| Age | Commit message (Collapse) | Author |
|---|
|
To simplify the creation of AArch64 SPL images for the RK3399, we
use the ENABLE_ARM_SOC_BOOT0_HOOK option and prepend 4 bytes of
padding at the start of the text section. This makes it easy for
mkimage to rewrite this word with the 'RK33' boot magic.
This change brings logic to calculate the header size and allocate
the header back in sync. For the RK3399 we now limit the header to
before the payload (i.e. the 'header0' and the padding up to the
actual image) and overwrite the first word (inserted by the
boot0-hook for this purpose) with the 'RK33' magic in-place.
X-AffectedPlatforms: RK3399-Q7
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Tested-by: Klaus Goger <klaus.goger@theobroma-systems.com>
|
|
The SPL binary needs to be prefixed with the boot magic ('RK33' for
the RK3399) on the Rockchip platform and starts execution of the
instruction word following immediately after this boot magic.
This poses a challenge for AArch64 (ARMv8) binaries, as the .text
section would need to start on the odd address, violating natural
alignment (and potentially triggering a fault for any code that
tries to access 64bit values embedded in the .text section).
A quick and easy fix is to have the .text section include the 'RK33'
magic and pad it with a boot0 hook to insert 4 bytes of padding at the
start of the section (with the intention of having mkimage overwrite
this padding with the appropriate boot magic). This avoids having to
modify the linker scripts or more complex logic in mkimage.
X-AffectedPlatforms: RK3399-Q7
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Tested-by: Klaus Goger <klaus.goger@theobroma-systems.com>
|
|
The RK3399 boot code (running as AArch64) poses a bit of a challenge
for SPL image generation:
* The BootROM will start execution right after the 4-byte header (at
the odd instruction word loaded into SRAM at 0xff8c2004, with the
'RK33' boot magic residing at 0xff8c2000).
* The default padding (during ELF generation) for AArch64 is 0x0,
which is an illegal instruction and the .text section needs to be
naturally aligned (someone might locate a 64bit constant relative
to the section start and unaligned loads trigger a fault for all
privileged modes of an ARMv8)... so we can't simply define the
CONFIG_SPL_TEXT_BASE option to the odd address (0xff8c2004).
* Finally, we don't want to change the values used for padding of
the SPL .text section for all ARMv8 targets to the instruction
word encoding 'nop', as this would affect all padding in this
section and might hide errors that would otherwise quickly trigger
an illegal insn exception.
To deal with this situation, we modify the rkimage generation to
- understand the fact that the RK3399 needs to pad the header to an
8 byte boundary using an AArch64 'nop'
- the necessary logic to adjust the header_size (which controls the
location where the payload is copied into the image) and to insert
this padding (AArch64 insn words are always little-endian) into
the image following the 4-byte header magic.
X-AffectedPlatforms: RK3399-Q7
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Tested-by: Klaus Goger <klaus.goger@theobroma-systems.com>
|
|
The RC4 encoding works on full blocks, but the calculation of the
starting offset and size are needlessly complicated by using a
reference value known to be offset into a block by the size of the
header and then correcting for the (hard-coded) size of the header
(i.e. 4 bytes).
We change this over to use the RK_SPL_HDR_START directly (which is
known to be on a block boundary).
X-AffectedPlatforms: RK3399-Q7
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Tested-by: Klaus Goger <klaus.goger@theobroma-systems.com>
|
|
we are using mmc alias , so mmc index have been changed.
now mmc dev 0 is emmc and mmc dev 1 is sdmmc.
Signed-off-by: Jacob Chen <jacob2.chen@rock-chips.com>
|
|
Signed-off-by: Tom Rini <trini@konsulko.com>
|
|
|
|
In this patch-set we add support of new AXS103 firmware as well as
troubleshoot unexpected execution by multiple cores simultaneously.
|
|
|
|
|
|
This includes Marvell mvpp2 patches with the ethernet support for the
ARMv8 Armada 7k/8k platforms. The ethernet patches are all acked by Joe
and he is okay with me pushing them via the Marvell tree.
|
|
|
|
|
|
Actually make use of that shiny new CONFIG_TEGRA124_MMC_DISABLE_EXT_LOOPBACK.
Signed-off-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Signed-off-by: Marcel Ziswiler <marcel@ziswiler.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
|
|
Introduce CONFIG_TEGRA124_MMC_DISABLE_EXT_LOOPBACK to disable the external clock
loopback and use the internal one on SDMMC3 as per the SDMMC_VENDOR_MISC_CNTRL_0
register's SDMMC_SPARE1 bits being set to 0xfffd according to the TRM.
Signed-off-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Acked-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Marcel Ziswiler <marcel@ziswiler.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
|
|
This patch adds board support for the Toradex Apalis TK1 a computer on
module which can be used on different carrier boards.
The module consists of a Tegra TK1 SoC, a PMIC solution, 2 GB of DDR3L
RAM, a bunch of level shifters, an eMMC, a TMP451 temperature sensor
chip, an I210 gigabit Ethernet controller and a SGTL5000 audio codec.
Furthermore, there is a Kinetis MK20DN512 companion micro controller for
analogue, CAN and resistive touch functionality.
For the sake of ease of use we do not distinguish between different
carrier boards for now as the base module features are deemed
sufficient enough for regular booting.
The following functionality is working so far:
- eMMC boot, environment storage and Toradex factory config block
- Gigabit Ethernet
- MMC/SD cards (both MMC1 as well as SD1 slot)
- USB client/host (dual role OTG port as client e.g. for DFU/UMS or host,
other two ports as host)
Signed-off-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Marcel Ziswiler <marcel@ziswiler.com>
Signed-off-by: Tom Warren <twarren@nvidia.com>
|
|
Even though we expect only master core to execute U-Boot code
let's make sure even if for some reason slave cores attempt to
execute U-Boot in parallel with master they get halted very early.
If platform wants it may kick-start slave cores before passing control
to say Linux kernel or any other application that want to see all cores
of SMP SoC up and running.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
|
|
In axs103 v1.1 procedure to kick-start slave cores has changed quite a bit
compared t previous implementation.
In particular:
* We used to have a generic START bit for all cores selected by CORE_SEL
mask. But now we don't touch CORE_SEL at all because we have a dedicated
START bit for each core:
bit 0: Core 0 (master)
bit 1: Core 1 (slave)
* Now there's no need to select "manual" mode of core start
Additional challenge for us is how to tell which axs103 firmware we're
dealing with. For now we'll rely on ARC core version which was bumped
from 2.1c to 3.0.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
|
|
* Rely on default pulse polarity value
* Don't mess with "multicore" value as it doesn't affect execution
In essence we now do a bare minimal stuff:
1) Select HS38x2_1 with CORE_SEL=1 bits
2) Select "manual" core start (via CREG) with START_MODE=0
3) Generate cpu_start pulse with START=1
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
|
|
Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
|
|
This is a preparation work for the support of CONFIG_BLK.
Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
|
|
Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
|
|
For consistency, use an accessor to access the private data. Also for the
same reason, rename all priv_data to priv.
Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
|
|
The Marvell PHY support is needed espescially for the A7040-DB with the
SGMII port (port 2). As without the marvell PHY driver configuration
for SGMII, ethernet won't work.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Stefan Chulski <stefanc@marvell.com>
Cc: Kostya Porotchkin <kostap@marvell.com>
Cc: Nadav Haklai <nadavh@marvell.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
The default configuration for the COMPHY-0 port should be 1G, as its
used as 1G SGMII connection. This change is necessary to get the
MAC2 port (SGMII) working on this DB.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Stefan Chulski <stefanc@marvell.com>
Cc: Kostya Porotchkin <kostap@marvell.com>
Cc: Nadav Haklai <nadavh@marvell.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
As pointed out by Stefan Chulski, this variable is unused and should be
removed.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Stefan Chulski <stefanc@marvell.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
On PPv2.2 we enable PHY polling, so we also need to configure the PHY
address in the specific PHY address rgisters.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Stefan Chulski <stefanc@marvell.com>
Cc: Kostya Porotchkin <kostap@marvell.com>
Cc: Nadav Haklai <nadavh@marvell.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
Testing shows, that PHY polling needs to be enabled on Armada 7k/8k.
Otherwise ethernet transfers will not work correctly. PHY polling
is enabled per default after reset, so we do not need to specifically
enable it, but this makes it clearer.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Stefan Chulski <stefanc@marvell.com>
Cc: Kostya Porotchkin <kostap@marvell.com>
Cc: Nadav Haklai <nadavh@marvell.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
Add a missing occurrance of PHY_INTERFACE_MODE_RGMII_ID, which should
be handled identical to PHY_INTERFACE_MODE_RGMII.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Stefan Chulski <stefanc@marvell.com>
Cc: Kostya Porotchkin <kostap@marvell.com>
Cc: Nadav Haklai <nadavh@marvell.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
This patch adds the GoP (Group of Ports) and NetC (Net Complex) setup to
the Marvell mvpp2 ethernet driver for the missing port 0. This code is
mostly copied from the Marvell U-Boot version and was written by Stefan
Chulski. Please note that only SFI support have been added, as this
is the only interface that this code has been tested with. XAUI and
RXAUI support might follow at a later stage.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Stefan Chulski <stefanc@marvell.com>
Cc: Kostya Porotchkin <kostap@marvell.com>
Cc: Nadav Haklai <nadavh@marvell.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
This patch adds the GoP (Group of Ports) and NetC (Net Complex) setup to
the Marvell mvpp2 ethernet driver. This code is mostly copied from the
Marvell U-Boot version and was written by Stefan Chulski. Please
note that only RGMII and SGMII support have been added, as these are
the only interfaces that this code has been tested with.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Stefan Chulski <stefanc@marvell.com>
Cc: Kostya Porotchkin <kostap@marvell.com>
Cc: Nadav Haklai <nadavh@marvell.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
Read the "phy-speed" DT property to differentiate between 1 and 2.5GB
SGMII operations. Please note that its unclear right now, if this
DT property will be accepted in mainline Linux. If not, we need to
revisit this code and change it to use the accepted property.
Signed-off-by: Stefan Roese <sr@denx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
This patch does a bit of restructuring of the probe / init functions,
mainly to allow earlier register access as it is needed for the upcoming
GoP (Group of Ports) and NetC (Net Complex) code.
Signed-off-by: Stefan Roese <sr@denx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
This patch adds the new PHY interface modes XAUI, RXAUI and SFI that will
be used by the PPv2.2 support in the Marvell mvpp2 ethernet driver.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Stefan Chulski <stefanc@marvell.com>
Cc: Kostya Porotchkin <kostap@marvell.com>
Cc: Nadav Haklai <nadavh@marvell.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
This patch adds the PPv2.2 specific FIFO configuration to the mvpp2
driver. The RX FIFO packet data size is changed to the recommended
FIFO sizes. The TX FIFO configuration is newly added.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Stefan Chulski <stefanc@marvell.com>
Cc: Kostya Porotchkin <kostap@marvell.com>
Cc: Nadav Haklai <nadavh@marvell.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
Currently, the naming of the ethernet ports is not handled correctly in
the multi-CP (Communication Processor) case. On Armada 8k, the slave-CP
also instantiates an ethernet controller with the same device ID's.
This patch now takes this into account and adds the required base-id
so that the slave-CP ethernet devices will be named "mvpp2-3 ...".
This patch also updates my Copyright notice to include 2017 as well.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Stefan Chulski <stefanc@marvell.com>
Cc: Kostya Porotchkin <kostap@marvell.com>
Cc: Nadav Haklai <nadavh@marvell.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
Since we've now integrated the A7k/8k support in the mvpp2 ethernet
driver, lets enable the support for both Marvell developments boards.
Signed-off-by: Stefan Roese <sr@denx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
This commit adds the description of the PPv2.2 hardware block for the
Marvell Armada 7K and Armada 8K processors, and their corresponding Armada
7040 and 8040 Development boards.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Stefan Roese <sr@denx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
Since Armada 7K/8K is also equipped with a newer version of the MVPP2
ethernet controller, lets enable compilation of this driver for these
platforms.
Signed-off-by: Stefan Roese <sr@denx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
This fixes problems noticed with the PPv2.2 A7k/8k port, when not all
elements of the descriptors had been cleared before use.
Signed-off-by: Stefan Roese <sr@denx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
This patch moves the base_probe function mvpp2_base_probe() from the
MISC driver to the ETH driver. When integrated in the MISC driver,
probe is called too early before the U-Boot ethernet infrastructure
(especially the MDIO / PHY interface) has been initialized. Resulting
in errors in mdio_register().
Signed-off-by: Stefan Roese <sr@denx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
In U-Boot the MDIO / SMI support is integrated in the mvpp2 driver,
currently only supporting the 32bit platforms (Armada 37x). This patch
now adds the A7k/8k PPv2.2 MDIO support to that the phy / mii IF
can be used as well on these platforms.
Signed-off-by: Stefan Roese <sr@denx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
Now that the mvpp2 driver has been modified to accommodate the support
for PPv2.2, we can finally advertise this support by adding the
appropriate compatible string.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Stefan Roese <sr@denx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
In PPv2.1, we have a maximum of 8 RXQs per port, with a default of 4
RXQs per port, and we were assigning RXQs 0->3 to the first port, 4->7
to the second port, 8->11 to the third port, etc.
In PPv2.2, we have a maximum of 32 RXQs per port, and we must allocate
RXQs from the range of 32 RXQs available for each port. So port 0 must
use RXQs in the range 0->31, port 1 in the range 32->63, etc.
This commit adapts the mvpp2 to this difference between PPv2.1 and
PPv2.2:
- The constant definition MVPP2_MAX_RXQ is replaced by a new field
'max_port_rxqs' in 'struct mvpp2', which stores the maximum number of
RXQs per port. This field is initialized during ->probe() depending
on the IP version.
- MVPP2_RXQ_TOTAL_NUM is removed, and instead we calculate the total
number of RXQs by multiplying the number of ports by the maximum of
RXQs per port. This was anyway used in only one place.
- In mvpp2_port_probe(), the calculation of port->first_rxq is adjusted
to cope with the different allocation strategy between PPv2.1 and
PPv2.2. Due to this change, the 'next_first_rxq' argument of this
function is no longer needed and is removed.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Stefan Roese <sr@denx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
This commit adjusts how the MVPP2_ISR_RXQ_GROUP_REG register is
configured, since it changed between PPv2.1 and PPv2.2.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Stefan Roese <sr@denx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
The PPv2.2 unit is connected to an AXI bus on Armada 7K/8K, so this
commit adds the necessary initialization of the AXI bridge.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Stefan Roese <sr@denx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
This commit handles a few miscellaneous differences between PPv2.1 and
PPv2.2 in different areas, where code done for PPv2.1 doesn't apply for
PPv2.2 or needs to be adjusted (getting the MAC address, disabling PHY
polling, etc.).
Changed by Stefan for U-Boot:
Since mvpp2_port_power_up() has multiple callers in U-Boot, the U-Boot
version of this patch does not remove this function but simply adds the
check for MVPP21 before the mvpp2_port_fc_adv_enable() call.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Stefan Roese <sr@denx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
This commit adjusts the mvpp2 driver register mapping and access logic
to support PPv2.2, to handle a number of differences.
Due to how the registers are laid out in memory, the Device Tree binding
for the "reg" property is different:
- On PPv2.1, we had a first area for the common registers, and then one
area per port.
- On PPv2.2, we have a first area for the common registers, and a
second area for all the per-ports registers.
In addition, on PPv2.2, the area for the common registers is split into
so-called "address spaces" of 64 KB each. They allow to access the same
registers, but from different CPUs. Hence the introduction of cpu_base[]
in 'struct mvpp2', and the modification of the mvpp2_write() and
mvpp2_read() register accessors. For PPv2.1, the compatibility is
preserved by using an "address space" size of 0.
Changed by Stefan for U-Boot:
Since we don't support multiple CPUs in U-Boot, I've removed all the
code, macros and variables introduced in the Linux patch version for this.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Stefan Roese <sr@denx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
In PPv2.2, the MVPP2_RXQ_DESC_ADDR_REG and MVPP2_TXQ_DESC_ADDR_REG
registers have a slightly different layout, because they need to contain
a 64-bit address for the RX and TX descriptor arrays. This commit
adjusts those functions accordingly.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Stefan Roese <sr@denx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
|
This commit modifies the mvpp2_defaults_set() function to not do the
loopback and FIFO threshold initialization, which are not needed for
PPv2.2.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Stefan Roese <sr@denx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
|
