1 <!DOCTYPE HTML PUBLIC "-//IETF//DTD HTML//EN">
4 <META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=ISO-8859-1">
5 <title>The GHC Commentary - The Evil Mangler</title>
8 <body BGCOLOR="FFFFFF">
9 <h1>The GHC Commentary - The Evil Mangler</h1>
11 The Evil Mangler (EM) is a Perl script invoked by the <a
12 href="driver.html">Glorious Driver</a> after the C compiler (gcc) has
13 translated the GHC-produced C code into assembly. Consequently, it is
14 only of interest if <code>-fvia-C</code> is in effect (either explicitly
19 The EM reads the assembly produced by gcc and re-arranges code blocks as
20 well as nukes instructions that it considers <em>non-essential.</em> It
21 derives it evilness from its utterly ad hoc, machine, compiler, and
22 whatnot dependent design and implementation. More precisely, the EM
23 performs the following tasks:
25 <li>The code executed when a closure is entered is moved adjacent to
26 that closure's infotable. Moreover, the order of the info table
28 <li>Function prologue and epilogue code is removed. (GHC generated code
29 manages its own stack and uses the system stack only for return
30 addresses and during calls to C code.)
31 <li>Certain code patterns are replaced by simpler code (eg, loads of
32 fast entry points followed by indirect jumps are replaced by direct
33 jumps to the fast entry point).
36 <h4>Implementation</h4>
38 The EM is located in the Perl script <a
39 href="http://cvs.haskell.org/cgi-bin/cvsweb.cgi/fptools/ghc/driver/mangler/ghc-asm.lprl"><code>ghc-asm.lprl</code></a>.
40 The script reads the <code>.s</code> file and chops it up into
41 <em>chunks</em> (that's how they are actually called in the script) that
42 roughly correspond to basic blocks. Each chunk is annotated with an
43 educated guess about what kind of code it contains (e.g., infotable,
44 fast entry point, slow entry point, etc.). The annotations also contain
45 the symbol introducing the chunk of assembly and whether that chunk has
46 already been processed or not.
48 The parsing of the input into chunks as well as recognising assembly
49 instructions that are to be removed or altered is based on a large
50 number of Perl regular expressions sprinkled over the whole code. These
51 expressions are rather fragile as they heavily rely on the structure of
52 the generated code - in fact, they even rely on the right amount of wide
53 space and thus on the formatting of the assembly.
55 Afterwards, the chunks are reordered, some of them purged, and some
56 stripped of some useless instructions. Moreover, some instructions are
57 manipulated (eg, loads of fast entry points followed by indirect jumps
58 are replaced by direct jumps to the fast entry point).
60 The EM knows which part of the code belongs to function prologues and
61 epilogues as <a href="../rts-libs/stgc.html">STG C</a> adds tags of the
62 form <code>--- BEGIN ---</code> and <code>--- END ---</code> the
63 assembler just before and after the code proper of a function starts.
64 It adds these tags using gcc's <code>__asm__</code> feature.
66 <strong>Update:</strong> Gcc 2.96 upwards performs more aggressive basic
67 block re-ordering and dead code elimination. This seems to make the
68 whole <code>--- END ---</code> tag business redundant -- in fact, if
69 proper code is generated, no <code>--- END ---</code> tags survive gcc
74 Last modified: Wed Aug 8 19:27:22 EST 2001