There's no need to have these as separate functions, they just take up
unnecessary code space and combining them allows to factor common code, and
also allows to support arbitrary string descriptor indices.
The routine waits for the DMA to finish, which is signalled from a DMA IRQ
handler. Using WFI makes the CPU sleep while waiting for the IRQ to arrive
which decreases power consumption. To make it work correctly the check for
the change in state must be atomic and so IRQs must be disabled during the
check. The key feature of the Cortex MCU that makes this possible is that
WFI will exit when an IRQ arrives even if IRQs are disabled.
CPython doesn't allow SEEK_CUR with non-zero offset for files in text mode,
and uPy inherited this behaviour for both text and binary files. It makes
sense to provide full support for SEEK_CUR of binary-mode files in uPy, and
to do this in a minimal way means also allowing to use SEEK_CUR with
non-zero offsets on text-mode files. That seems to be a fair compromise.
Build and test 32bit and 64bit versions of the windows port using gcc
from mingw-w64. Note a bunch of tests which rely on floating point
math/printing have been disabled for now since they fail.
all: Add .frequency read-only property for busio.SPI to return actual frequency.
Fix esp8266/posix_helpers.c, which was not up to date for the new
long-lived/short-lived heap allocation scheme.
There have been multiple uGame prototypes with differences in the pins
and other details. This updates the board definition to fit the final
uGame 10 design.
The parser attempts to allocate two large (~512 byte) chunks up
front. If it couldn't, then it would error out. This change will
cause it to try allocating half the previous attempt until its down
to two copies. This is ok upfront because later code checks bounds
and tries to extend the allocation if needed.
This adapts the allocation process to start from either end of the heap
when searching for free space. The default behavior is identical to the
existing behavior where it starts with the lowest block and looks higher.
Now it can also look from the highest block and lower depending on the
long_lived parameter to gc_alloc. As the heap fills, the two sections may
overlap. When they overlap, a collect may be triggered in order to keep
the long lived section compact. However, free space is always eligable
for each type of allocation.
By starting from either of the end of the heap we have ability to separate
short lived objects from long lived ones. This separation reduces heap
fragmentation because long lived objects are easy to densely pack.
Most objects are short lived initially but may be made long lived when
they are referenced by a type or module. This involves copying the
memory and then letting the collect phase free the old portion.
QSTR pools and chunks are always long lived because they are never freed.
The reallocation, collection and free processes are largely unchanged. They
simply also maintain an index to the highest free block as well as the lowest.
These indices are used to speed up the allocation search until the next collect.
In practice, this change may slightly slow down import statements with the
benefit that memory is much less fragmented afterwards. For example, a test
import into a 20k heap that leaves ~6k free previously had the largest
continuous free space of ~400 bytes. After this change, the largest continuous
free space is over 3400 bytes.
