a function return event
if $retval is used.
- selftests/ftrace: Add BTF data field access test cases.
- Documentation: Update fprobe event example with BTF data field.
* tag 'probes-v6.6' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
Documentation: tracing: Update fprobe event example with BTF field
selftests/ftrace: Add BTF fields access testcases
tracing/fprobe-event: Assume fprobe is a return event by $retval
tracing/probes: Add string type check with BTF
tracing/probes: Support BTF field access from $retval
tracing/probes: Support BTF based data structure field access
tracing/probes: Add a function to search a member of a struct/union
tracing/probes: Move finding func-proto API and getting func-param API to trace_btf
tracing/probes: Support BTF argument on module functions
tracing/eprobe: Iterate trace_eprobe directly
kernel: kprobes: Use struct_size()
Pull more tracing updates from Steven Rostedt:
"Tracing fixes and clean ups:
- Replace strlcpy() with strscpy()
- Initialize the pipe cpumask to zero on allocation
- Use within_module() instead of open coding it
- Remove extra space in hwlat_detectory/mode output
- Use LIST_HEAD() instead of open coding it
- A bunch of clean ups and fixes for the cpumask filter
- Set local da_mon_##name to static
- Fix race in snapshot buffer between cpu write and swap"
* tag 'trace-v6.6-2' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
tracing/filters: Fix coding style issues
tracing/filters: Change parse_pred() cpulist ternary into an if block
tracing/filters: Fix double-free of struct filter_pred.mask
tracing/filters: Fix error-handling of cpulist parsing buffer
tracing: Zero the pipe cpumask on alloc to avoid spurious -EBUSY
ftrace: Use LIST_HEAD to initialize clear_hash
ftrace: Use within_module to check rec->ip within specified module.
tracing: Replace strlcpy with strscpy in trace/events/task.h
tracing: Fix race issue between cpu buffer write and swap
tracing: Remove extra space at the end of hwlat_detector/mode
rv: Set variable 'da_mon_##name' to static
Pull pstore fix from Kees Cook:
- Adjust sizes of buffers just avoid uncompress failures (Ard
Biesheuvel)
* tag 'pstore-v6.6-rc1-fix' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux:
pstore: Base compression input buffer size on estimated compressed size
Pull x86 selftest fix from Ingo Molnar:
"Fix the __NR_map_shadow_stack syscall-renumbering fallout in the x86
self-test code.
[ Arguably the existing code was unnecessarily fragile, and tooling
should have picked up the new syscall number, and a wider fix is
being worked on - but meanwhile, let's not have the old syscall
number in the kernel tree. ]"
* tag 'x86-urgent-2023-09-02' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
selftests/x86: Update map_shadow_stack syscall nr
Pull timer fix from Ingo Molnar:
"Fix false positive 'softirq work is pending' messages on -rt kernels,
caused by a buggy factoring-out of existing code"
* tag 'timers-urgent-2023-09-02' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
tick/rcu: Fix false positive "softirq work is pending" messages
Pull CPU hotplug fix from Ingo Molnar:
"Fix a CPU hotplug related deadlock between the task which initiates
and controls a CPU hot-unplug operation vs. the CFS bandwidth timer"
* tag 'smp-urgent-2023-09-02' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
cpu/hotplug: Prevent self deadlock on CPU hot-unplug
Pull scheduler fixes from Ingo Molnar:
"Miscellaneous scheduler fixes: a reporting fix, a static symbol fix,
and a kernel-doc fix"
* tag 'sched-urgent-2023-09-02' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
sched/core: Report correct state for TASK_IDLE | TASK_FREEZABLE
sched/fair: Make update_entity_lag() static
sched/core: Add kernel-doc for set_cpus_allowed_ptr()
Mikhail reports early-6.6-based Fedora Rawhide not booting: "rcu_preempt
detected expedited stalls", minutes wait, and then hung_task splat while
kworker trying to synchronize_rcu_expedited(). Nothing logged to disk.
He bisected to my 6.6 a349d72fd9 ("mm/pgtable: add rcu_read_lock() and
rcu_read_unlock()s"): but the one to blame is my 6.5 commit to fix the
espfix "bad pmd" warnings when booting x86_64 with CONFIG_EFI_PGT_DUMP=y.
Gaah, that added an "addr >= TASK_SIZE" check to avoid pte_offset_map(),
but failed to add the equivalent check when choosing to pte_unmap().
It's not a problem on 6.5 (for different reasons, it's harmless on both
64-bit and 32-bit), but becomes a bootstopper on 6.6 with the unbalanced
rcu_read_unlock() - RCU has a WARN_ON_ONCE for that, but it would have
scrolled off Mikhail's console too quickly.
Reported-by: Mikhail Gavrilov <mikhail.v.gavrilov@gmail.com>
Closes: https://lore.kernel.org/linux-mm/CABXGCsNi8Tiv5zUPNXr6UJw6qV1VdaBEfGqEAMkkXE3QPvZuAQ@mail.gmail.com/
Fixes: 8b1cb4a2e8 ("mm/pagewalk: fix EFI_PGT_DUMP of espfix area")
Fixes: a349d72fd9 ("mm/pgtable: add rcu_read_lock() and rcu_read_unlock()s")
Signed-off-by: Hugh Dickins <hughd@google.com>
Tested-by: Mikhail Gavrilov <mikhail.v.gavrilov@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Sudip Mukherjee reports that the mips sb1250_swarm_defconfig build fails
with the current kernel. It isn't actually MIPS-specific, it's just
that that defconfig does not have CGROUP_SCHED enabled like most configs
do, and as such shows this error:
kernel/cgroup/cgroup.c: In function 'cgroup_local_stat_show':
kernel/cgroup/cgroup.c:3699:15: error: implicit declaration of function 'cgroup_tryget_css'; did you mean 'cgroup_tryget'? [-Werror=implicit-function-declaration]
3699 | css = cgroup_tryget_css(cgrp, ss);
| ^~~~~~~~~~~~~~~~~
| cgroup_tryget
kernel/cgroup/cgroup.c:3699:13: warning: assignment to 'struct cgroup_subsys_state *' from 'int' makes pointer from integer without a cast [-Wint-conversion]
3699 | css = cgroup_tryget_css(cgrp, ss);
| ^
because cgroup_tryget_css() only exists when CGROUP_SCHED is enabled,
and the cgroup_local_stat_show() function should similarly be guarded by
that config option.
Move things around a bit to fix this all.
Fixes: d1d4ff5d11 ("cgroup: put cgroup_tryget_css() inside CONFIG_CGROUP_SCHED")
Reported-by: Sudip Mukherjee <sudipm.mukherjee@gmail.com>
Cc: Miaohe Lin <linmiaohe@huawei.com>
Cc: Tejun Heo <tj@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
should_we_balance() is called in load_balance() to find out if the CPU that
is trying to do the load balance is the right one or not.
With commit:
b1bfeab9b002("sched/fair: Consider the idle state of the whole core for load balance")
the code tries to find an idle core to do the load balancing
and falls back on an idle sibling CPU if there is no idle core.
However, on larger SMT systems, it could be needlessly iterating to find a
idle by scanning all the CPUs in an non-idle core. If the core is not idle,
and first SMT sibling which is idle has been found, then its not needed to
check other SMT siblings for idleness
Lets say in SMT4, Core0 has 0,2,4,6 and CPU0 is BUSY and rest are IDLE.
balancing domain is MC/DIE. CPU2 will be set as the first idle_smt and
same process would be repeated for CPU4 and CPU6 but this is unnecessary.
Since calling is_core_idle loops through all CPU's in the SMT mask, effect
is multiplied by weight of smt_mask. For example,when say 1 CPU is busy,
we would skip loop for 2 CPU's and skip iterating over 8CPU's. That
effect would be more in DIE/NUMA domain where there are more cores.
Testing and performance evaluation
==================================
The test has been done on this system which has 12 cores, i.e 24 small
cores with SMT=4:
lscpu
Architecture: ppc64le
Byte Order: Little Endian
CPU(s): 96
On-line CPU(s) list: 0-95
Model name: POWER10 (architected), altivec supported
Thread(s) per core: 8
Used funclatency bcc tool to evaluate the time taken by should_we_balance(). For
base tip/sched/core the time taken is collected by making the
should_we_balance() noinline. time is in nanoseconds. The values are
collected by running the funclatency tracer for 60 seconds. values are
average of 3 such runs. This represents the expected reduced time with
patch.
tip/sched/core was at commit:
2f88c8e802 ("sched/eevdf/doc: Modify the documented knob to base_slice_ns as well")
Results:
------------------------------------------------------------------------------
workload tip/sched/core with_patch(%gain)
------------------------------------------------------------------------------
idle system 809.3 695.0(16.45)
stress ng – 12 threads -l 100 1013.5 893.1(13.49)
stress ng – 24 threads -l 100 1073.5 980.0(9.54)
stress ng – 48 threads -l 100 683.0 641.0(6.55)
stress ng – 96 threads -l 100 2421.0 2300(5.26)
stress ng – 96 threads -l 15 375.5 377.5(-0.53)
stress ng – 96 threads -l 25 635.5 637.5(-0.31)
stress ng – 96 threads -l 35 934.0 891.0(4.83)
Ran schbench(old), hackbench and stress_ng to evaluate the workload
performance between tip/sched/core and with patch.
No modification to tip/sched/core
TL;DR:
Good improvement is seen with schbench. when hackbench and stress_ng
runs for longer good improvement is seen.
------------------------------------------------------------------------------
schbench(old) tip +patch(%gain)
10 iterations sched/core
------------------------------------------------------------------------------
1 Threads
50.0th: 8.00 9.00(-12.50)
75.0th: 9.60 9.00(6.25)
90.0th: 11.80 10.20(13.56)
95.0th: 12.60 10.40(17.46)
99.0th: 13.60 11.90(12.50)
99.5th: 14.10 12.60(10.64)
99.9th: 15.90 14.60(8.18)
2 Threads
50.0th: 9.90 9.20(7.07)
75.0th: 12.60 10.10(19.84)
90.0th: 15.50 12.00(22.58)
95.0th: 17.70 14.00(20.90)
99.0th: 21.20 16.90(20.28)
99.5th: 22.60 17.50(22.57)
99.9th: 30.40 19.40(36.18)
4 Threads
50.0th: 12.50 10.60(15.20)
75.0th: 15.30 12.00(21.57)
90.0th: 18.60 14.10(24.19)
95.0th: 21.30 16.20(23.94)
99.0th: 26.00 20.70(20.38)
99.5th: 27.60 22.50(18.48)
99.9th: 33.90 31.40(7.37)
8 Threads
50.0th: 16.30 14.30(12.27)
75.0th: 20.20 17.40(13.86)
90.0th: 24.50 21.90(10.61)
95.0th: 27.30 24.70(9.52)
99.0th: 35.00 31.20(10.86)
99.5th: 46.40 33.30(28.23)
99.9th: 89.30 57.50(35.61)
16 Threads
50.0th: 22.70 20.70(8.81)
75.0th: 30.10 27.40(8.97)
90.0th: 36.00 32.80(8.89)
95.0th: 39.60 36.40(8.08)
99.0th: 49.20 44.10(10.37)
99.5th: 64.90 50.50(22.19)
99.9th: 143.50 100.60(29.90)
32 Threads
50.0th: 34.60 35.50(-2.60)
75.0th: 48.20 50.50(-4.77)
90.0th: 59.20 62.40(-5.41)
95.0th: 65.20 69.00(-5.83)
99.0th: 80.40 83.80(-4.23)
99.5th: 102.10 98.90(3.13)
99.9th: 727.10 506.80(30.30)
schbench does improve in general. There is some run to run variation with
schbench. Did a validation run to confirm that trend is similar.
------------------------------------------------------------------------------
hackbench tip +patch(%gain)
20 iterations, 50000 loops sched/core
------------------------------------------------------------------------------
Process 10 groups : 11.74 11.70(0.34)
Process 20 groups : 22.73 22.69(0.18)
Process 30 groups : 33.39 33.40(-0.03)
Process 40 groups : 43.73 43.61(0.27)
Process 50 groups : 53.82 54.35(-0.98)
Process 60 groups : 64.16 65.29(-1.76)
thread 10 Time : 12.81 12.79(0.16)
thread 20 Time : 24.63 24.47(0.65)
Process(Pipe) 10 Time : 6.40 6.34(0.94)
Process(Pipe) 20 Time : 10.62 10.63(-0.09)
Process(Pipe) 30 Time : 15.09 14.84(1.66)
Process(Pipe) 40 Time : 19.42 19.01(2.11)
Process(Pipe) 50 Time : 24.04 23.34(2.91)
Process(Pipe) 60 Time : 28.94 27.51(4.94)
thread(Pipe) 10 Time : 6.96 6.87(1.29)
thread(Pipe) 20 Time : 11.74 11.73(0.09)
hackbench shows slight improvement with pipe. Slight degradation in process.
------------------------------------------------------------------------------
stress_ng tip +patch(%gain)
10 iterations 100000 cpu_ops sched/core
------------------------------------------------------------------------------
--cpu=96 -util=100 Time taken : 5.30, 5.01(5.47)
--cpu=48 -util=100 Time taken : 7.94, 6.73(15.24)
--cpu=24 -util=100 Time taken : 11.67, 8.75(25.02)
--cpu=12 -util=100 Time taken : 15.71, 15.02(4.39)
--cpu=96 -util=10 Time taken : 22.71, 22.19(2.29)
--cpu=96 -util=20 Time taken : 12.14, 12.37(-1.89)
--cpu=96 -util=30 Time taken : 8.76, 8.86(-1.14)
--cpu=96 -util=40 Time taken : 7.13, 7.14(-0.14)
--cpu=96 -util=50 Time taken : 6.10, 6.13(-0.49)
--cpu=96 -util=60 Time taken : 5.42, 5.41(0.18)
--cpu=96 -util=70 Time taken : 4.94, 4.94(0.00)
--cpu=96 -util=80 Time taken : 4.56, 4.53(0.66)
--cpu=96 -util=90 Time taken : 4.27, 4.26(0.23)
Good improvement seen with 24 CPUs. In this case only one CPU is busy,
and no core is idle. Decent improvement with 100% utilization case. no
difference in other utilization.
Fixes: b1bfeab9b0 ("sched/fair: Consider the idle state of the whole core for load balance")
Signed-off-by: Shrikanth Hegde <sshegde@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20230902081204.232218-1-sshegde@linux.vnet.ibm.com
Micron 1100 drives lock up when encountering queued TRIM command. It is
a quite old hardware series, for past years we have been running our
machines with these drives using libata.force=noncqtrim.
[Damien] Move the "Crucial_CT*M500*" entry to keep Micron and Crucial
entries together.
Signed-off-by: Pawel Zmarzly <pzmarzly@meta.com>
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Elkhart Lake is the successor of Apollo Lake and Gemini Lake. These
CPUs and their PCHs are used in mobile and embedded environments.
With this patch I suggest that Elkhart Lake SATA controllers [1] should
use the default LPM policy for mobile chipsets.
The disadvantage of missing hot-plug support with this setting should
not be an issue, as those CPUs are used in embedded environments and
not in servers with hot-plug backplanes.
We discovered that the Elkhart Lake SATA controllers have been missing
in ahci.c after a customer reported the throttling of his SATA SSD
after a short period of higher I/O. We determined the high temperature
of the SSD controller in idle mode as the root cause for that.
Depending on the used SSD, we have seen up to 1.8 Watt lower system
idle power usage and up to 30°C lower SSD controller temperatures in
our tests, when we set med_power_with_dipm manually. I have provided a
table showing seven different SATA SSDs from ATP, Intel/Solidigm and
Samsung [2].
Intel lists a total of 3 SATA controller IDs (4B60, 4B62, 4B63) in [1]
for those mobile PCHs.
This commit just adds 0x4b63 as I do not have test systems with 0x4b60
and 0x4b62 SATA controllers.
I have tested this patch with a system which uses 0x4b63 as SATA
controller.
[1] https://sata-io.org/product/8803
[2] https://www.thomas-krenn.com/en/wiki/SATA_Link_Power_Management#Example_LES_v4
Signed-off-by: Werner Fischer <devlists@wefi.net>
Cc: stable@vger.kernel.org
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
The pipe cpumask used to serialize opens between the main and percpu
trace pipes is not zeroed or initialized. This can result in
spurious -EBUSY returns if underlying memory is not fully zeroed.
This has been observed by immediate failure to read the main
trace_pipe file on an otherwise newly booted and idle system:
# cat /sys/kernel/debug/tracing/trace_pipe
cat: /sys/kernel/debug/tracing/trace_pipe: Device or resource busy
Zero the allocation of pipe_cpumask to avoid the problem.
Link: https://lore.kernel.org/linux-trace-kernel/20230831125500.986862-1-bfoster@redhat.com
Cc: stable@vger.kernel.org
Fixes: c2489bb7e6 ("tracing: Introduce pipe_cpumask to avoid race on trace_pipes")
Reviewed-by: Zheng Yejian <zhengyejian1@huawei.com>
Reviewed-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Signed-off-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
