[llvm-bugs] [Bug 49364] New: llvm-objcopy should remove .symtab_shndx if enough symbols are removed

via llvm-bugs llvm-bugs at lists.llvm.org
Fri Feb 26 06:48:49 PST 2021


https://bugs.llvm.org/show_bug.cgi?id=49364

            Bug ID: 49364
           Summary: llvm-objcopy should remove .symtab_shndx if enough
                    symbols are removed
           Product: tools
           Version: trunk
          Hardware: PC
                OS: Windows NT
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: llvm-objcopy/strip
          Assignee: unassignedbugs at nondot.org
          Reporter: jh7370.2008 at my.bristol.ac.uk
                CC: alexander.v.shaposhnikov at gmail.com,
                    jake.h.ehrlich at gmail.com,
                    jh7370.2008 at my.bristol.ac.uk,
                    llvm-bugs at lists.llvm.org, rupprecht at google.com

llvm-objcopy is intended to emit the SHT_SYMTAB_SHNDX table if and only if it
is required. The table is required if at least one section with an index >=
SHN_LORESERVE has at least one symbol.

When adding/removing sections, or when adding new symbols, this process appears
to work correctly. However, if all symbols in sections with index >=
SHN_LORESERVE are removed (but not the sections themselves), llvm-objcopy still
emits the .symtab_shndx section. The problem is that the `HasSymbol` section
member, which is used to identify whether a section with index >= SHN_LORESERVE
should cause emission of the SHT_SYMTAB_SHNDX section, is not updated after a
symbol has been removed. Consequently, the check to see whether the table
should be added spuriously thinks it should.

Running llvm-objcopy without any switches on the output of an operation that
should have removed the table, but didn't, results in the table being
discarded.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20210226/6e08a487/attachment-0001.html>


More information about the llvm-bugs mailing list