[opaque pointer type] Add textual IR support for explicit type parameter to getelementptr instruction

One of several parallel first steps to remove the target type of pointers,
replacing them with a single opaque pointer type.

This adds an explicit type parameter to the gep instruction so that when the
first parameter becomes an opaque pointer type, the type to gep through is
still available to the instructions.

* This doesn't modify gep operators, only instructions (operators will be
  handled separately)

* Textual IR changes only. Bitcode (including upgrade) and changing the
  in-memory representation will be in separate changes.

* geps of vectors are transformed as:
    getelementptr <4 x float*> %x, ...
  ->getelementptr float, <4 x float*> %x, ...
  Then, once the opaque pointer type is introduced, this will ultimately look
  like:
    getelementptr float, <4 x ptr> %x
  with the unambiguous interpretation that it is a vector of pointers to float.

* address spaces remain on the pointer, not the type:
    getelementptr float addrspace(1)* %x
  ->getelementptr float, float addrspace(1)* %x
  Then, eventually:
    getelementptr float, ptr addrspace(1) %x

Importantly, the massive amount of test case churn has been automated by
same crappy python code. I had to manually update a few test cases that
wouldn't fit the script's model (r228970,r229196,r229197,r229198). The
python script just massages stdin and writes the result to stdout, I
then wrapped that in a shell script to handle replacing files, then
using the usual find+xargs to migrate all the files.

update.py:
import fileinput
import sys
import re

ibrep = re.compile(r"(^.*?[^%\w]getelementptr inbounds )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))")
normrep = re.compile(       r"(^.*?[^%\w]getelementptr )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))")

def conv(match, line):
  if not match:
    return line
  line = match.groups()[0]
  if len(match.groups()[5]) == 0:
    line += match.groups()[2]
  line += match.groups()[3]
  line += ", "
  line += match.groups()[1]
  line += "\n"
  return line

for line in sys.stdin:
  if line.find("getelementptr ") == line.find("getelementptr inbounds"):
    if line.find("getelementptr inbounds") != line.find("getelementptr inbounds ("):
      line = conv(re.match(ibrep, line), line)
  elif line.find("getelementptr ") != line.find("getelementptr ("):
    line = conv(re.match(normrep, line), line)
  sys.stdout.write(line)

apply.sh:
for name in "$@"
do
  python3 `dirname "$0"`/update.py < "$name" > "$name.tmp" && mv "$name.tmp" "$name"
  rm -f "$name.tmp"
done

The actual commands:
From llvm/src:
find test/ -name *.ll | xargs ./apply.sh
From llvm/src/tools/clang:
find test/ -name *.mm -o -name *.m -o -name *.cpp -o -name *.c | xargs -I '{}' ../../apply.sh "{}"
From llvm/src/tools/polly:
find test/ -name *.ll | xargs ./apply.sh

After that, check-all (with llvm, clang, clang-tools-extra, lld,
compiler-rt, and polly all checked out).

The extra 'rm' in the apply.sh script is due to a few files in clang's test
suite using interesting unicode stuff that my python script was throwing
exceptions on. None of those files needed to be migrated, so it seemed
sufficient to ignore those cases.

Reviewers: rafael, dexonsmith, grosser

Differential Revision: http://reviews.llvm.org/D7636

llvm-svn: 230786

5 files changed, 6 insertions, 6 deletions

diff --git a/llvm/test/Transforms/LoopUnswitch/2007-07-18-DomInfo.ll b/llvm/test/Transforms/LoopUnswitch/2007-07-18-DomInfo.ll
index dfca15403bf..c01bed142fc 100644
--- a/llvm/test/Transforms/LoopUnswitch/2007-07-18-DomInfo.ll
+++ b/llvm/test/Transforms/LoopUnswitch/2007-07-18-DomInfo.ll
@@ -52,7 +52,7 @@ cond_next1929:          ; preds = %cond_next1915
         br i1 false, label %cond_next1961, label %cond_next2009
 
 cond_next1961:          ; preds = %cond_next1929
-        %tmp1992 = getelementptr i8* %b.10.ph, i32 0            ; <i8*> [#uses=0]
+        %tmp1992 = getelementptr i8, i8* %b.10.ph, i32 0            ; <i8*> [#uses=0]
         br label %cond_next1915
 
 cond_next2009:          ; preds = %cond_next1929
diff --git a/llvm/test/Transforms/LoopUnswitch/2011-09-26-EHCrash.ll b/llvm/test/Transforms/LoopUnswitch/2011-09-26-EHCrash.ll
index e8feef383b0..5717dd89e2f 100644
--- a/llvm/test/Transforms/LoopUnswitch/2011-09-26-EHCrash.ll
+++ b/llvm/test/Transforms/LoopUnswitch/2011-09-26-EHCrash.ll
@@ -22,8 +22,8 @@ for.body:                                         ; preds = %for.cond
 
 if.then:                                          ; preds = %for.body
   %idxprom = sext i32 %inc1 to i64
-  %array_ = getelementptr inbounds %class.MyContainer.1.3.19.29* %this, i32 0, i32 0
-  %arrayidx = getelementptr inbounds [6 x %class.MyMemVarClass.0.2.18.28*]* %array_, i32 0, i64 %idxprom
+  %array_ = getelementptr inbounds %class.MyContainer.1.3.19.29, %class.MyContainer.1.3.19.29* %this, i32 0, i32 0
+  %arrayidx = getelementptr inbounds [6 x %class.MyMemVarClass.0.2.18.28*], [6 x %class.MyMemVarClass.0.2.18.28*]* %array_, i32 0, i64 %idxprom
   %tmp4 = load %class.MyMemVarClass.0.2.18.28** %arrayidx, align 8
   %isnull = icmp eq %class.MyMemVarClass.0.2.18.28* %tmp4, null
   br i1 %isnull, label %for.inc, label %delete.notnull
diff --git a/llvm/test/Transforms/LoopUnswitch/2012-04-30-LoopUnswitch-LPad-Crash.ll b/llvm/test/Transforms/LoopUnswitch/2012-04-30-LoopUnswitch-LPad-Crash.ll
index a6c0d83484b..80e4d782731 100644
--- a/llvm/test/Transforms/LoopUnswitch/2012-04-30-LoopUnswitch-LPad-Crash.ll
+++ b/llvm/test/Transforms/LoopUnswitch/2012-04-30-LoopUnswitch-LPad-Crash.ll
@@ -47,7 +47,7 @@ entry:
   %this.addr = alloca %class.D.22.42.66.102.138.158.178.198.238.242.246.250.262.294.302.338.346.379*, align 8
   store %class.D.22.42.66.102.138.158.178.198.238.242.246.250.262.294.302.338.346.379* %this, %class.D.22.42.66.102.138.158.178.198.238.242.246.250.262.294.302.338.346.379** %this.addr, align 8
   %this1 = load %class.D.22.42.66.102.138.158.178.198.238.242.246.250.262.294.302.338.346.379** %this.addr
-  %px = getelementptr inbounds %class.D.22.42.66.102.138.158.178.198.238.242.246.250.262.294.302.338.346.379* %this1, i32 0, i32 0
+  %px = getelementptr inbounds %class.D.22.42.66.102.138.158.178.198.238.242.246.250.262.294.302.338.346.379, %class.D.22.42.66.102.138.158.178.198.238.242.246.250.262.294.302.338.346.379* %this1, i32 0, i32 0
   %0 = load %class.C.23.43.67.103.139.159.179.199.239.243.247.251.263.295.303.339.347.376** %px, align 8
   %tobool = icmp ne %class.C.23.43.67.103.139.159.179.199.239.243.247.251.263.295.303.339.347.376* %0, null
   br i1 %tobool, label %cond.end, label %cond.false
diff --git a/llvm/test/Transforms/LoopUnswitch/basictest.ll b/llvm/test/Transforms/LoopUnswitch/basictest.ll
index 85e44eca9b6..2a0f5a5abb1 100644
--- a/llvm/test/Transforms/LoopUnswitch/basictest.ll
+++ b/llvm/test/Transforms/LoopUnswitch/basictest.ll
@@ -6,7 +6,7 @@ entry:
 no_exit:		; preds = %no_exit.backedge, %entry
 	%i.0.0 = phi i32 [ 0, %entry ], [ %i.0.0.be, %no_exit.backedge ]		; <i32> [#uses=3]
 	%gep.upgrd.1 = zext i32 %i.0.0 to i64		; <i64> [#uses=1]
-	%tmp.7 = getelementptr i32* %A, i64 %gep.upgrd.1		; <i32*> [#uses=4]
+	%tmp.7 = getelementptr i32, i32* %A, i64 %gep.upgrd.1		; <i32*> [#uses=4]
 	%tmp.13 = load i32* %tmp.7		; <i32> [#uses=2]
 	%tmp.14 = add i32 %tmp.13, 1		; <i32> [#uses=1]
 	store i32 %tmp.14, i32* %tmp.7
diff --git a/llvm/test/Transforms/LoopUnswitch/preserve-analyses.ll b/llvm/test/Transforms/LoopUnswitch/preserve-analyses.ll
index f79612bef51..2725745ea60 100644
--- a/llvm/test/Transforms/LoopUnswitch/preserve-analyses.ll
+++ b/llvm/test/Transforms/LoopUnswitch/preserve-analyses.ll
@@ -20,7 +20,7 @@ bb:                                               ; preds = %bb8
   br i1 %2, label %bb10, label %bb2
 
 bb2:                                              ; preds = %bb
-  %3 = getelementptr inbounds i8* %p_addr.0, i32 1 ; <i8*> [#uses=3]
+  %3 = getelementptr inbounds i8, i8* %p_addr.0, i32 1 ; <i8*> [#uses=3]
   switch i32 %ineq.0.ph, label %bb8.backedge [
     i32 0, label %bb3
     i32 1, label %bb6