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, 69 insertions, 7 deletions

diff --git a/src/arch/SConscript b/src/arch/SConscript
index e7d74ce51..e9b5c5365 100644
--- a/src/arch/SConscript
+++ b/src/arch/SConscript
@@ -30,6 +30,7 @@
 
 import sys
 import os
+import re
 
 Import('*')
 
@@ -97,18 +98,32 @@ def isa_desc_emitter(target, source, env):
     cpu_models = list(env['CPU_MODELS'])
     cpu_models.append('CheckerCPU')
 
-    # Several files are generated from the ISA description.
-    # We always get the basic decoder and header file.
-    target = [ 'decoder.cc', 'decoder.hh', 'max_inst_regs.hh' ]
-    # We also get an execute file for each selected CPU model.
-    target += [CpuModel.dict[cpu].filename for cpu in cpu_models]
-
     # List the isa parser as a source.
     source += [ isa_parser ]
     # Add in the CPU models.
     source += [ Value(m) for m in cpu_models ]
 
-    return [os.path.join("generated", t) for t in target], source
+    # Specify different targets depending on if we're running the ISA
+    # parser for its dependency information, or for the generated files.
+    # (As an optimization, the ISA parser detects the useless second run
+    # and skips doing any work, if the first run was performed, since it
+    # always generates all its files). The way we track this in SCons is the
+    # <arch>_isa_outputs value in the environment (env). If it's unset, we
+    # don't know what the dependencies are so we ask for generated/inc.d to
+    # be generated so they can be acquired. If we know what they are, then
+    # it's because we've already processed inc.d and then claim that our
+    # outputs (targets) will be thus.
+    isa = env['TARGET_ISA']
+    key = '%s_isa_outputs' % isa
+    if key in env:
+        targets = [ os.path.join('generated', f) for f in env[key] ]
+    else:
+        targets = [ os.path.join('generated','inc.d') ]
+
+    def prefix(s):
+        return os.path.join(target[0].dir.up().abspath, s)
+
+    return [ prefix(t) for t in targets ], source
 
 ARCH_DIR = Dir('.')
 
@@ -133,6 +148,53 @@ isa_desc_builder = Builder(action=isa_desc_action, emitter=isa_desc_emitter)
 
 env.Append(BUILDERS = { 'ISADesc' : isa_desc_builder })
 
+# The ISA is generated twice: the first time to find out what it generates,
+# and the second time to make scons happy by telling the ISADesc builder
+# what it will make before it builds it.
+def scan_isa_deps(target, source, env):
+    # Process dependency file generated by the ISA parser --
+    # add the listed files to the dependency tree of the build.
+    source = source[0]
+    archbase = source.dir.up().path
+
+    try:
+        depfile = open(source.abspath, 'r')
+    except:
+        print "scan_isa_deps: Can't open ISA deps file '%s' in %s" % \
+              (source.path,os.getcwd())
+        raise
+
+    # Scan through the lines
+    targets = {}
+    for line in depfile:
+        # Read the dependency line with the format
+        # <target file>: [ <dependent file>* ]
+        m = re.match(r'^\s*([^:]+\.([^\.:]+))\s*:\s*(.*)', line)
+        assert(m)
+        targ, extn = m.group(1,2)
+        deps = m.group(3).split()
+
+        files = [ targ ] + deps
+        for f in files:
+            targets[f] = True
+            # Eliminate unnecessary re-generation if we already generated it
+            env.Precious(os.path.join(archbase, 'generated', f))
+
+        files = [ os.path.join(archbase, 'generated', f) for f in files ]
+
+        if extn == 'cc':
+            Source(os.path.join(archbase,'generated', targ))
+    depfile.close()
+    env[env['TARGET_ISA'] + '_isa_outputs'] = targets.keys()
+
+    isa = env.ISADesc(os.path.join(archbase,'isa','main.isa'))
+    for t in targets:
+        env.Depends('#all-isas', isa)
+
+env.Append(BUILDERS = {'ScanISA' :
+                        Builder(action=MakeAction(scan_isa_deps,
+                                                  Transform("NEW DEPS", 1)))})
+
 DebugFlag('IntRegs')
 DebugFlag('FloatRegs')
 DebugFlag('CCRegs')