Ville Syrjälä [Wed, 28 Oct 2020 21:33:18 +0000 (23:33 +0200)]
drm/i915: Don't enable hpd detection logic from irq_postinstall()
No reason that I can see why we should enable the hpd detection logic
already during irq postinstall phase. We don't even do this on all
the platforms. We just need it before we actually enable the hotplug
interrupts in .hpd_irq_setup(), and in fact we already do it there as
well. Let's just eliminate the redundant early setup.
Ville Syrjälä [Wed, 28 Oct 2020 21:33:17 +0000 (23:33 +0200)]
drm/i915: Split gen11_hpd_detection_setup() into tc vs. tbt variants
No reason to stuff both type-c and tbt into the same function.
Let's split this so we may more easily handle platforms that
lack the tbt spefific bits.
Ville Syrjälä [Wed, 28 Oct 2020 21:33:11 +0000 (23:33 +0200)]
drm/i915: Parametrize BXT_DE_PORT_HP_DDI with hpd_pin
Use hpd_pin to parametrize BXT_DE_PORT_HP_DDI() to make it clear
these have nothing to do with DDI ports or PHYs as such. The only
thing that matters is the HPD pin assignment.
Ville Syrjälä [Wed, 28 Oct 2020 21:33:10 +0000 (23:33 +0200)]
drm/i915: Use AUX_CH_USBCn for the RKL VBT AUX CH setup
As with the VBT DVO port, RKL uses PHY based mapping for the
VBT AUX CH. Adjust the code to use the new AUX_USBCn names
and add a comment to explain the situation.
Ville Syrjälä [Wed, 28 Oct 2020 21:33:09 +0000 (23:33 +0200)]
drm/i915: Pimp AUX CH names
Let's make the AUX CH names match the spec (AUX A-F for pre-tgl,
AUX A-C or AUX USBC1-6 for tgl+). And while at it let's include
the full encoder name in the AUX CH name as well (as opposed to
just using port_name() which wouldn't give us the right thing on
tgl+).
Ville Syrjälä [Wed, 28 Oct 2020 21:33:07 +0000 (23:33 +0200)]
drm/i915: Give DDI encoders even better names
Let's pimp the DDI encoder->name to reflect what the spec calls them.
Ie. on pre-tgl DDI A-F, on tgl+ DDI A-C or DDI TC1-6.
Also since each encoder is really a combination of the DDI and the PHY
we include the PHY name as well.
ICL is a bit special since it already has the two different types
of DDIs (combo or TC) but it still calls them just DDI A-F regarless
of the type. For that let's add an extra "(TC)" note to remind
is which type of DDI it really is.
The code is darn ugly, but not sure there's much we can do about it.
Merge tag 'icc-5.10-rc2' of https://git.linaro.org/people/georgi.djakov/linux into char-misc-linus
Georgi writes:
interconnect fixes for v5.10
This contains one core fix and a few driver fixes.
- Fix the core to perform also aggregation when setting the initial
bandwidth with sync_state.
- Fixes in some drivers to make sure the correct sequence is used for
initialization when we use sync_state.
- Fix in the sdm845 driver to prevent a board hang that was hit when
bandwidth scaling for display and multimedia was enabled.
Signed-off-by: Georgi Djakov <[email protected]>
* tag 'icc-5.10-rc2' of https://git.linaro.org/people/georgi.djakov/linux:
interconnect: qcom: use icc_sync state for sm8[12]50
interconnect: qcom: Ensure that the floor bandwidth value is enforced
interconnect: qcom: sc7180: Init BCMs before creating the nodes
interconnect: qcom: sdm845: Init BCMs before creating the nodes
interconnect: Aggregate before setting initial bandwidth
interconnect: qcom: sdm845: Enable keepalive for the MM1 BCM
The ABI files are supposed to be unique. Yet,
in the specific case of hw_pattern, there are some duplicated
entries as warned by scripts/get_abi.pl:
Warning: /sys/class/leds/<led>/hw_pattern is defined 3 times: Documentation/ABI/testing/sysfs-class-led-trigger-pattern:14 Documentation/ABI/testing/sysfs-class-led-driver-sc27xx:0 Documentation/ABI/testing/sysfs-class-led-driver-el15203000:0
Drop the duplication from the ABI files, moving the specific
definitions to files inside Documentation/leds.
docs: ABI: sysfs-class-backlight: unify ABI documentation
Both adp8860 and adp8870 define some extensions to the
backlight class. This causes warnings:
Warning: /sys/class/backlight/<backlight>/ambient_light_level is defined 2 times: /sys/class/backlight/<backlight>/ambient_light_level:8 /sys/class/backlight/<backlight>/ambient_light_level:30
Warning: /sys/class/backlight/<backlight>/ambient_light_zone is defined 2 times: /sys/class/backlight/<backlight>/ambient_light_zone:18 /sys/class/backlight/<backlight>/ambient_light_zone:40
As ABI definitions shouldn't be duplicated.
Unfortunately, the ABI is dependent on the specific device
features. As such, ambient_light_level range is somewhat
different among the supported devices.
The ambient_light_zone is even worse: the meanings of each
preset are different, and there's no ABI to retrieve
the supported types nor their meanins. Unfortunately,
it is too late to fix it without causing regressions,
as this has been used since Kernel v2.6.35.
Rewrite those ABI documentation using the current documentation
as a reference, and double-checking at the datasheets:
docs: ABI: sysfs-c2port: remove a duplicated entry
As warned by scripts/get_abi.pl:
Warning: /sys/class/c2port/c2portX/flash_erase is defined 2 times: Documentation/ABI/testing/sysfs-c2port:60 Documentation/ABI/testing/sysfs-c2port:68
This entry was added twice at the same patch. Probalby a
cut-and paste issue.
The ABI is not supposed to have duplicated entries, as warned
by get_abi.pl:
$ ./scripts/get_abi.pl validate 2>&1|grep sysfs-class-power
Warning: /sys/class/power_supply/<supply_name>/current_avg is defined 2 times: Documentation/ABI/testing/sysfs-class-power:108 Documentation/ABI/testing/sysfs-class-power:391
Warning: /sys/class/power_supply/<supply_name>/current_max is defined 2 times: Documentation/ABI/testing/sysfs-class-power:121 Documentation/ABI/testing/sysfs-class-power:404
Warning: /sys/class/power_supply/<supply_name>/current_now is defined 2 times: Documentation/ABI/testing/sysfs-class-power:130 Documentation/ABI/testing/sysfs-class-power:414
Warning: /sys/class/power_supply/<supply_name>/temp is defined 2 times: Documentation/ABI/testing/sysfs-class-power:281 Documentation/ABI/testing/sysfs-class-power:493
Warning: /sys/class/power_supply/<supply_name>/temp_alert_max is defined 2 times: Documentation/ABI/testing/sysfs-class-power:291 Documentation/ABI/testing/sysfs-class-power:505
Warning: /sys/class/power_supply/<supply_name>/temp_alert_min is defined 2 times: Documentation/ABI/testing/sysfs-class-power:306 Documentation/ABI/testing/sysfs-class-power:521
Warning: /sys/class/power_supply/<supply_name>/temp_max is defined 2 times: Documentation/ABI/testing/sysfs-class-power:322 Documentation/ABI/testing/sysfs-class-power:537
Warning: /sys/class/power_supply/<supply_name>/temp_min is defined 2 times: Documentation/ABI/testing/sysfs-class-power:333 Documentation/ABI/testing/sysfs-class-power:547
Warning: /sys/class/power_supply/<supply_name>/voltage_max is defined 2 times: Documentation/ABI/testing/sysfs-class-power:356 Documentation/ABI/testing/sysfs-class-power:571
Warning: /sys/class/power_supply/<supply_name>/voltage_min is defined 2 times: Documentation/ABI/testing/sysfs-class-power:367 Documentation/ABI/testing/sysfs-class-power:581
Warning: /sys/class/power_supply/<supply_name>/voltage_now is defined 2 times: Documentation/ABI/testing/sysfs-class-power:378 Documentation/ABI/testing/sysfs-class-power:591
Yet, both USB and Battery share a common set of charging-related
properties.
Unify the entries for such properties in order to avoid
duplication, while preserving the battery and USB-specific
data properly documented.
Unfortunately, (R) and (W) are valid markups for enumerated
lists, as described at:
https://docutils.sourceforge.io/docs/ref/rst/restructuredtext.html#enumerated-lists
So, we ned to replace them by:
(R) -> (Read)
(W) -> (Write)
As otherwise, (R) will be displayed as R., with is not what
it is desired.
docs: Kconfig/Makefile: add a check for broken ABI files
The files under Documentation/ABI should follow the syntax
as defined at Documentation/ABI/README.
Allow checking if they're following the syntax by running
the ABI parser script on COMPILE_TEST.
With that, when there's a problem with a file under
Documentation/ABI, it would produce a warning like:
Warning: file ./Documentation/ABI/testing/sysfs-bus-pci-devices-aer_stats#14:
What '/sys/bus/pci/devices/<dev>/aer_stats/aer_rootport_total_err_cor' doesn't have a description
Warning: file ./Documentation/ABI/testing/sysfs-bus-pci-devices-aer_stats#21:
What '/sys/bus/pci/devices/<dev>/aer_stats/aer_rootport_total_err_fatal' doesn't have a description
docs: ABI: make it parse ABI/stable as ReST-compatible files
Now that the stable ABI files are compatible with ReST,
parse them without converting complex descriptions as literal
blocks nor escaping special characters.
Please notice that escaping special characters will probably
be needed at descriptions, at least for the asterisk character.
docs: ABI: README: specify that files should be ReST compatible
As we plan to remove the escaping code from the scripts/get_abi.pl,
specify at the ABI README file that the content of the file should
be ReST compatible.
docs: kernel_abi.py: Handle with a lazy Sphinx parser
The Sphinx docutils parser is lazy: if the content is bigger than
a certain number of lines, it silenlty stops parsing it,
producing an incomplete content. This seems to be worse on newer
Sphinx versions, like 2.0.
So, change the logic to parse the contents per input file.
scripts: get_abi.pl: detect duplicated ABI definitions
The ABI should define only once each What. The current script
logic assumes that.
However, that's not the case, currently: there are several
symbols with a generic definition, and per-driver ones.
Better handle such cases, by preserving the cross-references
with the files that define them, but also track such
cases, producing warnings, as they should be fixed.
scripts: get_abi.pl: improve its parser to better catch up indentation
The original parser for indentation were relying on having
just one description for each "what". However, that's not
the case: there are a number of ABI symbols that got defined
multiple times.
Improve the parser for it to better handle descriptions
if entries are duplicated.
scripts: get_abi.pl: Allow optionally record from where a line came from
The get_abi.pl reads a lot of files and can join them on a
single output file. Store where each "What:" output came from,
in order to be able to optionally display it.
This is useful for the Sphinx extension, with can now be
able to blame what ABI file has issues, and on what line
the What: description with problems begin.
When the source ABI file is using ReST notation, the script
should handle whitespaces and lines with care, as otherwise
the file won't be properly recognized.
Address the bugs that are on such part of the script.
Merge tag 'usb-v5.10-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/peter.chen/usb into usb-linus
Peter writes:
4 bug fixes for Cadence 3 driver.
The biggest fix is changing endpoint configuration method to
avoid FIFO overflow at multiple endpoints situation.
* tag 'usb-v5.10-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/peter.chen/usb:
usb: cdns3: gadget: own the lock wrongly at the suspend routine
usb: cdns3: Fix on-chip memory overflow issue
usb: cdns3: gadget: suspicious implicit sign extension
usb: cdns3: Variable 'length' set but not used
Mark Rutland [Mon, 26 Oct 2020 13:49:31 +0000 (13:49 +0000)]
arm64: cpufeature: upgrade hyp caps to final
We finalize caps before initializing kvm hyp code, and any use of
cpus_have_const_cap() in kvm hyp code generates redundant and
potentially unsound code to read the cpu_hwcaps array.
A number of helper functions used in both hyp context and regular kernel
context use cpus_have_const_cap(), as some regular kernel code runs
before the capabilities are finalized. It's tedious and error-prone to
write separate copies of these for hyp and non-hyp code.
So that we can avoid the redundant code, let's automatically upgrade
cpus_have_const_cap() to cpus_have_final_cap() when used in hyp context.
With this change, there's never a reason to access to cpu_hwcaps array
from hyp code, and we don't need to create an NVHE alias for this.
Mark Rutland [Mon, 26 Oct 2020 13:49:29 +0000 (13:49 +0000)]
KVM: arm64: Factor out is_{vhe,nvhe}_hyp_code()
Currently has_vhe() detects whether it is being compiled for VHE/NVHE
hyp code based on preprocessor definitions, and uses this knowledge to
avoid redundant runtime checks.
There are other cases where we'd like to use this knowledge, so let's
factor the preprocessor checks out into separate helpers.
There should be no functional change as a result of this patch.
Fangrui Song [Thu, 29 Oct 2020 18:19:51 +0000 (11:19 -0700)]
arm64: Change .weak to SYM_FUNC_START_WEAK_PI for arch/arm64/lib/mem*.S
Commit 39d114ddc682 ("arm64: add KASAN support") added .weak directives to
arch/arm64/lib/mem*.S instead of changing the existing SYM_FUNC_START_PI
macros. This can lead to the assembly snippet `.weak memcpy ... .globl
memcpy` which will produce a STB_WEAK memcpy with GNU as but STB_GLOBAL
memcpy with LLVM's integrated assembler before LLVM 12. LLVM 12 (since
https://reviews.llvm.org/D90108) will error on such an overridden symbol
binding.
Use the appropriate SYM_FUNC_START_WEAK_PI instead.
Qian Cai [Wed, 28 Oct 2020 18:26:14 +0000 (14:26 -0400)]
arm64/smp: Move rcu_cpu_starting() earlier
The call to rcu_cpu_starting() in secondary_start_kernel() is not early
enough in the CPU-hotplug onlining process, which results in lockdep
splats as follows:
Laurent Vivier [Thu, 29 Oct 2020 12:20:49 +0000 (13:20 +0100)]
vdpasim: fix MAC address configuration
vdpa_sim generates a ramdom MAC address but it is never used by upper
layers because the VIRTIO_NET_F_MAC bit is not set in the features list.
Because of that, virtio-net always regenerates a random MAC address each
time it is loaded whereas the address should only change on vdpa_sim
load/unload.
Fix that by adding VIRTIO_NET_F_MAC in the features list of vdpa_sim.
Zhu Lingshan [Fri, 23 Oct 2020 10:40:46 +0000 (18:40 +0800)]
vdpa: handle irq bypass register failure case
LKP considered variable 'ret' in vhost_vdpa_setup_vq_irq() as
a unused variable, so suggest we remove it. Actually it stores
return value of irq_bypass_register_producer(), but we did not
check it, we should handle the failure case.
This commit will print a message if irq bypass register producer
fail, in this case, vqs still remain functional.
debugfs: remove return value of debugfs_create_devm_seqfile()
No one checks the return value of debugfs_create_devm_seqfile(), as it's
not needed, so make the return value void, so that no one tries to do so
in the future.
Peter Chen [Thu, 29 Oct 2020 09:55:18 +0000 (17:55 +0800)]
usb: cdns3: gadget: own the lock wrongly at the suspend routine
When the system goes to suspend, if the controller is at device mode with
cable connecting to host, the call stack is: cdns3_suspend->
cdns3_gadget_suspend -> cdns3_disconnect_gadget, after cdns3_disconnect_gadget
is called, it owns lock wrongly, it causes the system being deadlock after
resume due to at cdns3_device_thread_irq_handler, it tries to get the lock,
but can't get it forever.
To fix it, we delete the unlock-lock operations at cdns3_disconnect_gadget,
and do it at the caller.
Pawel Laszczak [Thu, 22 Oct 2020 00:55:05 +0000 (08:55 +0800)]
usb: cdns3: Fix on-chip memory overflow issue
Patch fixes issue caused setting On-chip memory overflow bit in usb_sts
register. The issue occurred because EP_CFG register was set twice
before USB_STS.CFGSTS was set. Every write operation on EP_CFG.BUFFERING
causes that controller increases internal counter holding the number
of reserved on-chip buffers. First time this register was updated in
function cdns3_ep_config before delegating SET_CONFIGURATION request
to class driver and again it was updated when class wanted to enable
endpoint. This patch fixes this issue by configuring endpoints
enabled by class driver in cdns3_gadget_ep_enable and others just
before status stage.
Dave Airlie [Thu, 29 Oct 2020 23:33:51 +0000 (09:33 +1000)]
Merge tag 'drm-misc-fixes-2020-10-29' of git://anongit.freedesktop.org/drm/drm-misc into drm-fixes
First round of drm-misc-fixes with a couple of leftovers from
drm-misc-fixes next.
Some reset fixes for the mantix panel, some fixes for a scaler issue on
sun4i, many kernel-doc fixes and various fixes for vc4 (mostly HDMI audio
related)
Lyude Paul [Tue, 29 Sep 2020 22:31:32 +0000 (18:31 -0400)]
drm/nouveau/kms/nv50-: Fix clock checking algorithm in nv50_dp_mode_valid()
While I thought I had this correct (since it actually did reject modes
like I expected during testing), Ville Syrjala from Intel pointed out
that the logic here isn't correct. max_clock refers to the max data rate
supported by the DP encoder. So, limiting it to the output of ds_clock (which
refers to the maximum dotclock of the downstream DP device) doesn't make any
sense. Additionally, since we're using the connector's bpc as the canonical BPC
we should use this in mode_valid until we support dynamically setting the bpp
based on bandwidth constraints.
v2:
* Ville pointed out I mixed up the dotclock and the link rate. So fix that...
* ...and also rename all the variables in this function to be more appropriately
labeled so I stop mixing them up.
* Reuse the bpp from the connector for now until we have dynamic bpp selection.
* Use use DIV_ROUND_UP for calculating the mode rate like i915 does, which we
should also have been doing from the start
Lyude Paul [Tue, 29 Sep 2020 22:31:31 +0000 (18:31 -0400)]
drm/nouveau/kms/nv50-: Get rid of bogus nouveau_conn_mode_valid()
Ville also pointed out that I got a lot of the logic here wrong as well, whoops.
While I don't think anyone's likely using 3D output with nouveau, the next patch
will make nouveau_conn_mode_valid() make a lot less sense. So, let's just get
rid of it and open-code it like before, while taking care to move the 3D frame
packing calculations on the dot clock into the right place.
Lyude Paul [Fri, 4 Sep 2020 20:27:58 +0000 (16:27 -0400)]
drm/nouveau/kms/nv50-: Program notifier offset before requesting disp caps
Not entirely sure why this never came up when I originally tested this
(maybe some BIOSes already have this setup?) but the ->caps_init vfunc
appears to cause the display engine to throw an exception on driver
init, at least on my ThinkPad P72:
This is magic nvidia speak for "You need to have the DMA notifier offset
programmed before you can call NV507D_GET_CAPABILITIES." So, let's fix
this by doing that, and also perform an update afterwards to prevent
racing with the GPU when reading capabilities.
v2:
* Don't just program the DMA notifier offset, make sure to actually
perform an update
v3:
* Don't call UPDATE()
* Actually read the correct notifier fields, as apparently the
CAPABILITIES_DONE field lives in a different location than the main
NV_DISP_CORE_NOTIFIER_1 field. As well, 907d+ use a different
CAPABILITIES_DONE field then pre-907d cards.
v4:
* Don't forget to check the return value of core507d_read_caps()
v5:
* Get rid of NV50_DISP_CAPS_NTFY[14], use NV50_DISP_CORE_NTFY
* Disable notifier after calling GetCapabilities()
Signed-off-by: Lyude Paul <[email protected]> Fixes: 4a2cb4181b07 ("drm/nouveau/kms/nv50-: Probe SOR and PIOR caps for DP interlacing support") Cc: <[email protected]> # v5.8+ Signed-off-by: Ben Skeggs <[email protected]>
Ralph Campbell [Mon, 31 Aug 2020 20:31:11 +0000 (13:31 -0700)]
drm/nouveau/nouveau: fix the start/end range for migration
The user level OpenCL code shouldn't have to align start and end
addresses to a page boundary. That is better handled in the nouveau
driver. The npages field is also redundant since it can be computed
from the start and end addresses.
Dave Airlie [Thu, 29 Oct 2020 23:05:26 +0000 (09:05 +1000)]
Merge tag 'amd-drm-fixes-5.10-2020-10-29' of git://people.freedesktop.org/~agd5f/linux into drm-fixes
amd-drm-fixes-5.10-2020-10-29:
amdgpu:
- Add new navi1x PCI ID
- GPUVM reserved area fixes
- Misc display fixes
- Fix bad interactions between display code and CONFIG_KGDB
- Fixes for SMU manual fan control and i2c
scsi: target: tcmu: Replace zero-length array with flexible-array member
There is a regular need in the kernel to provide a way to declare having a
dynamically sized set of trailing elements in a structure. Kernel code should
always use “flexible array members”[1] for these cases. The older style of
one-element or zero-length arrays should no longer be used[2].
ima: Replace zero-length array with flexible-array member
There is a regular need in the kernel to provide a way to declare having a
dynamically sized set of trailing elements in a structure. Kernel code should
always use “flexible array members”[1] for these cases. The older style of
one-element or zero-length arrays should no longer be used[2].
enetc: Replace zero-length array with flexible-array member
There is a regular need in the kernel to provide a way to declare having a
dynamically sized set of trailing elements in a structure. Kernel code should
always use “flexible array members”[1] for these cases. The older style of
one-element or zero-length arrays should no longer be used[2].
fs: Replace zero-length array with flexible-array member
There is a regular need in the kernel to provide a way to declare having a
dynamically sized set of trailing elements in a structure. Kernel code should
always use “flexible array members”[1] for these cases. The older style of
one-element or zero-length arrays should no longer be used[2].
Bluetooth: Replace zero-length array with flexible-array member
There is a regular need in the kernel to provide a way to declare having a
dynamically sized set of trailing elements in a structure. Kernel code should
always use “flexible array members”[1] for these cases. The older style of
one-element or zero-length arrays should no longer be used[2].
params: Replace zero-length array with flexible-array member
There is a regular need in the kernel to provide a way to declare having a
dynamically sized set of trailing elements in a structure. Kernel code should
always use “flexible array members”[1] for these cases. The older style of
one-element or zero-length arrays should no longer be used[2].
tracepoint: Replace zero-length array with flexible-array member
There is a regular need in the kernel to provide a way to declare having a
dynamically sized set of trailing elements in a structure. Kernel code should
always use “flexible array members”[1] for these cases. The older style of
one-element or zero-length arrays should no longer be used[2].
platform/chrome: cros_ec_proto: Replace zero-length array with flexible-array member
There is a regular need in the kernel to provide a way to declare having a
dynamically sized set of trailing elements in a structure. Kernel code should
always use “flexible array members”[1] for these cases. The older style of
one-element or zero-length arrays should no longer be used[2].
platform/chrome: cros_ec_commands: Replace zero-length array with flexible-array member
There is a regular need in the kernel to provide a way to declare having a
dynamically sized set of trailing elements in a structure. Kernel code should
always use “flexible array members”[1] for these cases. The older style of
one-element or zero-length arrays should no longer be used[2].
mailbox: zynqmp-ipi-message: Replace zero-length array with flexible-array member
There is a regular need in the kernel to provide a way to declare having a
dynamically sized set of trailing elements in a structure. Kernel code should
always use “flexible array members”[1] for these cases. The older style of
one-element or zero-length arrays should no longer be used[2].
dmaengine: ti-cppi5: Replace zero-length array with flexible-array member
There is a regular need in the kernel to provide a way to declare having a
dynamically sized set of trailing elements in a structure. Kernel code should
always use “flexible array members”[1] for these cases. The older style of
one-element or zero-length arrays should no longer be used[2].
Tero Kristo [Thu, 29 Oct 2020 09:33:37 +0000 (11:33 +0200)]
soc: ti: ti_sci_pm_domains: check for proper args count in xlate
K2G devices still only use single parameter for power-domains property,
so check for this properly in the driver. Without this, every peripheral
fails to probe resulting in boot failure.
Santosh Shukla [Mon, 26 Oct 2020 11:24:07 +0000 (16:54 +0530)]
KVM: arm64: Force PTE mapping on fault resulting in a device mapping
VFIO allows a device driver to resolve a fault by mapping a MMIO
range. This can be subsequently result in user_mem_abort() to
try and compute a huge mapping based on the MMIO pfn, which is
a sure recipe for things to go wrong.
Instead, force a PTE mapping when the pfn faulted in has a device
mapping.
Linus Torvalds [Thu, 29 Oct 2020 20:02:52 +0000 (13:02 -0700)]
Merge tag 'fallthrough-fixes-clang-5.10-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gustavoars/linux
Pull fallthrough fix from Gustavo A. R. Silva:
"This fixes a ton of fall-through warnings when building with Clang
12.0.0 and -Wimplicit-fallthrough"
* tag 'fallthrough-fixes-clang-5.10-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gustavoars/linux:
include: jhash/signal: Fix fall-through warnings for Clang
Linus Torvalds [Thu, 29 Oct 2020 19:55:02 +0000 (12:55 -0700)]
Merge tag 'net-5.10-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Pull networking fixes from Jakub Kicinski:
"Current release regressions:
- r8169: fix forced threading conflicting with other shared
interrupts; we tried to fix the use of raise_softirq_irqoff from an
IRQ handler on RT by forcing hard irqs, but this driver shares
legacy PCI IRQs so drop the _irqoff() instead
- tipc: fix memory leak caused by a recent syzbot report fix to
tipc_buf_append()
Current release - bugs in new features:
- devlink: Unlock on error in dumpit() and fix some error codes
- net/smc: fix null pointer dereference in smc_listen_decline()
Previous release - regressions:
- tcp: Prevent low rmem stalls with SO_RCVLOWAT.
- net: protect tcf_block_unbind with block lock
- ibmveth: Fix use of ibmveth in a bridge; the self-imposed filtering
to only send legal frames to the hypervisor was too strict
- net: hns3: Clear the CMDQ registers before unmapping BAR region;
incorrect cleanup order was leading to a crash
- bnxt_en - handful of fixes to fixes:
- Send HWRM_FUNC_RESET fw command unconditionally, even if there
are PCIe errors being reported
- Check abort error state in bnxt_open_nic().
- Invoke cancel_delayed_work_sync() for PFs also.
- Fix regression in workqueue cleanup logic in bnxt_remove_one().
- mlxsw: Only advertise link modes supported by both driver and
device, after removal of 56G support from the driver 56G was not
cleared from advertised modes
- net/smc: fix suppressed return code
Previous release - always broken:
- netem: fix zero division in tabledist, caused by integer overflow
- bnxt_en: Re-write PCI BARs after PCI fatal error.
- cxgb4: set up filter action after rewrites
- net: ipa: command payloads already mapped
Misc:
- s390/ism: fix incorrect system EID, it's okay to change since it
was added in current release
- vsock: use ns_capable_noaudit() on socket create to suppress false
positive audit messages"
* tag 'net-5.10-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (36 commits)
r8169: fix issue with forced threading in combination with shared interrupts
netem: fix zero division in tabledist
ibmvnic: fix ibmvnic_set_mac
mptcp: add missing memory scheduling in the rx path
tipc: fix memory leak caused by tipc_buf_append()
gtp: fix an use-before-init in gtp_newlink()
net: protect tcf_block_unbind with block lock
ibmveth: Fix use of ibmveth in a bridge.
net/sched: act_mpls: Add softdep on mpls_gso.ko
ravb: Fix bit fields checking in ravb_hwtstamp_get()
devlink: Unlock on error in dumpit()
devlink: Fix some error codes
chelsio/chtls: fix memory leaks in CPL handlers
chelsio/chtls: fix deadlock issue
net: hns3: Clear the CMDQ registers before unmapping BAR region
bnxt_en: Send HWRM_FUNC_RESET fw command unconditionally.
bnxt_en: Check abort error state in bnxt_open_nic().
bnxt_en: Re-write PCI BARs after PCI fatal error.
bnxt_en: Invoke cancel_delayed_work_sync() for PFs also.
bnxt_en: Fix regression in workqueue cleanup logic in bnxt_remove_one().
...
Will Deacon [Thu, 29 Oct 2020 14:47:16 +0000 (14:47 +0000)]
KVM: arm64: Fix masks in stage2_pte_cacheable()
stage2_pte_cacheable() tries to figure out whether the mapping installed
in its 'pte' parameter is cacheable or not. Unfortunately, it fails
miserably because it extracts the memory attributes from the entry using
FIELD_GET(), which returns the attributes shifted down to bit 0, but then
compares this with the unshifted value generated by the PAGE_S2_MEMATTR()
macro.
A direct consequence of this bug is that cache maintenance is silently
skipped, which in turn causes 32-bit guests to crash early on when their
set/way maintenance is trapped but not emulated correctly.
Fix the broken masks by avoiding the use of FIELD_GET() altogether.
Marc Zyngier [Thu, 29 Oct 2020 17:24:09 +0000 (17:24 +0000)]
KVM: arm64: Fix AArch32 handling of DBGD{CCINT,SCRext} and DBGVCR
The DBGD{CCINT,SCRext} and DBGVCR register entries in the cp14 array
are missing their target register, resulting in all accesses being
targetted at the guard sysreg (indexed by __INVALID_SYSREG__).
Point the emulation code at the actual register entries.
Will Deacon [Mon, 26 Oct 2020 14:44:23 +0000 (14:44 +0000)]
KVM: arm64: Allocate stage-2 pgd pages with GFP_KERNEL_ACCOUNT
For consistency with the rest of the stage-2 page-table page allocations
(performing using a kvm_mmu_memory_cache), ensure that __GFP_ACCOUNT is
included in the GFP flags for the PGD pages.
Marc Zyngier [Mon, 26 Oct 2020 09:51:11 +0000 (09:51 +0000)]
KVM: arm64: Drop useless PAN setting on host EL1 to EL2 transition
Setting PSTATE.PAN when entering EL2 on nVHE doesn't make much
sense as this bit only means something for translation regimes
that include EL0. This obviously isn't the case in the nVHE case,
so let's drop this setting.
projects / linux.git / log