As pointed out by Şükrü Bahadır Arslan in PR765, function calls
inside DEBUGASSERT() will not be executed if DEBUGASSERT is
disabled. Following that PR, I searched for other instances of
similar errors and found four. As there are many thousands of
DEBUGASSERT in the code, this is *not* an exhaustive fix.
arch/arm/src/tms570/tms570_boot.c:
* In function arm_boot(), call tms570_memtest_complete()
outside of the DEBUGASSERT(). Otherwise, if DEBUGASSERT is
disabled, we will never rendezvous with completion of the
test. (Two instances of this fix.)
arch/arm/src/tms570/tms570_clockconfig.c:
* In function tms570_clockconfig(), call
tms570_efc_selftest_complete() outside of the DEBUGASSERT()
for the same reason as above.
boards/arm/sam34/sam4s-xplained-pro/src/sam_boot.c:
* In function board_late_initialize(), call
sam_watchdog_initialize() outside of the DEBUGASSERT() for
the same reason as above.
Modified boards/z80/z180/p112/README.txt and boards/z80/z80/z80sim/README.txt:
IMPORTANT NOTE as of 2020-4-11: Support for CONFIG_CAN_PASS_STRUCTS was removed in NuttX-9.1. This was necessary to enforce some POSIX interface compliance but also means that ALL older SDCC versions will no long build with NuttX. I have been told that the newest SDCC compilers can indeed pass structure and union parameters and return values. If that is correct, then perhaps the newer SDCC compilers will be used. Otherwise, it will be necessary to use some other, more compliant compiler.
So call 'make EXTRAFLAGS=-Wno-cpp' could suppress the warnings with pre-processor
directive #warning in GCC.
Change-Id: Iaa618238924c9969bf91db22117b39e6d2fc9bb6
Signed-off-by: liuhaitao <liuhaitao@xiaomi.com>
* boards/arm/stm32/nucleo-f429zi: Initial port to nucleo-f429zi board.
* boards/arm/stm32/nucleo-f429zi: Add two demo configs.
* Add the nucleo-f429zi to boards/Kconfig
boards/Kcondig: Fixed line alignment issue.
* boards/arm/stm32/nucleo-f429zi: Change the license header of all new files to an Apache 2.0 license.
* boards/arm/stm32/nucleo-f429zi: nxstyle fixed
Resolution of Issue 619 will require multiple steps, this part of the first step in that resolution: Every call to nxsem_wait_uninterruptible() must handle the return value from nxsem_wait_uninterruptible properly. This commit is only for those files under fs/driver, fs/aio, fs/nfs, crypto/, and boards/.
Please note: The modified file under fs/nfs generates several " Mixed case identifier found" errors. Please ignore these. These cannot be fixed without changes to numerous other files. They also follow a non-standard convention that is used many files: Using lower case structure names in custom SIZEOF_ definitions.
Namely these defconfig files:
boards/sim/sim/sim/configs/mtdpart/defconfig
boards/sim/sim/sim/configs/mtdrwb/defconfig
boards/sim/sim/sim/configs/rpproxy/defconfig
boards/sim/sim/sim/configs/rpserver/defconfig
boards/sim/sim/sim/configs/tcpblaster/defconfig
Because the recent versions of macOS is 64-bit only and thus
incompatible with CONFIG_SIM_M32=y.
The following defconfig files are left intact as these configs
are important for them:
boards/sim/sim/sim/configs/loadable/defconfig
boards/sim/sim/sim/configs/module32/defconfig
Because:
* A raw binary doesn't make sense for sim, where ./nuttx is
a host OS executable.
* It breaks test builds on macOS, where native objcopy
is not available.
* The appropriate size of stack varies among archs.
E.g. for 64-bit sim, 2048 is way too small, especially when the task
happens to use host OS functionalities.
I plan to allow an arch provide its own default.
* I plan to use this to replace hardcoded "STACKSIZE = 2048" in APPDIR.
This reverts commit b9ace36fcc.
This change was added by PR 625 but has a serious logic flaw. It removes all occurrences of INCDIROPT and replaces it with a definition in tools/Config.mk:
else ifeq ($(WINTOOL),y)
DEFINE = "$(TOPDIR)/tools/define.sh"
INCDIR = "$(TOPDIR)/tools/incdir.sh" -w
This logic flaw is the Config.mk is included in all Make.defs files BEFORE WINTOOL is defined. As a result, the definition is wrong in many places when building under Cygwin with a Windows native toolchain.
Necessary for dlfcn etc on ESP32, which has separate memory regions
for instruction and data.
known issues/todo
* consider something similar to dual heaps for PROTOECTED
* consider to adapt binfmt as well
