MGS-3832 [#imx-921] Fix the cl_convolution of ACL which cause MMU exception

Dump the error info only when there is an error. If the gcdALLOC_ON_FAULT is not enabled, still need to check
the HW status to decide whether to dump the exception info or not

Signed-off-by: Ella Feng <ella.feng@nxp.com>
Reviewed-by: Xianzhong Li <xianzhong.li@nxp.com>

MGS-3694 [#imx-913] enable fb fd and tile status fd in GPU and DCSS on wayland

Save the meta data info in the _gcsVIDMEM_NODE.
DCSS can query the meta data, and get the tile statust buffer info.

Date: Apr 13, 2018
Signed-off-by: Yong Gan yong.gan@nxp.com
(cherry picked from commit cea92256fec8380e6e185d65c7746988e7c6426e)

MLK-18031 video: mxsfb: enable global alpha when grayscale is 0 for 32bpp format

For overlay framebuffer, when its grayscale is '0' and the format
bpp is 32, enable the global alpha blending by default which can
make the overlay fb can display the framebuffer content as long as
it is unblanked.

Signed-off-by: Fancy Fang <chen.fang@nxp.com>
(cherry picked from commit fede70b9066b03c24347619775fd7d007c5ddafb)

MLK-18051 arm: imx: fix the audio bus hang when tee enabled

fix audio bus mode hang issue on imx6sl. The root cause of
this issue is that busfreq mode passed to TEE side is wrong,
it will lead to ccm setting is wrong in TEE.

Signed-off-by: Bai Ping <ping.bai@nxp.com>
Tested-by: Anson huang <anson.huang@nxp.com>

MLK-18036-3: ARM: dts: Remove imx optee dts files from Makefile

Fixes: a6fd1613cca4 ("MLK-18036-2 Delete *optee.dts files")

Signed-off-by: Leonard Crestez <leonard.crestez@nxp.com>
Reviewed-by: Clement Faure <clement.faure@nxp.com>
(cherry picked from commit 4706425bb745850c8d28608ebcaffd239a945e63)

MLK-18036-2 Delete *optee.dts files

A specific node for OCRAM mapping in optee as been added in
the device tree. These dedicated optee device trees can be
removed.

Signed-off-by: Clement Faure <clement.faure@nxp.com>
Acked-by: Peng Fan <peng.fan@nxp.com>
(cherry picked from commit a6fd1613cca4a5008c347d4473b92b119385644c)

MLK-18036-1 Add "fsl,optee-lpm-sram" node for optee os power management.

This node will be used by the OCRAM driver in optee to:
* Get the OCRAM start address for power management in optee.
* Add an entry that will overwrite ocrams nodes and dynamically reduce
the OCRAM available for mmio-sram in Linux.

That way we do not touch the legacy Linux boot and remove the dedicated
optee device tree.

Signed-off-by: Clement Faure <clement.faure@nxp.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
(cherry picked from commit e96a3bcd754dee0aef3519bc08979985493be52c)

MLK-16912 PL310: unlock ways during initialization

This change affects all i.MX 6 with PL310 L2 Cache controller.
When Linux runs in Non-secure World the PL310 has already
been initialized by the ARM secure World running OP-TEE os.
However, in order to have a proper Linux Initialization all the
L2 cache ways have been locked by the secure world.

This patch unlock all the ways during pl310 initialization.

Signed-off-by: Cedric Neveux <cedric.neveux@nxp.com>
(cherry picked from commit be7971b62e0c77cf70f828868a5d5a4184a926d2)

MLK-18088: arm: dts: enable the 7ulp mmdc profiling feature

7ULP uses the same mmdc profiling block as i.mx6q. Added the
"fsl,imx6q-mmdc" compatible string to enable the mmdc profiling
feature.

Signed-off-by: Shenwei Wang <shenwei.wang@nxp.com>

cfg80211: improve warnings in VHT rate calculation

Linus reported hitting the bandwidth warning, but it is indeed
pretty useless - improve it by printing the rate configuration
and make it only warn once, for both warnings here.

Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

MGS-3806: clocksource: tpm: make sure returning -ETIME case correct

Incorrect condition check causes -ETIME return only
happen when next event is equal to current counter, and
it would cause various system issue like RCU stalls etc.,
the correct case should be whenever next event is less
than current counter, -ETIME should be returned. Correct
the type cast during return condition check to make it
work right.

Signed-off-by: Anson Huang <Anson.Huang@nxp.com>
Reviewed-by: Bai Ping <ping.bai@nxp.com>

MLK-18011: imx7ulp: caam: Configure CAAM clocks

Signed-off-by: Franck LENORMAND <franck.lenormand@nxp.com>
(cherry picked from commit 0b122f882429a82274fc99439b5d73986b731672)

MLK-17992: caam: sm: Fix compilation warnings

Fix the following warnings in CAAM SM:

    drivers/crypto/caam/sm_store.c: In function 'blacken_key_jobdesc':
    drivers/crypto/caam/sm_store.c:141:19: warning: cast from pointer
     to integer of different size [-Wpointer-to-int-cast]
      tmpdesc[idx++] = (u32)key;
                       ^
    drivers/crypto/caam/sm_store.c:153:19: warning: cast from pointer
     to integer of different size [-Wpointer-to-int-cast]
      tmpdesc[idx++] = (u32)key;
                       ^
    drivers/crypto/caam/sm_store.c: In function 'blob_encap_jobdesc':
    drivers/crypto/caam/sm_store.c:274:19: warning: cast from pointer
     to integer of different size [-Wpointer-to-int-cast]
      tmpdesc[idx++] = (u32)secretbuf;
                       ^
    drivers/crypto/caam/sm_store.c: In function 'blob_decap_jobdesc':
    drivers/crypto/caam/sm_store.c:390:19: warning: cast from pointer
     to integer of different size [-Wpointer-to-int-cast]
      tmpdesc[idx++] = (u32)outbuf;
                       ^
    drivers/crypto/caam/sm_store.c: In function 'slot_get_base':
    drivers/crypto/caam/sm_store.c:569:9: warning: cast from pointer
     to integer of different size [-Wpointer-to-int-cast]
      return (u32)(ksdata->base_address);
             ^
    drivers/crypto/caam/sm_store.c: In function 'sm_keystore_slot_load':
    drivers/crypto/caam/sm_store.c:789:6:
     warning: unused variable 'i' [-Wunused-variable]
      u32 i;

Signed-off-by: Franck LENORMAND <franck.lenormand@nxp.com>
(cherry picked from commit b6bd87e624bbe30b9be19c3f8ccb8f5526e4186b)

MLK-18042-02 arm: dts: update the backlight brightness on imx7ulp evk

On the i.MX7ULP EVK Rev.B baord, the backlight brigntness driver circuit
is updated. A RC filter is added on the MP3301's EN pin. So the PWM's frequency
should be change to 20KHZ. for EN pin, A DC voltage from 0.7V to 1.4V can control
the LED current from 0% to 100%. the backlight brightness level also need to be
updated.

Signed-off-by: Bai Ping <ping.bai@nxp.com>
Reviewed-by: Anson Huang <Anson.Huang@nxp.com>
(cherry picked from commit 82555e15a5f958c09492d0103425dc30bc7cd927)

MLK-18042-01 driver: pwm: fix pwm pre-scale div config

In i.MX7ULP TPM PWM module, it has a pre-scale divider,
this divider setting is missed, so fix it.

Signed-off-by: Bai Ping <ping.bai@nxp.com>
Reviewed-by: Anson Huang <Anson.Huang@nxp.com>
(cherry picked from commit 3ffd915e44320a8142698ca3f6e19c30ec434f61)

MLK-17993: arm: dts: rename the dts to match i.MX7ULP board design

The default display interface on i.MX7ULP EVK board is the HDMI
interface, and a hardware rework is required to support the MIPI
panel. To match the current board design, added the HDMI node in
the imx7ulp-evk.dts and created a  new file named imx7ulp-evk-mipi.dts.

Signed-off-by: Shenwei Wang <shenwei.wang@nxp.com>
Reviewed-by: Andy Duan <fugang.duan@nxp.com>

MLK-18026 dts: imx7ulp-evk-sd1: delete non-removable property for sd1 slot

commit a56e6e190015 ("MLK-17961 dts: imx7ulp-evk: add non-removable
property for wifi sdio") add non-removable property, sd1 slot on
base board share the same usdhc with wifi, and the sd1 slot support
card detect, so for sd1 slot, need to remove the non-removable
property.

Signed-off-by: Haibo Chen <haibo.chen@nxp.com>
Reviewed-by: Andy Duan <fuguang.duan@nxp.com>
(cherry picked from commit 2a40d8123aff4b4fb7a5cbf286d0c308a42c2fc7)

MLK-18004: ARM: imx: pm-imx7ulp: fix resume failure in freeze mode

This patch fix resume failure in freeze suspend mode on i.mx7ULP
("echo freeze > /sys/power/state") while pressing onoff key or
enabling rtc alarm wakeup. In freeze mode, kernel can only be woken
up by drivers which register wakup source such as 'device_init_wakeup'
or 'irq_set_irq_wake', otherwise, kernel will wait for irq handler
freeze_wake(). Unfortunately, our NMI interrupt which used to wakeup
A7 by M4 is not a common device and request irq as 'IRQF_NO_SUSPEND'
which means feeze_wake() never get chance to run while wakeup by any
event from M4 such as RTC, ONOFF. In this case, use pm_system_wakeup()
instead in NMI interrupt handle to trigger freeze_wake() directly.

Signed-off-by: Robin Gong <yibin.gong@nxp.com>
Reviewed-by: Anson Huang <Anson.Huang@nxp.com>

MLK-17979: ASoC: rpmsg_wm8960: add mixer control for ADC input

The microphone only connect to left input, when record stereo channel
data, the right channel is mute. Add 'ADC Data Output Select' mixer
control that user can select the wanted configure. The default setting
is 'Left Data = Left ADC;  Right Data = Left ADC'.

Signed-off-by: Shengjiu Wang <shengjiu.wang@nxp.com>
(cherry picked from commit cce63c3e843b7d705df6e36adffc0226bfe40e42)

MLK-17971 clk: imx: fix pll set rate failure issue on imx7ulp

The logic of 'if' check for the mult is wrong, this will lead
to set rate to PLL type failed. Additionally, remove the
unnecessary 'CLK_IS_CRITICAL' flags.

Signed-off-by: Bai Ping <ping.bai@nxp.com>
Reviewed-by: Anson Huang <Anson.Huang@nxp.com>
(cherry picked from commit a67aa226b9d0d294b51cfc43371fe78a005dfae4)

MGS-3786 [#ccc] Cncrease the clock rate of GPU3D/GPU2D for 7ulp B0 board

The gpu3d/2d clock rate for 7ulp B0 board is 400M, increase it

Signed-off-by: yuchou gan <yuchou.gan@nxp.com>
(cherry picked from commit b51ae7e98ccfd9e25697d3e5b9795699917496ea)

MGS-3778 [#imx-989] fix GPU performance regression with 6.2.4.p1

the original patch will skip CMA memory allocation with CMA_LIMIT flag,
that enforces GPU memory allocation from virtual pool with MMU mapping,
then both 2D and 3D performance will have performance regression on i.MX6.

Revert "6.2.4.p1-0044-CL142820-check-flag-match-even-try-to-allocate-from-"

This reverts commit 8a8cbf389ad56dc49685ea078698087be867655a.

Signed-off-by: Xianzhong <xianzhong.li@nxp.com>
(cherry picked from commit 951e42c0ec05d4cdaf739eadd5fd40c2b8321b10)

MLK-17944: ARM: dts: imx7ulp-evk: add poweron key

Add poweron key support on i.mx7ulp-evk board since M4 take
over snvs on B0 chip.

Signed-off-by: Robin Gong <yibin.gong@nxp.com>
Reviewed-by: Anson Huang <Anson.Huang@nxp.com>

MLK-17961 dts: imx7ulp-evk: add non-removable property for wifi sdio

Add non-removable property for usdhc1 that is used as Murata
1PJ wifi sdio interface, which means wifi card always is present.

Signed-off-by: Shenwei Wang <shenwei.wang@nxp.com>
Signed-off-by: Fugang Duan <fugang.duan@nxp.com>
Tested-by: Fugang Duan <fugang.duan@nxp.com>

MLK-17779 input: egalax_ts: free irq resource before request the line as GPIO

If GPIO is connected to an IRQ then it should not request it as
GPIO function only when free its IRQ resouce.

Tested-by: Haibo Chen <haibo.chen@nxp.com>
Signed-off-by: Fugang Duan <fugang.duan@nxp.com>
Signed-off-by: Anson Huang <Anson.Huang@nxp.com>
Signed-off-by: Robin Gong <yibin.gong@nxp.com

MLK-17924 gpu: imx: imx8_dprc: Do not set FRAME_2P_PIX_X/Y_CTRL for updated IP

We've got some fixups for DPR IP in the new i.MX8QXP silicon.
To address the cropping issue(TKT344978), the new IP changes the
FRAME_2P_PIX_X/Y_CTRL(@F0h and @100h) register definitions to be
FRAME_PIX_X/Y_ULC_CTRL.  Thus, we should not set the two registers
for the new IP.  FRAME_PIX_X/Y_ULC_CTRL will be programmed after
we figure out how to use them to do fb x/y offset for tile formats.

Signed-off-by: Liu Ying <victor.liu@nxp.com>

MLK-17923 drm/imx: dpu: plane: Do not support fb x/y src offset for tile fmts

We don't have correct support for fb x/y source offset for tile formats.
The buffer address calculation is wrong when the offset is non-zero.
Also, finer offset needs a fix in silicon(TKT344978).  So, let's do not
support the offset currently.  We may add it back after we figure out
how the updated silicon supports the offset.

Signed-off-by: Liu Ying <victor.liu@nxp.com>

arm64: Kill PSCI_GET_VERSION as a variant-2 workaround

commit 3a0a397ff5ff upstream.

Now that we've standardised on SMCCC v1.1 to perform the branch
prediction invalidation, let's drop the previous band-aid.
If vendors haven't updated their firmware to do SMCCC 1.1, they
haven't updated PSCI either, so we don't loose anything.

Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
no falkor/thunderx2/vulcan in arch/arm64/kernel/cpu_errata.c

arm64: Add ARM_SMCCC_ARCH_WORKAROUND_1 BP hardening support

commit b092201e0020 upstream.

Add the detection and runtime code for ARM_SMCCC_ARCH_WORKAROUND_1.
It is lovely. Really.

Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
no qcom hyp functions in
arch/arm64/kernel/bpi.S
arch/arm64/kernel/cpu_errata.c

arm/arm64: smccc: Implement SMCCC v1.1 inline primitive

commit f2d3b2e8759a upstream.

One of the major improvement of SMCCC v1.1 is that it only clobbers
the first 4 registers, both on 32 and 64bit. This means that it
becomes very easy to provide an inline version of the SMC call
primitive, and avoid performing a function call to stash the
registers that would otherwise be clobbered by SMCCC v1.0.

Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm/arm64: smccc: Make function identifiers an unsigned quantity

commit ded4c39e93f3 upstream.

Function identifiers are a 32bit, unsigned quantity. But we never
tell so to the compiler, resulting in the following:

 4ac:   b26187e0        mov     x0, #0xffffffff80000001

We thus rely on the firmware narrowing it for us, which is not
always a reasonable expectation.

Cc: stable@vger.kernel.org
Reported-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Acked-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

firmware/psci: Expose SMCCC version through psci_ops

commit e78eef554a91 upstream.

Since PSCI 1.0 allows the SMCCC version to be (indirectly) probed,
let's do that at boot time, and expose the version of the calling
convention as part of the psci_ops structure.

Acked-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

firmware/psci: Expose PSCI conduit

commit 09a8d6d48499 upstream.

In order to call into the firmware to apply workarounds, it is
useful to find out whether we're using HVC or SMC. Let's expose
this through the psci_ops.

Acked-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: KVM: Add SMCCC_ARCH_WORKAROUND_1 fast handling

commit f72af90c3783 upstream.

We want SMCCC_ARCH_WORKAROUND_1 to be fast. As fast as possible.
So let's intercept it as early as we can by testing for the
function call number as soon as we've identified a HVC call
coming from the guest.

Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: KVM: Report SMCCC_ARCH_WORKAROUND_1 BP hardening support

commit 6167ec5c9145 upstream.

A new feature of SMCCC 1.1 is that it offers firmware-based CPU
workarounds. In particular, SMCCC_ARCH_WORKAROUND_1 provides
BP hardening for CVE-2017-5715.

If the host has some mitigation for this issue, report that
we deal with it using SMCCC_ARCH_WORKAROUND_1, as we apply the
host workaround on every guest exit.

Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
no sve support in arch/arm64/include/asm/kvm_host.h
mv changes from virt/kvm/arm/psci.c to arch/arm/kvm/psci.c
using cpus_have_cap instead of cpus_have_const_cap

arm/arm64: KVM: Turn kvm_psci_version into a static inline

commit a4097b351118 upstream.

We're about to need kvm_psci_version in HYP too. So let's turn it
into a static inline, and pass the kvm structure as a second
parameter (so that HYP can do a kern_hyp_va on it).

Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
mv changes from virt/kvm/arm/psci.c to arch/arm/kvm/psci.c

arm64: KVM: Make PSCI_VERSION a fast path

commit 90348689d500 upstream.

For those CPUs that require PSCI to perform a BP invalidation,
going all the way to the PSCI code for not much is a waste of
precious cycles. Let's terminate that call as early as possible.

Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm/arm64: KVM: Advertise SMCCC v1.1

commit 09e6be12effd upstream.

The new SMC Calling Convention (v1.1) allows for a reduced overhead
when calling into the firmware, and provides a new feature discovery
mechanism.

Make it visible to KVM guests.

Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
mv change from virt/kvm/arm/psci.c to arch/arm/kvm/psci.c

arm/arm64: KVM: Implement PSCI 1.0 support

commit 58e0b2239a4d upstream.

PSCI 1.0 can be trivially implemented by providing the FEATURES
call on top of PSCI 0.2 and returning 1.0 as the PSCI version.

We happily ignore everything else, as they are either optional or
are clarifications that do not require any additional change.

PSCI 1.0 is now the default until we decide to add a userspace
selection API.

Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
mv chagnes from virt/kvm/arm/psci.c to arch/arm/kvm/psci.c

arm/arm64: KVM: Add smccc accessors to PSCI code

commit 84684fecd7ea upstream.

Instead of open coding the accesses to the various registers,
let's add explicit SMCCC accessors.

Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
mv change from virt/kvm/arm/psci.c to arch/arm/kvm/psci.c

arm/arm64: KVM: Add PSCI_VERSION helper

commit d0a144f12a7c upstream.

As we're about to trigger a PSCI version explosion, it doesn't
hurt to introduce a PSCI_VERSION helper that is going to be
used everywhere.

Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
mv change form virt/kvm/arm/psci.c to arch/arm/kvm/psci.c

arm/arm64: KVM: Consolidate the PSCI include files

commit 1a2fb94e6a77 upstream.

As we're about to update the PSCI support, and because I'm lazy,
let's move the PSCI include file to include/kvm so that both
ARM architectures can find it.

Acked-by: Christoffer Dall <christoffer.dall@linaro.org>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
need kvm/arm_psci.h in files:
arch/arm64/kvm/handle_exit.c
arch/arm/kvm/psci.c and arch/arm/kvm/arm.c
no virt/kvm/arm/arm.c and virt/kvm/arm/psci.c

arm64: KVM: Increment PC after handling an SMC trap

commit f5115e8869e1 upstream.

When handling an SMC trap, the "preferred return address" is set
to that of the SMC, and not the next PC (which is a departure from
the behaviour of an SMC that isn't trapped).

Increment PC in the handler, as the guest is otherwise forever
stuck...

Cc: stable@vger.kernel.org
Fixes: acfb3b883f6d ("arm64: KVM: Fix SMCCC handling of unimplemented SMC/HVC calls")
Reviewed-by: Christoffer Dall <christoffer.dall@linaro.org>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: Implement branch predictor hardening for affected Cortex-A CPUs

commit aa6acde65e03 upstream.

Cortex-A57, A72, A73 and A75 are susceptible to branch predictor aliasing
and can theoretically be attacked by malicious code.

This patch implements a PSCI-based mitigation for these CPUs when available.
The call into firmware will invalidate the branch predictor state, preventing
any malicious entries from affecting other victim contexts.

Co-developed-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
no falkor in arch/arm64/kernel/cpu_errata.c

arm64: entry: Apply BP hardening for suspicious interrupts from EL0

commit 30d88c0e3ace upstream.

It is possible to take an IRQ from EL0 following a branch to a kernel
address in such a way that the IRQ is prioritised over the instruction
abort. Whilst an attacker would need to get the stars to align here,
it might be sufficient with enough calibration so perform BP hardening
in the rare case that we see a kernel address in the ELR when handling
an IRQ from EL0.

Reported-by: Dan Hettena <dhettena@nvidia.com>
Reviewed-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: entry: Apply BP hardening for high-priority synchronous exceptions

commit 5dfc6ed27710 upstream.

Software-step and PC alignment fault exceptions have higher priority than
instruction abort exceptions, so apply the BP hardening hooks there too
if the user PC appears to reside in kernel space.

Reported-by: Dan Hettena <dhettena@nvidia.com>
Reviewed-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
expand enable_da_f to 'msr daifclr, #(8 | 4 | 1)'
in arch/arm64/kernel/entry.S

arm64: KVM: Use per-CPU vector when BP hardening is enabled

commit 6840bdd73d07 upstream

Now that we have per-CPU vectors, let's plug then in the KVM/arm64 code.

Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
mv changes from virt/kvm/arm/arm.c to arch/arm/kvm/arm.c

arm64: Move BP hardening to check_and_switch_context

commit a8e4c0a919ae upstream.

We call arm64_apply_bp_hardening() from post_ttbr_update_workaround,
which has the unexpected consequence of being triggered on every
exception return to userspace when ARM64_SW_TTBR0_PAN is selected,
even if no context switch actually occured.

This is a bit suboptimal, and it would be more logical to only
invalidate the branch predictor when we actually switch to
a different mm.

In order to solve this, move the call to arm64_apply_bp_hardening()
into check_and_switch_context(), where we're guaranteed to pick
a different mm context.

Acked-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
no sw pan in arch/arm64/mm/context.c

arm64: Add skeleton to harden the branch predictor against aliasing attacks

commit 0f15adbb2861 upstream.

Aliasing attacks against CPU branch predictors can allow an attacker to
redirect speculative control flow on some CPUs and potentially divulge
information from one context to another.

This patch adds initial skeleton code behind a new Kconfig option to
enable implementation-specific mitigations against these attacks for
CPUs that are affected.

Co-developed-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
expand enable_da_f in entry.S
use 5 parameters ARM64_FTR_BITS()
add percpu.h in mm_types.h for percpu functions
use cpus_have_cap instead of cpus_have_const_cap
arch/arm64/Kconfig
arch/arm64/include/asm/cpucaps.h
arch/arm64/include/asm/mmu.h
arch/arm64/include/asm/sysreg.h
arch/arm64/kernel/cpufeature.c
arch/arm64/kernel/entry.S
arch/arm64/mm/fault.c

arm64: Move post_ttbr_update_workaround to C code

commit 95e3de3590e3 upstream.

We will soon need to invoke a CPU-specific function pointer after changing
page tables, so move post_ttbr_update_workaround out into C code to make
this possible.

Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
don't include PAN related changes
arch/arm64/include/asm/assembler.h
arch/arm64/kernel/entry.S
arch/arm64/mm/proc.S

arm64: cpufeature: Pass capability structure to ->enable callback

commit 0a0d111d40fd upstream.

In order to invoke the CPU capability ->matches callback from the ->enable
callback for applying local-CPU workarounds, we need a handle on the
capability structure.

This patch passes a pointer to the capability structure to the ->enable
callback.

Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
arch/arm64/kernel/cpufeature.c

arm64: Run enable method for errata work arounds on late CPUs

commit 55b35d070c25 upstream.

When a CPU is brought up after we have finalised the system
wide capabilities (i.e, features and errata), we make sure the
new CPU doesn't need a new errata work around which has not been
detected already. However we don't run enable() method on the new
CPU for the errata work arounds already detected. This could
cause the new CPU running without potential work arounds.
It is upto the "enable()" method to decide if this CPU should
do something about the errata.

Fixes: commit 6a6efbb45b7d95c84 ("arm64: Verify CPU errata work arounds on hotplugged CPU")
Cc: Will Deacon <will.deacon@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Andre Przywara <andre.przywara@arm.com>
Cc: Dave Martin <dave.martin@arm.com>
Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: cpu_errata: Allow an erratum to be match for all revisions of a core

commit 06f1494f837 upstream.

Some minor erratum may not be fixed in further revisions of a core,
leading to a situation where the workaround needs to be updated each
time an updated core is released.

Introduce a MIDR_ALL_VERSIONS match helper that will work for all
versions of that MIDR, once and for all.

Acked-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: cpufeature: __this_cpu_has_cap() shouldn't stop early

commit edf298cfce47 upstream.
Alex Shi rewrite this commit on func this_cpu_has_cap(). The following commit
log is still meaningful.

    this_cpu_has_cap() tests caps->desc not caps->matches, so it stops
    walking the list when it finds a 'silent' feature, instead of
    walking to the end of the list.

    Prior to v4.6's 644c2ae198412 ("arm64: cpufeature: Test 'matches' pointer
    to find the end of the list") we always tested desc to find the end of
    a capability list. This was changed for dubious things like PAN_NOT_UAO.
    v4.7's e3661b128e53e ("arm64: Allow a capability to be checked on
    single CPU") added this_cpu_has_cap() using the old desc style test.

CC: Suzuki K Poulose <suzuki.poulose@arm.com>
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Acked-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: James Morse <james.morse@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

drivers/firmware: Expose psci_get_version through psci_ops structure

commit d68e3ba5303f upstream.

Entry into recent versions of ARM Trusted Firmware will invalidate the CPU
branch predictor state in order to protect against aliasing attacks.

This patch exposes the PSCI "VERSION" function via psci_ops, so that it
can be invoked outside of the PSCI driver where necessary.

Acked-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: futex: Mask __user pointers prior to dereference

commit 91b2d3442f6a upstream.

The arm64 futex code has some explicit dereferencing of user pointers
where performing atomic operations in response to a futex command. This
patch uses masking to limit any speculative futex operations to within
the user address space.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
change on old futex_atomic_op_inuser function instead of
arch_futex_atomic_op_inuser in arch/arm64/include/asm/futex.h

arm64: uaccess: Mask __user pointers for __arch_{clear, copy_*}_user

Rewritting from commit f71c2ffcb20d upstream. On LTS 4.9, there has no
raw_copy_from/to_user, neither __copy_user_flushcache, and it isn't good
idead to pick up them. The following is origin commit log, that's also
applicable for the new patch.

    Like we've done for get_user and put_user, ensure that user pointers
    are masked before invoking the underlying __arch_{clear,copy_*}_user
    operations.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: uaccess: Don't bother eliding access_ok checks in __{get, put}_user

commit 84624087dd7e upstream.

access_ok isn't an expensive operation once the addr_limit for the current
thread has been loaded into the cache. Given that the initial access_ok
check preceding a sequence of __{get,put}_user operations will take
the brunt of the miss, we can make the __* variants identical to the
full-fat versions, which brings with it the benefits of address masking.

The likely cost in these sequences will be from toggling PAN/UAO, which
we can address later by implementing the *_unsafe versions.

Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
keep __{get/put}_user_unaligned in arch/arm64/include/asm/uaccess.h

arm64: uaccess: Prevent speculative use of the current addr_limit

commit c2f0ad4fc089 upstream.

A mispredicted conditional call to set_fs could result in the wrong
addr_limit being forwarded under speculation to a subsequent access_ok
check, potentially forming part of a spectre-v1 attack using uaccess
routines.

This patch prevents this forwarding from taking place, but putting heavy
barriers in set_fs after writing the addr_limit.

Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
no set_thread_flag(TIF_FSCHECK) in arch/arm64/include/asm/uaccess.h

arm64: entry: Ensure branch through syscall table is bounded under speculation

commit 6314d90e6493 upstream.

In a similar manner to array_index_mask_nospec, this patch introduces an
assembly macro (mask_nospec64) which can be used to bound a value under
speculation. This macro is then used to ensure that the indirect branch
through the syscall table is bounded under speculation, with out-of-range
addresses speculating as calls to sys_io_setup (0).

Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: syscallno is secretly an int, make it official

commit 35d0e6fb4d upstream.

The upper 32 bits of the syscallno field in thread_struct are
handled inconsistently, being sometimes zero extended and sometimes
sign-extended.  In fact, only the lower 32 bits seem to have any
real significance for the behaviour of the code: it's been OK to
handle the upper bits inconsistently because they don't matter.

Currently, the only place I can find where those bits are
significant is in calling trace_sys_enter(), which may be
unintentional: for example, if a compat tracer attempts to cancel a
syscall by passing -1 to (COMPAT_)PTRACE_SET_SYSCALL at the
syscall-enter-stop, it will be traced as syscall 4294967295
rather than -1 as might be expected (and as occurs for a native
tracer doing the same thing).  Elsewhere, reads of syscallno cast
it to an int or truncate it.

There's also a conspicuous amount of code and casting to bodge
around the fact that although semantically an int, syscallno is
stored as a u64.

Let's not pretend any more.

In order to preserve the stp x instruction that stores the syscall
number in entry.S, this patch special-cases the layout of struct
pt_regs for big endian so that the newly 32-bit syscallno field
maps onto the low bits of the stored value.  This is not beautiful,
but benchmarking of the getpid syscall on Juno suggests indicates a
minor slowdown if the stp is split into an stp x and stp w.

Signed-off-by: Dave Martin <Dave.Martin@arm.com>
Acked-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: Use pointer masking to limit uaccess speculation

commit 4d8efc2d5ee4 upstream.

Similarly to x86, mitigate speculation past an access_ok() check by
masking the pointer against the address limit before use.

Even if we don't expect speculative writes per se, it is plausible that
a CPU may still speculate at least as far as fetching a cache line for
writing, hence we also harden put_user() and clear_user() for peace of
mind.

Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>

arm64: Make USER_DS an inclusive limit

commit 51369e398d0d upstream.

Currently, USER_DS represents an exclusive limit while KERNEL_DS is
inclusive. In order to do some clever trickery for speculation-safe
masking, we need them both to behave equivalently - there aren't enough
bits to make KERNEL_DS exclusive, so we have precisely one option. This
also happens to correct a longstanding false negative for a range
ending on the very top byte of kernel memory.

Mark Rutland points out that we've actually got the semantics of
addresses vs. segments muddled up in most of the places we need to
amend, so shuffle the {USER,KERNEL}_DS definitions around such that we
can correct those properly instead of just pasting "-1"s everywhere.

Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
(cherry picked from commit 83b20dff71ea949431cf57c6aebaaf7ebd5c1991)
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
force replace __range_ok and add asm/processor.h
in arch/arm64/include/asm/uaccess.h
using old macro TI_ADDR_LIMIT instead of TSK_TI_ADDR_LIMIT
in arch/arm64/kernel/entry.S
manual change USER_DS to TASK_SIZE in arch/arm64/mm/fault.c

arm64: uaccess: consistently check object sizes

commit 76624175dca upstream.

Currently in arm64's copy_{to,from}_user, we only check the
source/destination object size if access_ok() tells us the user access
is permissible.

However, in copy_from_user() we'll subsequently zero any remainder on
the destination object. If we failed the access_ok() check, that applies
to the whole object size, which we didn't check.

To ensure that we catch that case, this patch hoists check_object_size()
to the start of copy_from_user(), matching __copy_from_user() and
__copy_to_user(). To make all of our uaccess copy primitives consistent,
the same is done to copy_to_user().

Cc: Catalin Marinas <catalin.marinas@arm.com>
Acked-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: Factor out TTBR0_EL1 post-update workaround into a specific asm macro

commit f33bcf03e6 upstream

This patch takes the errata workaround code out of cpu_do_switch_mm into
a dedicated post_ttbr0_update_workaround macro which will be reused in a
subsequent patch.

Cc: Will Deacon <will.deacon@arm.com>
Cc: James Morse <james.morse@arm.com>
Cc: Kees Cook <keescook@chromium.org>
Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: Factor out PAN enabling/disabling into separate uaccess_* macros

commit bd38967d406 upstream.

This patch moves the directly coded alternatives for turning PAN on/off
into separate uaccess_{enable,disable} macros or functions. The asm
macros take a few arguments which will be used in subsequent patches.

Note that any (unlikely) access that the compiler might generate between
uaccess_enable() and uaccess_disable(), other than those explicitly
specified by the user access code, will not be protected by PAN.

Cc: Will Deacon <will.deacon@arm.com>
Cc: James Morse <james.morse@arm.com>
Cc: Kees Cook <keescook@chromium.org>
Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: move TASK_* definitions to <asm/processor.h>

commit eef94a3d09aab upstream.

ILP32 series [1] introduces the dependency on <asm/is_compat.h> for
TASK_SIZE macro. Which in turn requires <asm/thread_info.h>, and
<asm/thread_info.h> include <asm/memory.h>, giving a circular dependency,
because TASK_SIZE is currently located in <asm/memory.h>.

In other architectures, TASK_SIZE is defined in <asm/processor.h>, and
moving TASK_SIZE there fixes the problem.

Discussion: https://patchwork.kernel.org/patch/9929107/

[1] https://github.com/norov/linux/tree/ilp32-next

CC: Will Deacon <will.deacon@arm.com>
CC: Laura Abbott <labbott@redhat.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: James Morse <james.morse@arm.com>
Suggested-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Yury Norov <ynorov@caviumnetworks.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
no ptrace.h in arch/arm64/kernel/entry.S

arm64: Implement array_index_mask_nospec()

commit 022620eed3d0 upstream.

Provide an optimised, assembly implementation of array_index_mask_nospec()
for arm64 so that the compiler is not in a position to transform the code
in ways which affect its ability to inhibit speculation (e.g. by introducing
conditional branches).

This is similar to the sequence used by x86, modulo architectural differences
in the carry/borrow flags.

Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: barrier: Add CSDB macros to control data-value prediction

commit 669474e772b9 upstream.

For CPUs capable of data value prediction, CSDB waits for any outstanding
predictions to architecturally resolve before allowing speculative execution
to continue. Provide macros to expose it to the arch code.

Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
arch/arm64/include/asm/assembler.h
no psb_csync in arch/arm64/include/asm/barrier.h

arm64: alternatives: apply boot time fixups via the linear mapping

commit 5ea5306c323 upstream.

One important rule of thumb when desiging a secure software system is
that memory should never be writable and executable at the same time.
We mostly adhere to this rule in the kernel, except at boot time, when
regions may be mapped RWX until after we are done applying alternatives
or making other one-off changes.

For the alternative patching, we can improve the situation by applying
the fixups via the linear mapping, which is never mapped with executable
permissions. So map the linear alias of .text with RW- permissions
initially, and remove the write permissions as soon as alternative
patching has completed.

Reviewed-by: Laura Abbott <labbott@redhat.com>
Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
replace update_mapping_prot with old create_mapping_late
arch/arm64/mm/mmu.c

mm: Introduce lm_alias

commit 568c5fe5a54 upstream.

Certain architectures may have the kernel image mapped separately to
alias the linear map. Introduce a macro lm_alias to translate a kernel
image symbol into its linear alias. This is used in part with work to
add CONFIG_DEBUG_VIRTUAL support for arm64.

Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Laura Abbott <labbott@redhat.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: idmap: Use "awx" flags for .idmap.text .pushsection directives

commit 439e70e27a51 upstream.

The identity map is mapped as both writeable and executable by the
SWAPPER_MM_MMUFLAGS and this is relied upon by the kpti code to manage
a synchronisation flag. Update the .pushsection flags to reflect the
actual mapping attributes.

Reported-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: entry: Reword comment about post_ttbr_update_workaround

commit f167211a93ac upstream.

We don't fully understand the Cavium ThunderX erratum, but it appears
that mapping the kernel as nG can lead to horrible consequences such as
attempting to execute userspace from kernel context. Since kpti isn't
enabled for these CPUs anyway, simplify the comment justifying the lack
of post_ttbr_update_workaround in the exception trampoline.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: Force KPTI to be disabled on Cavium ThunderX

commit 6dc52b15c4a4 upstream.

Cavium ThunderX's erratum 27456 results in a corruption of icache
entries that are loaded from memory that is mapped as non-global
(i.e. ASID-tagged).

As KPTI is based on memory being mapped non-global, let's prevent
it from kicking in if this erratum is detected.

Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
[will: Update comment]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
using old function read_system_reg/cpus_have_cap to replace
read_sanitised_ftr_reg/cpus_have_const_cap in
arch/arm64/kernel/cpufeature.c

arm64: kpti: Add ->enable callback to remap swapper using nG mappings

commit f992b4dfd58b upstream.

Defaulting to global mappings for kernel space is generally good for
performance and appears to be necessary for Cavium ThunderX. If we
subsequently decide that we need to enable kpti, then we need to rewrite
our existing page table entries to be non-global. This is fiddly, and
made worse by the possible use of contiguous mappings, which require
a strict break-before-make sequence.

Since the enable callback runs on each online CPU from stop_machine
context, we can have all CPUs enter the idmap, where secondaries can
wait for the primary CPU to rewrite swapper with its MMU off. It's all
fairly horrible, but at least it only runs once.

Nicolas Dechesne <nicolas.dechesne@linaro.org> found a bug on this commit
which cause boot failure on db410c etc board. Ard Biesheuvel found it
writting wrong contenct to ttbr1_el1 in __idmap_cpu_set_reserved_ttbr1
macro and fixed it by give it the right content.

Tested-by: Marc Zyngier <marc.zyngier@arm.com>
Reviewed-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
no get_thread_info/post_ttbr_update_workaround/pre_disable_mmu_workaround
in arch/arm64/include/asm/assembler.h and arch/arm64/mm/proc.S

arm64: kpti: Make use of nG dependent on arm64_kernel_unmapped_at_el0()

commit 41acec624087 upstream.

To allow systems which do not require kpti to continue running with
global kernel mappings (which appears to be a requirement for Cavium
ThunderX due to a CPU erratum), make the use of nG in the kernel page
tables dependent on arm64_kernel_unmapped_at_el0(), which is resolved
at runtime.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: Turn on KPTI only on CPUs that need it

commit 0ba2e29c7fc1 upstream.

Whitelist Broadcom Vulcan/Cavium ThunderX2 processors in
unmap_kernel_at_el0(). These CPUs are not vulnerable to
CVE-2017-5754 and do not need KPTI when KASLR is off.

Acked-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Jayachandran C <jnair@caviumnetworks.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: cputype: Add MIDR values for Cavium ThunderX2 CPUs

commit 0d90718871fe upstream.

Add the older Broadcom ID as well as the new Cavium ID for ThunderX2
CPUs.

Signed-off-by: Jayachandran C <jnair@caviumnetworks.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
no falkor support in arch/arm64/include/asm/cputype.h

arm64: cputype: Add missing MIDR values for Cortex-A72 and Cortex-A75

commit f0be3364335d47267aa1f7c5ed5faaa59c70db13 upstream

Hook up MIDR values for the Cortex-A72 and Cortex-A75 CPUs, since they
will soon need MIDR matches for hardening the branch predictor.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
add A73 type in arch/arm64/include/asm/cputype.h

arm64: capabilities: Handle duplicate entries for a capability

commit 67948af41f2e upstream.

Sometimes a single capability could be listed multiple times with
differing matches(), e.g, CPU errata for different MIDR versions.
This breaks verify_local_cpu_feature() and this_cpu_has_cap() as
we stop checking for a capability on a CPU with the first
entry in the given table, which is not sufficient. Make sure we
run the checks for all entries of the same capability. We do
this by fixing __this_cpu_has_cap() to run through all the
entries in the given table for a match and reuse it for
verify_local_cpu_feature().

Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Acked-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
arch/arm64/kernel/cpufeature.c

arm64: Allow checking of a CPU-local erratum

commit 8f4137588261d7504f4aa022dc9d1a1fd1940e8e upstream.

this_cpu_has_cap() only checks the feature array, and not the errata
one. In order to be able to check for a CPU-local erratum, allow it
to inspect the latter as well.

This is consistent with cpus_have_cap()'s behaviour, which includes
errata already.

Acked-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: Take into account ID_AA64PFR0_EL1.CSV3

commit 179a56f6f9fb upstream.

For non-KASLR kernels where the KPTI behaviour has not been overridden
on the command line we can use ID_AA64PFR0_EL1.CSV3 to determine whether
or not we should unmap the kernel whilst running at EL0.

Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
skip cpu features like SVE etc.
and use 5 paramaters function ARM64_FTR_BITS()
replace read_sanitised_ftr_reg with old name read_system_reg
arch/arm64/include/asm/sysreg.h
arch/arm64/kernel/cpufeature.c

arm64: Kconfig: Reword UNMAP_KERNEL_AT_EL0 kconfig entry

commit 0617052ddde3 upstream.

Although CONFIG_UNMAP_KERNEL_AT_EL0 does make KASLR more robust, it's
actually more useful as a mitigation against speculation attacks that
can leak arbitrary kernel data to userspace through speculation.

Reword the Kconfig help message to reflect this, and make the option
depend on EXPERT so that it is on by default for the majority of users.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: Kconfig: Add CONFIG_UNMAP_KERNEL_AT_EL0

commit 084eb77cd3a8 upstream.

Add a Kconfig entry to control use of the entry trampoline, which allows
us to unmap the kernel whilst running in userspace and improve the
robustness of KASLR.

Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Laura Abbott <labbott@redhat.com>
Tested-by: Shanker Donthineni <shankerd@codeaurora.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: use RET instruction for exiting the trampoline

commit be04a6d1126b upstream.

Speculation attacks against the entry trampoline can potentially resteer
the speculative instruction stream through the indirect branch and into
arbitrary gadgets within the kernel.

This patch defends against these attacks by forcing a misprediction
through the return stack: a dummy BL instruction loads an entry into
the stack, so that the predicted program flow of the subsequent RET
instruction is to a branch-to-self instruction which is finally resolved
as a branch to the kernel vectors with speculation suppressed.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: kaslr: Put kernel vectors address in separate data page

commit 6c27c4082f4f upstream.

The literal pool entry for identifying the vectors base is the only piece
of information in the trampoline page that identifies the true location
of the kernel.

This patch moves it into a page-aligned region of the .rodata section
and maps this adjacent to the trampoline text via an additional fixmap
entry, which protects against any accidental leakage of the trampoline
contents.

Suggested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Tested-by: Laura Abbott <labbott@redhat.com>
Tested-by: Shanker Donthineni <shankerd@codeaurora.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
remove ARM64_WORKAROUND_QCOM_FALKOR_E1003 fix
in arch/arm64/kernel/entry.S

arm64: entry: Add fake CPU feature for unmapping the kernel at EL0

commit ea1e3de85e94 upstream.

Allow explicit disabling of the entry trampoline on the kernel command
line (kpti=off) by adding a fake CPU feature (ARM64_UNMAP_KERNEL_AT_EL0)
that can be used to toggle the alternative sequences in our entry code and
avoid use of the trampoline altogether if desired. This also allows us to
make use of a static key in arm64_kernel_unmapped_at_el0().

Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Laura Abbott <labbott@redhat.com>
Tested-by: Shanker Donthineni <shankerd@codeaurora.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
skip non enabled cpu features in
arch/arm64/include/asm/cpucaps.h and
arch/arm64/kernel/cpufeature.c
using cpus_have_cap instead of cpus_have_const_cap in
arch/arm64/include/asm/mmu.h

arm64: tls: Avoid unconditional zeroing of tpidrro_el0 for native tasks

commit 18011eac28c7 upstream.

When unmapping the kernel at EL0, we use tpidrro_el0 as a scratch register
during exception entry from native tasks and subsequently zero it in
the kernel_ventry macro. We can therefore avoid zeroing tpidrro_el0
in the context-switch path for native tasks using the entry trampoline.

Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Laura Abbott <labbott@redhat.com>
Tested-by: Shanker Donthineni <shankerd@codeaurora.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
fold tls_preserve_current_state() in arch/arm64/kernel/process.c

arm64: entry: Hook up entry trampoline to exception vectors

commit 4bf3286d29f3 upstream.

Hook up the entry trampoline to our exception vectors so that all
exceptions from and returns to EL0 go via the trampoline, which swizzles
the vector base register accordingly. Transitioning to and from the
kernel clobbers x30, so we use tpidrro_el0 and far_el1 as scratch
registers for native tasks.

Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Laura Abbott <labbott@redhat.com>
Tested-by: Shanker Donthineni <shankerd@codeaurora.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: entry: Explicitly pass exception level to kernel_ventry macro

commit 5b1f7fe41909 upstream.

We will need to treat exceptions from EL0 differently in kernel_ventry,
so rework the macro to take the exception level as an argument and
construct the branch target using that.

Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Laura Abbott <labbott@redhat.com>
Tested-by: Shanker Donthineni <shankerd@codeaurora.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
no vmap_stack in arch/arm64/kernel/entry.S

arm64: mm: Map entry trampoline into trampoline and kernel page tables

commit 51a0048beb44 upstream.

The exception entry trampoline needs to be mapped at the same virtual
address in both the trampoline page table (which maps nothing else)
and also the kernel page table, so that we can swizzle TTBR1_EL1 on
exceptions from and return to EL0.

This patch maps the trampoline at a fixed virtual address in the fixmap
area of the kernel virtual address space, which allows the kernel proper
to be randomized with respect to the trampoline when KASLR is enabled.

Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Laura Abbott <labbott@redhat.com>
Tested-by: Shanker Donthineni <shankerd@codeaurora.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
no acpi apei in arch/arm64/include/asm/fixmap.h
no rodata in arch/arm64/mm/mmu.c

arm64: entry: Add exception trampoline page for exceptions from EL0

commit c7b9adaf85f8 upstream.

To allow unmapping of the kernel whilst running at EL0, we need to
point the exception vectors at an entry trampoline that can map/unmap
the kernel on entry/exit respectively.

This patch adds the trampoline page, although it is not yet plugged
into the vector table and is therefore unused.

Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Laura Abbott <labbott@redhat.com>
Tested-by: Shanker Donthineni <shankerd@codeaurora.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
add asm/mmu.h in entry.S for ASID marco
add kernel-pgtable.h in entry.S for SWAPPER_DIR_SIZE and
RESERVED_TTBR0_SIZE
no SW PAN in vmlinux.lds.S

module: extend 'rodata=off' boot cmdline parameter to module mappings

commit 39290b389ea upstream.

The current "rodata=off" parameter disables read-only kernel mappings
under CONFIG_DEBUG_RODATA:
    commit d2aa1acad22f ("mm/init: Add 'rodata=off' boot cmdline parameter
    to disable read-only kernel mappings")

This patch is a logical extension to module mappings ie. read-only mappings
at module loading can be disabled even if CONFIG_DEBUG_SET_MODULE_RONX
(mainly for debug use). Please note, however, that it only affects RO/RW
permissions, keeping NX set.

This is the first step to make CONFIG_DEBUG_SET_MODULE_RONX mandatory
(always-on) in the future as CONFIG_DEBUG_RODATA on x86 and arm64.

Suggested-by: and Acked-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Reviewed-by: Kees Cook <keescook@chromium.org>
Acked-by: Rusty Russell <rusty@rustcorp.com.au>
Link: http://lkml.kernel.org/r/20161114061505.15238-1-takahiro.akashi@linaro.org
Signed-off-by: Jessica Yu <jeyu@redhat.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
keeping kaiser.h in init/main.c

arm64: entry.S: move SError handling into a C function for future expansion

commit a92d4d1454ab upstream.

Today SError is taken using the inv_entry macro that ends up in
bad_mode.

SError can be used by the RAS Extensions to notify either the OS or
firmware of CPU problems, some of which may have been corrected.

To allow this handling to be added, add a do_serror() C function
that just panic()s. Add the entry.S boiler plate to save/restore the
CPU registers and unmask debug exceptions. Future patches may change
do_serror() to return if the SError Interrupt was notification of a
corrected error.

Signed-off-by: Xie XiuQi <xiexiuqi@huawei.com>
Signed-off-by: Wang Xiongfeng <wangxiongfengi2@huawei.com>
[Split out of a bigger patch, added compat path, renamed, enabled debug
 exceptions]
Signed-off-by: James Morse <james.morse@arm.com>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
no vmap_stack in arch/arm64/kernel/traps.c
using old enable_dbg_and_irq instead of enable_daif in
arch/arm64/kernel/entry.S

arm64: factor out entry stack manipulation

commit b11e5759bfac upstream.

In subsequent patches, we will detect stack overflow in our exception
entry code, by verifying the SP after it has been decremented to make
space for the exception regs.

This verification code is small, and we can minimize its impact by
placing it directly in the vectors. To avoid redundant modification of
the SP, we also need to move the initial decrement of the SP into the
vectors.

As a preparatory step, this patch introduces kernel_ventry, which
performs this decrement, and updates the entry code accordingly.
Subsequent patches will fold SP verification into kernel_ventry.

There should be no functional change as a result of this patch.

Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
[Mark: turn into prep patch, expand commit msg]
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Reviewed-by: Will Deacon <will.deacon@arm.com>
Tested-by: Laura Abbott <labbott@redhat.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: James Morse <james.morse@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: mm: Invalidate both kernel and user ASIDs when performing TLBI

commit 9b0de864b5bc upstream.

Since an mm has both a kernel and a user ASID, we need to ensure that
broadcast TLB maintenance targets both address spaces so that things
like CoW continue to work with the uaccess primitives in the kernel.

Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Laura Abbott <labbott@redhat.com>
Tested-by: Shanker Donthineni <shankerd@codeaurora.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: mm: Add arm64_kernel_unmapped_at_el0 helper

commit fc0e1299da54 upstream.

In order for code such as TLB invalidation to operate efficiently when
the decision to map the kernel at EL0 is determined at runtime, this
patch introduces a helper function, arm64_kernel_unmapped_at_el0, to
determine whether or not the kernel is mapped whilst running in userspace.

Currently, this just reports the value of CONFIG_UNMAP_KERNEL_AT_EL0,
but will later be hooked up to a fake CPU capability using a static key.

Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Laura Abbott <labbott@redhat.com>
Tested-by: Shanker Donthineni <shankerd@codeaurora.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>

arm64: mm: Allocate ASIDs in pairs

commit 0c8ea531b774 upstream.

In preparation for separate kernel/user ASIDs, allocate them in pairs
for each mm_struct. The bottom bit distinguishes the two: if it is set,
then the ASID will map only userspace.

Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Laura Abbott <labbott@redhat.com>
Tested-by: Shanker Donthineni <shankerd@codeaurora.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
no MMCF_AARCH32 in arch/arm64/include/asm/mmu.h

arm64: mm: Move ASID from TTBR0 to TTBR1

commit 7655abb95386 upstream.

In preparation for mapping kernelspace and userspace with different
ASIDs, move the ASID to TTBR1 and update switch_mm to context-switch
TTBR0 via an invalid mapping (the zero page).

Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Laura Abbott <labbott@redhat.com>
Tested-by: Shanker Donthineni <shankerd@codeaurora.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
no pre_ttbr0_update_workaround in arch/arm64/mm/proc.S

arm64: mm: Use non-global mappings for kernel space

commit e046eb0c9bf2 upstream.

In preparation for unmapping the kernel whilst running in userspace,
make the kernel mappings non-global so we can avoid expensive TLB
invalidation on kernel exit to userspace.

Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Laura Abbott <labbott@redhat.com>
Tested-by: Shanker Donthineni <shankerd@codeaurora.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Alex Shi <alex.shi@linaro.org>
Conflicts:
skip PTE_RDONLY of PAGE_NONE in arch/arm64/include/asm/pgtable-prot.h