Changes/DebugInfoStandardization

Debug Info Standardization

Summary

Fedora 18 implemented Features/DwarfCompressor. As the format did not get widespread and the tool is not much maintained it became burden to make existing debugging tools compatible with Fedora.

Owner

• Name: Jan Kratochvil
• Email: jan.kratochvil@redhat.com

Current status

• Targeted release: Fedora 34
• Last updated: 2020-09-23
• FESCo issue: <will be assigned by the Wrangler>
• Tracker bug: <will be assigned by the Wrangler>
• Release notes tracker: <will be assigned by the Wrangler>

Detailed Description

The is a refresh of Fedora 18 Features/DebugTypesSections.

Debug info files *.debug contained in *-debuginfo.rpm are very big in general (debug/ has 82GB while all the other x86_64 Fedora distribution files are 75GB). There are various methods how to make the *-debuginfo.rpms smaller. Fedora 18 started using DWZ tool (from Features/DwarfCompressor) while Google implemented a similar but different approach called -fdebug-types-section.

Almost nobody uses existing Fedora DWZ (only Fedora/CentOS/RHEL and SuSE OSes) and so its support is missing in tools like LLDB and llvm-dwarfdump.

• DWZ advantage: On the whole Fedora distro it saves 3.3% (5GB of the 157GB distribution size)

 * If the 3.3% size increase is a concern I can implement a different optimization (talk (2)) as GCC post-proccessing phase which requires no support in consumers.

• DWZ disadvantage: Compilation (linking) requires 2x as big disk space (as DWZ is processing files after linker and DWZ is incompatible with -fdebug-types-section)
• DWZ disadvantage: Compilation (linking) is slower
• DWZ disadvantage: DWZ requires 8x times more complicated (LoC count) support in consumers than -fdebug-types-section.
• DWZ disadvantage: DWZ has currently less support across consumers (LLDB, llvm-dwarfdump, binutils readelf)
• DWZ disadvantage: DWZ cannot update LLVM .debug_names index which can be generated only by clang (it cannot be regenerated later for DWZ-compressed file)
• DWZ disadvantage: DWZ DWARF-5 support is a work-in-progress and there is no plan to support DWARF-5 produced by clang

Scope

• Proposal owners:

It affects all packages generating *-debuginfo.rpm.

• Other developers:

Report any possible debuginfo incompatibility (unexpected).

• Release engineering: #Releng issue number (a check of an impact with Release Engineering is needed)

• Policies and guidelines:

All the needed changes should be done in redhat-rpm-config. The dwz package can be then retired.

• Trademark approval: N/A (not needed for this Change)

• Alignment with Objectives:

The size differences are only for *-debuginfo.rpm which is outside of scope of the listed objectives.

Upgrade/compatibility impact

As *-debuginfo.rpm must exactly match NVRA of its binary package the compatibility is not relevant. Existing tools supporting DWZ will still support the DWZ file format in packages which have not been rebuilt.

How To Test

It is not a big risk, -fdebug-types-section is more widespread and more simple than DWZ. Still I should run:

* libabigail DWARF comparisons
* testsuites of tools like: clang, lldb, gcc, binutils, gdb, elfutils, rpm

User Experience

* Better compatibility with existing debugging and tracing tools.
* Less resource-intensive rebuilds of C++ packages (both in disk space, memory requirements and compilation time).

Dependencies

none

Contingency Plan

• Contingency mechanism: Nothing needed, Fedora can continue using DWZ, just some debugging/tracing tools will continue to be incompatible.
• Contingency deadline: beta freeze
• Blocks release? No
• Blocks product? N/A

Documentation

* DWARF-5 E.2 Using Type Units
* GCC -fdebug-types-section