Subversion Repositories openrisc

What's LRS?
===========

LRS, or Live Range Splitting is an optimization technique which allows
a user variable to reside in different locations during different parts
of a function.

For example, a variable might reside in the stack for part of a function
and in a register during a loop and in a different register during
another loop.

Clearly, if a variable may reside in different locations, then the
compiler must describe to the debugger where the variable resides for
any given part of the function.

This document describes the debug format for encoding these extensions
in stabs.

Since these extensions are gcc specific, these additional symbols and
stabs can be disabled by the gcc command option -gstabs.


GNU extensions for LRS under stabs:
===================================


range symbols:
-------------

    A range symbol will be used to mark the beginning or end of a
    live range (the range which describes where a symbol is active,
    or live).  These symbols will later be referenced in the stabs for
    debug purposes.  For simplicity, we'll use the terms "range_start"
    and "range_end" to identify the range symbols which mark the beginning
    and end of a live range respectively.

    Any text symbol which would normally appear in the symbol table
    (eg. a function name) can be used as range symbol.  If an address
    is needed to delimit a live range and does not match any of the
    values of symbols which would normally appear in the symbol table,
    a new symbol will be added to the table whose value is that address.

    The three new symbol types described below have been added for this 
    purpose.

    For efficiency, the compiler should use existing symbols as range
    symbols whenever possible; this reduces the number of additional
    symbols which need to be added to the symbol table.

    
New debug symbol type for defining ranges:
------------------------------------------

    range_off - contains PC function offset for start/end of a live range.
                Its location is relative to the function start and therefore 
                eliminates the need for additional relocation.

    This symbol has a values in the text section, and does not have a name.

            NOTE: the following may not be needed but are included here just 
                in case.
            range - contains PC value of beginning or end of a live range
                (relocs required).

            NOTE: the following will be required if we desire LRS debugging
                to work with old style a.out stabs.
            range_abs - contains absolute PC value of start/end of a live 
                range.  The range_abs debug symbol is provided for 
                completeness, in case there is a need to describe addresses 
                in ROM, etc.


Live range:
-----------

    The compiler and debugger view a variable with multiple homes as
    a primary symbol and aliases for that symbol.  The primary symbol
    describes the default home of the variable while aliases describe
    alternate homes for the variable.

    A live range defines the interval of instructions beginning with
    range_start and ending at range_end-1, and is used to specify a
    range of instructions where an alias is active or "live".  So,
    the actual end of the range will be one less than the value of the
    range_end symbol.

    Ranges do not have to be nested. Eg. Two ranges may intersect while 
        each range contains subranges which are not in the other range.

    There does not have to be a 1-1 mapping from range_start to
        range_end symbols.  Eg. Two range_starts can share the same 
        range_end, while one symbol's range_start can be another symbol's
        range_end.

    When a variable's storage class changes (eg. from stack to register,
        or from one register to another), a new symbol entry will be
        added to the symbol table with stabs describing the new type,
        and appropriate live ranges refering to the variable's initial
        symbol index.

    For variables which are defined in the source but optimized away,
        a symbol should be emitted with the live range l(0,0).

    Live ranges for aliases of a particular variable should always
        be disjoint.  Overlapping ranges for aliases of the same variable
        will be treated as an error by the debugger, and the overlapping
        range will be ignored.

    If no live range information is given, the live range will be assumed to 
        span the symbol's entire lexical scope.


New stabs string identifiers:
-----------------------------

  "id" in "#id" in the following section refers to a numeric value.

  New stab syntax for live range:  l(<ref_from>,<ref_to>) 

    <ref_from> - "#id" where #id identifies the text symbol (range symbol) to
                use as the start of live range (range_start).  The value for 
                the referenced text symbol is the starting address of the 
                live range.

    <ref_to> - "#id" where #id identifies the text symbol (range symbol) to
                use as the end of live range (range_end).  The value for 
                the referenced text symbol is ONE BYTE PAST the ending 
                address of the live range.


  New stab syntax for identifying symbols.

    <def> - "#id="

            Uses:
              <def><name>:<typedef1>...
                  When used in front of a symbol name, "#id=" defines a
                  unique reference number for this symbol.  The reference
                  number can be used later when defining aliases for this
                  symbol.
              <def>
                  When used as the entire stab string, "#id=" identifies this 
                  nameless symbol as being the symbol for which "#id" refers to.


    <ref> - "#id" where "#id" refers to the symbol for which the string 
                "#id=" identifies.
            Uses:
              <ref>:<typedef2>;<liverange>;<liverange>...
                  Defines an alias for the symbol identified by the reference
                  number ID.
              l(<ref1>,<ref2>)
                  When used within a live range, "#id" refers to the text 
                  symbol identified by "#id=" to use as the range symbol.

    <liverange> - "l(<ref_from>,<ref_to>)" - specifies a live range for a 
                symbol.  Multiple "l" specifiers can be combined to represent 
                mutiple live ranges, separated by semicolons.




Example:
========

Consider a program of the form:

    void foo(){
      int a = ...;
      ...
      while (b--)
       c += a;
      ..
      d = a;
      ..
    }

Assume that "a" lives in the stack at offset -8, except for inside the
loop where "a" resides in register "r5".

The way to describe this is to create a stab for the variable "a" which
describes "a" as living in the stack and an alias for the variable "a"
which describes it as living in register "r5" in the loop.

Let's assume that "#1" and "#2" are symbols which bound the area where
"a" lives in a register.

The stabs to describe "a" and its alias would look like this:

        .stabs "#3=a:1",128,0,8,-8
        .stabs "#3:r1;l(#1,#2)",64,0,0,5


This design implies that the debugger will keep a chain of aliases for
any given variable with aliases and that chain will be searched first
to find out if an alias is active.  If no alias is active, then the
debugger will assume that the main variable is active.