<?xml version="1.0" encoding="UTF-8" standalone="yes" ?>
<!DOCTYPE bugzilla SYSTEM "https://bugs.webkit.org/page.cgi?id=bugzilla.dtd">

<bugzilla version="5.0.4.1"
          urlbase="https://bugs.webkit.org/"
          
          maintainer="admin@webkit.org"
>

    <bug>
          <bug_id>237841</bug_id>
          
          <creation_ts>2022-03-14 10:42:09 -0700</creation_ts>
          <short_desc>[JSC] Eliminate dead bytecode locals</short_desc>
          <delta_ts>2022-03-21 10:43:16 -0700</delta_ts>
          <reporter_accessible>1</reporter_accessible>
          <cclist_accessible>1</cclist_accessible>
          <classification_id>1</classification_id>
          <classification>Unclassified</classification>
          <product>WebKit</product>
          <component>JavaScriptCore</component>
          <version>WebKit Nightly Build</version>
          <rep_platform>Unspecified</rep_platform>
          <op_sys>Unspecified</op_sys>
          <bug_status>NEW</bug_status>
          <resolution></resolution>
          
          
          <bug_file_loc></bug_file_loc>
          <status_whiteboard></status_whiteboard>
          <keywords>InRadar</keywords>
          <priority>P2</priority>
          <bug_severity>Normal</bug_severity>
          <target_milestone>---</target_milestone>
          
          
          <everconfirmed>1</everconfirmed>
          <reporter name="Robin Morisset">rmorisset</reporter>
          <assigned_to name="Nobody">webkit-unassigned</assigned_to>
          <cc>webkit-bug-importer</cc>
          

      

      

      

          <comment_sort_order>oldest_to_newest</comment_sort_order>  
          <long_desc isprivate="0" >
    <commentid>1850935</commentid>
    <comment_count>0</comment_count>
    <who name="Robin Morisset">rmorisset</who>
    <bug_when>2022-03-14 10:42:09 -0700</bug_when>
    <thetext>The bytecode generator tends to allocate more locals than strictly necessary. I saw quite a few examples, e.g. allocating two locals for the scope even when the function is simple enough to only need one:
[   0] enter              
[   1] get_scope          dst:loc4
[   3] mov                dst:loc5, src:loc4
[   6] check_traps  
Where loc5 is completely unused afterwards

I can think of two ways to solve this:
- either going carefully through a bunch of examples and for each one finding all dead locals, tracking where it came from and fixing this particular part of the bytecode generator
- or just adding a generic dead local elimination phase to the bytecode generator.

The latter seems easier, but the former is less likely to cause a perf regression.

The memory benefit here should be tiny, but I would expect some significant win in DFG compilation time, considering that most DFG phases constantly manipulate mappings from Local -&gt; some kind of data, and materialize these mappings at each BasicBlock boundary.</thetext>
  </long_desc><long_desc isprivate="0" >
    <commentid>1853246</commentid>
    <comment_count>1</comment_count>
    <who name="Radar WebKit Bug Importer">webkit-bug-importer</who>
    <bug_when>2022-03-21 10:43:16 -0700</bug_when>
    <thetext>&lt;rdar://problem/90576249&gt;</thetext>
  </long_desc>
      
      

    </bug>

</bugzilla>