elf: Handle SHN_XINDEX for strtab section index (#304) #316
Conversation
| @@ -51,7 +51,12 @@ fn parse_text_section_size_lazy(base: &[u8]) -> Result<u64, &'static str> { | |||
| SectionHeader::parse(base, header.e_shoff as usize, header.e_shnum as usize, ctx) | |||
| .map_err(|_| "parse section headers error")?; | |||
|
|
|||
| let strtab_idx = header.e_shstrndx as usize; | |||
| let strtab_idx = if header.e_shstrndx as u32 == SHN_XINDEX { | |||
There was a problem hiding this comment.
weird i feel like this was added already, long time ago, surprised to see it now
| // FIXME: warn! here | ||
| return Ok(Strtab::default()) | ||
| } | ||
| section_idx = section_headers[0].sh_link as usize; |
There was a problem hiding this comment.
just curious (can't remember why) it's always the first section header's sh_link that is used?
There was a problem hiding this comment.
It seems to. I actually looked what Gimli was doing, but the ELF docs at Oracle says:
If the section name string table section index is greater than or equal to SHN_LORESERVE (0xff00), this member has the value SHN_XINDEX (0xffff) and the actual index of the section name string table section is contained in the sh_link field of the section header at index 0. Otherwise, the sh_link member of the initial section header entry contains the value zero. See Table 12-6 and Table 12-7.
|
And thank you for stepping up and fixing that issue! |
This happens when the number of sections in an ELF object reaches 2^16, at which point the u16 shstrndx takes the special value SHN_XINDEX. Fix based on algorithm from: github.com/gimli-rs/object/blob/c476071/src/read/elf/file.rs#L582-L600
|
released in 0.5.3 on crates; thanks for fixing this @tux3! |
This attempts to fix #304.
With this patch, I no longer reproduce the issue using the same large object file that triggered it before.
I've also updated the test code to handle XINDEX for the strtab, however I don't know a good way to generate a test case for an ELF file with >=2^16 sections without embedding a very large blob, so I've left it for future work.
(If I've missed a simple way to do this, happy to do it now)