We only need to initialise sctlr_el1 if we're installing an EL2 stub, so
we may as well defer this until we're doing so. Similarly, we can defer
intialising CPTR_EL2 until then, as we do not access any trapped
functionality as part of el2_setup.
This patch modified el2_setup accordingly, allowing us to remove a
branch and simplify the code flow.
Acked-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
The early el2_setup code is a little convoluted, with two branches where
one would do. This makes the code more painful to read than is
necessary.
We can remove a branch and simplify the logic by moving the early return
in the booted-at-EL1 case earlier in the function. This separates it
from all the setup logic that only makes sense for EL2.
Acked-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Check if CONFIG_HAVE_HW_BREAKPOINT is enabled before compiling in extra
data required for hardware breakpoints. Compiling out this code when hw
breakpoints are disabled saves about 272 bytes per struct task_struct.
Signed-off-by: Chris Redmon <credmonster@gmail.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Add support for allocating physically contiguous DMA buffers on arm64
systems with an IOMMU. This can be useful when two or more devices
with different memory requirements are involved in buffer sharing.
Note that as this uses the CMA allocator, setting the
DMA_ATTR_FORCE_CONTIGUOUS attribute has a runtime-dependency on
CONFIG_DMA_CMA, just like on arm32.
For arm64 systems using swiotlb, no changes are needed to support the
allocation of physically contiguous DMA buffers:
- swiotlb always uses physically contiguous buffers (up to
IO_TLB_SEGSIZE = 128 pages),
- arm64's __dma_alloc_coherent() already calls
dma_alloc_from_contiguous() when CMA is available.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
This mirrors commit e9c38ceba8 ("ARM: 8455/1: define __BUG as
asm(BUG_INSTR) without CONFIG_BUG") to make the behavior of
arm64 consistent with arm and x86, and avoids lots of warnings in
randconfig builds, such as:
kernel/seccomp.c: In function '__seccomp_filter':
kernel/seccomp.c:666:1: error: no return statement in function returning non-void [-Werror=return-type]
Acked-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Cache related issues with DMA rings and performance issues related to
caching are being caused by not properly setting the "dma-coherent" flag
in the device tree entries. Adding it here to correct the issue.
Signed-off-by: Jon Mason <jon.mason@broadcom.com>
Fixes: fd5e5dd56 ("arm64: dts: Add PCIe0 and PCIe4 DT nodes for NS2")
Fixes: dddc3c9d7 ("arm64: dts: NS2: add AMAC ethernet support")
Fixes: e79249143 ("arm64: dts: Add Broadcom Northstar2 device tree entries for PDC driver")
Fixes: ac9aae00f ("arm64: dts: Add SATA3 AHCI and SATA3 PHY DT nodes for NS2")
Fixes: efc877676 ("arm64: dts: Add SDHCI DT node for NS2")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Rockchip finally named the SOC as RV1108, so change it.
Signed-off-by: Andy Yan <andy.yan@rock-chips.com>
[include rename in rk1108.dtsi to prevent compile errors]
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
It's suggested to fix the domain number for all PCIe
host bridges or not set it at all. However, if we don't
fix it, the domain number will keep increasing ever when
doing unbind/bind test, which makes the bus tree of lspci
introduce pointless domain hierarchy. More investigation shows
the domain number allocater of PCI doesn't consider the conflict
of domain number if we have more than one PCIe port belonging to
different domains. So once unbinding/binding one of them and keep
others would going to overflow the domain number so that finally
it will share the same domain as others, but actually it shouldn't.
We should fix the domain number for PCIe or invent new indexing
ID mechanisms. However it isn't worth inventing new indexing ID
mechanisms personlly, Just look at how other Root Complex drivers
did, for instance, broadcom and qualcomm, it seems fixing the domain
number was more popular. So this patch gonna fix the domain number
of PCIe for rk3399.
Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
Reviewed-by: Brian Norris <briannorris@chromium.org>
Tested-by: Brian Norris <briannorris@chromium.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
This adds baud rate, parity & number of data bits. It's required to get
serial working correctly.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
So far every Northstar device we have seen was using the same serial
console params (115200n8). It probably make the most sense to put it in
some proper dtsi files instead of repeating over and over for every
single device. As different boards may use different bootloaders it
seems the safest idea is to use board specific dtsi files.
Just in case some vendor decides to use different UART (parameters) this
can be always easily overwritten.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Droid 4 has two modems, mdm6600 and w3glte. Both are on the HCI USB
controller.
Let's add a configuration for the HCI so the modems can be enabled.
Note that the modems still need additional GPIO based configuration.
Cc: devicetree@vger.kernel.org
Cc: Marcel Partap <mpartap@gmx.net>
Cc: Michael Scott <michael.scott@linaro.org>
Tested-by: Sebastian Reichel <sre@kernel.org>
[tony@atomide.com: left out url]
Signed-off-by: Tony Lindgren <tony@atomide.com>
We can get HDMI working as long as the 5V regulator is on. There is
probably an encoder chip there too, but so far no idea what it might be.
Let's keep the 5V HDMI regulator always enabled for now as otherwise we
cannot detect the monitor properly.
Cc: devicetree@vger.kernel.org
Cc: Marcel Partap <mpartap@gmx.net>
Cc: Michael Scott <michael.scott@linaro.org>
Cc: Tomi Valkeinen <tomi.valkeinen@ti.com>
Tested-By: Sebastian Reichel <sre@kernel.org>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Add tmp105 sensor for droid 4. This can be used with modprobe
lm75.ko and running sensors from lm-sensors package. Note that
the lm75.c driver does not yet support alert interrupt but
droid 4 seems to be wired for it.
Cc: devicetree@vger.kernel.org
Cc: Marcel Partap <mpartap@gmx.net>
Cc: Michael Scott <michael.scott@linaro.org>
Tested-By: Sebastian Reichel <sre@kernel.org>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Droid 4 has a GPIO line that we can use with CONFIG_POWER_RESET_GPIO.
It is probably connected to the CPCAP PMIC, and seems to power down
the whole device taking power consumption to zero based on what
I measured.
Cc: devicetree@vger.kernel.org
Cc: Marcel Partap <mpartap@gmx.net>
Cc: Michael Scott <michael.scott@linaro.org>
Tested-By: Sebastian Reichel <sre@kernel.org>
Signed-off-by: Tony Lindgren <tony@atomide.com>
This commit enable DMA-based transfers for SD/eMMC card adapters
and reduce number of interrupts produced by SD-card/eMMC-card
adapters.
Sometimes interrupts from SD-card/eMMC-card adapters running in
PIO mode blocks execution of hrtimers and I2S DMA callbacks for
a long periods (100 ms or more).
Signed-off-by: Alexander Kochetkov <al.kochet@gmail.com>
[moved dma properties to rk3xxx.dtsi and added sdio dma]
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
If kernel image extends across alignment boundary, existing
code increases the KASLR offset by size of kernel image. The
offset is masked after resizing. There are cases, where after
masking, we may still have kernel image extending across
boundary. This eventually results in only 2MB block getting
mapped while creating the page tables. This results in data aborts
while accessing unmapped regions during second relocation (with
kaslr offset) in __primary_switch. To fix this problem, round up the
kernel image size, by swapper block size, before adding it for
correction.
For example consider below case, where kernel image still crosses
1GB alignment boundary, after masking the offset, which is fixed
by rounding up kernel image size.
SWAPPER_TABLE_SHIFT = 30
Swapper using section maps with section size 2MB.
CONFIG_PGTABLE_LEVELS = 3
VA_BITS = 39
_text : 0xffffff8008080000
_end : 0xffffff800aa1b000
offset : 0x1f35600000
mask = ((1UL << (VA_BITS - 2)) - 1) & ~(SZ_2M - 1)
(_text + offset) >> SWAPPER_TABLE_SHIFT = 0x3fffffe7c
(_end + offset) >> SWAPPER_TABLE_SHIFT = 0x3fffffe7d
offset after existing correction (before mask) = 0x1f37f9b000
(_text + offset) >> SWAPPER_TABLE_SHIFT = 0x3fffffe7d
(_end + offset) >> SWAPPER_TABLE_SHIFT = 0x3fffffe7d
offset (after mask) = 0x1f37e00000
(_text + offset) >> SWAPPER_TABLE_SHIFT = 0x3fffffe7c
(_end + offset) >> SWAPPER_TABLE_SHIFT = 0x3fffffe7d
new offset w/ rounding up = 0x1f38000000
(_text + offset) >> SWAPPER_TABLE_SHIFT = 0x3fffffe7d
(_end + offset) >> SWAPPER_TABLE_SHIFT = 0x3fffffe7d
Fixes: f80fb3a3d5 ("arm64: add support for kernel ASLR")
Cc: <stable@vger.kernel.org>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Neeraj Upadhyay <neeraju@codeaurora.org>
Signed-off-by: Srinivas Ramana <sramana@codeaurora.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
The vpif display driver uses a static i2c adapter ID of 1 but on the
da850-evm board in DT boot mode the i2c adapter ID is actually 0.
Make the adapter ID configurable like it already is for vpif capture.
Signed-off-by: Bartosz Golaszewski <bgolaszewski@baylibre.com>
Acked-by: Kevin Hilman <khilman@baylibre.com>
Acked-by: Lad, Prabhakar <prabhakar.csengg@gmail.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@s-opensource.com>
We want to select the GPIO_FTGPIO010 symbol for the generic
Faraday FTGPIO010 driver, not the old driver.
Cc: Jonas Jensen <jonas.jensen@gmail.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
The binding should state "cortina,gemini-gpio", "faraday,ftgpio010"
stating the full name of the IP part.
Cc: Jonas Jensen <jonas.jensen@gmail.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
dw-mmc got its reset-properties specified, so add the softresets
for it on the rk3368.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Shawn Lin <shawn.lin@rock-chips.com>
I2S of RK3368 SoCs keep same as RK3066 SoCs found on Rockchip,
add nodes to support them.
Signed-off-by: Jianqun Xu <jay.xu@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Add dmac bus and dmac peri dts nodes for peripherals,
such as I2S, SPI, UART and so on.
Signed-off-by: Huibin Hong <huibin.hong@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
As reported by Lorenzo, the residency/latency values defined in the
idle-state for rk3368 "make no sense". When introducing them I
simply took the idle-state node from the vendor kernel in error
as I didn't look up if these values were sane in the first place.
Talking to people and determining why they were used in this way
showed that it was meant to make sure the cpu_suspend callback
got initialized which at the 3.10 time was somehow required even
for wfi-based idle handling.
Of course the generic arch_cpu_idle() now does wfi-based idle-handling
already and the rk3368 does not implement any other idle states than
the default WFI, so these wrong idle-states should go away.
Reported-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Acked-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Used for Gru/Kevin only, as they're the only ones which have a described
CPU regulator. Also, I'm not sure we've validated this table non-Gru
boards.
At the same time, partially describe PWM regulators for Gru, so cpufreq
doesn't think it can crank up the clock speed without changing the
voltage. However, we don't yet have the DT bindings to fully describe
the Over Voltage Protection (OVP) circuits on these boards. Without that
description, we might end up changing the voltage too much, too fast.
Add the pwm-regulator descriptions and associate the CPU OPPs, but leave
them disabled.
Signed-off-by: Brian Norris <briannorris@chromium.org>
[shared gru/kevin parts on a gru device]
Tested-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
[with a bit of reordering]
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Kevin is part of a family of boards called Gru. As best as possible, the
properties shared by the Gru family are placed in rk3399-gru.dtsi, while
Kevin-specific bits are in rk3399-gru-kevin.dts. This does not add full
support for the base Gru board.
Working and tested (to some extent):
* EC support -- including keyboard, battery, PWM, and probably more
* UART / console
* Thermal
* Touchscreen
* Touchpad
* Digitizer (regulator still WIP)
* PCIe / Wifi
* Bluetooth / Webcam
* SD card
* eMMC
* USB2 on TypeC
- This works much of the time, but USB3 devices may or may not detect
properly. Waiting on proper extcon support for USB3 over TypeC.
- Depends on XHCI/DWC3 fixes for ARM64 that still haven't landed
* Backlight
Not working:
* CPUFreq -- relies on special OVP support for our PWM regulator
circuits
* EC / extcon support -- and with it, USB3/TypeC/DP
* DRM -- won't even build on ARM64, so all display, eDP, etc. is not
enabled
Not tested:
* Audio
Signed-off-by: Brian Norris <briannorris@chromium.org>
[shared gru/kevin parts on a gru device]
Tested-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
[with a bit of reordering]
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
When the separate IRQ stack was introduced, stack unwinding only
proceeded as far as the top of the IRQ stack, leading to kernel
backtraces being less useful, lacking the trace of what was interrupted.
Fix this by providing a means for the kernel to unwind the IRQ stack
onto the interrupted task stack. The processor state is saved to the
kernel task stack on interrupt. The IRQ_STACK_START macro reserves an
unsigned long at the top of the IRQ stack where the interrupted task
stack pointer can be saved. After the active stack is switched to the
IRQ stack, save the interrupted tasks stack pointer to the reserved
location.
Fix the stack unwinding code to look for the frame being the top of the
IRQ stack and if so get the next frame from the saved location. The
existing test does not work with the separate stack since the ra is no
longer pointed at ret_from_{irq,exception}.
The test to stop unwinding the stack 32 bytes from the top of a stack
must be modified to allow unwinding to continue up to the location of
the saved task stack pointer when on the IRQ stack. The low / high marks
of the stack are set depending on whether the sp is on an irq stack or
not.
Signed-off-by: Matt Redfearn <matt.redfearn@imgtec.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Marcin Nowakowski <marcin.nowakowski@imgtec.com>
Cc: Masanari Iida <standby24x7@gmail.com>
Cc: Chris Metcalf <cmetcalf@mellanox.com>
Cc: James Hogan <james.hogan@imgtec.com>
Cc: Paul Burton <paul.burton@imgtec.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jason A. Donenfeld <jason@zx2c4.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/15788/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
According to [0] pointed out by Marc Zyngier in a report about a
similar error message, PPIs 11 and 13 are edge triggered on
Cortex-A9 socs including the rk3066 and rk3188 which currently
mark them as level triggered.
Until some time ago the gic did not care but commit 992345a58e
("irqchip/gic: WARN if setting the interrupt type for a PPI fails")
introduced a warning for that case.
Fix the warning on these socs by describing the interrupts correctly
and also using the binding constants for easier reading in the future.
[0] http://infocenter.arm.com/help/topic/com.arm.doc.ddi0407f/CCHEIGIC.html
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
There is no need for the __ASSEMBLY__ ifdefery anymore since the
architecture level set code that deals with facility bits was
converted to C in the meantime.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Use the actual size of the facility list array within the lowcore
structure for the MAX_FACILITY_BIT define instead of a comment which
states what this is good for. This makes it a bit harder to break
things.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Provide the raw stsi 15,1,x data contents via debugfs. This makes it
much easier to debug unexpected scheduling domains on machines that
provide cpu topology information.
Therefore this file adds a new 's390/stsi' debugfs directory with a
file for each possible topology nesting level that is allowed by the
architecture. The files will be created regardless if the machine
supports all, or any, level. If a level is not supported, or no data
is available, user space can recognize this with a -EINVAL error code
when trying to read such data.
In addition a 'topology' symlink is created that points to the file
that contains the data that is used to create the scheduling domains.
Acked-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Introduce a top-level 's390' directory which should be used when
adding new s390 specific debug feature files and/or directories.
This makes hopefully sure that the contents of the s390 directory will
be a bit more structured. Right now we have a couple of top-level
files where it is not easy to tell to which subsystem they belong to.
Acked-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
User space needs some information about the secure key(s)
before actually invoking the pkey and/or paes funcionality.
This patch introduces a new ioctl API and in kernel API to
verify the the secure key blob and give back some
information about the key (type, bitsize, old MKVP).
Both APIs are described in detail in the header files
arch/s390/include/asm/pkey.h and arch/s390/include/uapi/asm/pkey.h.
Signed-off-by: Harald Freudenberger <freude@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
If running within a level 3 hypervisor, the hypervisor provides a
SYSIB block which contains a control program indentifier string. Use
this string instead of the simple KVM and z/VM strings only. In case
of z/VM this provides addtional information: the z/VM version.
The new string looks similar to this:
Hardware name: IBM 2964 N96 702 (z/VM 6.4.0)
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
The arch description provided for the "Hardware name:" contains lots
of extra whitespace due to the way the SYSIB contents are defined
(strings aren't zero terminated).
This looks a bit odd and therefore remove the extra whitespace
characters. This also gives the opportunity to add more information,
if required, without hitting the magic 80 characters per line limit.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Allow compiler warnings again for the sysinfo file. Compiler warnings
were disabled when the bogomips calculation with math-emu code was
introduced ("[S390] Calibrate delay and bogomips.").
Since that code is gone, we can enable warnings again.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Use MACHINE_FLAG_TOPOLOGY instead of MACHINE_HAS_TOPOLOGY when
clearing the bit that indicates if the machine provides topology
information (and if it should be used). Currently works anyway.
Fixes: 68cc795d19 ("s390/topology: make "topology=off" parameter work")
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Use a single long value instead of a single element array to represent
the core mask. The array is a leftover from 32/31 bit code so we were
able to use bitops helper functions.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
The s390 specific little endian bitop macros are gone since a long time.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
__GFP_REPEAT has a rather weak semantic but since it has been introduced
around 2.6.12 it has been ignored for low order allocations.
page_table_alloc then uses the flag for a single page allocation. This
means that this flag has never been actually useful here because it has
always been used only for PAGE_ALLOC_COSTLY requests.
An earlier attempt to remove the flag 10d58bf297 ("s390: get rid of
superfluous __GFP_REPEAT") has missed this one but the situation is very
same here.
Signed-off-by: Michal Hocko <mhocko@suse.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
