deletions | additions       diff --git a/related.tex b/related.tex  index a13956e..959680a 100644  --- a/related.tex  +++ b/related.tex 
...
McOSR~\cite{lameed2013modular} is a library for LLVM's legacy JIT compiler to insert OSR points at loop headers. When an OSR transition is fired, the live state is saved into a set of global variables (one per live variable) and a helper method is invoked to modify the IR of the function using a code transformer provided by the front-end and a copy of the original code saved as control version. The library generates a new entrypoint for the function to check a global condition and discriminate whether the function is being invoked through an OSR transition or a regular call: in the first case, values for live variables are read from the associated global variables, and the execution jumps to the block to resume the execution at. The SSAUpdater component of LLVM is then used to restore the SSA form after the update. When the helper method returns, the updated function is invoked and the OSR is thus performed in a new stack frame. When the updated function returns, a second helper method is called to recompile the updated function to remove the new entrypoint inserted for the OSR transition, as it can disrupt LLVM optimizations and lead to poorer performance on subsequent invocations of the function.  \paragraph{Related work}  Dynamic Software Updating (DSU) is a methodology for permitting programs to be updated while they run and is thus useful for systems that cannot afford to halt service. DSU techniques (e.g. ~\cite{neamtiu2006dsu, makris2009dsu}) ~\cite{neamtiu2006dsu,makris2009dsu})  are required to update all functions active on the call stack at the same time, so their code should be instrumented and involved data types wrapped to support future extensions. In tracing JIT compilers deoptimization techniques are used to safely leave an optimized trace when a guard fails and to continue the execution in the interpreter or with a different piece of code. SPUR~\cite{bebenita2010spur} is a trace-based JIT compiler for Microsoft's Common Intermediate Language (CIL) with four levels of JIT-ting: profiling, tracing, optimizing and transfer-tail; transfer-tail JIT is used to bridge the execution from an instruction in a block from tracing or optimizing mode to the next safe point for deoptimization to profiling code. In RPython guards are implemented as a conditional jump to a trampoline that analyzes resume information for the guard and executes compensation code to leave the trace. Resume data is compactly encoded by sharing parts of the data structure between subsequent guards~\cite{schneider2012rpython}; a similar approach is used in LuaJIT, where sparse snapshots are taken to enable state restoration when leaving a trace~\cite{luajit}.