Change se->load.weight to se_weight(se) in the calculation for the
initial util_avg to avoid unnecessarily inflating the util_avg by 1024
times.
The reason is that se->load.weight has the unit/scale as the scaled-up
load, while cfs_rg->avg.load_avg has the unit/scale as the true task
weight (as mapped directly from the task's nice/priority value). With
CONFIG_32BIT, the scaled-up load is equal to the true task weight. With
CONFIG_64BIT, the scaled-up load is 1024 times the true task weight.
Thus, the current code may inflate the util_avg by 1024 times. The
follow-up capping will not allow the util_avg value to go wild. But the
calculation should have the correct logic.
Signed-off-by: Dawei Li <daweilics@gmail.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Reviewed-by: Vishal Chourasia <vishalc@linux.ibm.com>
Link: https://lore.kernel.org/r/20240315015916.21545-1-daweilics@gmail.com
Change relax_domain_level checks so that it would be possible
to include or exclude all domains from newidle balancing.
This matches the behavior described in the documentation:
-1 no request. use system default or follow request of others.
0 no search.
1 search siblings (hyperthreads in a core).
"2" enables levels 0 and 1, level_max excludes the last (level_max)
level, and level_max+1 includes all levels.
Fixes: 1d3504fcf5 ("sched, cpuset: customize sched domains, core")
Signed-off-by: Vitalii Bursov <vitaly@bursov.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Reviewed-by: Valentin Schneider <vschneid@redhat.com>
Link: https://lore.kernel.org/r/bd6de28e80073c79466ec6401cdeae78f0d4423d.1714488502.git.vitaly@bursov.com
Commit in Fixes moved the optimize_nops() call inside apply_relocation()
and made it a second optimization pass after the relocations have been
done.
Since optimize_nops() works only on NOPs, that is fine and it'll simply
jump over instructions which are not NOPs.
However, it made that call with repl_len as the buffer length to
optimize.
However, it can happen that there are alternatives calls like this one:
alternative("mfence; lfence", "", ALT_NOT(X86_FEATURE_APIC_MSRS_FENCE));
where the replacement length is 0. And using repl_len is wrong because
apply_alternatives() expands the buffer size to the length of the source
insn that is being patched, by padding it with one-byte NOPs:
for (; insn_buff_sz < a->instrlen; insn_buff_sz++)
insn_buff[insn_buff_sz] = 0x90;
Long story short: pass the length of the original instruction(s) as the
length of the temporary buffer which to optimize.
Result:
SMP alternatives: feat: 11*32+27, old: (lapic_next_deadline+0x9/0x50 (ffffffff81061829) len: 6), repl: (ffffffff89b1cc60, len: 0) flags: 0x1
SMP alternatives: ffffffff81061829: old_insn: 0f ae f0 0f ae e8
SMP alternatives: ffffffff81061829: final_insn: 90 90 90 90 90 90
=>
SMP alternatives: feat: 11*32+27, old: (lapic_next_deadline+0x9/0x50 (ffffffff81061839) len: 6), repl: (ffffffff89b1cc60, len: 0) flags: 0x1
SMP alternatives: ffffffff81061839: [0:6) optimized NOPs: 66 0f 1f 44 00 00
SMP alternatives: ffffffff81061839: old_insn: 0f ae f0 0f ae e8
SMP alternatives: ffffffff81061839: final_insn: 66 0f 1f 44 00 00
Fixes: da8f9cf7e7 ("x86/alternatives: Get rid of __optimize_nops()")
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20240515104804.32004-1-bp@kernel.org
trafgen performance considerably sank on hosts with many cores
after the blamed commit.
packet_read_pending() is very expensive, and calling it
in af_packet fast path defeats Daniel intent in commit
b013840810 ("packet: use percpu mmap tx frame pending refcount")
tpacket_destruct_skb() makes room for one packet, we can immediately
wakeup a producer, no need to completely drain the tx ring.
Fixes: 89ed5b5190 ("af_packet: Block execution of tasks waiting for transmit to complete in AF_PACKET")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Neil Horman <nhorman@tuxdriver.com>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Link: https://lore.kernel.org/r/20240515163358.4105915-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Disabling tcp-data-split on idpf silently fails:
# ethtool -G $NETDEV tcp-data-split off
# ethtool -g $NETDEV | grep 'TCP data split'
TCP data split: on
But it works if you also change 'tx' or 'rx':
# ethtool -G $NETDEV tcp-data-split off tx 256
# ethtool -g $NETDEV | grep 'TCP data split'
TCP data split: off
The bug is in idpf_set_ringparam, where it takes a shortcut out if the
TX and RX sizes are not changing. Fix it by checking also if the
tcp-data-split setting remains unchanged. Only then can the soft reset
be skipped.
Fixes: 9b1aa3ef23 ("idpf: add get/set for Ethtool's header split ringparam")
Reported-by: Xu Du <xudu@redhat.com>
Closes: https://issues.redhat.com/browse/RHEL-36182
Signed-off-by: Michal Schmidt <mschmidt@redhat.com>
Reviewed-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Link: https://lore.kernel.org/r/20240515092414.158079-1-mschmidt@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
"rmmod bonding" causes an oops ever since commit cc317ea3d9 ("bonding:
remove redundant NULL check in debugfs function"). Here are the relevant
functions being called:
bonding_exit()
bond_destroy_debugfs()
debugfs_remove_recursive(bonding_debug_root);
bonding_debug_root = NULL; <--------- SET TO NULL HERE
bond_netlink_fini()
rtnl_link_unregister()
__rtnl_link_unregister()
unregister_netdevice_many_notify()
bond_uninit()
bond_debug_unregister()
(commit removed check for bonding_debug_root == NULL)
debugfs_remove()
simple_recursive_removal()
down_write() -> OOPS
However, reverting the bad commit does not solve the problem completely
because the original code contains a race that could cause the same
oops, although it was much less likely to be triggered unintentionally:
CPU1
rmmod bonding
bonding_exit()
bond_destroy_debugfs()
debugfs_remove_recursive(bonding_debug_root);
CPU2
echo -bond0 > /sys/class/net/bonding_masters
bond_uninit()
bond_debug_unregister()
if (!bonding_debug_root)
CPU1
bonding_debug_root = NULL;
So do NOT revert the bad commit (since the removed checks were racy
anyway), and instead change the order of actions taken during module
removal. The same oops can also happen if there is an error during
module init, so apply the same fix there.
Fixes: cc317ea3d9 ("bonding: remove redundant NULL check in debugfs function")
Cc: stable@vger.kernel.org
Signed-off-by: Tony Battersby <tonyb@cybernetics.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Acked-by: Jay Vosburgh <jay.vosburgh@canonical.com>
Link: https://lore.kernel.org/r/641f914f-3216-4eeb-87dd-91b78aa97773@cybernetics.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Problem
=========
After commit 67f6951347 ("ipv6: Move setting default metric for routes"),
we noticed that the logic of assigning the default value of fc_metirc
changed in the ioctl process. That is, when users use ioctl(fd, SIOCADDRT,
rt) with a non-zero metric to add a route, then they may fail to delete a
route with passing in a metric value of 0 to the kernel by ioctl(fd,
SIOCDELRT, rt). But iproute can succeed in deleting it.
As a reference, when using iproute tools by netlink to delete routes with
a metric parameter equals 0, like the command as follows:
ip -6 route del fe80::/64 via fe81::5054:ff:fe11:3451 dev eth0 metric 0
the user can still succeed in deleting the route entry with the smallest
metric.
Root Reason
===========
After commit 67f6951347 ("ipv6: Move setting default metric for routes"),
When ioctl() pass in SIOCDELRT with a zero metric, rtmsg_to_fib6_config()
will set a defalut value (1024) to cfg->fc_metric in kernel, and in
ip6_route_del() and the line 4074 at net/ipv3/route.c, it will check by
if (cfg->fc_metric && cfg->fc_metric != rt->fib6_metric)
continue;
and the condition is true and skip the later procedure (deleting route)
because cfg->fc_metric != rt->fib6_metric. But before that commit,
cfg->fc_metric is still zero there, so the condition is false and it
will do the following procedure (deleting).
Solution
========
In order to keep a consistent behaviour across netlink() and ioctl(), we
should allow to delete a route with a metric value of 0. So we only do
the default setting of fc_metric in route adding.
CC: stable@vger.kernel.org # 5.4+
Fixes: 67f6951347 ("ipv6: Move setting default metric for routes")
Co-developed-by: Fan Yu <fan.yu9@zte.com.cn>
Signed-off-by: Fan Yu <fan.yu9@zte.com.cn>
Signed-off-by: xu xin <xu.xin16@zte.com.cn>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/20240514201102055dD2Ba45qKbLlUMxu_DTHP@zte.com.cn
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Daniel Borkmann says:
====================
pull-request: bpf 2024-05-17
We've added 7 non-merge commits during the last 2 day(s) which contain
a total of 8 files changed, 20 insertions(+), 9 deletions(-).
The main changes are:
1) Fix KASAN slab-out-of-bounds in percpu_array_map_gen_lookup and add
BPF selftests to cover this case, from Andrii Nakryiko.
(Report https://lore.kernel.org/bpf/20240514231155.1004295-1-kuba@kernel.org/)
2) Fix two BPF selftests to adjust for kernel changes after fast-forwarding
Linus' tree to make BPF CI all green again, from Martin KaFai Lau.
3) Fix libbpf feature detectors when using token_fd by adjusting the
attribute size for memset to cover the former, also from Andrii Nakryiko.
4) Fix the description of 'src' in ALU instructions for the BPF ISA
standardization doc, from Puranjay Mohan.
* tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf:
selftests/bpf: Adjust btf_dump test to reflect recent change in file_operations
selftests/bpf: Adjust test_access_variable_array after a kernel function name change
selftests/bpf: add more variations of map-in-map situations
bpf: save extended inner map info for percpu array maps as well
MAINTAINERS: Update ARM64 BPF JIT maintainer
bpf, docs: Fix the description of 'src' in ALU instructions
libbpf: fix feature detectors when using token_fd
====================
Link: https://lore.kernel.org/r/20240517001600.23703-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The btf_dump test fails:
test_btf_dump_struct_data:FAIL:file_operations unexpected file_operations: actual '(struct file_operations){
.owner = (struct module *)0xffffffffffffffff,
.fop_flags = (fop_flags_t)4294967295,
.llseek = (loff_t (*)(struct f' != expected '(struct file_operations){
.owner = (struct module *)0xffffffffffffffff,
.llseek = (loff_t (*)(struct file *, loff_t, int))0xffffffffffffffff,'
The "fop_flags" is a recent addition to the struct file_operations in
commit 210a03c9d5 ("fs: claw back a few FMODE_* bits")
This patch changes the test_btf_dump_struct_data() to reflect
this change.
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Alan Maguire <alan.maguire@oracle.com>
Link: https://lore.kernel.org/bpf/20240516164310.2481460-1-martin.lau@linux.dev
- Simplify pci_epf_test_alloc_space() by using pci_epc_get_next_free_bar()
as other similar iterators do (Niklas Cassel)
- Configure endpoint BARs as 64-bit if that's all the hardware supports, in
addition to doing it if the BAR size is larger than 2GB (Niklas Cassel)
- Remove superfluous pci_epf_configure_bar(), since pci_epf_alloc_space()
now contains that functionality (Niklas Cassel)
- Simplify pci_epf_test_set_bar() (Niklas Cassel)
- Clean up pci_epf_test_unbind() to reduce indentation level (Niklas
Cassel)
* pci/endpoint:
PCI: endpoint: pci-epf-test: Clean up pci_epf_test_unbind()
PCI: endpoint: pci-epf-test: Simplify pci_epf_test_set_bar() loop
PCI: endpoint: pci-epf-test: Remove superfluous code
PCI: endpoint: Allocate a 64-bit BAR if that is the only option
PCI: endpoint: pci-epf-test: Simplify pci_epf_test_alloc_space() loop
- Return success from endpoint probe before incorrectly dropping the
reference to the BPMP (Vidya Sagar)
* pci/controller/tegra194:
PCI: tegra194: Fix probe path for Endpoint mode
- Configure endpoint BAR to be 64-bit if the PCI_BASE_ADDRESS_MEM_TYPE_64
flag is set instead of depending on the new BAR value itself (Niklas
Cassel)
- Set Subsystem Vendor ID correctly (Rick Wertenbroek)
* pci/controller/rockchip:
PCI: rockchip-ep: Remove wrong mask on subsys_vendor_id
PCI: rockchip-ep: Set a 64-bit BAR if requested
- Move DBI accesses from dw_pcie_ep_init() to dw_pcie_ep_init_complete() so
drivers for endpoints that require Refclk for DBI access, e.g., qcom and
tegra194, can control when this happens (Manivannan Sadhasivam)
- Add endpoint API kernel-doc (Manivannan Sadhasivam)
- Remove .deinit() callback and instead call rcar_gen4_pcie_ep_deinit()
explicitly from rcar-gen4, which was the only user (Manivannan
Sadhasivam)
- Rename dw_pcie_ep_exit() to dw_pcie_ep_deinit() to correspond with
dw_pcie_ep_init() (Manivannan Sadhasivam)
- Add dw_pcie_ep_cleanup() for drivers that need to clean up eDMA resources
when PERST# is asserted, e.g., qcom, tegra194 (Manivannan Sadhasivam)
- Rename dw_pcie_ep_init_complete() to dw_pcie_ep_init_registers() to
better reflect the functionality (Manivannan Sadhasivam)
- Call dw_pcie_ep_init_registers() directly from drivers instead of from
dw_pcie_ep_init() so drivers, e.g., qcom and tegra194, can do it when
Refclk is available (Manivannan Sadhasivam)
- Remove the "core_init_notifier" flag, which previously identified drivers
that required Refclk before DBI access, because it's now unnecessary
(Manivannan Sadhasivam)
* pci/controller/dwc:
PCI: endpoint: Remove "core_init_notifier" flag
PCI: dwc: ep: Call dw_pcie_ep_init_registers() API directly from all glue drivers
PCI: dwc: ep: Rename dw_pcie_ep_init_complete() to dw_pcie_ep_init_registers()
PCI: dwc: ep: Introduce dw_pcie_ep_cleanup() API for drivers supporting PERST#
PCI: dwc: ep: Rename dw_pcie_ep_exit() to dw_pcie_ep_deinit()
PCI: dwc: ep: Remove deinit() callback from struct dw_pcie_ep_ops
PCI: dwc: ep: Add Kernel-doc comments for APIs
PCI: dwc: ep: Fix DBI access failure for drivers requiring refclk from host
- Configure endpoint BAR to be 64-bit if the PCI_BASE_ADDRESS_MEM_TYPE_64
flag is set instead of depending on the new BAR value itself (Niklas
Cassel)
* pci/controller/cadence:
PCI: cadence: Set a 64-bit BAR if requested
- Avoid D3cold for HP Pavilion 17 PC/1972 PCIe Ports because we can't get
them back out of D3cold (Mario Limonciello)
* pci/pm:
PCI/PM: Avoid D3cold for HP Pavilion 17 PC/1972 PCIe Ports
- Update coding style to "mainline is normal path, errors are the
exceptions" (Andy Shevchenko)
* pci/msi:
PCI/MSI: Make error path handling follow the standard pattern
- Clear bridge Secondary Status errors after enumeration since enumeration
causes many errors (Vidya Sagar)
- Wait for Link Training==0 before starting Link retrain to avoid a race;
this was done previously but broken by a faulty merge (Ilpo Järvinen)
- Rename PCI_IRQ_LEGACY to PCI_IRQ_INTX to be more specific about what
"LEGACY" means (Damien Le Moal)
- Update return types of pci_find_capability() stubs to match the extern
declarations for the actual implementations (Bjorn Helgaas)
- Drop unnecessary pci_enable_device_io() from pata_cs5520 (Heiner
Kallweit)
- Drop unused pci_enable_device_io() (Heiner Kallweit)
- On 2016 and newer BIOSes, skip early E820 check for ECAM regions
described in ACPI MCFG; there's no spec requirement for E820
reservations, and some machines don't provide them (Bjorn Helgaas)
- If devices were disconnected while suspended, don't wait for them when
resuming (Ilpo Järvinen)
* pci/enumeration:
PCI: Do not wait for disconnected devices when resuming
x86/pci: Skip early E820 check for ECAM region
PCI: Remove unused pci_enable_device_io()
ata: pata_cs5520: Remove unnecessary call to pci_enable_device_io()
PCI: Update pci_find_capability() stub return types
PCI: Remove PCI_IRQ_LEGACY
scsi: vmw_pvscsi: Do not use PCI_IRQ_LEGACY instead of PCI_IRQ_LEGACY
scsi: pmcraid: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
scsi: mpt3sas: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
scsi: megaraid_sas: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
scsi: ipr: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
scsi: hpsa: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
scsi: arcmsr: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
wifi: rtw89: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
wifi: rtw88: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
wifi: ath10k: Refer to INTX instead of LEGACY
net: wangxun: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
r8169: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
net: alx: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
net: atlantic: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
net: amd-xgbe: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
VMCI: Use PCI_IRQ_ALL_TYPES to remove PCI_IRQ_LEGACY use
RDMA/vmw_pvrdma: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
IB/qib: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
drm/amdgpu: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
mfd: intel-lpss: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
ntb: idt: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
platform/x86: intel_ips: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
tty: 8250_pci: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
usb: hcd-pci: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
ASoC: Intel: avs: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
Documentation: PCI: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
PCI/portdrv: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
PCI/MSI: Use PCI_IRQ_INTX instead of PCI_IRQ_LEGACY
PCI: Clarify intent of LT wait
PCI: Wait for Link Training==0 before starting Link retrain
PCI: Clear Secondary Status errors after enumeration
- Specify Revision 6 of the "Enable DPC" _DSM function to match the
implementation (Kuppuswamy Sathyanarayanan)
- Check for failure of the "Locate Port" _DSM function (Kuppuswamy
Sathyanarayanan)
* pci/edr:
PCI/EDR: Align EDR_PORT_LOCATE_DSM with PCI Firmware r3.3
PCI/EDR: Align EDR_PORT_DPC_ENABLE_DSM with PCI Firmware r3.3
- Lock the upstream bridge while using it to perform a Secondary Bus Reset
(Dave Jiang)
- Return failure when attempting Secondary Bus Reset below a CXL Port that
has SBR masked (Dave Jiang)
- Add a "cxl_bus" reset method that temporarily unmasks SBR (Dave Jiang)
- Add a warning if we reset a CXL type 3 memory device that was in use
while being reset (Dave Jiang)
* pci/cxl:
cxl: Add post-reset warning if reset results in loss of previously committed HDM decoders
PCI/CXL: Add 'cxl_bus' reset method for devices below CXL Ports
PCI/CXL: Fail bus reset if upstream CXL Port has SBR masked
PCI: Lock upstream bridge for pci_reset_function()
PCI/CXL: Move CXL Vendor ID to pci_ids.h
- Mask reporting of Mask Replay Timer Timeout Correctable Errors in the
bridge above Genesys GL975x SD host controllers; the errors are caused by
a GL975x hardware defect and they may lead to AER interrupts that prevent
system suspend (Kai-Heng Feng)
- Update URL of aer-inject tool (Kuppuswamy Sathyanarayanan)
* pci/aer:
PCI/AER: Update aer-inject tool source URL
PCI: Mask Replay Timer Timeout errors for Genesys GL975x SD host controller
TASK_SIZE_MAX should be set to a constant value, at least the largest
valid userspace address under any runtime configuration. This optimizes
the check in __access_ok(), which no longer needs to compute the runtime
value of TASK_SIZE. The check does not need to be exact, as long as it
accepts all valid userspace addresses and rejects all valid kernel
addresses; well-behaved programs will never fail the access_ok() check.
For RISC-V, which requires all virtual addresses to be sign extended,
the optimal choice is LONG_MAX because it simplifies the limit
comparison to a sign bit test.
This removes about half of the references to pgtable_l[45]_enabled.
Signed-off-by: Samuel Holland <samuel.holland@sifive.com>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Alexandre Ghiti <alexghiti@rivosinc.com>
Link: https://lore.kernel.org/r/20240327143858.711792-3-samuel.holland@sifive.com
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
Dawei Li
Vitalii Bursov
Borislav Petkov (AMD)
Cong Yang
Nathan Chancellor
Jacek Lawrynowicz
Wachowski, Karol
Eric Dumazet
Daniel Jurgens
Michal Schmidt
Sagar Cheluvegowda
Tony Battersby
xu xin
Hangbin Liu
Stephen Boyd
Jakub Kicinski
Martin KaFai Lau
Bjorn Helgaas
Samuel Holland