[llvm-objcopy] Add --compress-sections

--compress-sections is similar to --compress-debug-sections but applies
to arbitrary sections.

* `--compress-sections <section>=none`: decompress sections
* `--compress-sections <section>=[zlib|zstd]`: compress sections with zlib/zstd

Like `--remove-section`, the pattern is by default a glob, but a regex
when --regex is specified.

For `--remove-section` like options, `!` prevents matches and is not
dependent on ordering (see `ELF/wildcard-syntax.test`). Since
`--compress-sections a=zlib --compress-sections a=none` naturally allows
overriding, having an order-independent `!` would be confusing.
Therefore, `!` is disallowed.

Sections within a segment are effectively immutable. Report an error for
an attempt to (de)compress them. `SHF_ALLOC` sections in a relocatable
file can be compressed, but linkers usually reject them.

Note: Before this patch, a compressed relocation section is recognized
as a `RelocationSectionBase` as well and `removeSections` `!ToRemove(*ToRelSec)`
may incorrectly interpret a `CompressedSections` as `RelocationSectionBase`,
leading to ubsan failure for the new test. Fix this by setting
`OriginalFlags` in CompressedSection::CompressedSection.

Link: https://discourse.llvm.org/t/rfc-compress-arbitrary-sections-with-ld-lld-compress-sections/71674

Pull Request: https://github.com/llvm/llvm-project/pull/85036

1 files changed, 36 insertions, 0 deletions

diff --git a/llvm/tools/llvm-objcopy/ObjcopyOptions.cpp b/llvm/tools/llvm-objcopy/ObjcopyOptions.cpp
index 7269c51..70e8546 100644
--- a/llvm/tools/llvm-objcopy/ObjcopyOptions.cpp
+++ b/llvm/tools/llvm-objcopy/ObjcopyOptions.cpp
@@ -736,6 +736,42 @@ objcopy::parseObjcopyOptions(ArrayRef<const char *> RawArgsArr,
       return createStringError(errc::invalid_argument, Reason);
   }
 
+  for (const auto *A : InputArgs.filtered(OBJCOPY_compress_sections)) {
+    SmallVector<StringRef, 0> Fields;
+    StringRef(A->getValue()).split(Fields, '=');
+    if (Fields.size() != 2 || Fields[1].empty()) {
+      return createStringError(
+          errc::invalid_argument,
+          A->getSpelling() +
+              ": parse error, not 'section-glob=[none|zlib|zstd]'");
+    }
+
+    auto Type = StringSwitch<DebugCompressionType>(Fields[1])
+                    .Case("zlib", DebugCompressionType::Zlib)
+                    .Case("zstd", DebugCompressionType::Zstd)
+                    .Default(DebugCompressionType::None);
+    if (Type == DebugCompressionType::None && Fields[1] != "none") {
+      return createStringError(
+          errc::invalid_argument,
+          "invalid or unsupported --compress-sections format: %s",
+          A->getValue());
+    }
+
+    auto &P = Config.compressSections.emplace_back();
+    P.second = Type;
+    auto Matcher =
+        NameOrPattern::create(Fields[0], SectionMatchStyle, ErrorCallback);
+    // =none allows overriding a previous =zlib or =zstd. Reject negative
+    // patterns, which would be confusing.
+    if (Matcher && !Matcher->isPositiveMatch()) {
+      return createStringError(
+          errc::invalid_argument,
+          "--compress-sections: negative pattern is unsupported");
+    }
+    if (Error E = P.first.addMatcher(std::move(Matcher)))
+      return std::move(E);
+  }
+
   Config.AddGnuDebugLink = InputArgs.getLastArgValue(OBJCOPY_add_gnu_debuglink);
   // The gnu_debuglink's target is expected to not change or else its CRC would
   // become invalidated and get rejected. We can avoid recalculating the