Rename the Rockchip RK356x SoC dtsi files and, consequently, adjust their
contents appropriately, to prepare them for the ability to specify different
CPU and GPU OPPs for each of the supported RK356x SoC variants.
The first new RK356x SoC variant to be introduced is the RK3566T, which the
Pine64 Quartz64 Zero SBC is officially based on. [1] Some other SBCs are
also based on the RK3566T variant, including Radxa ROCK 3C and ZERO 3E/3W,
but the slight trouble is that Radxa doesn't state that officially. Though,
it's rather easy to spot the RK3566T on such boards, because their official
specifications state that the maximum frequency for the Cortex-A55 cores is
lower than the "full-fat" RK3566's 1.8 GHz. [2][3][4]
These changes follow the approach used for the Rockchip RK3588 SoC variants,
which was introduced and described further in commit def88eb4d836 ("arm64:
dts: rockchip: Prepare RK3588 SoC dtsi files for per-variant OPPs"). Please
see that commit for a more detailed explanation.
No functional changes are introduced, which was validated by decompiling and
comparing all affected board dtb files before and after these changes. In
more detail, the affected dtb files have some of their blocks shuffled around
a bit and some of their phandles have different values, as a result of the
changes to the order in which the building blocks from the parent dtsi files
are included, but they effectively remain the same as the originals.
As a side note, due to the nature of introduced changes, this commit is a bit
more readable when viewed using the --break-rewrites option for git-log(1).
Dragan Simic [Sat, 2 Nov 2024 03:08:59 +0000 (04:08 +0100)]
arm64: dts: rockchip: Update CPU OPP voltages in RK356x SoC dtsi
Update the lower/upper voltage limits and the exact voltages for the Rockchip
RK356x CPU OPPs, using the most conservative values (i.e. the highest per-OPP
voltages) found in the vendor kernel source. [1]
Using the most conservative per-OPP voltages ensures reliable CPU operation
regardless of the actual CPU binning, with the downside of possibly using
a bit more power for the CPU cores than absolutely needed.
Additionally, fill in the missing "clock-latency-ns" CPU OPP properties, using
the values found in the vendor kernel source. [1]
Dragan Simic [Wed, 6 Nov 2024 08:33:08 +0000 (09:33 +0100)]
arm64: dts: rockchip: Add OPP voltage ranges to RK3399 OP1 SoC dtsi
Add support for voltage ranges to the CPU, GPU and DMC OPPs defined in the
SoC dtsi for Rockchip OP1, as a variant of the Rockchip RK3399. This may be
useful if there are any OP1-based boards whose associated voltage regulators
are unable to deliver the exact voltages; otherwise, it causes no functional
changes to the resulting OPP voltages at runtime.
These changes cannot cause stability issues or any kind of damage, because
it's perfectly safe to use the highest voltage from an OPP group for each OPP
in the same group. The only possible negative effect of using higher voltages
is wasted energy in form of some additionally generated heat.
Chris Morgan [Thu, 31 Oct 2024 15:05:03 +0000 (10:05 -0500)]
arm64: dts: rockchip: correct analog audio name on Indiedroid Nova
Correct the audio name for the Indiedroid Nova from
rockchip,es8388-codec to rockchip,es8388. This name change corrects a
kernel log error of "ASoC: driver name too long 'rockchip,es8388-codec'
-> 'rockchip_es8388'".
Cenk Uluisik [Thu, 24 Oct 2024 09:50:22 +0000 (09:50 +0000)]
arm64: dts: rockchip: Add rk3588-orangepi-5b device tree
Implements a slightly modified rk3588s-orangepi-5b.dts from the vendor.
Unfortunately the &wireless_bluetooth and &wireless_wlan are not
implemented yet.
Therefore add the sdhci alias to be mmc0 on the rk3588s-orangepi-5b.dts.
How is the Orange Pi 5B unique?
- the Orange Pi 5b uses combphy0_ps for the WiFi.
- the Orange Pi 5B has GPIO0_C5 hooked to BT_WAKE_HOST.
- builtin eMMC storage
- ap6275p Wifi module (like the Orange Pi 5 Plus)
- builtin BlueTooth module
Cenk Uluisik [Thu, 24 Oct 2024 09:50:21 +0000 (09:50 +0000)]
dt-bindings: arm: rockchip: Add Orange Pi 5b enum to Orange Pi 5 entry
This extends the Xunlong Orange Pi 5 device tree binding
with an enum for the Orange Pi 5b, which is implemented
before the device tree.
How does this board differ from the original Orange Pi 5?
- the Orange Pi 5 has a M.2 NVMe M-key PCI 2.0x1
slot (hooked to combphy0_ps) whereas the Orange Pi 5b uses combphy0_ps
for the WiFi.
- The Orange Pi 5 with the M.2 socket has a regulator defined hooked to
"GPIO0_C5" (i.e. PCIE_PWREN_H) whereas the Orange Pi 5B has GPIO0_C5
hooked to BT_WAKE_HOST.
- builtin eMMC storage
- no SPI NOR flash (u-boot, preboot etc. initiates
from within the eMMC
storage)
- ap6275p Wifi module (like the Orange Pi 5 Plus)
- builtin BlueTooth module
Heiko Stuebner [Tue, 8 Oct 2024 20:39:35 +0000 (22:39 +0200)]
arm64: dts: rockchip: Remove 'enable-active-low' from two boards
The 'enable-active-low' property is not a valid, because it is the
default behaviour of the fixed regulator.
Only 'enable-active-high' is valid, and when this property is absent
the fixed regulator will act as active low by default.
Both the rk3588-orange-pi-5 and the Wolfvision pf5 io expander overlay
smuggled those enable-active-low properties in, so remove them to
make dtbscheck happier.
Heiko Stuebner [Thu, 24 Oct 2024 15:14:01 +0000 (17:14 +0200)]
arm64: dts: rockchip: add HDMI pinctrl to rk3588-tiger SoM
The Tiger SoM routes all relevant HDMI pins to its Q7 connector.
Some from the M0 and some from the M1 set of pins.
Add the necessary pinctrl entry to the hdmi controller for the SoM.
Not all baseboards may use all pins, but even for them it'll serve
documentation purposes.
Tianling Shen [Tue, 22 Oct 2024 19:35:30 +0000 (03:35 +0800)]
arm64: dts: rockchip: reorder mmc aliases for NanoPi R3S
Typically any non-removable storage (emmc) is listed before removable
storage (sd-card) options. Also U-Boot will try to override and use
mmc0=sdhci and mmc1=sdmmc0 for all rk356x boards.
Jimmy Hon [Thu, 24 Oct 2024 04:18:51 +0000 (23:18 -0500)]
arm64: dts: rockchip: reorder audio/hdmi nodes in Orange Pi 5
Fix the node order so analog-audio is before hdmi0-con
Audio was submitted first, and it wanted to live above the leds node.
Next, the HDMI was submitted, but it wanted to live above the leds node.
However, HDMI was approved first, so the Audio node ended up living above
the leds node.
Dragan Simic [Fri, 11 Oct 2024 07:40:51 +0000 (09:40 +0200)]
arm64: dts: rockchip: Add dtsi file for RK3399S SoC variant
Following the hierarchical representation of the SoC data that's been already
established in the commit 296602b8e5f7 ("arm64: dts: rockchip: Move RK3399
OPPs to dtsi files for SoC variants"), add new SoC dtsi file for the Rockchip
RK3399S SoC, which is yet another variant of the Rockchip RK3399 SoC.
The only perceivable differences between the RK3399S and the RK3399 are in
the supported CPU DVFS OPPs, which result from the RK3399S being binned for
lower maximum CPU frequencies than the regular RK3399 variant.
The RK3399S variant is used in the Pine64 PinePhone Pro only, [1] whose board
dts file included the necessary adjustments to the CPU DVFS OPPs. This commit
effectively moves those adjustments into the separate RK3399S SoC dtsi file,
following the above-mentioned "encapsulation" approach.
No functional changes are introduced, which was validated by decompiling and
comparing the affected dtb file before and after these changes.
Dragan Simic [Mon, 14 Oct 2024 20:21:57 +0000 (22:21 +0200)]
arm64: dts: rockchip: Convert dts files used as parents to dtsi files
Including a board dts file is not the right way to represent the hierarchical
nature of the board dts files and to create a dts file for another variant of
an ancestor board. However, a few boards and their variants (ab)used this
approach, so let's clean that up by converting the common ancestors into dtsi
files, and by adding separate board-variant dts files.
No functional changes are introduced, which was validated by decompiling and
comparing all affected board dtb files before and after these changes. In
more detail, the affected dtb files have some of their blocks shuffled around
a bit and some of their phandles have different values, as a result of the
changes to the order in which the building blocks from the parent dtsi files
are included, but they effectively remain the same as the originals.
The only perceivable introduced change is the turning of "roc-rk3328-cc" into
"ROC-RK3328-CC", which is the model name of one of the affected boards, which
was performed to match the styling of the official board name.
As a side note, due to the nature of introduced changes, this commit is best
viewed using "-B80%/80% -M20% -C5%" as the set of options for git-log(1).
arm64: dts: rockchip: fix the pcie refclock oscillator on Rock 5 ITX
The Rock 5 ITX uses two PCIe controllers to drive both a M.2 slot and its
SATA controller with 2 lanes each. The supply for the refclk oscillator is
the same that supplies the M.2 slot, but the SATA controller port is
supplied by a different rail.
This leads to the effect that if the PCIe30x4 controller for the M.2
probes first, everything works normally. But if the PCIe30x2 controller
that is connected to the SATA controller probes first, it will hang on
the first DBI read as nothing will have enabled the refclock before.
Fix this by describing the clock generator with its supplies so that
both controllers can reference it as needed.
Yao Zi [Thu, 29 Aug 2024 09:27:05 +0000 (09:27 +0000)]
arm64: dts: rockchip: Add Radxa e20c board
Add board-level device tree for Radxa e20c board[1]. This basic
implementation supports boot into a kernel with only UART console.
Other features will be added later.
Yao Zi [Thu, 29 Aug 2024 09:27:04 +0000 (09:27 +0000)]
arm64: dts: rockchip: Add base DT for rk3528 SoC
This initial device tree describes CPU, interrupts and UART on the chip
and is able to boot into basic kernel with only UART. Cache information
is omitted for now as there is no precise documentation. Support for
other features will be added later.
Add board file for the rk3576 based ArmSoM Sige5 board. While the hardware
offers plenty of peripherals and connectivity this basic implementation
just handles things required to successfully boot Linux from SD card and
connect via UART or Ethernet.
This device tree contains all devices necessary for booting from network
or SD Card.
It supports CPU, CRU, PM domains, dma, interrupts, timers, UART, I2C
and SDHCI (everything necessary to boot Linux on this system on chip)
as well as Ethernet, SPI, GPU and RTC.
The cru node references undocumented compatibles of "rockchip,cru" and also
marks it as syscon.
A general rockchip,cru is way too generic to ever be used anywhere, so
needs to go away, similarly the cru should not be written to from other
places, instead regular clock routines should be used.
Both mainline Linux as well as the vendor-kernel up to their 6.1 branch
only reference the cru via the normal assigned-clocks, clocks and resets
properties and do not get a syscon from the node.
Similarly, there is no syscon access by compatible both in mainline
nor the vendor-kernel up to their 6.1 branch, through either the
rockchip,rk3328-cru nor rockchip,cru compatibles.
So these two really are unused in all publically visible places.
Sidenote: the vendor-kernel does pretty crazy stuff in the camera interface
and tdm driver, where they map the cru separately and set clock muxes and
gates directly. This should of course never reach mainline anyway.
Sam Edwards [Mon, 30 Sep 2024 21:06:52 +0000 (14:06 -0700)]
arm64: dts: rockchip: Enable all 3 USBs on Turing RK1
The Turing RK1 contains 3 different USBs:
- USB0: USB 2.0, OTG
- USB1: USB 3.0, host
- USB2: USB 2.0, host
This patch activates the necessary DT nodes to enable all 3 buses.
Future work will be needed on USB0: it is not USB3-capable, so the USB0
controller needs to be told that there is no USB3 port. Per Jonas's
suggestion, the USBDP0 node is given a `rockchip,dp-lane-mux` property
that tells the USBDP driver that USBDP0 is not involved in USB so that
it can make the necessary configuration changes in hardware.
Technically, this is USB *controller* configuration, not *PHY*
configuration, so the underlying code may be moved in the future to the
USB controller driver instead, freeing up the (software) dependency on
USBDP0. A TODO comment is added to explain this.
Chris Morgan [Tue, 1 Oct 2024 15:40:16 +0000 (10:40 -0500)]
arm64: dts: rockchip: Add Powkiddy RGB20SX
The Powkiddy RGB20SX is a portable game console extremely similar to
the existing RGB30 console. The key differences from the RGB30 are
as follows:
- Realtek RTW8723DS WiFi and Bluetooth.
- UART pads for debug console (UART2).
- A function button (ADC channel 0).
- A much larger battery (5000 mAh).
Otherwise, the device is identical to the RGB30, including the
hard-coded value on ADC channel 1 used to identify the device at
runtime.
arm64: dts: rockchip: Add power button for puma-haikou
There is a PWRBTN# input pin exposed on a Q7 connector. The pin
is routed to a GPIO0_A1 through a diode. Q7 specification describes
the PWRBTN# pin as a Power Button signal.
Configure the pin as KEY_POWER, so it can function as power button and
trigger device shutdown.
arm64: dts: rockchip: Add AP6275P wireless support to Khadas Edge 2
Khadas Edge2 uses the PCI-e Ampak AP6275P 2T2R Wi-Fi 6 module. The
pcie@0 node can be used as Bridge1, so the wifi@0 node is used as a
device under the Bridge1.
Sam Edwards [Thu, 12 Sep 2024 02:50:34 +0000 (19:50 -0700)]
arm64: dts: rockchip: Enable GPU on Turing RK1
Enable the Mali GPU in the Turing RK1.
This patch also sets the external GPU voltage regulator in the RK806-1
to "always-on" because it is necessary for this regulator to be active
when enabling the GPU power domain or the kernel will fail with:
rockchip-pm-domain fd8d8000.power-management:power-controller: \
failed to set domain 'gpu', val=0
rockchip-pm-domain fd8d8000.power-management:power-controller: \
failed to get ack on domain 'gpu', val=0x1bffff
...followed by a panic when it attempts to access unavailable QoS
registers.
Since there is currently no `domain-supply` or similar to express this
dependency, the only way to ensure that the regulator is never off when
the GPU power domain is brought up is to ensure that the regulator is
never off.
Sam Edwards [Thu, 12 Sep 2024 02:50:32 +0000 (19:50 -0700)]
arm64: dts: rockchip: Enable automatic fan control on Turing RK1
This patch adds thermal trip points and cooling maps to the Turing RK1
in order to enable automatic control of the external PWM fan. The fan is
not active below 45C, as the heatsink alone can generally keep the chip
in this temperature region at idle load. This cooling profile errs on
the side of quietness, since the RK1 is commonly deployed in a Turing
Pi 2 clusterboard alongside three others, with additional cooling
provided at the chassis level.
Sam Edwards [Thu, 12 Sep 2024 02:50:31 +0000 (19:50 -0700)]
arm64: dts: rockchip: Fix Turing RK1 PCIe3 hang
The PCIe 3 PHY in the RK3588 requires a running external reference clock
for both external bus transfers and some internal PIPE operations.
Without this clock, the PCIe3 controller fails to initialize and ignores
DBI transactions indefinitely, which stalls the Linux boot process.
On most RK3588 boards, this is evidently not an issue. But on some "SoM"
designs (Turing RK1, Mixtile Core 3588E, ArmSoM AIM7, to name a few),
this clock is only provided when the CLKREQ# signal is asserted.
The PCIe 3 PHY generates the CLKREQ# signal when it knows it needs the
reference clock for proper operation. Unfortunately, the current DT for
Turing RK1 does not mux out these low-speed signals, resulting in broken
boots and potentially other issues.
This patch, following the previous one that split up the PCIe pinctrls,
resolves this problem for Turing RK1 by explicitly muxing all of the
signals needed for PCIe 2 and 3 support.
Sam Edwards [Thu, 12 Sep 2024 02:50:30 +0000 (19:50 -0700)]
arm64: dts: rockchip: Split up RK3588's PCIe pinctrls
These pinctrls manage the low-speed PCIe signals:
- CLKREQ#: An output on the RK3588 (both RC or EP modes), used to
request that external clock-generation circuitry provide a clock.
- PERST#: An input on the RK3588 in EP mode, used to detect a reset
signal from the RC. In RC mode, the hardware does not use this signal:
Linux itself generates it by putting the pin in GPIO mode.
- WAKE#: In EP mode, this is an output; in RC mode, this is an input.
Each of these signals serves a distinct purpose, and more importantly,
PERST# should not be muxed when the RK3588 is in the RC role. Bundling
them together in pinctrl groups prevents proper use: indeed, almost none
of the current board-specific .dts files make any use of them.
(Exception: Rock 5A recently had a patch land that misuses _pins; this
patch corrects that.)
However, on some RK3588 boards, the PCIe 3 controller will indefinitely
stall the boot if CLKREQ# is not muxed (details in the next patch).
This patch unbundles the signals to allow them to be used.
Jianfeng Liu [Wed, 18 Sep 2024 16:50:08 +0000 (00:50 +0800)]
arm64: dts: rockchip: Add ArmSoM W3 board
W3 is the carrier board for LM7 System on Module.
W3 features:
- 1x 2.5GbE Realtek RTL8125 Ethernet
- 2x HDMI Type A out
- 1x HDMI Type A in
- 1x USB 3.1 Type C
- 2x USB 2.0 Type A
- 2x USB 3.0 Type A
- 1x PCIE 2.0 M.2 E Key (1 lane)
- 1x PCIE 3.0 PCIe (4 lanes)
- 1x TF scard slot
- 1x MIPI CSI
- 1x MIPI DSI
- 1x ES8316 audio jack
- 1x FAN connector
- 1x RTC
- 40-pin expansion header
arm64: dts: rockchip: enable automatic fan control on Orange Pi 5+
This links the PWM fan on Orange Pi 5+ as an active cooling device
managed automatically by the thermal subsystem, with a target SoC
temperature of 65C and a minimum-spin interval from 55C to 65C to
ensure airflow when the system gets warm.
This is pretty much the same as '4a152231b050 ("arm64: dts: rockchip:
enable automatic fan control on Rock 5B")', except for the Orange Pi
5+ board.
arm64: dts: rockchip: add attiny_rst_gate to Ringneck
Ringneck v1.4 can contain (placement option) an on-board ATtiny
microcontroller instead of an STM32. In normal operation, this
is transparent to the software, as both microcontrollers emulate
the same ICs (amc6821 and isl1208).
For flashing the ATtiny, the SWITCH_REG1 regulator of the board's PMIC is
used to enable the ATtiny UPDI debug interface. If the STM32 is placed, or if
we are running on an older Ringneck revision, SWITCH_REG1 is not connected
and has no effect.
Add attiny-updi-gate-regulator so userspace can control it via sysfs
(needs CONFIG_REGULATOR_USERSPACE_CONSUMER):
leading to a build error when neither KVM_INTEL nor KVM_AMD support is
enabled:
arch/x86/kvm/x86.c: In function ‘kvm_arch_enable_virtualization’:
arch/x86/kvm/x86.c:12517:9: error: implicit declaration of function ‘cpu_emergency_register_virt_callback’ [-Wimplicit-function-declaration]
12517 | cpu_emergency_register_virt_callback(kvm_x86_ops.emergency_disable_virtualization_cpu);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
arch/x86/kvm/x86.c: In function ‘kvm_arch_disable_virtualization’:
arch/x86/kvm/x86.c:12522:9: error: implicit declaration of function ‘cpu_emergency_unregister_virt_callback’ [-Wimplicit-function-declaration]
12522 | cpu_emergency_unregister_virt_callback(kvm_x86_ops.emergency_disable_virtualization_cpu);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Fix the build by defining empty helper functions the same way the old
cpu_emergency_disable_virtualization() function was dealt with for the
same situation.
Maybe we could instead have made the call sites conditional, since the
callers (kvm_arch_{en,dis}able_virtualization()) have an empty weak
fallback. I'll leave that to the kvm people to argue about, this at
least gets the build going for that particular config.
Merge tag 'mailbox-v6.12' of git://git.kernel.org/pub/scm/linux/kernel/git/jassibrar/mailbox
Pull mailbox updates from Jassi Brar:
- fix kconfig dependencies (mhu-v3, omap2+)
- use devie name instead of genereic imx_mu_chan as interrupt name
(imx)
- enable sa8255p and qcs8300 ipc controllers (qcom)
- Fix timeout during suspend mode (bcm2835)
- convert to use use of_property_match_string (mailbox)
- enable mt8188 (mediatek)
- use devm_clk_get_enabled helpers (spreadtrum)
- fix device-id typo (rockchip)
* tag 'mailbox-v6.12' of git://git.kernel.org/pub/scm/linux/kernel/git/jassibrar/mailbox:
mailbox, remoteproc: omap2+: fix compile testing
dt-bindings: mailbox: qcom-ipcc: Document QCS8300 IPCC
dt-bindings: mailbox: qcom-ipcc: document the support for SA8255p
dt-bindings: mailbox: mtk,adsp-mbox: Add compatible for MT8188
mailbox: Use of_property_match_string() instead of open-coding
mailbox: bcm2835: Fix timeout during suspend mode
mailbox: sprd: Use devm_clk_get_enabled() helpers
mailbox: rockchip: fix a typo in module autoloading
mailbox: imx: use device name in interrupt name
mailbox: ARM_MHU_V3 should depend on ARM64
Merge tag 'i2c-for-6.12-rc1-additional_fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux
Pull i2c fixes from Wolfram Sang:
- fix DesignWare driver ENABLE-ABORT sequence, ensuring ABORT can
always be sent when needed
- check for PCLK in the SynQuacer controller as an optional clock,
allowing ACPI to directly provide the clock rate
- KEBA driver Kconfig dependency fix
- fix XIIC driver power suspend sequence
* tag 'i2c-for-6.12-rc1-additional_fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:
i2c: xiic: Fix pm_runtime_set_suspended() with runtime pm enabled
i2c: keba: I2C_KEBA should depend on KEBA_CP500
i2c: synquacer: Deal with optional PCLK correctly
i2c: designware: fix controller is holding SCL low while ENABLE bit is disabled
Merge tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Pull more SCSI updates from James Bottomley:
"These are mostly minor updates.
There are two drivers (lpfc and mpi3mr) which missed the initial
pull and a core change to retry a start/stop unit which affect
suspend/resume"
* tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: (32 commits)
scsi: lpfc: Update lpfc version to 14.4.0.5
scsi: lpfc: Support loopback tests with VMID enabled
scsi: lpfc: Revise TRACE_EVENT log flag severities from KERN_ERR to KERN_WARNING
scsi: lpfc: Ensure DA_ID handling completion before deleting an NPIV instance
scsi: lpfc: Fix kref imbalance on fabric ndlps from dev_loss_tmo handler
scsi: lpfc: Restrict support for 32 byte CDBs to specific HBAs
scsi: lpfc: Update phba link state conditional before sending CMF_SYNC_WQE
scsi: lpfc: Add ELS_RSP cmd to the list of WQEs to flush in lpfc_els_flush_cmd()
scsi: mpi3mr: Update driver version to 8.12.0.0.50
scsi: mpi3mr: Improve wait logic while controller transitions to READY state
scsi: mpi3mr: Update MPI Headers to revision 34
scsi: mpi3mr: Use firmware-provided timestamp update interval
scsi: mpi3mr: Enhance the Enable Controller retry logic
scsi: sd: Fix off-by-one error in sd_read_block_characteristics()
scsi: pm8001: Do not overwrite PCI queue mapping
scsi: scsi_debug: Remove a useless memset()
scsi: pmcraid: Convert comma to semicolon
scsi: sd: Retry START STOP UNIT commands
scsi: mpi3mr: A performance fix
scsi: ufs: qcom: Update MODE_MAX cfg_bw value
...
Merge tag 'bcachefs-2024-09-28' of git://evilpiepirate.org/bcachefs
Pull more bcachefs updates from Kent Overstreet:
"Assorted minor syzbot fixes, and for bigger stuff:
Fix two disk accounting rewrite bugs:
- Disk accounting keys use the version field of bkey so that journal
replay can tell which updates have been applied to the btree.
This is set in the transaction commit path, after we've gotten our
journal reservation (and our time ordering), but the
BCH_TRANS_COMMIT_skip_accounting_apply flag that journal replay
uses was incorrectly skipping this for new updates generated prior
to journal replay.
This fixes the underlying cause of an assertion pop in
disk_accounting_read.
- A couple of fixes for disk accounting + device removal.
Checking if acocunting replicas entries were marked in the
superblock was being done at the wrong point, when deltas in the
journal could still zero them out, and then additionally we'd try
to add a missing replicas entry to the superblock without checking
if it referred to an invalid (removed) device.
A whole slew of repair fixes:
- fix infinite loop in propagate_key_to_snapshot_leaves(), this fixes
an infinite loop when repairing a filesystem with many snapshots
- fix incorrect transaction restart handling leading to occasional
"fsck counted ..." warnings
- fix warning in __bch2_fsck_err() for bkey fsck errors
- check_inode() in fsck now correctly checks if the filesystem was
clean
- there shouldn't be pending logged ops if the fs was clean, we now
check for this
- remove_backpointer() doesn't remove a dirent that doesn't actually
point to the inode
- many more fsck errors are AUTOFIX"
* tag 'bcachefs-2024-09-28' of git://evilpiepirate.org/bcachefs: (35 commits)
bcachefs: check_subvol_path() now prints subvol root inode
bcachefs: remove_backpointer() now checks if dirent points to inode
bcachefs: dirent_points_to_inode() now warns on mismatch
bcachefs: Fix lost wake up
bcachefs: Check for logged ops when clean
bcachefs: BCH_FS_clean_recovery
bcachefs: Convert disk accounting BUG_ON() to WARN_ON()
bcachefs: Fix BCH_TRANS_COMMIT_skip_accounting_apply
bcachefs: Check for accounting keys with bversion=0
bcachefs: rename version -> bversion
bcachefs: Don't delete unlinked inodes before logged op resume
bcachefs: Fix BCH_SB_ERRS() so we can reorder
bcachefs: Fix fsck warnings from bkey validation
bcachefs: Move transaction commit path validation to as late as possible
bcachefs: Fix disk accounting attempting to mark invalid replicas entry
bcachefs: Fix unlocked access to c->disk_sb.sb in bch2_replicas_entry_validate()
bcachefs: Fix accounting read + device removal
bcachefs: bch_accounting_mode
bcachefs: fix transaction restart handling in check_extents(), check_dirents()
bcachefs: kill inode_walker_entry.seen_this_pos
...
Merge tag 'x86-urgent-2024-09-29' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 fixes from Ingo Molnar:
"Fix TDX MMIO #VE fault handling, and add two new Intel model numbers
for 'Pantherlake' and 'Diamond Rapids'"
* tag 'x86-urgent-2024-09-29' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/cpu: Add two Intel CPU model numbers
x86/tdx: Fix "in-kernel MMIO" check
Merge tag 'locking-urgent-2024-09-29' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull locking updates from Ingo Molnar:
"lockdep:
- Fix potential deadlock between lockdep and RCU (Zhiguo Niu)
- Use str_plural() to address Coccinelle warning (Thorsten Blum)
- Add debuggability enhancement (Luis Claudio R. Goncalves)
static keys & calls:
- Fix static_key_slow_dec() yet again (Peter Zijlstra)
- Handle module init failure correctly in static_call_del_module()
(Thomas Gleixner)
- Replace pointless WARN_ON() in static_call_module_notify() (Thomas
Gleixner)
<linux/cleanup.h>:
- Add usage and style documentation (Dan Williams)
rwsems:
- Move is_rwsem_reader_owned() and rwsem_owner() under
CONFIG_DEBUG_RWSEMS (Waiman Long)
atomic ops, x86:
- Redeclare x86_32 arch_atomic64_{add,sub}() as void (Uros Bizjak)
- Introduce the read64_nonatomic macro to x86_32 with cx8 (Uros
Bizjak)"
Signed-off-by: Ingo Molnar <[email protected]>
* tag 'locking-urgent-2024-09-29' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
locking/rwsem: Move is_rwsem_reader_owned() and rwsem_owner() under CONFIG_DEBUG_RWSEMS
jump_label: Fix static_key_slow_dec() yet again
static_call: Replace pointless WARN_ON() in static_call_module_notify()
static_call: Handle module init failure correctly in static_call_del_module()
locking/lockdep: Simplify character output in seq_line()
lockdep: fix deadlock issue between lockdep and rcu
lockdep: Use str_plural() to fix Coccinelle warning
cleanup: Add usage and style documentation
lockdep: suggest the fix for "lockdep bfs error:-1" on print_bfs_bug
locking/atomic/x86: Redeclare x86_32 arch_atomic64_{add,sub}() as void
locking/atomic/x86: Introduce the read64_nonatomic macro to x86_32 with cx8
Merge tag 'linux_kselftest-next-6.12-rc1-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest
Pull kselftest fix from Shuah Khan:
"One urgent fix to vDSO as automated testing is failing due to this
bug"
* tag 'linux_kselftest-next-6.12-rc1-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest:
selftests: vDSO: align stack for O2-optimized memcpy
Julia Lawall [Sat, 28 Sep 2024 19:26:22 +0000 (21:26 +0200)]
Reduce Coccinelle choices in string_choices.cocci
The isomorphism neg_if_exp negates the test of a ?: conditional,
making it unnecessary to have an explicit case for a negated test
with the branches inverted.
At the same time, we can disable neg_if_exp in cases where a
different API function may be more suitable for a negated test.
Finally, in the non-patch cases, E matches an expression with
parentheses around it, so there is no need to mention ()
explicitly in the pattern. The () are still needed in the patch
cases, because we want to drop them, if they are present.
Hongbo Li [Wed, 11 Sep 2024 01:09:27 +0000 (09:09 +0800)]
coccinelle: Remove unnecessary parentheses for only one possible change.
The parentheses are only needed if there is a disjunction, ie a
set of possible changes. If there is only one pattern, we can
remove these parentheses. Just like the format:
Hongbo Li [Wed, 11 Sep 2024 01:09:18 +0000 (09:09 +0800)]
coccinelle: Add rules to find str_true_false() replacements
After str_true_false() has been introduced in the tree,
we can add rules for finding places where str_true_false()
can be used. A simple test can find over 10 locations.
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Pull x86 kvm updates from Paolo Bonzini:
"x86:
- KVM currently invalidates the entirety of the page tables, not just
those for the memslot being touched, when a memslot is moved or
deleted.
This does not traditionally have particularly noticeable overhead,
but Intel's TDX will require the guest to re-accept private pages
if they are dropped from the secure EPT, which is a non starter.
Actually, the only reason why this is not already being done is a
bug which was never fully investigated and caused VM instability
with assigned GeForce GPUs, so allow userspace to opt into the new
behavior.
- Advertise AVX10.1 to userspace (effectively prep work for the
"real" AVX10 functionality that is on the horizon)
- Rework common MSR handling code to suppress errors on userspace
accesses to unsupported-but-advertised MSRs
This will allow removing (almost?) all of KVM's exemptions for
userspace access to MSRs that shouldn't exist based on the vCPU
model (the actual cleanup is non-trivial future work)
- Rework KVM's handling of x2APIC ICR, again, because AMD (x2AVIC)
splits the 64-bit value into the legacy ICR and ICR2 storage,
whereas Intel (APICv) stores the entire 64-bit value at the ICR
offset
- Fix a bug where KVM would fail to exit to userspace if one was
triggered by a fastpath exit handler
- Add fastpath handling of HLT VM-Exit to expedite re-entering the
guest when there's already a pending wake event at the time of the
exit
- Fix a WARN caused by RSM entering a nested guest from SMM with
invalid guest state, by forcing the vCPU out of guest mode prior to
signalling SHUTDOWN (the SHUTDOWN hits the VM altogether, not the
nested guest)
- Overhaul the "unprotect and retry" logic to more precisely identify
cases where retrying is actually helpful, and to harden all retry
paths against putting the guest into an infinite retry loop
- Add support for yielding, e.g. to honor NEED_RESCHED, when zapping
rmaps in the shadow MMU
- Refactor pieces of the shadow MMU related to aging SPTEs in
prepartion for adding multi generation LRU support in KVM
- Don't stuff the RSB after VM-Exit when RETPOLINE=y and AutoIBRS is
enabled, i.e. when the CPU has already flushed the RSB
- Trace the per-CPU host save area as a VMCB pointer to improve
readability and cleanup the retrieval of the SEV-ES host save area
- Remove unnecessary accounting of temporary nested VMCB related
allocations
- Set FINAL/PAGE in the page fault error code for EPT violations if
and only if the GVA is valid. If the GVA is NOT valid, there is no
guest-side page table walk and so stuffing paging related metadata
is nonsensical
- Fix a bug where KVM would incorrectly synthesize a nested VM-Exit
instead of emulating posted interrupt delivery to L2
- Add a lockdep assertion to detect unsafe accesses of vmcs12
structures
- Harden eVMCS loading against an impossible NULL pointer deref
(really truly should be impossible)
- Minor SGX fix and a cleanup
- Misc cleanups
Generic:
- Register KVM's cpuhp and syscore callbacks when enabling
virtualization in hardware, as the sole purpose of said callbacks
is to disable and re-enable virtualization as needed
- Enable virtualization when KVM is loaded, not right before the
first VM is created
Together with the previous change, this simplifies a lot the logic
of the callbacks, because their very existence implies
virtualization is enabled
- Fix a bug that results in KVM prematurely exiting to userspace for
coalesced MMIO/PIO in many cases, clean up the related code, and
add a testcase
- Fix a bug in kvm_clear_guest() where it would trigger a buffer
overflow _if_ the gpa+len crosses a page boundary, which thankfully
is guaranteed to not happen in the current code base. Add WARNs in
more helpers that read/write guest memory to detect similar bugs
Selftests:
- Fix a goof that caused some Hyper-V tests to be skipped when run on
bare metal, i.e. NOT in a VM
- Add a regression test for KVM's handling of SHUTDOWN for an SEV-ES
guest
- Explicitly include one-off assets in .gitignore. Past Sean was
completely wrong about not being able to detect missing .gitignore
entries
- Verify userspace single-stepping works when KVM happens to handle a
VM-Exit in its fastpath
- Misc cleanups"
* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm: (127 commits)
Documentation: KVM: fix warning in "make htmldocs"
s390: Enable KVM_S390_UCONTROL config in debug_defconfig
selftests: kvm: s390: Add VM run test case
KVM: SVM: let alternatives handle the cases when RSB filling is required
KVM: VMX: Set PFERR_GUEST_{FINAL,PAGE}_MASK if and only if the GVA is valid
KVM: x86/mmu: Use KVM_PAGES_PER_HPAGE() instead of an open coded equivalent
KVM: x86/mmu: Add KVM_RMAP_MANY to replace open coded '1' and '1ul' literals
KVM: x86/mmu: Fold mmu_spte_age() into kvm_rmap_age_gfn_range()
KVM: x86/mmu: Morph kvm_handle_gfn_range() into an aging specific helper
KVM: x86/mmu: Honor NEED_RESCHED when zapping rmaps and blocking is allowed
KVM: x86/mmu: Add a helper to walk and zap rmaps for a memslot
KVM: x86/mmu: Plumb a @can_yield parameter into __walk_slot_rmaps()
KVM: x86/mmu: Move walk_slot_rmaps() up near for_each_slot_rmap_range()
KVM: x86/mmu: WARN on MMIO cache hit when emulating write-protected gfn
KVM: x86/mmu: Detect if unprotect will do anything based on invalid_list
KVM: x86/mmu: Subsume kvm_mmu_unprotect_page() into the and_retry() version
KVM: x86: Rename reexecute_instruction()=>kvm_unprotect_and_retry_on_failure()
KVM: x86: Update retry protection fields when forcing retry on emulation failure
KVM: x86: Apply retry protection to "unprotect on failure" path
KVM: x86: Check EMULTYPE_WRITE_PF_TO_SP before unprotecting gfn
...
Merge tag 's390-6.12-2' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux
Pull more s390 updates from Vasily Gorbik:
- Clean up and improve vdso code: use SYM_* macros for function and
data annotations, add CFI annotations to fix GDB unwinding, optimize
the chacha20 implementation
- Add vfio-ap driver feature advertisement for use by libvirt and
mdevctl
* tag 's390-6.12-2' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux:
s390/vfio-ap: Driver feature advertisement
s390/vdso: Use one large alternative instead of an alternative branch
s390/vdso: Use SYM_DATA_START_LOCAL()/SYM_DATA_END() for data objects
tools: Add additional SYM_*() stubs to linkage.h
s390/vdso: Use macros for annotation of asm functions
s390/vdso: Add CFI annotations to __arch_chacha20_blocks_nostack()
s390/vdso: Fix comment within __arch_chacha20_blocks_nostack()
s390/vdso: Get rid of permutation constants
Merge tag 'modules-6.12-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/mcgrof/linux
Pull module updates from Luis Chamberlain:
"There are a few fixes / cleanups from Vincent, Chunhui, and Petr, but
the most important part of this pull request is the Rust community
stepping up to help maintain both C / Rust code for future Rust module
support. We grow the set of modules maintainers by three now, and with
this hope to scale to help address what's needed to properly support
future Rust module support.
A lot of exciting stuff coming in future kernel releases.
This has been on linux-next for ~ 3 weeks now with no issues"
* tag 'modules-6.12-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/mcgrof/linux:
module: Refine kmemleak scanned areas
module: abort module loading when sysfs setup suffer errors
MAINTAINERS: scale modules with more reviewers
module: Clean up the description of MODULE_SIG_<type>
module: Split modules_install compression and in-kernel decompression
Merge tag 'fbdev-for-6.12-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/linux-fbdev
Pull fbdev fixes from Helge Deller:
- crash fix in fbcon_putcs
- avoid a possible string memory overflow in sisfb
- minor code optimizations in omapfb and fbcon
* tag 'fbdev-for-6.12-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/linux-fbdev:
fbdev: sisfb: Fix strbuf array overflow
fbcon: break earlier in search_fb_in_map and search_for_mapped_con
fbdev: omapfb: Call of_node_put(ep) only once in omapdss_of_find_source_for_first_ep()
fbcon: Fix a NULL pointer dereference issue in fbcon_putcs
Merge tag 'drm-next-2024-09-28' of https://gitlab.freedesktop.org/drm/kernel
Pull drm fixes from Dave Airlie:
"Regular fixes for the week to end the merge window, i915 and xe have a
few each, amdgpu makes up most of it with a bunch of SR-IOV related
fixes amongst others.
i915:
- Fix BMG support to UHBR13.5
- Two PSR fixes
- Fix colorimetry detection for DP
xe:
- Fix macro for checking minimum GuC version
- Fix CCS offset calculation for some BMG SKUs
- Fix locking on memory usage reporting via fdinfo and BO destroy
- Fix GPU page fault handler on a closed VM
- Fix overflow in oa batch buffer
amdgpu:
- MES 12 fix
- KFD fence sync fix
- SR-IOV fixes
- VCN 4.0.6 fix
- SDMA 7.x fix
- Bump driver version to note cleared VRAM support
- SWSMU fix
- CU occupancy logic fix
- SDMA queue fix"
* tag 'drm-next-2024-09-28' of https://gitlab.freedesktop.org/drm/kernel: (79 commits)
drm/amd/pm: update workload mask after the setting
drm/amdgpu: bump driver version for cleared VRAM
drm/amdgpu: fix vbios fetching for SR-IOV
drm/amdgpu: fix PTE copy corruption for sdma 7
drm/amdkfd: Add SDMA queue quantum support for GFX12
drm/amdgpu/vcn: enable AV1 on both instances
drm/amdkfd: Fix CU occupancy for GFX 9.4.3
drm/amdkfd: Update logic for CU occupancy calculations
drm/amdgpu: skip coredump after job timeout in SRIOV
drm/amdgpu: sync to KFD fences before clearing PTEs
drm/amdgpu/mes12: set enable_level_process_quantum_check
drm/i915/dp: Fix colorimetry detection
drm/amdgpu/mes12: reduce timeout
drm/amdgpu/mes11: reduce timeout
drm/amdgpu: use GEM references instead of TTMs v2
drm/amd/display: Allow backlight to go below `AMDGPU_DM_DEFAULT_MIN_BACKLIGHT`
drm/amd/display: Fix kdoc entry for 'tps' in 'dc_process_dmub_dpia_set_tps_notification'
drm/amdgpu: update golden regs for gfx12
drm/amdgpu: clean up vbios fetching code
drm/amd/display: handle nulled pipe context in DCE110's set_drr()
...
Merge tag 'ceph-for-6.12-rc1' of https://github.com/ceph/ceph-client
Pull ceph updates from Ilya Dryomov:
"Three CephFS fixes from Xiubo and Luis and a bunch of assorted
cleanups"
* tag 'ceph-for-6.12-rc1' of https://github.com/ceph/ceph-client:
ceph: remove the incorrect Fw reference check when dirtying pages
ceph: Remove empty definition in header file
ceph: Fix typo in the comment
ceph: fix a memory leak on cap_auths in MDS client
ceph: flush all caps releases when syncing the whole filesystem
ceph: rename ceph_flush_cap_releases() to ceph_flush_session_cap_releases()
libceph: use min() to simplify code in ceph_dns_resolve_name()
ceph: Convert to use jiffies macro
ceph: Remove unused declarations
projects / linux.git / log