Instead of always starting a new goroutine for the main goroutine, run
the main goroutine on the system stack.
The system stack is not occupied with scheduling, instead each goroutine
that wants to pause itself calls into the scheduler which will switch to
the next task (goroutine) to run, or sleeps.
There are various advantages of this over the previous system:
* When the program doesn't start a goroutine, the code size and RAM
consumption is close to what you'd get with `-scheduler=none`.
* When the program does start a goroutine, there is still a reduction
in RAM consumption because only one extra stack is needed.
* Because tasks directly switch to the next task to run, only a single
task switch is needed instead of two (goroutine -> scheduler ->
goroutine). This should improve task switching performance.
I kept the current behavior for WebAssembly/Asyncify. I looked into how
the same benefits can be realized for WebAssembly but couldn't easily
find how to do that. Maybe this can be done separately, or maybe we'll
just wait for the stack switching proposal to finish.
The code for Cortex-M is currently more complicated than I'd like, and
therefore can sometimes result in a slight increase in code size. I'd
like to fix this eventually but am still looking into good ways to do
this. I still think this change is generally beneficial because many
programs see big reductions in code size when compiling for Cortex-M.
For some reason, these aren't lowered when a generic function is
instantiated by the SSA package.
I've left unsafe.Offsetof to be implemented later, it's a bit difficult
to do correctly the way the code is currently structured.
Without this change, the compiler would probably have worked just fine
but the generated types would look odd.
You can see in the test case that it now doesn't use `main.Point` but
rather the correct `main.Poin[float32]` etc.
For write-only operations (in SPI displays for example), the transmit
speed is doubled with this relatively small change.
In the future, we should try to use DMA instead for larger buffers. But
this is already a significant improvement and will always be an
improvement for small buffer sizes.
These are reserved registers on MacOS so don't need to be clobbered, and
result in a compiler waring:
ld.lld-14: warning: inline asm clobber list contains reserved registers: X18, FP
Reserved registers on the clobber list may not be preserved across the asm statement, and clobbering them may lead to undefined behaviour.
This makes nrf51 consistent with nrf52 and other chips, which do provide
constants for hardware pin numbers.
I've also added the microbit to the smoketest because it is used on
play.tinygo.org. And removed PCA10040 and PCA10056 because they aren't
provided on play.tinygo.org anymore.
This makes it easier to move the TinyGo compiler between Linux versions
because it doesn't depend on any system libraries anymore. For example,
binaries should be able to run on old Linux versions and on
distributions without glibc (such as Alpine Linux).
Ayke van Laethem
Phil Kedy
Kenneth Bell
deadprogram
sago35
Dan Kegel
Federico G. Schwindt
Daniel Esteban