stable-symbolic: Properly initialize immutable malloc'd memory

[RyanGlScott]

Previously, stable-symbolic would only initialize mutable memory with symbolic bytes when calling malloc. malloc can also be used to initialize immutable memory, however, so we must also account for this possibility when stable-symbolic is enabled.

To this end, we create a new doConstStoreStableSymbolic function that is like doStoreStableSymbolic, but does not check if the memory being written to is mutable. (This is very much in the spirit of the existing doArrayStore/doArrayConstStore, storeRaw/storeConstRaw, and mallocRaw/mallocConstRaw split.) We then pick between doStoreStableSymbolic and doConstStoreStableSymbolic in the override for malloc depending on whether the memory is mutable or immutable, respectively.

See GaloisInc/saw-script#1691 for the motivation behind this bugfix.

[RyanGlScott]

There is a code organization decision to make here: should we have separate doStoreStableSymbolic and doConstStoreStableSymbolic functions, or should we combine them into a single function that distinguishes between the two based on a Mutability argument? We very well could do the latter, since the implementations of the two functions are very similar. On the other hand, the existing style in this module seems to be the former, since we already have separate doArrayStore/doArrayConstStore, storeRaw/storeConstRaw, and mallocRaw/mallocConstRaw functions. For this reason, I chose the former convention here, but do let me know if you'd prefer the latter instead.

[robdockins]

I think having separate functions is deliberate. The Const variants are only ever intended to be used in initialization, and the intention is to make it hard to accidentally chose the one that is abstraction-breaking (can write to immutable memory).

[RyanGlScott]

Adding a test case for this is surprisingly tricky, since the failure observed in GaloisInc/saw-script#1691 seemingly only arises during execution. The crucible-llvm test suite, on the other hand, adds the relevant assertions from doMalloc but doesn't go as far as executing them. We could, I suppose, modify the crucible-llvm test suite to invoke executeCrucible somewhere, but I'm unclear if the additional effort is worth it in this case.