This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
[PATCH][BZ #17078] ARM: R_ARM_TLS_DESC prelinker support
- From: "Maciej W. Rozycki" <macro at codesourcery dot com>
- To: <libc-alpha at sourceware dot org>
- Cc: "Joseph S. Myers" <joseph at codesourcery dot com>
- Date: Tue, 1 Jul 2014 14:28:04 +0100
- Subject: [PATCH][BZ #17078] ARM: R_ARM_TLS_DESC prelinker support
- Authentication-results: sourceware.org; auth=none
Hello,
Here is a change to the dynamic linker to add prelinker support for the
R_ARM_TLS_DESC relocation. Two cases can be considered here, the usual
one where lazy binding is in use and the less frequent one, where
immediate binding is requested via the use of the DF_BIND_NOW dynamic flag
(e.g. by using the GNU linker's "-z now" option).
The change below only handle the first case. In this scenario the
prelinker does what the dynamic linker would do, that is preinitialises
R_ARM_TLS_DESC relocations with a pointer to the lazy specialisation as
provided with the DT_TLSDESC_PLT dynamic tag. A conflict is additionally
created and in the conflict resolution path the dynamic linker complements
the work by initialising the object's pointer as indicated by the
DT_TLSDESC_GOT dynamic tag to the linker's internal lazy specialisation
worker function and also providing the associated link map in the second
entry of the GOT. This step is required, because if prelinking is
successful at the run time, then the dynamic linker's
elf_machine_runtime_setup() function isn't called that would normally do
so.
The second case remains unresolved, because support for that scenario has
not been implemented in the prelinker. In this case the lazy
specialisation is unavailable and the DT_TLSDESC_PLT dynamic tag is not
present.
The prelinker could assume the common case of static specialisation and
resolve the relocation, but that would require the exposure of dynamic
linker's specialisation worker function. Furthermore the dynamic linker
would have to handle the relocation in the conflict resoultion path and
see if the dynamic specialisation shouldn't be used instead. This however
would require access to data structures currently not made available to
the conflict resoultion path and therefore a redesign of this part of the
dynamic linker.
Alternatively the prelinker could defer all processing to the dynamic
linker's conflict resolution path, but that would require similar access
to the said data structures.
Therefore the prelinker issues an error instead and the dynamic linker
has assertions to check that DT_TLSDESC_PLT and DT_TLSDESC_GOT are in use
in its conflict resolution path.
The changes below resolve all TLS failures in the prelinker testsuite,
as noted in the bug report, as well as the small test case provided there.
Unfortunately we don't seem to have any hooks to factor in the prelinker
(if present on a system) to testing, so unless someone has an idea how to
do that, this fix will have to rely on using the prelinker test suite and
enabling TLS descriptors there for coverage.
This has been verified across a number of multilibs with no regressions
in our test suite. OK to apply?
2014-07-01 Maciej W. Rozycki <macro@codesourcery.com>
[BZ #17078]
* sysdeps/arm/dl-machine.h (elf_machine_rela)
[RESOLVE_CONFLICT_FIND_MAP]: Handle R_ARM_TLS_DESC relocation.
(elf_machine_lazy_rel): Handle prelinked R_ARM_TLS_DESC entries.
Maciej
glibc-arm-tlsdesc-prelink.patch
Index: glibc-fsf-trunk-quilt/sysdeps/arm/dl-machine.h
===================================================================
--- glibc-fsf-trunk-quilt.orig/sysdeps/arm/dl-machine.h 2014-06-23 03:36:48.621961686 +0100
+++ glibc-fsf-trunk-quilt/sysdeps/arm/dl-machine.h 2014-06-23 03:37:04.621789475 +0100
@@ -573,6 +573,32 @@ elf_machine_rela (struct link_map *map,
case R_ARM_ABS32:
*reloc_addr = value + reloc->r_addend;
break;
+# ifdef RESOLVE_CONFLICT_FIND_MAP
+ case R_ARM_TLS_DESC:
+ {
+ struct tlsdesc volatile *td =
+ (struct tlsdesc volatile *) reloc_addr;
+
+ RESOLVE_CONFLICT_FIND_MAP (map, reloc_addr);
+
+ /* Make sure we know what's going on. */
+ assert (td->entry
+ == (void *) (D_PTR (map, l_info[ADDRIDX (DT_TLSDESC_PLT)])
+ + map->l_addr));
+ assert (map->l_info[ADDRIDX (DT_TLSDESC_GOT)]);
+
+ /* Set up the lazy resolver and store the pointer to our link
+ map in _GLOBAL_OFFSET_TABLE[1] now as for a prelinked
+ binary elf_machine_runtime_setup() is not called and hence
+ neither has been initialized. */
+ *(Elf32_Addr *) (D_PTR (map, l_info[ADDRIDX (DT_TLSDESC_GOT)])
+ + map->l_addr)
+ = (Elf32_Addr) &_dl_tlsdesc_lazy_resolver;
+ ((Elf32_Addr *) D_PTR (map, l_info[DT_PLTGOT]))[1]
+ = (Elf32_Addr) map;
+ }
+ break;
+# endif /* RESOLVE_CONFLICT_FIND_MAP */
case R_ARM_PC24:
relocate_pc24 (map, value, reloc_addr, reloc->r_addend);
break;
@@ -652,9 +678,11 @@ elf_machine_lazy_rel (struct link_map *m
(struct tlsdesc volatile *)reloc_addr;
/* The linker must have given us the parameter we need in the
- first GOT entry, and left the second one empty. We fill the
- last with the resolver address */
- assert (td->entry == 0);
+ first GOT entry, and left the second one empty. The latter
+ will have been preset by the prelinker if used though.
+ We fill it with the resolver address. */
+ assert (td->entry == 0
+ || map->l_info[VALIDX (DT_GNU_PRELINKED)] != NULL);
td->entry = (void*)(D_PTR (map, l_info[ADDRIDX (DT_TLSDESC_PLT)])
+ map->l_addr);
}