[LLVMdev] [question] IR<-->native program state mode switch how / where is it done

[ant. ts.]

Hi, all I'm new here so I hope I'm not breaching etiquette or posting in
the wrong list, if so please redirect me.

I'm a student studying the LLVM JIT-Interpreter source and i have a
question:

where and how is the switch/mapping from infinite register IR to native
arch dependent register set managed, to preserve the program state, during
jumps from the Interpreter to the JIT and vice versa?

all i can tell currently is that it happens in ExecutionEngine somewhere
using these methods, but i cant find their wrapper, the one that would use
these and say "now map everything and switch state" and somehow pass that
mapping to the JIT and backwards.
emitGlobals(), EmitGlobalVariable(),
llvm::JIT::getOrEmitGlobalVariable()
llvm::ExecutionEngineState::getGlobalAddressMap(),
llvm::ExecutionEngineState::getGlobalAddressReverseMap()

I hope I'm not misguided by an incorrect understanding of the JIT
architecture, in any case any links to documentation, Doxygen tree, etc are
absolutely welcome

thanks,
(and thanks for all of LLVM actually :) )
Antonis
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20120603/eb0b765a/attachment.html>