All instructions with associated encodings are now annotated with
encoding information in a column before the code.
When write_function() is givan a TargetIsa reference, the annotations use
ISA-specific names. Otherwise everything is numeric.
The # is a more conventional prefix for hexadecimal, and when ISA
information is not available, there may be a decimal number in front
which would be confusing.
So prefer [1#10c] for the ISA-less encoding format. Here '1' is decimal
and '#10c' is hexadecimal.
Add a `needs_isa()` method to the SubTest trait, and pass a TargetIsa
trait object to those sub-tests that request it.
When multiple sub-tests and ISAs are specified, test the cross product.
If a sub-test requires an ISA, but none are specified, fail the test. In
the future, it may be a good idea to generate a default set of ISAs and
test against those.
Some tests are only applicable to specific ISAs. This can be indicated
with an ISA specification:
test legalizer
isa riscv
function foo() {
....
}
The ISA specifications have the same format as the test lines: The name
of the ISA following by optional settings until the end of the line.
Also parse `set` commands mixed in with the `isa` commands. These are
used to set ISA-independent settings as defined in
meta/cretonne/settings.py.
When constructing the Flags object from the Builder, don't consume it,
but take a reference instead.
This makes it possible for the parser to accept multiple 'set' lines and
apply them to different ISA specifications.
Use the source map to track instruction locations instead.
The rewrite methods now take an AnyEntity argument as the location to
use for errors. This means that bad EBB references in jump tables are
now reported correctly.
The slow tests are computed as those that would be printed as outliers
on a boxplot of all the test runtimes. These are more than 1.5
inter-quartile range away from the 75% quartile.
Spin up one worker thread per cpu, and run filetests on all of them. Use
a reorder buffer in Runner to make sure results are still reported in
order.
Individual test files given as command line arguments are still run
synchronously for easier debugging. Only directories are run on worker
threads. The recursive directory traversal is still happening on the
main thread.
Use a heartbeat thread to send ticks on the reply channel every second,
and use the ticks to detect tests that are stuck. When
Receiver::recv_timeout() is stabilized, we can probably get rid of the
heartbeat thread.
Catch panics on the worker threads and report them as test failures.
Always keep PathBufs for every entry in the test list. When concurrent
testing is enabled, we'll want to clone the path for the worker threads.
Remove the Job struct for the same reason.
The main libcretonne crate should not have any external dependencies if
at all possible. Use simple substring matching instead of regular
expressions in the verifier tests to achieve this.
The tools crate no longer depends directly on glob and regex. It still
has an indirect dependency on regex through libfilecheck.
This test runs the verifier on each function and matches the resulting
verifier error against the "error:" annotation.
Move the existing verifier test into filetests/verifier/ and use the new
syntex.
The verifier reports the 'location' of an error message as a reference
to the entity that has a problem. This uses the 'AnyEntity' type to
refer to instructions/values/ebbs etc.
Also add an err! macro similar to the one used by the parser.
This makes it possible to refer to entities defined in the source file,
using the source names prefixed with $.
For example, $v20 refers to the value by that name in the sources, even
if it was renumbered to 'vx0' in the parsed file.
Jakob Stoklund Olesen