Use the correct constants when setting the section type of HPPA unwind sections.

* elf-hppa.h (elf_hppa_fake_sections): Use SHT_PARISC_UNWIND as
	the section type of the .PARISC.unwind section on 64-bit binaries
	and SHT_PROGBITS for 32-bit binaries.  Add a comment about it.
	Add comment about the sh_entsize value.
This commit is contained in:
Helge Deller 2018-08-22 10:32:53 +01:00 committed by Nick Clifton
parent 3da64fe404
commit 6f4286c4f7
2 changed files with 16 additions and 3 deletions

View File

@ -1,3 +1,11 @@
2018-08-22 Helge Deller <deller@gmx.de>
Nick Clifton <nickc@redhat.com>
* elf-hppa.h (elf_hppa_fake_sections): Use SHT_PARISC_UNWIND as
the section type of the .PARISC.unwind section on 64-bit binaries
and SHT_PROGBITS for 32-bit binaries. Add a comment about it.
Add comment about the sh_entsize value.
2018-08-22 Rafeal Auler <rafaelauler@gmail.com>
PR 23560

View File

@ -889,9 +889,11 @@ elf_hppa_fake_sections (bfd *abfd, Elf_Internal_Shdr *hdr, asection *sec)
asection *asec;
#if ARCH_SIZE == 64
hdr->sh_type = SHT_LOPROC + 1;
hdr->sh_type = SHT_PARISC_UNWIND;
#else
hdr->sh_type = 1;
/* Note - it is not clear why this is not SHT_PARISC_UNWIND as well.
Presumably it is a historical constraint, so leave it as it is. */
hdr->sh_type = SHT_PROGBITS;
#endif
/* ?!? How are unwinds supposed to work for symbols in arbitrary
sections? Or what if we have multiple .text sections in a single
@ -912,7 +914,10 @@ elf_hppa_fake_sections (bfd *abfd, Elf_Internal_Shdr *hdr, asection *sec)
}
}
/* I have no idea if this is really necessary or what it means. */
/* The unwind table entries are 16 bytes long, so it is not clear
why this field is set to 4. (The ELF spec says that the sh_entsize
field is a byte quantity, but this is a processor specific section,
so it is allowed to change the rules). Leave as it is for now. */
hdr->sh_entsize = 4;
}
return TRUE;