In Mbed OS, page size is program unit, which is different than FMC definition.
After fixing page size, we can pass NVSTORE test (mbed-os-features-nvstore-tests-nvstore-functionality).
- Move macro definition to mbed_toolchain.h
- Remove double underscores from macro which are reserved.
- Fix macro for IAR until compiler flags to disable path are added again.
The use of __FILE__ macro to get a usable identifier from the driver path
causes the path of the file to be stored in the .text region of the binary.
Given that this remains for the entire duration of the program, storing a
pointer to this string as an identifier is more efficient than copying the
contents of the string during lookup/insertion.
Sleep manager tracing strips the path from filenames and uses the result as an
identifier to track drivers that unlock/lock sleep tracing. Replace the function
that strips the path from the string, replace this function with a new macro,
__FILENAME__ which performs the same action in a compiler specific manner.
- GCC_ARM, use __builtin_strrchr which is optimized out at compile time.
- ARM, use __MODULE__ which returns the filename without path.
- IAR, specifiy the --no_path_in_file_macros compiler flag.
The atomic functions preserve volatile semantics - they only perform the
accesses specified. Add the volatile qualifier to the value pointer to
reflect this. This does not change existing caller code - it's
equivalent to adding a const qualifier to indicate we don't write to
a pointer - it means people can pass us qualified pointers without
casts, letting the compile check const- or volatile-correctness.
This is consistent with C11 <stdatomic.h>, which volatile-qualifies its
equivalent functions.
Note that this useage of volatile has nothing to do with the atomicity -
objects accessed via the atomic functions do not need to be volatile.
But it does permit these calls to be used on objects which have been
declared volatile.
The volatile qualifier on the __LDREX/__STREX prototypes only means that
it's safe to use them on volatile objects. Doesn't mean you actually
have to pass them volatile pointers.
Adding the volatile is a bit like doing strlen((const char *) ptr)
because you've got a non-const pointer.
This patch changes the linker files and defines to use the ZBT SSRAM
instead of the FPGA Block RAM for code and data.
The section 4.1.1, Code and RAM memory map, of the CM3DS Eval RTL and
testbench user guide explains the available memories.
This switch improves code memory from 256 kB to 4 MB and data memory
from 128 kB to 4 MB.
However, the ZBT SSRAM1 for code memory begins at 0x00400000 while the
processor can only boot at address 0x00000000 which means that it
expects the vector table to be at that address. That is why we have to
create 2 load regions in the linker scripts: one with only the vector
table at address 0x0 and one with code + data at address 0x00400000.
Because of these two load regions, linker will produce different
behaviours:
* GCC_ARM and IAR will only create 1 binary with both load regions
padding with 0 in between. The binary will then be very large (at least
4 MB) and the flash process will take longer.
* ARM and ARMC6 will create 2 binaries for the two load regions. The
load addresses of the two binaries can be written in the images.txt file
on the MPS2 board. You can also use the --bincombined option of fromelf
utility to produce only 1 large binary.
This patch also adds the memory_zones.h file to try to put in common all
the memory addresses that were previously hard coded in the linker
scripts / startup files.
With that patch in, the simplest option is to directly use the .elf file
with the MPS2, which is only possible with mbb_v225.ebf and more recent
firmwares. It will now be the default for CM3DS.
This commit works with greentea thanks to the now merged pull request
ARMmbed/htrun#181 in order to copy .elf file to the MPS2 board.
Signed-off-by: Hugues de Valon <hugues.devalon@arm.com>
Mbed-os 5.4.7 was the last unofficial working support for this target.
Since Mbed-os 5.6.0, the support is now official and VK_RZ_A1H is now "codebase aligned" with GR_PEACH (RZ_A1H) & GR_LYCHEE (RZ_A1LU) !
Like most of the lfs_dir_t functions, lfs_dir_append is responsible for
updating the lfs_dir_t struct if the underlying directory block is
moved. This property makes handling worn out blocks much easier by
removing the amount of state that needs to be considered during a
directory update.
However, extending the dir chain is a bit of a corner case. It's not
changing the old block, but callers of lfs_dir_append do assume the
"entry" will reside in "dir" after lfs_dir_append completes.
This issue only occurs when creating files, since mkdir does not use
the entry after lfs_dir_append. Unfortunately, the tests against
extending the directory chain were all made using mkdir.
Found by schouleu
When exporting to a uvision project, the include flags are not put in
the assembly compilation line. When assembling the files containing
includes, the search path will then fail. This patch adds the include
paths to the Assembly sequence, as it is done for compilation.
This issue was found in the pull request ARMmbed/mbed-os#6168.
Signed-off-by: Hugues de Valon <hugues.devalon@arm.com>
When exporting a mbed project to make_armc5, the include options (-I)
of the ASM flags are not pointing to the good folder. It should be
pointing to the root mbed-os folder and not the one in BUILD.
This issue was found in the pull request ARMmbed/mbed-os#6168.
Signed-off-by: Hugues de Valon <hugues.devalon@arm.com>
- sector size is 0x800 bytes
- writeable unit size is 0x8 bytes
- flash start address is 0x0
- total ADuCM3029 on chip flash size is 0x40000 bytes
- total ADuCM4050 on chip flash size is 0x7F000 bytes
Martin Kojtal
zzw
Antti Yli-Tokola
adbridge
Cruz Monrreal II
ccli8
Jimmy Brisson
Steven Cartmell
jeromecoutant
Kevin Bracey
Christopher Haster
Amanda Butler
Hugues de Valon
Tony Wu
mbedNoobNinja
bcostm
Mika Leppänen
Mudassar Hussain
Deepika
Mirela Chirica
li-ho
Bartek Szatkowski