This patch does following things.
1. Added Cache support.
2. Modified TXD_DEPTH to 64.
3. Enabled RXEMPTY interrupt.
4. Modified the param passed to XQspiPsu_ReadRxFifo API to
RxThreshold*4, since it is expecting number of bytes to read.
5. Used Qspipsu Instance buffers in Interrupt handler rather
than Msg buffers.
Signed-off-by: P L Sai Krishna <lakshmis@xilinx.com>
This patch removes the status register bit definations
that are not available on the RTL.
Signed-off-by: Venkata Naga Sai Krishna Kolapalli <venkatan@xilinx.com>
Removed the configuration of DMA Buffer Boundary register in WriteSpareBytes API
as the driver uses oob size as pkt size and thus configuring DMA Buffer
Boundary Register has no effect.
Signed-off-by: Shakti Bhatnagar <shaktib@xilinx.com>
Since the check for the block to have reached the end block was at the end of the loop,
thus assert call triggering could happen if the block crosses the max block limit,
making the driver to hang. Thus to fix the triggering of assert call,
the check for the block to have reached the max block is being done at the start of
the loop in XNandPsu_CalculateLength function.
The u32 typecasting for the variable OffsetVar has been removed and the data type for
the variable TempLen has been changed to u64 as for higher block numbers the value for
these variables will require u64 datatype otherwise corrupted value will get stored
which will lead to wrong calculation of block.
Signed-off-by: Shakti Bhatnagar <shaktib@xilinx.com>
Corresponding fields in the devcfg.STATUS register are written to,
for clearing DMA done count.
Signed-off-by: Sarat Chand Savitala <saratcha@xilinx.com>
These changes were present in master-sdk-beta2 and also need to be
applied to master as well.
Changes done are:
- Macro name change in QSPI for PSU naming change
- Workaround for QEMU in QSPI32 dummy mode
Signed-off-by: Sarat Chand Savitala <saratcha@xilinx.com>
This patch initialises RSAKeyReadback value with zero's since
if RSA key is read with XSK_EFUSEPS_ENABLE_RSA_KEY_HASH as FALSE
then it will return zero.
Signed-off-by: P L Sai Krishna <lakshmis@xilinx.com>
This patch adds call to check the supported OS platform when
application is Zynq MP FSBL.
This enables correct error message to be displayed when trying
to create FSBL project with OS platform other than Standalone.
Signed-off-by: Sarat Chand Savitala <saratcha@xilinx.com>
This patch modifies files to fix warnings that got
generated when -Wextra flag was used.
Signed-off-by: Venkata Naga Sai Krishna Kolapalli <venkatan@xilinx.com>
This patch deprecates the old v2.0 version and created
new v2.1 version for xilskey library.
Signed-off-by: Venkata Naga Sai Krishna Kolapalli <venkatan@xilinx.com>
Removing compiler optimization level and debug level flags from FSBL tcl file.
These flags can be passed on as arguments while compiling using make.
O2 optimization will be required while building FSBL to reduce the
size of FSBL and hence to accomodate ATF in OCM.
Signed-off-by: Sarat Chand Savitala <saratcha@xilinx.com>
Removed the workaround for qemu in driver code as qemu now supports
ecc and bad block table handling.
Signed-off-by: Shakti Bhatnagar <shaktib@xilinx.com>
Modified the number of bytes of data to be read for signature and version when
NO OOB option is enabled.
Signed-off-by: Shakti Bhatnagar <shaktib@xilinx.com>
Modified number of bytes to be read for bbt in oob and bbt with signature
and version for no oob option. This change is done kept in mind that it
should support flashes with small page size as well.
Signed-off-by: Shakti Bhatnagar <shaktib@xilinx.com>
For file system based devices, flash offset address
(location of image) should be independednt of multiboot
and should always be 0.
Signed-off-by: Sarat Chand Savitala <saratcha@xilinx.com>
Setting BBT option at the beginning and checking afterwards to set
the offset for signature and version.
Signed-off-by: Shakti Bhatnagar <shaktib@xilinx.com>
Performing Read modify write operation for the interrupt status enable
and interrupt status register in readpage function instead of directly
writing to fix the masking error interrupts in case ecc enabled.
Signed-off-by: Shakti Bhatnagar <shaktib@xilinx.com>
If oob size is less than 64 bytes, ecc position should start from 0 position
of oob area, else should be at the end of the oob area.
Keeping ecc start position in alligned with the Linux nand base ecc layout.
Signed-off-by: Shakti Bhatnagar <shaktib@xilinx.com>
This patch modifies the code as per the opensource git changes and
updates the license information
Signed-off-by: Kinjal Pravinbhai Patel <patelki@xilinx.com>
This patch modifies the code as per the opensource git changes and
updates the license information
Signed-off-by: Kinjal Pravinbhai Patel <patelki@xilinx.com>
This patch modifies the code as per the opensource git changes and
updates the license information
Signed-off-by: Kinjal Pravinbhai Patel <patelki@xilinx.com>
This patch modifies the code according to wendy's openamp changes
and also updates the license information
Signed-off-by: Kinjal Pravinbhai Patel <patelki@xilinx.com>
This patch deletes the *.mss file and modified the tcl file
to prevent the error in SDK log when any other processor or OS
is selected in create application project GUI of SDK
Signed-off-by: Kinjal Pravinbhai Patel <patelki@xilinx.com>
QSPI HW doesn't actually care what the SPI mode is for the dummy phase
of flash commands. We have confirmed this with randomized testing.
However due to a core limitation of QEMU it is expected to match the
mode of the address phase for QEMU.
So since QEMU is the only platform that cares, set the dummy mode to
QEMUs expectation.
Signed-off-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
QSPI HW doesn't actually care what the SPI mode is for the dummy phase
of flash commands. We have confirmed this with randomized testing.
However due to a core limitation of QEMU it is expected to match the
mode of the address phase for QEMU.
So since QEMU is the only platform that cares, set the dummy mode to
QEMUs expectation.
Signed-off-by: Peter Crosthwaite <peter.crosthwaite@xilinx.com>
Provided support to pass handoff parameters to ATF
which it will use to load further partitions.
Provided support to load PMU firmware from FSBL.
Signed-off-by: Sarat Chand Savitala <saratcha@xilinx.com>
This patch modifies the eeprom page size to 32 for zynq ultrascale MP
in polled and interrupt examples.
Signed-off-by: P L Sai Krishna <lakshmis@xilinx.com>
This patch modifies the baremetal code to include ipi related code.
It moves the baremetal specific platform API from platform.c to
baremetal.c. It also moves the platform specific information from
baremetal.h to platform.h. The patch also updates license information
in linker script
Signed-off-by: Kinjal Pravinbhai Patel <patelki@xilinx.com>
