GitLab

This patch updates t194_nvg.h to v6.1 and does not issue NVG
commands for unsupported platforms.

Change-Id: I506b594a70a3651d01a412ab79b3c8919b1d66f1
Signed-off-by: Steven Kao <skao@nvidia.com>

This patch updates the cache ops to use system registers in
order to trigger cache flush/clean operations.

Change-Id: I888abad22f22b8a33c7193b991fad8c4a78030d0
Signed-off-by: Steven Kao <skao@nvidia.com>

Introduce platform handlers to reprogram the MSS settings.

Change-Id: Ibb9a5457d1bad9ecccea619d69a62bed3bf7d861
Signed-off-by: Puneet Saxena <puneets@nvidia.com>
Signed-off-by: Krishna Reddy <vdumpa@nvidia.com>
Signed-off-by: Varun Wadekar <vwadekar@nvidia.com>

This patch updates the total number of CPU clusters and number
of cores per cluster, in the platform makefile.

Change-Id: I569ebc1bb794ecab09a1043511b3d936bf450428
Signed-off-by: Varun Wadekar <vwadekar@nvidia.com>

This patch corrects the TEGRA_CAR_RESET_BASE macro value to
0x20000000 from 0x200000000.

Change-Id: Iba25394ea99237df85395c39059926c5a8b26a84
Signed-off-by: Steven Kao <skao@nvidia.com>

This patch adds masks for the TZDRAM base/size registers.

Change-Id: I5f688793be8cace28d2aa2d177a295e4faffd666
Signed-off-by: Harvey Hsieh <hhsieh@nvidia.com>

This patch updates the wake mask and wake time to indicate to the
mce/mts that the cpu is powering down. Wake time is set to highest
possible value and wake mask is set to zero.

Change-Id: Ic5abf15e7b98f911def6aa610d300b0668cd287e
Signed-off-by: Krishna Sitaraman <ksitaraman@nvidia.com>

T194 XUSB has support for XUSB virtualization. It will have one
physical function (PF) and four Virtual function (VF)

There were below two SIDs for XUSB until T186.
1) #define TEGRA_SID_XUSB_HOST    0x1bU
2) #define TEGRA_SID_XUSB_DEV    0x1cU

We have below four new SIDs added for VF(s)
3) #define TEGRA_SID_XUSB_VF0    0x5dU
4) #define TEGRA_SID_XUSB_VF1    0x5eU
5) #define TEGRA_SID_XUSB_VF2    0x5fU
6) #define TEGRA_SID_XUSB_VF3    0x60U

When virtualization is enabled then we have to disable SID override
and program above SIDs in below newly added SID registers in XUSB
PADCTL MMIO space. These registers are TZ protected and so need to
be done in ATF.
a) #define XUSB_PADCTL_HOST_AXI_STREAMID_PF_0 (0x136cU)
b) #define XUSB_PADCTL_DEV_AXI_STREAMID_PF_0  (0x139cU)
c) #define XUSB_PADCTL_HOST_AXI_STREAMID_VF_0 (0x1370U)
d) #define XUSB_PADCTL_HOST_AXI_STREAMID_VF_1 (0x1374U)
e) #define XUSB_PADCTL_HOST_AXI_STREAMID_VF_2 (0x1378U)
f) #define XUSB_PADCTL_HOST_AXI_STREAMID_VF_3 (0x137cU)

This change disables SID override and programs XUSB SIDs in
above registers to support both virtualization and non-virtualization.

Change-Id: I38213a72999e933c44c5392441f91034d3b47a39
Signed-off-by: Ajay Gupta <ajayg@nvidia.com>

This patch adds proper checks for the cpu c-stats. It checks both
cpu id and stat id before sending the nvg request to ccplex.

Change-Id: I732957d1e10d6ce6cffb2c6f5963ca614aadd948
Signed-off-by: Krishna Sitaraman <ksitaraman@nvidia.com>

Need to use bitwise & instead of condition &&.

Change-Id: I8f70aac95d116188ba972f3d38b02e1d3dd32acb
Signed-off-by: Pritesh Raithatha <praithatha@nvidia.com>

The FPGA configuration is encoded in the high byte of
MISCREG_EMU_REVID. Configs GPU and MAX (encoded as
2 and 3) support the ISO SMMU, while BASE (encoded as 1)
does not. This patch implements this encoding and returns
the proper number of SMMU instances.

Change-Id: I024286b6091120c7602f63065d20ce48bcfd13fe
Signed-off-by: Steven Kao <skao@nvidia.com>

System suspend sequence involves initializing the SMMU
as a part of the system suspend exit, which is currently
not present for Tegra194 platform.

Thus call tegra_smmu_init() as a part of system suspend
exit.

Change-Id: I3086301743019e05a40fd221372e7f8713f286ae
Signed-off-by: Vignesh Radhakrishnan <vigneshr@nvidia.com>

This patch updates the cpu core id calculation to match with
internal numbering method used by the MTS.

Change-Id: I5fbe9c8685c23017edc796e114d07c5e979e0d3d
Signed-off-by: Krishna Sitaraman <ksitaraman@nvidia.com>

This patch changes direct writes to ACTLR_ELx registers to use
read-modify-write instead.

Change-Id: I536dce75c01356ce054dd2edee80875e56164439
Signed-off-by: Steven Kao <skao@nvidia.com>

Fake system suspend for Tegra194, calls the routine
tegra_secure_entrypoint() instead of calling WFI.
In essence, this is a debug mode that ensures
that the code path of kernel->ATF and back to kernel
is executed without depending on other components
involved in the system suspend path.

This is for ensuring that verification of system suspend
can be done on pre-silicon platforms without depending on
the rest of the layers being enabled.

Change-Id: I18572b169b7ef786f9029600dad9ef5728634f2b
Signed-off-by: Vignesh Radhakrishnan <vigneshr@nvidia.com>

This patch converts the 'target_cpu' and 'target_cluster' variables from
the tegra_soc_pwr_domain_on() handler to 32-bits. This fixes the signed
comparison warning flagged by the compiler.

Change-Id: Idfd7ad2a62749bb0dd032eb9eb5f4b28df32bba0
Signed-off-by: Varun Wadekar <vwadekar@nvidia.com>

This patch adds platform support for the Memory Controller and
SMMU drivers, for the Tegra194 SoC.

Change-Id: Id8b482de70f1f93bedbca8d124575c39b469927f
Signed-off-by: Varun Wadekar <vwadekar@nvidia.com>

This patch adds support for cpu suspend in T19x soc.

Change-Id: I8ef1d3e03ee9c528dec34eaff6dcbbfa43941484
Signed-off-by: Krishna Sitaraman <ksitaraman@nvidia.com>

* changes:
  mediatek: mt8183: Fix AARCH64 init fail on CPU0
  mediatek: mt8183: refine GIC driver for low power scenarios

CPU0 is default on, so it doesn't need to run pwr_domain_on() at
boot. CPU0 AARCH64 will be set in pwr_domain_suspend(), but it may
encounter race condition with other CPUs.

Now AARCH64 will be set with cluster on in pwr_domain_on(), and
all CPUs on this cluster will be set together. It doesn't need to
set AARCH64 again in pwr_domain_suspend(), so the race condition
can be avoided.

Change-Id: I5693ad56e4901f82badb0fc0d8d13e4c9acfe648
Signed-off-by: James Liao <jamesjj.liao@mediatek.com>

Implement rdist save/resore functions to support low power scenarios.
Signed-off-by: kenny liang <kenny.liang@mediatek.com>
Change-Id: I9ddc077a04f843275fbe2e868cdd0bd00d622de7

Change-Id: Ia120bcaacea3a462ab78db13f84ed23493033601
Signed-off-by: Max Shvetsov <maksims.svecovs@arm.com>

* changes:
  Tegra194: remove L2 ECC parity protection setting
  Tegra194: sip_calls: mark unused parameter as const
  Tegra194: implement handler to retrieve power domain tree
  Tegra194: mce: fix function declaration conflicts
  Tegra194: add macros to read GPU reset status
  Tegra194: skip notifying MCE in fake system suspend
  Tegra194: Enable system suspend

* changes:
  DOC: Update ROMLIB page with memory impact info
  ROMLIB: Optimize memory layout when ROMLIB is used

Complete the Library at ROM documentation with information regarding
the memory impact of the feature.

Change-Id: I5a10620a8e94f123021bb19523a36d558b330deb
Signed-off-by: Louis Mayencourt <louis.mayencourt@arm.com>

ROMLIB extract functions code from BL images to put them inside ROM.
This has for effect to reduce the size of the BL images.

This patch take this size reduction into consideration to optimize the
memory layout of BL2.
A new "PLAT_ARM_BL2_ROMLIB_OPTIMIZATION" macro is defined and used to
reduce "PLAT_ARM_MAX_BL2_SIZE". This allows to remove the gap between
BL1 and BL2 when ROMLIB is used and provides more room for BL31.

The current memory gain is 0x6000 for fvp and 0x8000 for juno.

Change-Id: I71c2c2c63b57bce5b22a125efaefc486ff3e87be
Signed-off-by: Louis Mayencourt <louis.mayencourt@arm.com>

The number of levels in the topology has not changed but the count of
processing elements on the lowest layer is now multiplied by the value
of FVP_MAX_PE_PER_CPU.
Signed-off-by: Imre Kis <imre.kis@arm.com>
Change-Id: Ia1568a40ea33dbbbcdfab6c8ab6d19f4db0b8eb4

Arm's GIC-600 features a Power Register (GICR_PWRR),
which needs to be programmed to enable redistributor
operation. Section 3.6.1 in the GIC-600 TRM describes
the power-up and power-down sequence in pseudo code,
which deviates from the current TF-A implementation
in drivers/arm/gic/v3/gic600.c.
For powering on a redistributor, the pseudo code suggests
to loop over the whole sequence (check for transition,
write request bit) instead of just looping over the
ready bit read as TF-A does in gic600_pwr_on().
This patch fixes GIC-600 power up sequence according
to the TRM.

Change-Id: I445c480e96ba356b69a2d8e5308ffe6c0a97f45b
Signed-off-by: Alexei Fedorov <Alexei.Fedorov@arm.com>

Change-Id: Ibca94eae1a9a89c98b4d7cb5b4fd8943bf854030
Signed-off-by: Paul Beesley <paul.beesley@arm.com>

Previous implementation of timers assumed that clk_div has pretty
representation in MHz (10MHz, 100MHz, etc). Unusual frequencies
(99.99MHz) were causing assertion error and made udelay unusable.
Signed-off-by: Max Shvetsov <maksims.svecovs@arm.com>
Change-Id: Ic915fff224369d113fd9f8edbcfff169fca8beac

The pre-processor directives make it hard to read the non-TBB version of
this function. Refactor the code to improve readability. No functional
change introduced.

In particular, introduce a new helper function load_image_flush(),
that simply loads an image and flushes it out to main memory. This is
the only thing load_auth_image_internal() needs to do when TBB is
disabled or when authentication is dynamically disabled.

In other cases, we need to recursively authenticate the parent images up
to the root of trust. To make this clearer, this code gets moved to a
TBB-specific helper function called load_auth_image_recursive().

As a result, load_auth_image_internal() now boils down to calling the
right helper function (depending on TBB enablement and dynamic
authentication status).

Change-Id: I20a39a3b833810b97ecf4219358e7d2cac263890
Signed-off-by: Sandrine Bailleux <sandrine.bailleux@arm.com>