Age | Commit message (Collapse) | Author |
|
Change-Id: Ib6e10f237906baf35657a1b1699036611d4fc4e0
Signed-off-by: Konstantin Porotchkin <kostap@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/40025
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
|
|
This patch sets the binary extension size close to its maximum as
defined by the specification.
The upper limit may be set to 0xffe7c000 (0xffe1c000 + 384KB), but
actually set to 0xffe7800 to provide 16KB for the secure extension
of 9,956 bytes size.
Change-Id: Id6e78eba95ab9d0e5241c7b70d750cb5b39230b1
Signed-off-by: Victor Axelrod <victora@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/37789
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/40002
|
|
Change-Id: I50303cf6870de1d9480ed66a8916ec93cf36db91
Signed-off-by: Konstantin Porotchkin <kostap@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39704
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Igal Liberman <igall@marvell.com>
|
|
according to design guidelines, enablement of register-mode
access (mci_enable_phy_regs_access) is redundant and should not
performed as a part of MCI initialization for A80x0-A1.
Change-Id: I7b8fb70a4d6fa518ee56c64556305210a183f7a6
Signed-off-by: Omri Itach <omrii@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39746
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
|
|
At AP and CP
//reg[15:8] – RX NQ threshold reduced to 0x7
//reg[23:16] – RX PQ threshold reduced to 0x7
//reg[31:24] – RX RQ threshold reduced to 0x7
//reg[7:4] – RX delta threshold reduced to 0x2
At AP:
//reg[4:0] – WR outstanding reduced to 0x6 (0x5 written to reg)
//reg[10:6] – RD outstanding reduced to 0x6 (0x5 written to reg)
At CP:
//reg[10:6] – RD outstanding reduced to d’14 (d’13 written to reg)
Change-Id: I00d1f22501d22171015873a1b12224e0c0443fbe
Signed-off-by: Omri Itach <omrii@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39526
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
|
|
After commit "f82b49e fix: pcie: cp110: fix pcie clock selection"
we encountered some instabilities in PCIe.
This patch disables the PCIe clock fix temporarily,
until we figure out the root cause for this this issue.
Change-Id: I521e4495118fab2bcbe1e99e6080d1cbb2b08f39
Signed-off-by: Igal Liberman <igall@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39663
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
|
|
When boot source is from MCI, bootROM is already enabling MCI simultaneous
transactions (ID assignment), so in that case we must avoid enabling it
for the 2nd time (according to design guidelines).
Change-Id: I3967fc3526e27356f9e8ebf184b52cc21003e3c1
Signed-off-by: Omri Itach <omrii@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39557
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
|
|
This patch fixes incorrect handling of pcie clock:
If pcie_clk is set to input (by the Sample-at-Reset),
PCIE refclk buffer 0/1 source field in
PCIe_Reference_Clock_Buffer_Control register should be
set to 1 and not 0 as it done now.
Change-Id: I14fedce9d3299b5b85eb742048f03b17f1f9ab14
Signed-off-by: Igal Liberman <igall@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39456
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
|
|
Currently, incorrect preset4 value is used.
This patch fixes the preset4 according to HW measurments.
Change-Id: I839a6c843e912182bb0c71e3ab836dac27ce8a97
Signed-off-by: Igal Liberman <igall@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39421
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
|
|
Add PCIE analog parameters initialization values according to
latest ETP.
Change-Id: I3f5892702610e46ebaba7485e3cef5817f1518b5
Signed-off-by: Igal Liberman <igall@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/36995
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39411
|
|
Change-Id: I11ee81c5741cbe5b7d1f07f57aed095a3f31c182
Signed-off-by: Konstantin Porotchkin <kostap@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39323
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Wilson Ding <dingwei@marvell.com>
|
|
The 'opthash.txt' contains the KAK key digest, which
is used by the customer to program the fuse for the
trusted boot.
Change-Id: I5d7e095a6b48a0c5cb8587c2a8e0063e7988ab7e
Signed-off-by: zachary <zhangzg@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39356
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Wilson Ding <dingwei@marvell.com>
|
|
DOIMAGEPATH used to get its path by 'dir' command when
Makefile assigns the WTP path to it. The 'dir' command
returns the left string from the last "/" of the file
or directory path. Therefore, the actual DOIMAGEPATH is
missing one level directory without one extra "/" in
WTP export env.
This patch fixed this issue by assign the WTP path to
DOIMAGEPATH directly.
Change-Id: I2b18b13f2d3283e714ec6bfa47db6d477295f9c1
Signed-off-by: zachary <zhangzg@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39355
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Wilson Ding <dingwei@marvell.com>
|
|
Added the intermediate files "CSK*-KeyHash.txt" and "TIMHash.txt" and
"Tim_msg.txt" files to gitignore list, which are generated by boot image
generated tool
Change-Id: I3c880f898d4c46a20e85ee3202677a9f682615dc
Signed-off-by: zachary <zhangzg@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39286
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Wilson Ding <dingwei@marvell.com>
|
|
Remove "CSKN-KeyHash.txt" and Hash.txt after building trusted image
Remove "TIMHash.txt" and "Tim_msg.txt" after building untrusted image
Change-Id: I6ce675f39d49c3cd6538205acefd5e64685d6fbf
Signed-off-by: zachary <zhangzg@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39358
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Terry Zhou <bjzhou@marvell.com>
Reviewed-by: Wilson Ding <dingwei@marvell.com>
|
|
Armada37x0's ATF compilation includes native ATF code and
external WTP component. Thus, WTP component should also
be cleaned up along with ATF. However, WTP path is now
tied to DOIMAGEPATH, which is different from the default
for other Marvell SoCs. Thus, PLAT option is required in
order to clean up the WTP component:
make PLAT=a3700 WTP=<path/to/wtp> distclean/clean
ATF always invokes make clean for "CRTTOOLPATH". When
specifying PLAT option in "make clean" command, it will
validate the "include" folder under "PLAT_DIR", which
expects "include" folder located in the same directory
of "platform.mk". Since the existing Armada37x0's PLAT
directory doesn't compiles with ATF convention, it caused
the error in "make clean".
The temporary solution is to add a symbolic link named as
'include' under "PLAT_DIR", which points to itself. This
WA will be removed after 'plat/marvell/a3700' folder is
reorganized.
Change-Id: I95bea27f3999b6490103aec56197eecee2cdd8b6
Signed-off-by: Terry Zhou <bjzhou@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39357
Reviewed-by: Wilson Ding <dingwei@marvell.com>
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
|
|
Issue found when enable 48-bit Virtual Address in kernel
In ARMv8 the CPU can work with 48-bit virtual address,
at this case for propagating TLB maintenance 44-bit of
Physical address are needed.
Marvell interconnect is configured to be 40-bit address
by default, therefore AxAddr[43:40] are not propagated
so the DVM is not working.
Change-Id: I7b8a3ec7960697814079f2c932fd06962bfa4c75
Signed-off-by: Marcin Wojtas <mw@semihalf.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39185
Reviewed-by: Igal Liberman <igall@marvell.com>
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Hanna Hawa <hannah@marvell.com>
|
|
This patch adds static startup XDB scripts for Armada37x0:
- The DDR will be initialized, but only the first 512MB, which is
good enough for board recovery
- The first dram cpu decoder window size will be set;
- BL1 and FIP will be loaded and executed by default.
Change-Id: I4ec4cdb7dcded4d9757fa647d00c977ab4de3b63
Signed-off-by: Ken Ma <make@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39249
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
Reviewed-by: Wilson Ding <dingwei@marvell.com>
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
|
|
In addition, add a call to cp110_pcie_clk_cfg().
This function is responsible for configuring PCIe
ref clock and must me called before configuring
PCIe interface.
Change-Id: I807a4466d9c4459f9403b2542d771c961f07a46e
Signed-off-by: Igal Liberman <igall@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39257
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
|
|
This patch does the following:
- Disable auto flush
If auto flash is enabled, we can't enable/disable
the link and perform "hot reset"
- Set link equalization training to preset4
According to ETP, this is the best preset that our
receiver can handle.
- Remove VPD capability from the capability list
It's not supported
- Remove SRIOV capability from the capability list
It's not supported
- In end point mode: unmask the reset request:
The hot reset and link disable/enable must penetrate
and reset the MAC configurations.
Change-Id: Ia189aaef29b175eb351ee9a861ae3883b3677fa9
Signed-off-by: Igal Liberman <igall@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39226
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
|
|
There's a difference between PCIe clock configuration.
in CP110 revision A1 and CP110 revision A2
CP110 revision A1:
The SatR that decides if the PCIE uses internal refclock or external
refclock does not work properly.
When the clock is set to input we need to do extra configuration
in order to make the input clock functional.
(Iboth PCIe clocks must be aligned, so if one set to input, we must
Performe the configuration).
CP110 revision A2:
PCIe Reference Clock Buffer Control register must be set according to
the clock direction (input/output), there's a field for each refclock
(pcie0 and pcie1) which are set accordingly to the clock configuration.
Change-Id: I6b0150293bfd2d7595d19016c1fd82c4d4ed326a
Signed-off-by: Igal Liberman <igall@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39297
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
|
|
Change-Id: I8288e56954c3d31af5ffe7184c5b3a10ae491dee
Signed-off-by: Igal Liberman <igall@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39296
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
|
|
The patch updated the TBB binary file name from
tbb_linux.exe and ntbb_linux.exe to TBB_linux.
The new bianry file of TBB_linux supports both
security boot and nonsecurity.
The new binary is generated by the latest TBB
source code.
Change-Id: Ifb251a10062dbcdff09aca5b7f92a1ea08f0e262
Signed-off-by: Evan Wang <xswang@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39050
Reviewed-by: Wilson Ding <dingwei@marvell.com>
Tested-by: Wilson Ding <dingwei@marvell.com>
|
|
- There was a wrong DDR Topology option number description issue for DDR4 2CS.
- The patch changes its option number from 4 to 3 for DDR4 2CS.
Change-Id: If37b885b231f57215ea1b5bc51faa5b0e3d234eb
Signed-off-by: zachary <zhangzg@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39136
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Wilson Ding <dingwei@marvell.com>
Reviewed-by: Gina Tadmore <gina@marvell.com>
Reviewed-by: Igal Liberman <igall@marvell.com>
|
|
- By default, several IOs have decode window to rWTM:
- SATA: window 2, from 0x1fff0000 to 0x2000ffff.
- SDIO/eMMC: window 2, from 0x1fff0000 to 0x2000ffff.
- USB32: window 2, from 0xC0000000 to 0xC000ffff.
- These windows should be disabled in non-DAS mode, since:
- These IOs should not access rWTM.
- These decode windows overlap with DRAM decode window.
For example, in 512MB DRAM case, SATA and SDIO/eMMC
rWTM window overlap with DRAM decode window in the
last 64KB DRAM address.
- This patch sets these decode window to be used for DRAM,
so software would disable them by default, and configure
them to be DRAM decode window if needed.
- JIRA: A3700-1163
Change-Id: I6d35a98b14f3ddc575930f83ea5e700995a29519
Signed-off-by: jinghua <jinghua@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39247
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Wilson Ding <dingwei@marvell.com>
|
|
In order to support 4GB DRAM the CPU decode window changed
from default configuration, and the PCIe decode window is
also changed with less than 2GB DRAM to align with 4GB
DRAM. So PCIe will not work well with default CPU decode
window configuration and it needs init again in callback
of pwr_domain_suspend_finish().
JIRA number: A3700-1157
Change-Id: I938070589890215b8d8aaffed4f23ba592184c3e
Signed-off-by: Evan Wang <xswang@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39167
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
Reviewed-by: Wilson Ding <dingwei@marvell.com>
Reviewed-by: Ken Ma <make@marvell.com>
|
|
EIP97 (crypto hardware) DMA address decoder register
is not configured. This patch configures the register
and fixes the EIP97 driver operation in the Kernel.
Change-Id: I5a3d1acfce0b08956534cfbde6cae3d208ecf2b4
Signed-off-by: Igal Liberman <igall@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39184
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
|
|
Rename Macciatobin platform from a80x0_cust to a80x0_mcbin
Change-Id: I715be80891f7c2d5d68a84f22a0568eb94b26c18
Signed-off-by: Konstantin Porotchkin <kostap@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38411
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
|
|
- This patch adds build option DDR_TOPOLOGY number to support DDR4
2cs 4GB build.
- Example:
make DEBUG=1 USE_COHERENT_MEM=0 DDR_TOPOLOGY=3 PLAT=a3700 all fip
Change-Id: I966dc6630fda5ebab906ba669ef93f4614bd2a36
Signed-off-by: zachary <zhangzg@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38864
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Gina Tadmore <gina@marvell.com>
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
|
|
Change-Id: I0143ffbeae1d48e2700b35107588b4d6de893485
Signed-off-by: Ken Ma <make@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39046
Reviewed-by: Hua Jing <jinghua@marvell.com>
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
|
|
- Update build.txt for the build option - WTMI_IMG. WTMI_IMG can point
to a image which does nothing, a image which supports EFUSE or a
customized CM3 firmware image.
Change-Id: I63304fe39c823d67db39c7ee111972fef063cecd
Signed-off-by: Ken Ma <make@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39047
Reviewed-by: Hua Jing <jinghua@marvell.com>
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
|
|
In addition, this patch fixes DRAM initialization issue
introduced in:
"6942e0c fix: a8k: Fix the platform dependency issue"
The root cause of the DRAM failure is the pcac naming.
All other A7040 devices names A70x0_XX but pcac was
named A7040_pcac.
Change-Id: Ic204a3d5272a3563a15f955bb8ce70e2e7390ba1
Signed-off-by: Igal Liberman <igall@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39021
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
|
|
plat_get_pcie_hw_data() has weak implementation in pci_ep_setup.c
and another implementation for A70x0 PCAC in marvell_plat_config.c.
Currently, there's an issue with A70x0 PCAC linkage and the weak
implementation is chosen.
The issue is caused by the fact that the A70x0 PCAC implementation
is located under #ifndef BLE but the PCIe code is located in the BLE.
This patch fixes the linkage issue (by moving the
plat_get_pcie_hw_data() code out if the #ifndef).
Change-Id: I53cd32af6f858b0bfd63a0fffd562c34e17eb2e3
Signed-off-by: Igal Liberman <igall@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/39020
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
|
|
- Not all configurable decode windows could be used for dram, some
units have to reserve one decode window for other unit they have
to communicate with; for example, DMA engineer has 3 configurable
windows, but only two could be for dram while the last one has to
be for pcie, so for DMA, its max dram decode windows number is 2;
- Only io dram decode windows needs to be reconfigured for different
dram size boards(512MB, 2GB and 4GB), this patch renames the field
member max_win to max_dram_win and updates the old io max decode
windows number to max dram decode windows number.
Change-Id: I61ed0c936baa5b2f73c5e7d0c41dfc81c7033b90
Signed-off-by: Ken Ma <make@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38956
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Hua Jing <jinghua@marvell.com>
|
|
- When dram size is 4GB, there are 4 cpu dram windows, the number 4
exceeds the limit of io decode windows number, so in this case, only
one io decode window whose size is 4GB is used for dram;
- In order to have the same view on dram for cpu and ios, cpu dram
window must not use remap function, this patch removes remap function
for cpu dram windows.
Change-Id: I328b7338ee464935917c80857be208c40369d4ba
Signed-off-by: Ken Ma <make@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38925
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Hua Jing <jinghua@marvell.com>
|
|
- The TIMN image is missing in Armada3700 build folder,
which is one of the image components needed by the
recovery procedure. Makefile is supposed to get the
image filename from atf_timN.txt and it defines a
variable to parse the filename in the atf_timN.txt.
But the atf_timN.txt is dynamically generated. It
doesn't exist when this variable is initialized (if
ATF performs a clean build).
- This patch is to change this global variable as the
second-expansion of shell command in Makefile. So
that the shell will execute the enclosed command only
when the Makefile is referring this variable
Change-Id: Iba39377cc2752c42bcb0ecbd854f8a9e57e1353d
Signed-off-by: Terry Zhou <bjzhou@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38699
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Hua Jing <jinghua@marvell.com>
|
|
only support AES_CBC_256 algorithm
build uart image without encryption
Change-Id: I66caf134b6e159433561e34c0f775d6e177dde31
Signed-off-by: Terry Zhou <bjzhou@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38698
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Hua Jing <jinghua@marvell.com>
|
|
- Now in ATF IO decode windows are configured according to CPU DRAM
decode window; For 4GB dram board, there are 4 CPU DRAM decode
windows, the IO decode window number is smaller than 4(some IO has 2
decode windows while other IOs has 3 decode windows), then 4 CPU DRAM
windows can not be configured for IO all;
- This patch only fill the first IO decode window with base(0) and size
(4GB) when CPU DRAM windows number exceeds the IO decode windows max
number.
Change-Id: Ic386c85fae11c3d6381c738e37ade261382cfce0
Signed-off-by: Ken Ma <make@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38868
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Hua Jing <jinghua@marvell.com>
|
|
- In order to support 2 chips dram 4Gb, a3700 needs to use more than
1 dram cpu decode window configuration.
- A3700 uboot used to read the first dram cpu decode window and size,
it is not enough for more than 2 cpu decode window.
- This patch writes dram cpu decode windows information to DDR for
passing these information to uboot for using.
Change-Id: I721cf0ca0eeb634c7e860dd4a6e15a56c2ed1605
Signed-off-by: zachary <zhangzg@marvell.com>
Signed-off-by: Ken Ma <make@marvell.com>
Signed-off-by: zachary <zhangzg@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38695
Reviewed-by: Hua Jing <jinghua@marvell.com>
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
|
|
- In order to pass cpu decode window information from atf to uboot,
add more system information types for it.
Change-Id: I83032a9ab26c7438e542e1c9a04ecbeaebcc3bc1
Signed-off-by: zachary <zhangzg@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38833
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
Reviewed-by: Hua Jing <jinghua@marvell.com>
|
|
- To support 4GB DRAM, CPU decode window needs to be updated: there are
at most 5 configurable cpu decode windows and each window could only
supports the size with 2 powers of N; since the address range is 4GB
and Internal Regs, CCI-400 and Boot Rom windows hold fixed ranges, so
it means that the configurable window size could be 2GB, 1GB, 512MB,
256MB...; and cpu dram windows number must increase to utilize memory
as much as possible, they will take place of some unused cpu io
windows;
- If total dram size is more than 2GB, now there is only one case - 4GB
dram, the cpu windows configuration is as below:
- Internal Regs, CCI-400 and Boot Rom windows are kept as default;
- Use 4 CPU decode windows for DRAM, which cover 3.75GB DRAM; since
DDR window 0 is configured in tim header with 2GB size, this patch
only configures cpu dram windows 1/2/3 with sizes of
1GB/256MB/512MB;
- The only one CPU decode window left is for PCIe, which has 64MB
address;
- If total dram size is no more than 2GB, then pcie cpu window is
modified to be aligned with 4GB dram's configuration, other cpu
windows are kept as default.
Change-Id: I178df27fd5a38d18472f35e7ba89227ef269572a
Signed-off-by: Ken Ma <make@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38762
Reviewed-by: Hua Jing <jinghua@marvell.com>
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
|
|
- Update Marvell platform porting documents.
Change-Id: I6a8d3a0d527dcce342f358e8f0d49ea2598306ac
Signed-off-by: allen yan <yanwei@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38689
Reviewed-by: Victor Gu <xigu@marvell.com>
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
Reviewed-by: Hua Jing <jinghua@marvell.com>
|
|
- The patch adds macro prints for relevant GIC and CCI registers
registers whenever an unhandled exception is taken in BL31.
Change-Id: Ic9519fba41d45e6e47a615aea242d1b240aa3ce6
Signed-off-by: zachary <zhangzg@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38637
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Hua Jing <jinghua@marvell.com>
|
|
- Use plat cci reg print to repalce plat interconnect reg print to
align mainline code.
Change-Id: If0df0049f8efb3ac1cf54117b431ae11b2827b85
Signed-off-by: zachary <zhangzg@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38666
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
|
|
- For uart1, the first step is setting uart tx and rx pin
as gpio mode, and then set gpio 18/19 as wake up source.
- For uart0, the first step is setting uart tx and rx pin
as gpio mode, and then set gpio 25/26 as wake up source.
- Default config uart1 as wake up source.
Change-Id: I8deb16266300af5b1d72d7a13ebd7a2e5cecc451
Signed-off-by: wei yan <yanwei@marvell.com>
Signed-off-by: allen yan <yanwei@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38685
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Hua Jing <jinghua@marvell.com>
|
|
The GPIO wake up enable registers(0xD0013C18 and 0xD0013C1C) are used
to enable the GPIO pin to wake up the CPU from partial powerdown and
step down mode. According to the suspend to RAM testing, the GPIO
wake up do not need to set it, and there is crash if some of its bits
are set.
This patch removes the setting for GPIO wake up bits.
Change-Id: Iae5ab2aac1fe5b8a4023158d4a4fe763c2723f4c
Signed-off-by: Victor Gu <xigu@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38683
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Hua Jing <jinghua@marvell.com>
|
|
It is possible to trigger extra interrupts during setting the wake
up option and sources, which affects the suspend operation.
Thus the wake up option and source setting should be done prior to
ack the irq.
Change-Id: I83f79193b935e98a29443c1efaaec7075b8eb01f
Signed-off-by: Victor Gu <xigu@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38644
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Hua Jing <jinghua@marvell.com>
|
|
- In UART boot mode initialization, SPI CS and CLK pins are used as
downloading image purpose; But after downloading, these bits should
be switch back to SPI mode. Otherwise SPI functionality is broken.
- This patch sets bit 28 of north_bridge_pin_out_en_high register 13804,
to switch SPI CS and CLK pins back from UART download feature.
Change-Id: I30e2c0202da7f7b0fba3916af121da9017e1cce4
Signed-off-by: jinghua <jinghua@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38690
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
(cherry picked from commit eaa8a65da8af00d3443068418955f9396becd00f)
Reviewed-on: http://vgitil04.il.marvell.com:8080/38691
|
|
The SPI and PPI irq states are saved during core power off or suspend,
and resumed during core up or resume.
Change-Id: Ib41a28b453d07cc718c61ab2246dfa382c2d87bf
Signed-off-by: Victor Gu <xigu@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/37432
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Hua Jing <jinghua@marvell.com>
Signed-off-by: zachary <zhangzg@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38078
Tested-by: Hua Jing <jinghua@marvell.com>
Signed-off-by: zachary <zhangzg@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38610
|
|
This patch wrap the ARM GICv3 SPI and PPI irq state PM APIs
for Marvell platform, which are used in suspend and resume
process.
Marvell plat_marvell_gic_irq_save/restore APIs are used to
save and restore SPI irq states.
Marvell plat_marvell_gic_irq_pcpu_save/restore APIs are used
to save and restore per-cpu irq states.
Change-Id: I7a466fef369241b3ce7f727c0c7179ec7023a09e
Signed-off-by: Victor Gu <xigu@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/37431
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Hua Jing <jinghua@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38077
Tested-by: Hua Jing <jinghua@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38609
|