Skip to content

Commit 6ac71a0

Browse files
rlavaeerlavaee
authored
[BasicBlockSections] Introduce the basic block sections profile version 1. (#65506)
This patch introduces a new version for the basic block sections profile as was requested in D158442, while keeping backward compatibility for the old version. The new encoding is as follows: ``` m <module_name> f <function_name_1> <function_name_2>... c <bb_id_1> <bb_id_2> <bb_id_3> c <bb_id_4> <bb_id_5> ... ``` Module name specifier (starting with 'm') is optional and allows distinguishing profiles for internal-linkage functions with the same name. If not specified, profile will be applied to any function with the same name. Function name specifier (starting with 'f') can specify multiple function name aliases. Finally, basic block clusters are specified by 'c' and specify the cluster of basic blocks, and the internal order in which they must be placed in the same section.
1 parent 46d5d26 commit 6ac71a0

File tree

7 files changed

+286
-62
lines changed

7 files changed

+286
-62
lines changed

‎llvm/include/llvm/CodeGen/BasicBlockSectionsProfileReader.h‎

Lines changed: 20 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -15,7 +15,6 @@
1515
#ifndef LLVM_CODEGEN_BASICBLOCKSECTIONSPROFILEREADER_H
1616
#define LLVM_CODEGEN_BASICBLOCKSECTIONSPROFILEREADER_H
1717

18-
#include "llvm/ADT/SmallSet.h"
1918
#include "llvm/ADT/SmallString.h"
2019
#include "llvm/ADT/SmallVector.h"
2120
#include "llvm/ADT/StringMap.h"
@@ -26,7 +25,6 @@
2625
#include "llvm/Support/Error.h"
2726
#include "llvm/Support/LineIterator.h"
2827
#include "llvm/Support/MemoryBuffer.h"
29-
using namespace llvm;
3028

3129
namespace llvm {
3230

@@ -47,7 +45,8 @@ class BasicBlockSectionsProfileReader : public ImmutablePass {
4745
static char ID;
4846

4947
BasicBlockSectionsProfileReader(const MemoryBuffer *Buf)
50-
: ImmutablePass(ID), MBuf(Buf) {
48+
: ImmutablePass(ID), MBuf(Buf),
49+
LineIt(*Buf, /*SkipBlanks=*/true, /*CommentMarker=*/'#') {
5150
initializeBasicBlockSectionsProfileReaderPass(
5251
*PassRegistry::getPassRegistry());
5352
};
@@ -84,12 +83,30 @@ class BasicBlockSectionsProfileReader : public ImmutablePass {
8483
return R == FuncAliasMap.end() ? FuncName : R->second;
8584
}
8685

86+
// Returns a profile parsing error for the current line.
87+
Error createProfileParseError(Twine Message) const {
88+
return make_error<StringError>(
89+
Twine("invalid profile " + MBuf->getBufferIdentifier() + " at line " +
90+
Twine(LineIt.line_number()) + ": " + Message),
91+
inconvertibleErrorCode());
92+
}
93+
8794
// Reads the basic block sections profile for functions in this module.
8895
Error ReadProfile();
8996

97+
// Reads version 0 profile.
98+
// TODO: Remove this function once version 0 is deprecated.
99+
Error ReadV0Profile();
100+
101+
// Reads version 1 profile.
102+
Error ReadV1Profile();
103+
90104
// This contains the basic-block-sections profile.
91105
const MemoryBuffer *MBuf = nullptr;
92106

107+
// Iterator to the line being parsed.
108+
line_iterator LineIt;
109+
93110
// Map from every function name in the module to its debug info filename or
94111
// empty string if no debug info is available.
95112
StringMap<SmallString<128>> FunctionNameToDIFilename;

‎llvm/lib/CodeGen/BasicBlockSectionsProfileReader.cpp‎

Lines changed: 171 additions & 37 deletions
Original file line numberDiff line numberDiff line change
@@ -18,9 +18,10 @@
1818
#include "llvm/ADT/SmallVector.h"
1919
#include "llvm/ADT/StringMap.h"
2020
#include "llvm/ADT/StringRef.h"
21-
#include "llvm/CodeGen/Passes.h"
2221
#include "llvm/IR/DebugInfoMetadata.h"
22+
#include "llvm/Pass.h"
2323
#include "llvm/Support/Error.h"
24+
#include "llvm/Support/ErrorHandling.h"
2425
#include "llvm/Support/LineIterator.h"
2526
#include "llvm/Support/MemoryBuffer.h"
2627
#include "llvm/Support/Path.h"
@@ -46,34 +47,121 @@ BasicBlockSectionsProfileReader::getBBClusterInfoForFunction(
4647
: std::pair(false, SmallVector<BBClusterInfo>{});
4748
}
4849

49-
// Basic Block Sections can be enabled for a subset of machine basic blocks.
50-
// This is done by passing a file containing names of functions for which basic
51-
// block sections are desired. Additionally, machine basic block ids of the
52-
// functions can also be specified for a finer granularity. Moreover, a cluster
53-
// of basic blocks could be assigned to the same section.
54-
// Optionally, a debug-info filename can be specified for each function to allow
55-
// distinguishing internal-linkage functions of the same name.
56-
// A file with basic block sections for all of function main and three blocks
57-
// for function foo (of which 1 and 2 are placed in a cluster) looks like this:
58-
// (Profile for function foo is only loaded when its debug-info filename
59-
// matches 'path/to/foo_file.cc').
60-
// ----------------------------
61-
// list.txt:
62-
// !main
63-
// !foo M=path/to/foo_file.cc
64-
// !!1 2
65-
// !!4
66-
Error BasicBlockSectionsProfileReader::ReadProfile() {
67-
assert(MBuf);
68-
line_iterator LineIt(*MBuf, /*SkipBlanks=*/true, /*CommentMarker=*/'#');
50+
// Reads the version 1 basic block sections profile. Profile for each function
51+
// is encoded as follows:
52+
// m <module_name>
53+
// f <function_name_1> <function_name_2> ...
54+
// c <bb_id_1> <bb_id_2> <bb_id_3>
55+
// c <bb_id_4> <bb_id_5>
56+
// ...
57+
// Module name specifier (starting with 'm') is optional and allows
58+
// distinguishing profile for internal-linkage functions with the same name. If
59+
// not specified, it will apply to any function with the same name. Function
60+
// name specifier (starting with 'f') can specify multiple function name
61+
// aliases. Basic block clusters are specified by 'c' and specify the cluster of
62+
// basic blocks, and the internal order in which they must be placed in the same
63+
// section.
64+
Error BasicBlockSectionsProfileReader::ReadV1Profile() {
65+
auto FI = ProgramBBClusterInfo.end();
66+
67+
// Current cluster ID corresponding to this function.
68+
unsigned CurrentCluster = 0;
69+
// Current position in the current cluster.
70+
unsigned CurrentPosition = 0;
71+
72+
// Temporary set to ensure every basic block ID appears once in the clusters
73+
// of a function.
74+
SmallSet<unsigned, 4> FuncBBIDs;
75+
76+
// Debug-info-based module filename for the current function. Empty string
77+
// means no filename.
78+
StringRef DIFilename;
79+
80+
for (; !LineIt.is_at_eof(); ++LineIt) {
81+
StringRef S(*LineIt);
82+
char Specifier = S[0];
83+
S = S.drop_front().trim();
84+
SmallVector<StringRef, 4> Values;
85+
S.split(Values, ' ');
86+
switch (Specifier) {
87+
case '@':
88+
break;
89+
case 'm': // Module name speicifer.
90+
if (Values.size() != 1) {
91+
return createProfileParseError(Twine("invalid module name value: '") +
92+
S + "'");
93+
}
94+
DIFilename = sys::path::remove_leading_dotslash(Values[0]);
95+
continue;
96+
case 'f': { // Function names specifier.
97+
bool FunctionFound = any_of(Values, [&](StringRef Alias) {
98+
auto It = FunctionNameToDIFilename.find(Alias);
99+
// No match if this function name is not found in this module.
100+
if (It == FunctionNameToDIFilename.end())
101+
return false;
102+
// Return a match if debug-info-filename is not specified. Otherwise,
103+
// check for equality.
104+
return DIFilename.empty() || It->second.equals(DIFilename);
105+
});
106+
if (!FunctionFound) {
107+
// Skip the following profile by setting the profile iterator (FI) to
108+
// the past-the-end element.
109+
FI = ProgramBBClusterInfo.end();
110+
DIFilename = "";
111+
continue;
112+
}
113+
for (size_t i = 1; i < Values.size(); ++i)
114+
FuncAliasMap.try_emplace(Values[i], Values.front());
115+
116+
// Prepare for parsing clusters of this function name.
117+
// Start a new cluster map for this function name.
118+
auto R = ProgramBBClusterInfo.try_emplace(Values.front());
119+
// Report error when multiple profiles have been specified for the same
120+
// function.
121+
if (!R.second)
122+
return createProfileParseError("duplicate profile for function '" +
123+
Values.front() + "'");
124+
FI = R.first;
125+
CurrentCluster = 0;
126+
FuncBBIDs.clear();
127+
// We won't need DIFilename anymore. Clean it up to avoid its application
128+
// on the next function.
129+
DIFilename = "";
130+
continue;
131+
}
132+
case 'c': // Basic block cluster specifier.
133+
// Skip the profile when we the profile iterator (FI) refers to the
134+
// past-the-end element.
135+
if (FI == ProgramBBClusterInfo.end())
136+
break;
137+
// Reset current cluster position.
138+
CurrentPosition = 0;
139+
for (auto BBIDStr : Values) {
140+
unsigned long long BBID;
141+
if (getAsUnsignedInteger(BBIDStr, 10, BBID))
142+
return createProfileParseError(Twine("unsigned integer expected: '") +
143+
BBIDStr + "'");
144+
if (!FuncBBIDs.insert(BBID).second)
145+
return createProfileParseError(
146+
Twine("duplicate basic block id found '") + BBIDStr + "'");
147+
if (BBID == 0 && CurrentPosition)
148+
return createProfileParseError(
149+
"entry BB (0) does not begin a cluster");
69150

70-
auto invalidProfileError = [&](auto Message) {
71-
return make_error<StringError>(
72-
Twine("Invalid profile " + MBuf->getBufferIdentifier() + " at line " +
73-
Twine(LineIt.line_number()) + ": " + Message),
74-
inconvertibleErrorCode());
75-
};
151+
FI->second.emplace_back(
152+
BBClusterInfo{((unsigned)BBID), CurrentCluster, CurrentPosition++});
153+
}
154+
CurrentCluster++;
155+
continue;
156+
default:
157+
return createProfileParseError(Twine("invalid specifier: '") +
158+
Twine(Specifier) + "'");
159+
}
160+
}
161+
return Error::success();
162+
}
76163

164+
Error BasicBlockSectionsProfileReader::ReadV0Profile() {
77165
auto FI = ProgramBBClusterInfo.end();
78166

79167
// Current cluster ID corresponding to this function.
@@ -105,13 +193,14 @@ Error BasicBlockSectionsProfileReader::ReadProfile() {
105193
for (auto BBIDStr : BBIDs) {
106194
unsigned long long BBID;
107195
if (getAsUnsignedInteger(BBIDStr, 10, BBID))
108-
return invalidProfileError(Twine("Unsigned integer expected: '") +
109-
BBIDStr + "'.");
196+
return createProfileParseError(Twine("unsigned integer expected: '") +
197+
BBIDStr + "'");
110198
if (!FuncBBIDs.insert(BBID).second)
111-
return invalidProfileError(Twine("Duplicate basic block id found '") +
112-
BBIDStr + "'.");
199+
return createProfileParseError(
200+
Twine("duplicate basic block id found '") + BBIDStr + "'");
113201
if (BBID == 0 && CurrentPosition)
114-
return invalidProfileError("Entry BB (0) does not begin a cluster.");
202+
return createProfileParseError(
203+
"entry BB (0) does not begin a cluster");
115204

116205
FI->second.emplace_back(
117206
BBClusterInfo{((unsigned)BBID), CurrentCluster, CurrentPosition++});
@@ -126,10 +215,10 @@ Error BasicBlockSectionsProfileReader::ReadProfile() {
126215
DIFilename =
127216
sys::path::remove_leading_dotslash(DIFilenameStr.substr(2));
128217
if (DIFilename.empty())
129-
return invalidProfileError("Empty module name specifier.");
218+
return createProfileParseError("empty module name specifier");
130219
} else if (!DIFilenameStr.empty()) {
131-
return invalidProfileError("Unknown string found: '" + DIFilenameStr +
132-
"'.");
220+
return createProfileParseError("unknown string found: '" +
221+
DIFilenameStr + "'");
133222
}
134223
// Function aliases are separated using '/'. We use the first function
135224
// name for the cluster info mapping and delegate all other aliases to
@@ -160,8 +249,8 @@ Error BasicBlockSectionsProfileReader::ReadProfile() {
160249
// Report error when multiple profiles have been specified for the same
161250
// function.
162251
if (!R.second)
163-
return invalidProfileError("Duplicate profile for function '" +
164-
Aliases.front() + "'.");
252+
return createProfileParseError("duplicate profile for function '" +
253+
Aliases.front() + "'");
165254
FI = R.first;
166255
CurrentCluster = 0;
167256
FuncBBIDs.clear();
@@ -170,6 +259,51 @@ Error BasicBlockSectionsProfileReader::ReadProfile() {
170259
return Error::success();
171260
}
172261

262+
// Basic Block Sections can be enabled for a subset of machine basic blocks.
263+
// This is done by passing a file containing names of functions for which basic
264+
// block sections are desired. Additionally, machine basic block ids of the
265+
// functions can also be specified for a finer granularity. Moreover, a cluster
266+
// of basic blocks could be assigned to the same section.
267+
// Optionally, a debug-info filename can be specified for each function to allow
268+
// distinguishing internal-linkage functions of the same name.
269+
// A file with basic block sections for all of function main and three blocks
270+
// for function foo (of which 1 and 2 are placed in a cluster) looks like this:
271+
// (Profile for function foo is only loaded when its debug-info filename
272+
// matches 'path/to/foo_file.cc').
273+
// ----------------------------
274+
// list.txt:
275+
// !main
276+
// !foo M=path/to/foo_file.cc
277+
// !!1 2
278+
// !!4
279+
Error BasicBlockSectionsProfileReader::ReadProfile() {
280+
assert(MBuf);
281+
282+
unsigned long long Version = 0;
283+
StringRef FirstLine(*LineIt);
284+
if (FirstLine.consume_front("v")) {
285+
if (getAsUnsignedInteger(FirstLine, 10, Version)) {
286+
return createProfileParseError(Twine("version number expected: '") +
287+
FirstLine + "'");
288+
}
289+
if (Version > 1) {
290+
return createProfileParseError(Twine("invalid profile version: ") +
291+
Twine(Version));
292+
}
293+
++LineIt;
294+
}
295+
296+
switch (Version) {
297+
case 0:
298+
// TODO: Deprecate V0 once V1 is fully integrated downstream.
299+
return ReadV0Profile();
300+
case 1:
301+
return ReadV1Profile();
302+
default:
303+
llvm_unreachable("Invalid profile version.");
304+
}
305+
}
306+
173307
bool BasicBlockSectionsProfileReader::doInitialization(Module &M) {
174308
if (!MBuf)
175309
return false;

‎llvm/test/CodeGen/X86/basic-block-sections-clusters-error.ll‎

Lines changed: 25 additions & 8 deletions
Original file line numberDiff line numberDiff line change
@@ -1,27 +1,44 @@
1-
; BB cluster sections error handling
1+
;; BB cluster sections error handling
2+
;; Error handling for version 0:
23
; RUN: echo '!dummy1' > %t1
34
; RUN: echo '!!1 4' >> %t1
45
; RUN: echo '!!1' >> %t1
56
; RUN: not --crash llc < %s -O0 -mtriple=x86_64-pc-linux -function-sections -basic-block-sections=%t1 2>&1 | FileCheck %s --check-prefix=CHECK-ERROR1
6-
; CHECK-ERROR1: LLVM ERROR: Invalid profile {{.*}} at line 3: Duplicate basic block id found '1'.
7+
; CHECK-ERROR1: LLVM ERROR: invalid profile {{.*}} at line 3: duplicate basic block id found '1'
78
; RUN: echo '!dummy1' > %t2
89
; RUN: echo '!!4 0' >> %t2
910
; RUN: not --crash llc < %s -O0 -mtriple=x86_64-pc-linux -function-sections -basic-block-sections=%t2 2>&1 | FileCheck %s --check-prefix=CHECK-ERROR2
10-
; CHECK-ERROR2: LLVM ERROR: Invalid profile {{.*}} at line 2: Entry BB (0) does not begin a cluster.
11+
; CHECK-ERROR2: LLVM ERROR: invalid profile {{.*}} at line 2: entry BB (0) does not begin a cluster
1112
; RUN: echo '!dummy1' > %t3
1213
; RUN: echo '!!-1' >> %t3
1314
; RUN: not --crash llc < %s -O0 -mtriple=x86_64-pc-linux -function-sections -basic-block-sections=%t3 2>&1 | FileCheck %s --check-prefix=CHECK-ERROR3
14-
; CHECK-ERROR3: LLVM ERROR: Invalid profile {{.*}} at line 2: Unsigned integer expected: '-1'.
15+
; CHECK-ERROR3: LLVM ERROR: invalid profile {{.*}} at line 2: unsigned integer expected: '-1'
1516
; RUN: echo '!dummy1 /path/to/filename' > %t4
1617
; RUN: not --crash llc < %s -O0 -mtriple=x86_64 -function-sections -basic-block-sections=%t4 2>&1 | FileCheck %s --check-prefix=CHECK-ERROR4
17-
; CHECK-ERROR4: LLVM ERROR: Invalid profile {{.*}} at line 1: Unknown string found: '/path/to/filename'.
18+
; CHECK-ERROR4: LLVM ERROR: invalid profile {{.*}} at line 1: unknown string found: '/path/to/filename'
1819
; RUN: echo '!dummy2 M=test_dir/test_file' > %t5
1920
; RUN: echo '!dummy2 M=test_dir/test_file' >> %t5
2021
; RUN: not --crash llc < %s -O0 -mtriple=x86_64 -function-sections -basic-block-sections=%t5 2>&1 | FileCheck %s --check-prefix=CHECK-ERROR5
21-
; CHECK-ERROR5: LLVM ERROR: Invalid profile {{.*}} at line 2: Duplicate profile for function 'dummy2'.
22-
; RUN: echo '!dummy1 M=' >> %t6
22+
; CHECK-ERROR5: LLVM ERROR: invalid profile {{.*}} at line 2: duplicate profile for function 'dummy2'
23+
; RUN: echo '!dummy1 M=' > %t6
2324
; RUN: not --crash llc < %s -O0 -mtriple=x86_64 -function-sections -basic-block-sections=%t6 2>&1 | FileCheck %s --check-prefix=CHECK-ERROR6
24-
; CHECK-ERROR6: LLVM ERROR: Invalid profile {{.*}} at line 1: Empty module name specifier.
25+
; CHECK-ERROR6: LLVM ERROR: invalid profile {{.*}} at line 1: empty module name specifier
26+
;;
27+
;; Error handling for version 1:
28+
; RUN: echo 'v2' > %t7
29+
; RUN: not --crash llc < %s -O0 -mtriple=x86_64 -function-sections -basic-block-sections=%t7 2>&1 | FileCheck %s --check-prefix=CHECK-ERROR7
30+
; CHECK-ERROR7: LLVM ERROR: invalid profile {{.*}} at line 1: invalid profile version: 2
31+
; RUN: echo 'v1' > %t8
32+
; RUN: echo '!dummy1' >> %t8
33+
; RUN: not --crash llc < %s -O0 -mtriple=x86_64 -function-sections -basic-block-sections=%t8 2>&1 | FileCheck %s --check-prefix=CHECK-ERROR8
34+
; CHECK-ERROR8: LLVM ERROR: invalid profile {{.*}} at line 2: invalid specifier: '!'
35+
; RUN: echo 'v1' > %t0
36+
; RUN: echo 'm dummy1/module1 dummy1/module2'
37+
; RUN: echo 'f dummy1' >> %t9
38+
; RUN: not --crash llc < %s -O0 -mtriple=x86_64 -function-sections -basic-block-sections=%t8 2>&1 | FileCheck %s --check-prefix=CHECK-ERROR8
39+
; CHECK-ERROR9: LLVM ERROR: invalid profile {{.*}} at line 2: invalid module name value: 'dummy1/module dummy1/module2'
40+
41+
2542

2643
define i32 @dummy1(i32 %x, i32 %y, i32 %z) {
2744
entry:

0 commit comments

Comments
 (0)