There are couple issues in set_ios function which are introduced when
xlnx code is rebased to mainline U-Boot version 2021.1 with
'commit 880b655d660e ("mmc: zynq_sdhci: Fix UHS 1.8v switching with 5ms
delay")'
Mainline 'commit f12341a95295 ("mmc: sdhci: Fix HISPD bit handling")' is
pulled partially when rebased to 2021.1. Fix it by adding back missing
HISPD bit setting code.
Remove extra line which writes back to sdhc host control register.
set_control_reg call got duplicated after rebase, remove it.
Since, voltage setting function (set_control_reg) is moved to the
beginning of the function, no need to return if the set_ios is called
with clock disable, so remove return.
mmc: zynq_sdhci: Change variable deviceid to node_id
Change deviceid to node_id in arasan_zynqmp_dll_reset() and also in
tapdelay related static inline functions to reflect proper name and
for consistency.
mmc: zynq_sdhci: Move setting tapdelay code to driver
tap_delays.c just has calls to xilinx_pm_request() for setting tapdelays.
Simply move these calls to zynq_sdhci.c and make them static inline.
Similarly zynqmp_tap_delay.h also has call to xilinx_pm_request() for
dll reset. Do the same for this file as well.
arm64: zynqmp: Move USB3 PHY properties from DWC3 node to USB node
Move the PHY properties from DWC3 node to USB node in ZynqMP DTs as here
the USB3 PHY used is PSGTR, which is connected to Xilinx USB core. This
PHY initialization should be handled from Xilinx USB core as the
prerequisite register configurations are done here only.
Add kernel_comp_addr_r, kernel_comp_size env variables for zynqmp and
versal to be able to use the compressed kernel Image(.gz,.bz2,.lzma,.lzo)
using booti command.
mmc: zynq_sdhci: Use pm_node id instead of device sequence id
When programming tapdelays & resetting dll, we are using device
sequence id, which is not programming tapelays & dll reset.
So use pm_node id instead of device sequence id to fix above issue.
Also use u8 instead of u32 for node_id variable as we have pm_node
ids less than 128.
Signed-off-by: T Karthik Reddy <t.karthik.reddy@xilinx.com> Reviewed-by: Ashok Reddy Soma <ashok.reddy.soma@xilinx.com>
Michal Simek [Mon, 12 Jul 2021 08:56:33 +0000 (10:56 +0200)]
arm64: zynqmp: Remove gpio from aliases list
It is not recommended to have aliases for gpio. In past it was used in
Linux for assigning numbers via sysfs which is deprecated and libgpiod
should be used instead.
In U-Boot this number is used for seq number but gpio offset are not
counted from this number. That's why having these aliases only for seq
number is not needed. As is done in Linux it is the best to use full gpio
name instead of sequence number which depends on sequence in binding.
The same changes was done by commit d458db3ce900 ("arm64: zynqmp: Remove
gpio from aliases list").
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
arm64: zynqmp: Use new compatible strings for DP snd pcm
To remove a stack trace when removing and reprobing the DP snd pcm,
use the new DP pcm compatible strings to allow the driver to create
standardized names for each DP snd pcm.
xilinx: Add jedec compatible string for QSPI (Zynq/ZynqMP)
Add missing "jedec, spi-nor" compatible string for QSPI flash node.
Spi-nor framework uses this compatiblity string to probe &
initilize flash. With missing compatibility string we are observing
below error:
Zynq> sf probe 0 0 0
jedec_spi_nor spi_flash@0:0: unrecognized JEDEC id bytes:
00, 00, 00 Failed to initialize SPI flash at 0:0 (error -2)
mtd: spi-nor: Fix read SFDP data in dual parallel mode
Reading the SFDP data is broken in two ways in dual parallel mode in the
current implementation.
1) It reads the data interleaved from both chips, which results in data
that can not be parsed. Since both chips have to be identical for dual
parallel mode the SFDP data will also be identical. So only read the data
from one of the chips.
2) The page, erase and chip size are not update to reflect that they should
be twice as large in dual parallel mode. Fix this by moving the update of
those parameter after running the SFDP detection rather than doing it
before.
In the current implementation dual parallel mode only works for limited
corner cases, but is mostly broken.
For example
* Erase will only erase one of the chips leaving every second byte
non-erased
* Lock/unlock only works if a read is done before it, otherwise it will
timeout
* Using bank/extended address mode (CONFIG_SPI_FLASH_BAR) fails in
`sf probe` with a timeout
This is primarily due to incorrect setting of the striped access. Some
commands that should set it don't set it, while others that do set it
shouldn't set it.
To fix this make sure that the SPI_XFER_STRIPE flag gets set in
spi_nor_init() when operating in dual parallel mode. This means the flag is
always set before issuing any commands that need it.
And then in addition use a list of opcodes for which to disable striped
access (e.g. erase).
This is similar to what the Linux driver does to support dual parallel
mode.
Kunihiko Hayashi [Tue, 15 Jun 2021 06:33:02 +0000 (15:33 +0900)]
arm64: Fix relocation of env_addr if POSITION_INDEPENDENT=y
If both POSITION_INDEPENDENT and SYS_RELOC_GD_ENV_ADDR are enabled,
wherever original env is placed anywhere, it should be relocated to
the right address.
Relocation offset gd->reloc_off is calculated with SYS_TEXT_BASE in
setup_reloc() and env address gd->env_addr is relocated by the offset in
initr_reloc_global_data().
However, SYS_TEXT_BASE isn't always runtime base address when
POSITION_INDEPENDENT is enabled. So the relocated env_addr might point to
wrong address. For example, if SYS_TEXT_BASE is zero, gd->env_addr is
out of memory location and memory exception will occur.
There is a difference between linked address such as SYS_TEXT_BASE and
runtime base address. In _main, the difference is calculated as
"run-vs-link" offset. The env_addr should also be added to the offset
to fix the address.
Add support for xilinx multirate(MRMAC) ethernet driver.
This driver uses multichannel DMA(MCDMA) for data transfers of MRMAC.
Added support for 4 ports of MRMAC for speeds 10G and 25G.
MCDMA supports upto 16 channels but in this driver we have setup only
one channel which is enough.
net: ethtool: Add ethernet speed macros for higher speeds
Add speed macro's for higher ethernet speeds to be used in u-boot
networking drivers. Added Macros for speeds 14G, 20G, 25G, 40G, 50G,
56G, 100G and 200G inline with linux.
There are lot of accesses to priv data in ofdata_to_platdata, which is
incorrect. Create a platform data structure and use it in
ofdata_to_platdata, then copy all platform data to priv data in probe.
Michal Simek [Mon, 14 Jun 2021 13:12:27 +0000 (15:12 +0200)]
arm64: zynqmp: Remove unused snps,enable-hibernation flag
This property is used only in gadget mode and there were some upstreaming
commnents about disabling it. Also this property is commented in usb0 but
not in usb1. That's why remove it.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Michal Simek [Thu, 3 Jun 2021 08:47:04 +0000 (10:47 +0200)]
arm64: zynqmp: Remove gpio from aliases list
It is not recommended to have aliases for gpio. In past it was used in
Linux for assigning numbers via sysfs which is deprecated and libgpiod
should be used instead.
In U-Boot this number is used for seq number but gpio offset are not
counted from this number. That's why having these aliases only for seq
number is not needed. As is done in Linux it is the best to use full gpio
name instead of sequence number which depends on sequence in binding.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
spi: zynqmp_gqspi: Fix issue of reading more than 32bits length
As the flash sizes are increasing day by day, QSPI in dual-parallel mode
can have devices of size > 512MB. In qspi driver we are trying to read
all the data at once using DMA.
The DMA descriptor destination size is only 29bits long.
QSPIDMA_DST_SIZE 0xFF0F0804
BITS: 1:0 Reserved to keep word alignment
BITS: 28:2 Number of 4-byte words the DMA will transfer
BITS: 31:29 Reserved: Returns 0 when read, writes ignored
So we can only transfer data of 0x1FFFFFF0(512MB minus 4bytes) bytes.
Anything above will overflow this register and will ignore higher bits
above 29 bits.
Change the DMA functionality if the requested size is greater than or
equal to 512MB to read 256MB chunks.
Srinivas Neeli [Mon, 8 Mar 2021 08:35:19 +0000 (14:05 +0530)]
arm64: dts: zynqmp: rtc: Update rtc calibration value
As per the design specification
"The 16-bit Seconds Calibration Value represents the number of
Oscillator Ticks that are required to measure the largest time
period that is less than or equal to 1 second.
For an oscillator that is 32.768 KHz, this value will be 0x7FFF."
Michal Simek [Tue, 26 Jan 2021 14:59:59 +0000 (15:59 +0100)]
arm64: versal: zynqmp: Add support for VPXA2785
VPXA2785(vp-x-a2785-00) is evaluation board which contains two PCIe-Edge
fingers, one for PCIe-B(gen5x8) and one for CPM(dual gen5x8, gen5x16).
Each of the ports can operate in endpoint or root port mode. This allows
the single card to be used for both root port, endpoint, and switch modes.
The board is designed in the similar manner as others Versal boards. It
means board also have ZynqMP Zu4 System Controller which is described in a
separate file.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Michal Simek [Tue, 19 Jan 2021 14:00:45 +0000 (15:00 +0100)]
arm64: versal: zynqmp: Add support for vpk120
Add support for Versal premium evaluation board vpk120.
Board contains two systems. The primary is Versal VP1202 ACAP device and
the secondary is ZynqMP zu4 which acts as system controller.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Michal Simek [Thu, 26 Nov 2020 13:25:03 +0000 (14:25 +0100)]
ARM: zynq: Fix OCM mapping to be aligned with binding on zc702
The Linux commit f69629919942 ("dt-bindings: sram: Convert SRAM bindings to
json-schema") converted binding to yaml and some missing required
properties started to be reported. Align binding based on it.
The patch is fixing these warnings:
.../zynq-zc702.dt.yaml: sram@fffc0000: '#address-cells' is a required property
.../zynq-zc702.dt.yaml: sram@fffc0000: '#size-cells' is a required property
.../zynq-zc702.dt.yaml: sram@fffc0000: 'ranges' is a required property
>From schema: .../Documentation/devicetree/bindings/sram/sram.yaml
Michal Simek [Thu, 26 Nov 2020 13:25:04 +0000 (14:25 +0100)]
ARM: zynq: Convert at25 binding to new description on zc770-xm013
The Linux commit f8f79fa6bb25 ("dt-bindings: at25: convert the binding
document to yaml") converted binding to yaml and 3 deprecated properties
pop up.
The patch is fixing these warnings:
.../zynq-zc770-xm013.dt.yaml: eeprom@2: 'pagesize' is a required property
.../zynq-zc770-xm013.dt.yaml: eeprom@2: 'size' is a required property
.../zynq-zc770-xm013.dt.yaml: eeprom@2: 'address-width' is a required property
>From schema: .../Documentation/devicetree/bindings/eeprom/at25.yaml
T Karthik Reddy [Tue, 18 May 2021 14:28:02 +0000 (08:28 -0600)]
mmc: zynq_sdhci: wait till sd card detect state is stable
As per SD spec when SD host controller is reset, it takes 1000msec
to detect the card state. In case, if we enable the sd bus voltage &
card detect state is not stable, then host controller will disable
the sd bus voltage.
In case of warm/subsystem reboot, due to unstable card detect state
host controller is disabling the sd bus voltage to sd card causing
sd card timeout error. So we wait for a maximum of 1000msec to get
the card detect state stable before we enable the sd bus voltage.
This current fix is workaround for now, this needs to be analysed
further. Zynqmp platform should behave the same as Versal, but we
did not encounter this issue as of now. So we are fixing it for
Versal only.
Signed-off-by: T Karthik Reddy <t.karthik.reddy@xilinx.com>
Michal Simek [Thu, 13 May 2021 11:58:58 +0000 (13:58 +0200)]
arm64: zynqmp: Enable EFI secure boot
Enabling EFI secure boot which is required for EBBR specification.
Enabling this will fix
"RT.SetVariable - Create one Time Base Auth Variable, the expect return
status should be EFI_SUCCESS"
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
The EFI_LOAD_FILE_PROTOCOL_GUID and EFI_LOAD_FILE2_PROTOCOL_GUID are needed
to complement the implementation of the LoadFile() boot service.
Remove a duplicate declaration of a variable for the
EFI_LOAD_FILE2_PROTOCOL_GUID.
Move the remaining declaration to efi_boottime.c.
Add a variable for the EFI_LOAD_FILE_PROTOCOL_GUID.
efi_loader: resequence functions in efi_boottime.c
For implementing support for the EFI_LOAD_FILE_PROTOCOL in the LoadImage()
service we will have to call the LocateDevicePath() service. To avoid a
forward declaration resequence the functions.
U-Boot expects to be linked to a specific hard-coded address and to
be loaded to and run from that address. CONFIG_POSITION_INDEPENDENT
config lifts that restriction & allowing the code to be loaded to
and executed from almost any address.
As we enabled CONFIG_POSITION_INDEPENDENT, CONFIG_INIT_SP_RELATIVE
is enabled by default, where it will set the early stack pointer at
runtime by adding an offset value to &_bss_start. The offset value
is taken from SYS_INIT_SP_BSS_OFFSET.
SYS_INIT_SP_BSS_OFFSET offset should be large enough so that the
early malloc region, global data (gd), and early stack should fit.
With commit d8fabcc424bd ("arm64: versal: Increase SYS_MALLOC_F_LEN")
SYS_MALLOC_F_LEN is increased from 32KB to 1MB, so we need to
accommodate this space with SYS_INIT_SP_BSS_OFFSET. Hence increasing
SYS_INIT_SP_BSS_OFFSET to 1.5MB.
Signed-off-by: T Karthik Reddy <t.karthik.reddy@xilinx.com> Acked-by: Ashok Reddy Soma <ashok.reddy.soma@xilinx.com>
arm64: versal: Disable unnecessary configs from OSPI mini config
Disable ZYNQMP_FIRMWARE and GPIO from the OSPI mini config. Also there
are no OSPI parts from couple of manufacturers like SPANSION, SST and
WINBOND. Disable them to save some size of the mini u-boot.
spi: cadence_qspi: Enable linear mode in mini u-boot
OSPI writes are done using direct access(DAC) mode with AHB bus. This needs
linear mode to be enabled. In case of full U-Boot linear mode is enabled
using xilinx_pm_request() calls. But in mini u-boot it will not work
since ZYNQMP_FIRMWARE will not be enabled.
Tried enabling ZYNQMP_FIRMWARE, ZYNQMP_IPI and MAILBOX to make
xilinx_pm_request() working for mini U-Boot as well but it is getting
hung somewhere before it comes to the prompt. This needs to be debugged
later.
For now add condition to call xilinx_pm_request() only if ZYNQMP_FIRMWARE
is enabled in config.
Enable linear mode using register writes. Also remove ZYNQMP_FIRMWARE from
Kconfig as a dependency.
arm64: zynqmp: Add missing SMID for pcie to zynqmp.dtsi
The SMMU is disabled in device tree so this change has no impact.
The benefit is that this way it is in sync with xen.dtsi. Xen enables
the SMMU and makes use of it.
arm64: zynqmp: Uncomment iommus for zynqmp_dpdma and zynqmp_dpsub
The SMMU is disabled in device tree so this change has no impact.
The benefit is that this way it is in sync with xen.dtsi. Xen enables
the SMMU and makes use of it.
spi: cadence_qspi: Add support for ospi dual stacked
Add support for ospi dual stacked flash configuration.
Read "is-stacked" property from dt and export it to spi-nor framework
via slave->option. Based on the address/offset spi-nor framework will
populate CS through flags for read, write and erase functions.
configure chip select in cadence_qspi driver based on the flags
received from spi-nor framework.
mmc: zynq_sdhci: Use xilinx pm request instead of mmio_write call
Currently xilinx sdhci driver is using zynqmp_mmio_write() to set
tapdelay values. Use xilinx_pm_request() using appropriate arguments
to set input/output tapdelays for zynqmp. Where tapdelay setting is
done by firmware. Host driver should explicitly request DLL reset
before ITAP (assert DLL) and after OTAP (release DLL) to avoid issues
in some cases. Also handle error return where possible.
Signed-off-by: T Karthik Reddy <t.karthik.reddy@xilinx.com>
set_delay() function is from sdhci host ops, which does not return
any error case due to void return type. Where
arasan_sdhci_set_tapdelay() from arasan sdhci driver returns error
when there is a failure. So set return type to set_delay().
Signed-off-by: T Karthik Reddy <t.karthik.reddy@xilinx.com>
Simon Glass [Tue, 16 Feb 2021 00:08:11 +0000 (17:08 -0700)]
libfdt: Check for multiple/invalid root nodes
It is possible to construct a devicetree blob with multiple root nodes.
Update fdt_check_full() to check for this, along with a root node with an
invalid name.
CVE-2021-27097
Signed-off-by: Simon Glass <sjg@chromium.org> Reported-by: Bruce Monroe <bruce.monroe@intel.com> Reported-by: Arie Haenel <arie.haenel@intel.com> Reported-by: Julien Lenoir <julien.lenoir@intel.com>
State: upstream (124c255731c76a2b09587378b2bcce561bcd3f2d)
Simon Glass [Tue, 16 Feb 2021 00:08:09 +0000 (17:08 -0700)]
image: Adjust the workings of fit_check_format()
At present this function does not accept a size for the FIT. This means
that it must be read from the FIT itself, introducing potential security
risk. Update the function to include a size parameter, which can be
invalid, in which case fit_check_format() calculates it.
For now no callers pass the size, but this can be updated later.
Also adjust the return value to an error code so that all the different
types of problems can be distinguished by the user.
Signed-off-by: Simon Glass <sjg@chromium.org> Reported-by: Bruce Monroe <bruce.monroe@intel.com> Reported-by: Arie Haenel <arie.haenel@intel.com> Reported-by: Julien Lenoir <julien.lenoir@intel.com>
State: upstream (c5819701a3de61e2ba2ef7ad0b616565b32305e5)
Simon Glass [Tue, 16 Feb 2021 00:08:06 +0000 (17:08 -0700)]
fit: Don't allow verification of images with @ nodes
When searching for a node called 'fred', any unit address appended to the
name is ignored by libfdt, meaning that 'fred' can match 'fred@1'. This
means that we cannot be sure that the node originally intended is the one
that is used.
Disallow use of nodes with unit addresses.
Update the forge test also, since it uses @ addresses.
CVE-2021-27138
Signed-off-by: Simon Glass <sjg@chromium.org> Reported-by: Bruce Monroe <bruce.monroe@intel.com> Reported-by: Arie Haenel <arie.haenel@intel.com> Reported-by: Julien Lenoir <julien.lenoir@intel.com>
State: upstream (79af75f7776fc20b0d7eb6afe1e27c00fdb4b9b4)
Simon Glass [Tue, 16 Feb 2021 00:08:05 +0000 (17:08 -0700)]
fdt_region: Check for a single root node of the correct name
At present fdt_find_regions() assumes that the FIT is a valid devicetree.
If the FIT has two root nodes this is currently not detected in this
function, nor does libfdt's fdt_check_full() notice. Also it is possible
for the root node to have a name even though it should not.
Add checks for these and return -FDT_ERR_BADSTRUCTURE if a problem is
detected.
CVE-2021-27097
Signed-off-by: Simon Glass <sjg@chromium.org> Reported-by: Bruce Monroe <bruce.monroe@intel.com> Reported-by: Arie Haenel <arie.haenel@intel.com> Reported-by: Julien Lenoir <julien.lenoir@intel.com>
State: upstream (8a7d4cf9820ea16fabd25a6379351b4dc291204b)
Some ARMv7 boards using PSCI require to be in secure-mode when booted via
'bootz' or 'bootm'. During distro-boot 'bootefi bootmgr' is called to check
if booting via UEFI is possible.
With the change we change the switch from secure mode to non-secure mode is
moved from the UEFI subsystem setup to just before calling StartImage().
Cc: Jernej Škrabec <jernej.skrabec@gmail.com>
Reported by: Andre Przywara <andre.przywara@arm.com> Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
State: upstream (82d01f04facef1276cede067efd02d2a731ffe83)
* apply U-Boot's fix-ups to the a device-tree
* let U-Boot make memory reservations according to the device-tree
* install the device-tree as a configuration table
Up to now the bootefi command used the last file loaded to determine the
boot partition. This has led to errors when the fdt had been loaded from
another partition after the EFI binary.
Before setting the boot device from a loaded file check if it is a PE-COFF
image or a FIT image.
For a PE-COFF image remember address and size, boot device and path.
For a FIT image remember boot device and path.
If the PE-COFF image is overwritten by loading another file, forget it.
Do not allow to start an image via bootefi which is not the last loaded
PE-COFF image.
Currently when executing 'bootefi hello' we copy helloworld.efi to the
address identified by environment variable loadaddr. This is unexected
behavior for a user. There is no need to copy helloworld.efi before
executing it after relocation.
arm64: zynqmp: Fix application loading on R5 core1
From U-Boot, loading application on RPU core 0 is fine but loading on
core 1 is not handled properly. Lock-step mode needs both the R5 cores
to be initialized and it is working fine. Whereas in SPLIT mode individual
R5 cores needs to be initialized as they need to execute differenet
applications. Handle both these lock-step and split modes by propagating
mode and RPU core number(4 for RPU0 and 5 for RPU1) for various functions
and by adding conditions in those functions.
With the updates in framework, 'regulator-boot-on' property is not enough
to keep the regulator on, when no one is registered through the regulator
framework. We need to keep the outputs on, regardless of the consumers
as these ouptuts are controlled through a separate device called
'Sequencer' on these boards.
So, add 'regulator-always-on' which makes the num_users count to '1'
and prevents disabling the output during cleanup call of framework.
arm64: zynqmp: Add label to all GPIO lines for VCK190 SC
Add label to GPIO lines so the user-level applications can find any line
without knowing its physical path on System Controller on VCK190/VMK180.
These labels are describing EMIO gpio connection which depends on PL which
we normally don't describe but that's only way to go for now. Lately this
should be done out of this source code.
Signed-off-by: Saeed Nowshadi <saeed.nowshadi@xilinx.com> Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Michal Simek [Wed, 7 Apr 2021 15:08:13 +0000 (17:08 +0200)]
arm64: zynqmp: Separate SMK-K26 from SM-K26
Starter kit has sdhci0 disabled in HW that's why create separate it from
each other. Issue is if HW disable clk for this IP but IP is enabled in DT
u-boot is trying to access that regs and fails.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Now it results in 2 slave devices with same address (0x50)
on the I2C bus, making the communication un-reliable.
When ' i2c-mux-idle-disconnect' is in DT, after '1', the Mux
channel output is disconnected, making none of the channels
available to the I2C1. So, there is no question of having the
same addressed slave (0x50) present on the bus when we are doing '2'.
Michal Simek [Tue, 6 Apr 2021 12:52:58 +0000 (14:52 +0200)]
arm64: zynqmp: Squashed revB for k26 with revA
K26 has two variants with and without EMMC. SM-K26 is board with EMMC on it
and SMK-K26 is board without EMMC. The same DT can be used for both. The
only difference is if you can access that memory or not.
Compatible string is extended to also cover SMK versions of this board.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
projects / thirdparty / u-boot.git / log
