Solves a problem when building in strict C++ mode, as GCC defines
uint32_t as unsigned long on systems where sizeof(long) == 4. This
causes GCC to complain that uint32 != uint32_t as unsigned int !=
unsigned long.
Signed-off-by: Michael Hope <michaelh@juju.net.nz>
Since toolchains other then older Code Sourcery (Mentor Graphics)
might not use the arm-none-eabi triplet, make it possible to use
any triplet by just setting the CROSS_COMPILE environment variable
accordingly.
Signed-off-by: Daniel Nyström <daniel@nystrom.st>
caddr_t is obsolete in POSIX and thus unavailable in most modern
toolchains. caddr_t usage should be replaced by void *.
stddef.h includes the size_t typedef which was missing.
Signed-off-by: Daniel Nyström <daniel@nystrom.st>
This resolves issues related to sending full (64B) packets via USB
2.0. In this case, some hosts continue to expect more data. Add
infrastructure for sending 0-byte packets to signal end of
transmission, and use it in SerialUSB.
Signed-off-by: Marti Bolivar <mbolivar@leaflabs.com>
- Check if we're transmitting (instead of how many bytes are left
untransmitted) before bailing, in case we're transmitting a
zero-length packet.
- Set transmitting=1 before setting the endpoint valid to avoid races
with the USB interrupt.
- Eliminate some duplicated code.
Signed-off-by: Marti Bolivar <mbolivar@leaflabs.com>
Added a flag to see if we are currently waiting on an interrupt to acknowledge the sending of the current
IN packet.
Added a method usb_cdcacm_is_transmitting() to check for that flag.
Signed-off-by: Manuel Odendahl <wesen@ruinwesen.com>
I am not sure why this would work for most DMA transfers but I ran into trouble when doing
SDIO DMA.
Signed-off-by: Manuel Odendahl <wesen@ruinwesen.com>
It is possible to receive empty bulk transfers. This would lead to the endpoint never being reenabled as this was handled by the read routine. This caused the USB to hang.
Signed-off-by: Manuel Odendahl <wesen@ruinwesen.com>
stm32f1/performance: load stm32.h in order to have STM32_HIGH_DENSITY when necessary
STM32_HIGH_DENSITY was not defined in isrs.S and vector_table.S for a high density STM32. This was because stm32.h where the flag is now defined (in contrast to being defined on the command line in an earlier version of libmaple) was not included. This led to crashes when using one of the high density peripherals.
Updated the ASFLAGS to have the correct include path in rules.mk.
Include stm32.h in the assembler files.
Signed-off-by: Manuel Odendahl <wesen@ruinwesen.com>
Add libraries directory to global include path, so Wire sub-headers
can be included in a directory-independent way.
Tweak Wire global declaration and definition so it builds.
Signed-off-by: Marti Bolivar <mbolivar@leaflabs.com>
This causes the warning to be emitted at most once. As is, it can be
emitted each time HardwareSPI.h is included, which is annoying.
Signed-off-by: Marti Bolivar <mbolivar@leaflabs.com>
This is a developer's doc explaining how interrupts are handled in
libmaple. It's time we exposed this information to enable people who
know what they're doing to move around more easily.
Signed-off-by: Marti Bolivar <mbolivar@leaflabs.com>
This reverts commit 715a63dccb. It fails
to build using the 2010 CodeSourcery toolchain, which is still the latest
officially supported version.
Signed-off-by: Marti Bolivar <mbolivar@leaflabs.com>
Move magic bytes from stack to Flash. De-magic-number sizeof magic
byte buffer. Use uintptr_t instead of unsigned int when casting
address of wait_reset().
Based on a patch by Michael Hope.
Signed-off-by: Marti Bolivar <mbolivar@leaflabs.com>