The new option is `NoRegisterArgumentsCanComeAfterStackOnes`
If it's set to `true`, then no register argument can follow any stack
argument under the said abi, otherwise such "shuffling" of arguments
is allowed.
This makes the conversions more robust, since everything can rely on
the same tools for filling in missing registers and ordering them based
on the ABI requirements.
This introduces `emplace` and `emplace_or_assign` members
allowing population of the container without copying values in.
It also reworks the batch insertion. On top of adding the ability
to move the elements in (without making a copy) and constructing
them in place, it also makes the non-assign version of the inserter
a lot safer by explicitly asserting the fact that while it was
in use no duplicates were added to the vector. Before now,
the duplicates were just silently removed from the container.
It also optimizes the batch insertion somewhat by using non-stable
`std::sort` for unique case as well as replacing `std::unique`
invocation with `std::find_adjacent`.
* Let GraphQL accept zero or more parameters for analyses
* Handle the general case of analyses using zero or more containers with
a variable ammount of targets
`getUniqueString` returns a dedup'd char array `GlobalVariable` given a
string. We used to rely on named metadata to dedup them. However, this
lead to issues with `LLVMContainer` cloning and linking.
The new implementation uses the name of the variable as the
de-decuplication key, circumventing the problem entirely and improving
performance.
We used to turn `internal` linkage to the equivalent of C++'s inline.
We need to do the same for `private`. If we don't do this, the number of
`private` variables grows exponentially.
This commit introduces `BasicBlockID` as the unique identifier for a
`efa::BasicBlock` into the CFG. A `BasicBlockID` is defined by a
`MetaAddress` plus an incremental integer. This enables us to have
multiple instances of the same block in a single function, which is
particularly useful when inlining multiple times the same function.
Apart from this, the commit also does the following:
* It drops representing `MetaAddress`es a `structs` in the IR. This created
several issues related to ABI. We now represent them as strings.
* It defines more functions in `support.h`, instead of defining prototypes
by hand in `CodeGenerator.cpp` and the like. Specifically, `unknownPC`
and `raise_exception_helper`. We also introduce a C "constructor" for
`PlainMetaAddress`.
* It significantly reduces the API of `GeneratedCodeBasicInfo`, which
was supposed to be put on a diet since a long time. Specifically,
many jump target related methods have been moved to free functions in
`IRHelpers.h`. Also `GCBI::getSuccessors` has been pushed into its
only user, `PruneRetSuccessors`, to prevent further usage of a
deprecated API. In the future, it would be nice to drop it entirely.
* It introduces `efa::BasicBlock::InlinedFrom`.
* Introduce an enum to represent named argument indices for `newpc`.
This enables us to more effectively manipulate its argument list.
* It improves the verification and error reporting for
`efa::FunctionMetadata`.
* Update tests.
This commit is preliminary to another piece of work to improve the
generality of inlining beyond the simple "fake function" scenario, for
which the feature was originally conceived.
TrackingContainer is a wrapper around KeyedObjectContainers (notably,
SortedVector) intended to figure out which elements of the underlying
container have been accessed.
Ivan Krysak