[LLVM][DWARF] Make dwarf::getDebugNamesBucketCount return a pair. #83047
Conversation
llvm::dwarf::getDebugNamesBucketCount directly returns the bucket count, via return statement, but it also returns the hash count via a parameter. This changes the function to return them both as a std::pair, in the return statement. It also changes the name of the function to make it clear it returns both values.
|
@llvm/pr-subscribers-debuginfo @llvm/pr-subscribers-llvm-binary-utilities Author: None (cmtice) Changesllvm::dwarf::getDebugNamesBucketCount directly returns the bucket count, via return statement, but it also returns the hash count via a parameter. This changes the function to return them both as a std::pair, in the return statement. It also changes the name of the function to make it clear it returns both values. Full diff: https://github.com/llvm/llvm-project/pull/83047.diff 2 Files Affected:
diff --git a/llvm/include/llvm/BinaryFormat/Dwarf.h b/llvm/include/llvm/BinaryFormat/Dwarf.h
index 44c0030251b37e..e2d79785162e37 100644
--- a/llvm/include/llvm/BinaryFormat/Dwarf.h
+++ b/llvm/include/llvm/BinaryFormat/Dwarf.h
@@ -616,20 +616,21 @@ enum AcceleratorTable {
// Uniquify the string hashes and calculate the bucket count for the
// DWARF v5 Accelerator Table. NOTE: This function effectively consumes the
// 'hashes' input parameter.
-inline uint32_t getDebugNamesBucketCount(MutableArrayRef<uint32_t> hashes,
- uint32_t &uniqueHashCount) {
- uint32_t BucketCount = 0;
+inline std::pair<uint32_t, uint32_t> getDebugNamesBucketAndHashCount(
+ MutableArrayRef<uint32_t> hashes) {
+ uint32_t uniqueHashCount = 0;
+ uint32_t bucketCount = 0;
sort(hashes);
uniqueHashCount = llvm::unique(hashes) - hashes.begin();
if (uniqueHashCount > 1024)
- BucketCount = uniqueHashCount / 4;
+ bucketCount = uniqueHashCount / 4;
else if (uniqueHashCount > 16)
- BucketCount = uniqueHashCount / 2;
+ bucketCount = uniqueHashCount / 2;
else
- BucketCount = std::max<uint32_t>(uniqueHashCount, 1);
+ bucketCount = std::max<uint32_t>(uniqueHashCount, 1);
- return BucketCount;
+ return {bucketCount, uniqueHashCount};
}
// Constants for the GNU pubnames/pubtypes extensions supporting gdb index.
diff --git a/llvm/lib/CodeGen/AsmPrinter/AccelTable.cpp b/llvm/lib/CodeGen/AsmPrinter/AccelTable.cpp
index 23fc9b2e0410e0..94e9d5e9cb77ce 100644
--- a/llvm/lib/CodeGen/AsmPrinter/AccelTable.cpp
+++ b/llvm/lib/CodeGen/AsmPrinter/AccelTable.cpp
@@ -38,7 +38,9 @@ void AccelTableBase::computeBucketCount() {
for (const auto &E : Entries)
Uniques.push_back(E.second.HashValue);
- BucketCount = llvm::dwarf::getDebugNamesBucketCount(Uniques, UniqueHashCount);
+ auto counts = llvm::dwarf::getDebugNamesBucketAndHashCount(Uniques);
+ BucketCount = counts.first;
+ UniqueHashCount = counts.second;
}
void AccelTableBase::finalize(AsmPrinter *Asm, StringRef Prefix) {
|
|
✅ With the latest revision this PR passed the C/C++ code formatter. |
| uint32_t BucketCount = 0; | ||
| inline std::pair<uint32_t, uint32_t> getDebugNamesBucketAndHashCount( | ||
| MutableArrayRef<uint32_t> hashes) { | ||
| uint32_t uniqueHashCount = 0; |
There was a problem hiding this comment.
we should avoid declaring-then-initializing as much as possible, this should be moved to where the variable is actually computed:
uint32_t uniqueHashCount = llvm::unique(hashes) - hashes.begin();
Fix clang-format issue. Avoid unnecessary declaration before initialization.
| uint32_t BucketCount = 0; | ||
| inline std::pair<uint32_t, uint32_t> | ||
| getDebugNamesBucketAndHashCount(MutableArrayRef<uint32_t> hashes) { | ||
| uint32_t bucketCount = 0; |
There was a problem hiding this comment.
Why did you change the case? Doesn't LLVM use UpperCamelCase for variables?
There was a problem hiding this comment.
good point, I had missed this
There was a problem hiding this comment.
I thought LLVM variables were supposed to start with lower case? So I thought I was fixing a problem? I can switch it back if you think I should...
There was a problem hiding this comment.
You're probably thinking of functions, which (unlink LLDB) use lowerCamelCase: https://llvm.org/docs/CodingStandards.html#name-types-functions-variables-and-enumerators-properly
|
I suggested |
Make local variables start with uppercase letter.
Make local variables start with uppercase letter.
| @@ -38,7 +38,9 @@ void AccelTableBase::computeBucketCount() { | |||
| for (const auto &E : Entries) | |||
| Uniques.push_back(E.second.HashValue); | |||
|
|
|||
| BucketCount = llvm::dwarf::getDebugNamesBucketCount(Uniques, UniqueHashCount); | |||
| auto Counts = llvm::dwarf::getDebugNamesBucketAndHashCount(Uniques); | |||
There was a problem hiding this comment.
This can be simplified to std::tie(BucketCount, UniqueHashCount)
There was a problem hiding this comment.
+1 to this - @cmtice perhaps you could make this change as a follow-up commit. No need to send it for review. (but please follow up on this comment thread and mention the commit hash of the change once it's upstream)
There was a problem hiding this comment.
I made the simplification already. Sometimes reviewers may click "Accept" and expect some minor code suggestions will be adopted before landing. It's worth paying attention to these comments.
(For both #82153 and this #83047, my comments were created before the patch landed. Don't worry. We all made mistakes.)
There was a problem hiding this comment.
ah, 7b0b64a cool, thanks!
Fix case of parameter in comment.
|
The buildkite build failure appears to have no relation to this change. Does anybody know how I can force the build test to re-run, so that I can get a green "Squash and merge" button? |
You can still click the button, the failure seems unrelated indeed |
The code suggestion was neglected when the patch landed.
llvm::dwarf::getDebugNamesBucketCount directly returns the bucket count, via return statement, but it also returns the hash count via a parameter. This changes the function to return them both as a std::pair, in the return statement. It also changes the name of the function to make it clear it returns both values.