arch: teach ISA parser how to split code across files

This patch encompasses several interrelated and interdependent changes
to the ISA generation step.  The end goal is to reduce the size of the
generated compilation units for instruction execution and decoding so
that batch compilation can proceed with all CPUs active without
exhausting physical memory.

The ISA parser (src/arch/isa_parser.py) has been improved so that it can
accept 'split [output_type];' directives at the top level of the grammar
and 'split(output_type)' python calls within 'exec {{ ... }}' blocks.
This has the effect of "splitting" the files into smaller compilation
units.  I use air-quotes around "splitting" because the files themselves
are not split, but preprocessing directives are inserted to have the same
effect.

Architecturally, the ISA parser has had some changes in how it works.
In general, it emits code sooner.  It doesn't generate per-CPU files,
and instead defers to the C preprocessor to create the duplicate copies
for each CPU type.  Likewise there are more files emitted and the C
preprocessor does more substitution that used to be done by the ISA parser.

Finally, the build system (SCons) needs to be able to cope with a
dynamic list of source files coming out of the ISA parser. The changes
to the SCons{cript,truct} files support this. In broad strokes, the
targets requested on the command line are hidden from SCons until all
the build dependencies are determined, otherwise it would try, realize
it can't reach the goal, and terminate in failure. Since build steps
(i.e. running the ISA parser) must be taken to determine the file list,
several new build stages have been inserted at the very start of the
build. First, the build dependencies from the ISA parser will be emitted
to arch/$ISA/generated/inc.d, which is then read by a new SCons builder
to finalize the dependencies. (Once inc.d exists, the ISA parser will not
need to be run to complete this step.) Once the dependencies are known,
the 'Environments' are made by the makeEnv() function. This function used
to be called before the build began but now happens during the build.
It is easy to see that this step is quite slow; this is a known issue
and it's important to realize that it was already slow, but there was
no obvious cause to attribute it to since nothing was displayed to the
terminal. Since new steps that used to be performed serially are now in a
potentially-parallel build phase, the pathname handling in the SCons scripts
has been tightened up to deal with chdir() race conditions. In general,
pathnames are computed earlier and more likely to be stored, passed around,
and processed as absolute paths rather than relative paths.  In the end,
some of these issues had to be fixed by inserting serializing dependencies
in the build.

Minor note:
For the null ISA, we just provide a dummy inc.d so SCons is never
compelled to try to generate it. While it seems slightly wrong to have
anything in src/arch/*/generated (i.e. a non-generated 'generated' file),
it's by far the simplest solution.

1 files changed, 57 insertions, 11 deletions

diff --git a/SConstruct b/SConstruct
index 4df529b83..f20492a0b 100755
--- a/SConstruct
+++ b/SConstruct
@@ -109,6 +109,7 @@ For more details, see:
     raise
 
 # Global Python includes
+import itertools
 import os
 import re
 import subprocess
@@ -1160,16 +1161,21 @@ main.SConscript('ext/dramsim2/SConscript',
 ###################################################
 
 main['ALL_ISA_LIST'] = all_isa_list
+all_isa_deps = {}
 def make_switching_dir(dname, switch_headers, env):
     # Generate the header.  target[0] is the full path of the output
     # header to generate.  'source' is a dummy variable, since we get the
     # list of ISAs from env['ALL_ISA_LIST'].
     def gen_switch_hdr(target, source, env):
         fname = str(target[0])
-        f = open(fname, 'w')
         isa = env['TARGET_ISA'].lower()
-        print >>f, '#include "%s/%s/%s"' % (dname, isa, basename(fname))
-        f.close()
+        try:
+            f = open(fname, 'w')
+            print >>f, '#include "%s/%s/%s"' % (dname, isa, basename(fname))
+            f.close()
+        except IOError:
+            print "Failed to create %s" % fname
+            raise
 
     # Build SCons Action object. 'varlist' specifies env vars that this
     # action depends on; when env['ALL_ISA_LIST'] changes these actions
@@ -1180,8 +1186,37 @@ def make_switching_dir(dname, switch_headers, env):
     # Instantiate actions for each header
     for hdr in switch_headers:
         env.Command(hdr, [], switch_hdr_action)
+
+    isa_target = Dir('.').up().name.lower().replace('_', '-')
+    env['PHONY_BASE'] = '#'+isa_target
+    all_isa_deps[isa_target] = None
+
 Export('make_switching_dir')
 
+# all-isas -> all-deps -> all-environs -> all_targets
+main.Alias('#all-isas', [])
+main.Alias('#all-deps', '#all-isas')
+
+# Dummy target to ensure all environments are created before telling
+# SCons what to actually make (the command line arguments).  We attach
+# them to the dependence graph after the environments are complete.
+ORIG_BUILD_TARGETS = list(BUILD_TARGETS) # force a copy; gets closure to work.
+def environsComplete(target, source, env):
+    for t in ORIG_BUILD_TARGETS:
+        main.Depends('#all-targets', t)
+
+# Each build/* switching_dir attaches its *-environs target to #all-environs.
+main.Append(BUILDERS = {'CompleteEnvirons' :
+                        Builder(action=MakeAction(environsComplete, None))})
+main.CompleteEnvirons('#all-environs', [])
+
+def doNothing(**ignored): pass
+main.Append(BUILDERS = {'Dummy': Builder(action=MakeAction(doNothing, None))})
+
+# The final target to which all the original targets ultimately get attached.
+main.Dummy('#all-targets', '#all-environs')
+BUILD_TARGETS[:] = ['#all-targets']
+
 ###################################################
 #
 # Define build environments for selected configurations.
@@ -1290,14 +1325,25 @@ for variant_path in variant_paths:
     # The src/SConscript file sets up the build rules in 'env' according
     # to the configured variables.  It returns a list of environments,
     # one for each variant build (debug, opt, etc.)
-    envList = SConscript('src/SConscript', variant_dir = variant_path,
-                         exports = 'env')
-
-    # Set up the regression tests for each build.
-    for e in envList:
-        SConscript('tests/SConscript',
-                   variant_dir = joinpath(variant_path, 'tests', e.Label),
-                   exports = { 'env' : e }, duplicate = False)
+    SConscript('src/SConscript', variant_dir = variant_path, exports = 'env')
+
+def pairwise(iterable):
+    "s -> (s0,s1), (s1,s2), (s2, s3), ..."
+    a, b = itertools.tee(iterable)
+    b.next()
+    return itertools.izip(a, b)
+
+# Create false dependencies so SCons will parse ISAs, establish
+# dependencies, and setup the build Environments serially. Either
+# SCons (likely) and/or our SConscripts (possibly) cannot cope with -j
+# greater than 1. It appears to be standard race condition stuff; it
+# doesn't always fail, but usually, and the behaviors are different.
+# Every time I tried to remove this, builds would fail in some
+# creative new way. So, don't do that. You'll want to, though, because
+# tests/SConscript takes a long time to make its Environments.
+for t1, t2 in pairwise(sorted(all_isa_deps.iterkeys())):
+    main.Depends('#%s-deps'     % t2, '#%s-deps'     % t1)
+    main.Depends('#%s-environs' % t2, '#%s-environs' % t1)
 
 # base help text
 Help('''