tailcall analysis passes return_addr as first arg

Version and Platform (required):

• Binary Ninja Version: 3.2.3814
• OS: Windows 10
• OS Version: 22H2
• CPU Architecture x64

Bug Description:
When analysing functions with tailcalls the HLIL adds the return address as first arg (makes sense, as it's the top of the stack, but is wrong). I've seen it with 0 args, but also when there are multiple args it will look like this

func1(void* arg1, void* arg2, void* arg3)
do_things
other_things
func2(return_addr, arg1, arg2)

even when the underlying ASM doesn't touch the stack frame

Steps To Reproduce:
I can find some more detailed ones if you like

Expected Behavior:
Tailcalls should not pass return_addr as arg1

Screenshots:

Additional Information:

Start by reproducing the reported x64 tailcall case in HLIL and compare the generated call arguments with the underlying assembly. The issue does not name source files or tests, so locate the tailcall analysis entry point and any related tests first. Done means tailcalls no longer include the return address as the first argument.