It has been determined that the threaded IRQ API accomplishes effectively
the same performance metrics as blk_irq_poll. As blk_irq_poll is mostly
scheduled by the softirqd and handled in softirq context, this is not
entirely desired from a Fibre Channel driver context. A threaded IRQ model
fits cleaner. This patch replaces the blk_irq_poll logic with threaded
IRQ.
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Link: https://lore.kernel.org/r/20230417191558.83100-7-justintee8345@gmail.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
A new RCQE status value indicating DMA failure when transferring
asynchronously received data to an RQE is introduced. Such errors are
unexpected and handlers are updated to log KERN_ERR and dump lpfc's debug
trace buffer to kmsg.
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Link: https://lore.kernel.org/r/20230417191558.83100-6-justintee8345@gmail.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The CMF_SYNC_WQE command is updated to use an 8-bit field sync period. All
related variables used to calculate congestion warning notifications are
updated to 8-bit fields accordingly.
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Link: https://lore.kernel.org/r/20230417191558.83100-5-justintee8345@gmail.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The SCSI version of the abort handler routine, lpfc_abort_handler(), takes
the lpfc_cmd->buf_lock and then phba->hbalock.
Make the same change for the NVMe abort path, lpfc_nvme_fcp_abort(), to
have consistent lock ordering logic between the two abort paths.
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Link: https://lore.kernel.org/r/20230417191558.83100-4-justintee8345@gmail.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Smatch detected a double free path because lpfc_nlp_not_used() releases an
ndlp object before reaching lpfc_nlp_put() at the end of
lpfc_cmpl_els_logo_acc().
Remove the outdated lpfc_nlp_not_used() routine. In
lpfc_mbx_cmpl_ns_reg_login(), replace the call with lpfc_nlp_put(). In
lpfc_cmpl_els_logo_acc(), replace the call with lpfc_unreg_rpi() and keep
the lpfc_nlp_put() at the end of the routine. If ndlp's rpi was
registered, then lpfc_unreg_rpi()'s completion routine performs the final
ndlp clean up after lpfc_nlp_put() is called from lpfc_cmpl_els_logo_acc().
Otherwise if ndlp has no rpi registered, the lpfc_nlp_put() at the end of
lpfc_cmpl_els_logo_acc() is the final ndlp clean up.
Fixes: 4430f7fd09 ("scsi: lpfc: Rework locations of ndlp reference taking")
Cc: <stable@vger.kernel.org> # v5.11+
Reported-by: Dan Carpenter <error27@gmail.com>
Link: https://lore.kernel.org/all/Y3OefhyyJNKH%2Fiaf@kili/
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Link: https://lore.kernel.org/r/20230417191558.83100-3-justintee8345@gmail.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
For SES LUNs with scsi_device sector_size member set to zero, there is no
point to log an LBA. When verbose FCP driver logging is enabled, sanity
check sector_size before calling scsi_get_lba() on a scsi_cmnd.
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Link: https://lore.kernel.org/r/20230417191558.83100-2-justintee8345@gmail.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Add a wait timeout to prevent the kernel from waiting for the GET_NVMD
response forever during probe. Add a check for the controller state before
issuing GET_NVMD request.
Signed-off-by: Changyuan Lyu <changyuanl@google.com>
Signed-off-by: Pranav Prasad <pranavpp@google.com>
Link: https://lore.kernel.org/r/20230419175502.919999-1-pranavpp@google.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Update phy_attached, phy_state, and port_state to correct values after a
hard rest. Without this patch, after a successful hard reset, phy_attached
is still 0, as a result, any following hard reset will cause a PHY START to
be issued first.
Signed-off-by: Changyuan Lyu <changyuanl@google.com>
Signed-off-by: Pranav Prasad <pranavpp@google.com>
Link: https://lore.kernel.org/r/20230418190101.696345-7-pranavpp@google.com
Acked-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Log port state during PHY_DOWN event to understand reasoning for PHY_DOWNs.
Signed-off-by: Akshat Jain <akshatzen@google.com>
Signed-off-by: Pranav Prasad <pranavpp@google.com>
Link: https://lore.kernel.org/r/20230418190101.696345-6-pranavpp@google.com
Acked-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Print phy_id and port_id sent as part of device registration request.
Signed-off-by: Akshat Jain <akshatzen@google.com>
Signed-off-by: Pranav Prasad <pranavpp@google.com>
Link: https://lore.kernel.org/r/20230418190101.696345-5-pranavpp@google.com
Acked-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Log port_id and phy_id along with the PHY_UP event.
Signed-off-by: Akshat Jain <akshatzen@google.com>
Signed-off-by: Pranav Prasad <pranavpp@google.com>
Link: https://lore.kernel.org/r/20230418190101.696345-4-pranavpp@google.com
Acked-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Even when urgent BKOPS fails, the consumer will get stuck in runtime
suspend status. Like commit 1a5665fc8d ("scsi: ufs: core: WLUN suspend
SSU/enter hibern8 fail recovery"), trigger the error handler and return
-EBUSY to break the suspend.
Fixes: b294ff3e34 ("scsi: ufs: core: Enable power management for wlun")
Signed-off-by: Keoseong Park <keosung.park@samsung.com>
Link: https://lore.kernel.org/r/20230425031721epcms2p5d4de65616478c967d466626e20c42a3a@epcms2p5
Reviewed-by: Avri Altman <avri.altman@wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
When CONFIG_OF is disabled, the 'data' variable is not used at all
because of_match_node() turns into a dummy macro:
drivers/media/platform/mediatek/mdp3/mtk-mdp3-comp.c: In function 'mdp_comp_sub_create':
drivers/media/platform/mediatek/mdp3/mtk-mdp3-comp.c:1038:36: error: unused variable 'data' [-Werror=unused-variable]
1038 | const struct mtk_mdp_driver_data *data = mdp->mdp_data;
| ^~~~
Remove the variable again by moving the pointer dereference into the
of_match_node call.
Fixes: b385b991ef ("media: platform: mtk-mdp3: chip config split about subcomponents")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
gcc warns about some functions being unused when CONFIG_PM_SLEEP
is disabled:
drivers/media/platform/nxp/imx8-isi/imx8-isi-core.c:328:12: error: 'mxc_isi_pm_resume' defined but not used [-Werror=unused-function]
328 | static int mxc_isi_pm_resume(struct device *dev)
| ^~~~~~~~~~~~~~~~~
drivers/media/platform/nxp/imx8-isi/imx8-isi-core.c:314:12: error: 'mxc_isi_pm_suspend' defined but not used [-Werror=unused-function]
314 | static int mxc_isi_pm_suspend(struct device *dev)
| ^~~~~~~~~~~~~~~~~~
Use the modern SYSTEM_SLEEP_PM_OPS()/RUNTIME_PM_OPS() helpers in place
of the old SET_SYSTEM_SLEEP_PM_OPS()/SET_RUNTIME_PM_OPS() ones.
By convention, use pm_ptr() to guard the reference to the operations.
This makes no difference as long as the driver requires CONFIG_PM,
but is what users of SET_RUNTIME_PM_OPS() are supposed to do.
Fixes: cf21f328fc ("media: nxp: Add i.MX8 ISI driver")
Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
When adding proper support for V4L2_FIELD_ALTERNATE it was missed that
this field format should trigger an interrupt for each field, not just
for the whole frame. Fix this by marking it as progressive in the
capture setup, which will then select the correct interrupt mode.
Tested on both Gen2 and Gen3 with the result of a doubling of the frame
rate for V4L2_FIELD_ALTERNATE. From a PAL video source the frame rate is
now 50, which is expected for alternate field capture.
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
When doing format validation for NV12 the width and height should be
aligned to 32 pixels.
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
The VIN modules on Gen3 can not scale NV12, fail format validation if
the user tries. Currently no frames are produced if this is attempted.
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
maxItems is already globally set to 3. To make the binding easier to read
and remove redundancy, set minItems to 3 for sa8775p as this platform
requires exactly three clocks.
Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
Acked-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20230419120914.173715-1-brgl@bgdev.pl
Signed-off-by: Vinod Koul <vkoul@kernel.org>
At least on SC8280XP the eDP PHY is part of power domain:
sc8280xp-crd.dtb: phy@220c2a00: 'power-domains' does not match any of the regexes: 'pinctrl-[0-9]+'
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Bjorn Andersson <andersson@kernel.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20230416151233.346336-1-krzysztof.kozlowski@linaro.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
According to the datasheets, the Strobe Timing Adjustment bit (STRTIM)
setting is different on R-Car SoCs, i.e.
R-Car M3 ES1.* : STRTIM[2:0] is set to 0x6
other R-Car Gen3: STRTIM[2:0] is set to 0x7
other R-Car Gen4: STRTIM[3:0] is set to 0xf
To fix this issue, a DT match data was added to specify the setting
for special use cases.
Signed-off-by: Cong Dang <cong.dang.xn@renesas.com>
Signed-off-by: Hai Pham <hai.pham.ud@renesas.com>
[wsa: rebased, restructured, added Gen4 support]
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/20230419130234.44321-1-wsa+renesas@sang-engineering.com
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
A SEV-ES guest is active on AMD when CC_ATTR_GUEST_STATE_ENCRYPT is set.
I.e., MSR_AMD64_SEV, bit 1, SEV_ES_Enabled. So no need for a special
static key.
No functional changes.
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Acked-by: Tom Lendacky <thomas.lendacky@amd.com>
Link: https://lore.kernel.org/r/20230328201712.25852-3-bp@alien8.de
Commit de614ac319 ("MAINTAINERS: Add entries for Apple PWM driver") adds
an entry for Documentation/devicetree/bindings/pwm/pwm-apple.yaml, but
commit 87a3a3929c ("dt-bindings: pwm: Add Apple PWM controller") from
the same patch series actually adds the devicetree binding file with the
name apple,s5l-fpwm.yaml.
Adjust the file entry to the file actually added.
Fixes: de614ac319 ("MAINTAINERS: Add entries for Apple PWM driver")
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@gmail.com>
Link: https://lore.kernel.org/r/20230424114043.22475-1-lukas.bulwahn@gmail.com
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Those will be used in code regions where instrumentation is not allowed
so mark them as such.
No functional changes.
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Acked-by: Tom Lendacky <thomas.lendacky@amd.com>
Link: https://lore.kernel.org/r/20230328201712.25852-2-bp@alien8.de
The SA8775P platform has 5nm USB3 UNI phy attached to the USB0 and USB1
controllers.
Add QMP PHY config, pcs entries and support for the new compatible for
SA8775P platform.
Signed-off-by: Shazad Hussain <quic_shazhuss@quicinc.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Link: https://lore.kernel.org/r/20230428130824.23803-5-quic_shazhuss@quicinc.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Add compatible string for Qualcomm QMP Super Speed (SS) UNI PHY found
in SA8775P.
Signed-off-by: Shazad Hussain <quic_shazhuss@quicinc.com>
Acked-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20230428130824.23803-4-quic_shazhuss@quicinc.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Document the compatible string for USB phy found in Qualcomm SA8775P SoC
Signed-off-by: Shazad Hussain <quic_shazhuss@quicinc.com>
Acked-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20230428130824.23803-3-quic_shazhuss@quicinc.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Convert Broadcom Kona family USB 2.0 PHY bindings to DT schema.
Signed-off-by: Stanislav Jakubek <stano.jakubek@gmail.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20230427190725.GA7730@standask-GA-A55M-S2HP
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Previously, all Amlogic boards used the XTAL clock as the default board
clock for the USB PHY input, so there was no need to enable it.
However, with the introduction of new Amlogic SoCs like the A1 family,
the USB PHY now uses a gated clock. Hence, it is necessary to enable
this gated clock during the PHY initialization sequence, or disable it
during the PHY exit, as appropriate.
Signed-off-by: Dmitry Rokosov <ddrokosov@sberdevices.ru>
Reviewed-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Link: https://lore.kernel.org/r/20230426102922.19705-2-ddrokosov@sberdevices.ru
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Commit
310e782a99 ("platform/x86/amd: pmc: Utilize SMN index 0 for driver probe")
switched to using amd_smn_read() which relies upon the misc PCI ID used
by DF function 3 being included in a table. The ID for model 78h is
missing in that table, so amd_smn_read() doesn't work.
Add the missing ID into amd_nb, restoring s2idle on this system.
[ bp: Simplify commit message. ]
Fixes: 310e782a99 ("platform/x86/amd: pmc: Utilize SMN index 0 for driver probe")
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Acked-by: Bjorn Helgaas <bhelgaas@google.com> # pci_ids.h
Acked-by: Guenter Roeck <linux@roeck-us.net>
Link: https://lore.kernel.org/r/20230427053338.16653-2-mario.limonciello@amd.com
By my own mistake, after adding !IS_DG2 into wrong branch,
bxt_set_cdclk started to execute code intended for platforms
gen < 11, which is wrong.
Move IS_DG2 check to better place.
Fixes: ceb0cc3b42 ("drm/i915: Communicate display power demands to pcode")
Signed-off-by: Stanislav Lisovskiy <stanislav.lisovskiy@intel.com>
Reviewed-by: Vinod Govindapillai <vinod.govindapillai@intel.com
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230506144217.26075-1-stanislav.lisovskiy@intel.com
The init counter is not decremented on initialisation errors, which
prevents retrying initialisation.
Add the missing decrement on initialisation errors so that the counter
reflects the state of the device.
Fixes: e78f3d15e1 ("phy: qcom-qmp: new qmp phy driver for qcom-chipsets")
Cc: stable@vger.kernel.org # 4.12
Signed-off-by: Johan Hovold <johan+linaro@kernel.org>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Link: https://lore.kernel.org/r/20230502103810.12061-3-johan+linaro@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
The init counter is not decremented on initialisation errors, which
prevents retrying initialisation and can lead to the runtime suspend
callback attempting to disable resources that have never been enabled.
Add the missing decrement on initialisation errors so that the counter
reflects the state of the device.
Fixes: e78f3d15e1 ("phy: qcom-qmp: new qmp phy driver for qcom-chipsets")
Cc: stable@vger.kernel.org # 4.12
Signed-off-by: Johan Hovold <johan+linaro@kernel.org>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Link: https://lore.kernel.org/r/20230502103810.12061-2-johan+linaro@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Commit e6fe3f422b ("sched: Make multiple runqueue task counters
32-bit") changed the type for rq->nr_uninterruptible from "unsigned
long" to "unsigned int", but left wrong cast print to
/sys/kernel/debug/sched/debug and to the console.
For example, nr_uninterruptible's value is fffffff7 with type
"unsigned int", (long)nr_uninterruptible shows 4294967287 while
(int)nr_uninterruptible prints -9. So using int cast fixes wrong
printing.
Signed-off-by: Yan Yan <yanyan.yan@antgroup.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20230506074253.44526-1-yanyan.yan@antgroup.com
When a degenerate cluster domain for core with SMT CPUs is removed,
the SD_SHARE_CPUCAPACITY flag in the local child sched group was not
propagated to the new parent. We need this flag to properly determine
whether the local sched group is SMT. Set the flag in the local
child sched group of the new parent sched domain.
Signed-off-by: Tim Chen <tim.c.chen@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com>
Link: https://lkml.kernel.org/r/73cf0959eafa53c02e7ef6bf805d751d9190e55d.1683156492.git.tim.c.chen@linux.intel.com
Current 500ms min window size for psi triggers limits polling interval
to 50ms to prevent polling threads from using too much cpu bandwidth by
polling too frequently. However the number of cgroups with triggers is
unlimited, so this protection can be defeated by creating multiple
cgroups with psi triggers (triggers in each cgroup are served by a single
"psimon" kernel thread).
Instead of limiting min polling period, which also limits the latency of
psi events, it's better to limit psi trigger creation to authorized users
only, like we do for system-wide psi triggers (/proc/pressure/* files can
be written only by processes with CAP_SYS_RESOURCE capability). This also
makes access rules for cgroup psi files consistent with system-wide ones.
Add a CAP_SYS_RESOURCE capability check for cgroup psi file writers and
remove the psi window min size limitation.
Suggested-by: Sudarshan Rajagopalan <quic_sudaraja@quicinc.com>
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Michal Hocko <mhocko@suse.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lore.kernel.org/all/cover.1676067791.git.quic_sudaraja@quicinc.com/
Intel Meteor Lake hybrid processors have cores in two separate dies. The
cores in one of the dies have higher maximum frequency. Use the SD_ASYM_
PACKING flag to give higher priority to the die with CPUs of higher maximum
frequency.
Suggested-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com>
Signed-off-by: Chen Yu <yu.c.chen@intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20230406203148.19182-13-ricardo.neri-calderon@linux.intel.com
X86 does not have the SD_ASYM_PACKING flag in the SMT domain. The scheduler
knows how to handle SMT and non-SMT cores of different priority. There is
no reason for SMT siblings of a core to have different priorities.
Signed-off-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Len Brown <len.brown@intel.com>
Tested-by: Zhang Rui <rui.zhang@intel.com>
Link: https://lore.kernel.org/r/20230406203148.19182-12-ricardo.neri-calderon@linux.intel.com
There is no difference between any of the SMT siblings of a physical core.
Do not do asym_packing load balancing at this level.
Signed-off-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Zhang Rui <rui.zhang@intel.com>
Link: https://lore.kernel.org/r/20230406203148.19182-11-ricardo.neri-calderon@linux.intel.com
Only x86 and Power7 use ASYM_PACKING. They use it differently.
Power7 has cores of equal priority, but the SMT siblings of a core have
different priorities. Parent scheduling domains do not need (nor have) the
ASYM_PACKING flag. SHARED_CHILD is not needed. Using SHARED_PARENT would
cause the topology debug code to complain.
X86 has cores of different priority, but all the SMT siblings of the core
have equal priority. It needs ASYM_PACKING at the MC level, but not at the
SMT level (it also needs it at upper levels if they have scheduling groups
of different priority). Removing ASYM_PACKING from the SMT domain causes
the topology debug code to complain.
Remove SHARED_CHILD for now. We still need a topology check that satisfies
both architectures.
Suggested-by: Valentin Schneider <vschneid@redhat.com>
Signed-off-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Zhang Rui <rui.zhang@intel.com>
Link: https://lore.kernel.org/r/20230406203148.19182-10-ricardo.neri-calderon@linux.intel.com
Do not assume that all the children of a scheduling domain have a given
flag. Check whether it has the SDF_SHARED_CHILD meta flag.
Suggested-by: Ionela Voinescu <ionela.voinescu@arm.com>
Signed-off-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lore.kernel.org/r/20230406203148.19182-9-ricardo.neri-calderon@linux.intel.com
Now that find_busiest_group() triggers load balancing between a fully_
busy SMT2 core and an idle non-SMT core, it is no longer needed to force
balancing via asym_packing. Use asym_packing only as intended: when there
is high-priority CPU that is idle.
After this change, the same logic apply to SMT and non-SMT local groups.
It makes less sense having a separate function to deal specifically with
SMT. Fold the logic in asym_smt_can_pull_tasks() into sched_asym().
Signed-off-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Zhang Rui <rui.zhang@intel.com>
Link: https://lore.kernel.org/r/20230406203148.19182-8-ricardo.neri-calderon@linux.intel.com
The prefer_sibling setting acts on the busiest group to move excess tasks
to the local group. This should be done as per request of the child of the
busiest group's sched domain, not the local group's.
Using the flags of the child domain of the local group works fortuitously
if both groups have child domains.
There are cases, however, in which the busiest group's sched domain has
child but the local group's does not. Consider, for instance a non-SMT
core (or an SMT core with only one online sibling) doing load balance with
an SMT core at the MC level. SD_PREFER_SIBLING of the busiest group's child
domain will not be honored. We are left with a fully busy SMT core and an
idle non-SMT core.
Suggested-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Signed-off-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Zhang Rui <rui.zhang@intel.com>
Link: https://lore.kernel.org/r/20230406203148.19182-7-ricardo.neri-calderon@linux.intel.com
When comparing two fully_busy scheduling groups, keep the current busiest
group if it represents an SMT core. Tasks in such scheduling group share
CPU resources and need more help than tasks in a non-SMT fully_busy group.
Signed-off-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Zhang Rui <rui.zhang@intel.com>
Link: https://lore.kernel.org/r/20230406203148.19182-6-ricardo.neri-calderon@linux.intel.com
Using asym_packing priorities within an SMT core is straightforward. Just
follow the priorities that hardware indicates.
When balancing load from an SMT core, also consider the idle state of its
siblings. Priorities do not reflect that an SMT core divides its throughput
among all its busy siblings. They only makes sense when exactly one sibling
is busy.
Indicate that active balance is needed if the destination CPU has lower
priority than the source CPU but the latter has busy SMT siblings.
Make find_busiest_queue() not skip higher-priority SMT cores with more than
busy sibling.
Suggested-by: Valentin Schneider <vschneid@redhat.com>
Signed-off-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Zhang Rui <rui.zhang@intel.com>
Link: https://lore.kernel.org/r/20230406203148.19182-5-ricardo.neri-calderon@linux.intel.com
Callers of asym_smt_can_pull_tasks() check the idle state of the
destination CPU and its SMT siblings, if any. No extra checks are needed
in such function.
Since SMT cores divide capacity among its siblings, priorities only really
make sense if only one sibling is active. This is true for SMT2, SMT4,
SMT8, etc. Do not use asym_packing load balance for this case. Instead,
let find_busiest_group() handle imbalances.
When balancing non-SMT cores or at higher scheduling domains (e.g.,
between MC scheduling groups), continue using priorities.
Signed-off-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Len Brown <len.brown@intel.com>
Tested-by: Zhang Rui <rui.zhang@intel.com>
Link: https://lore.kernel.org/r/20230406203148.19182-4-ricardo.neri-calderon@linux.intel.com
