mirror/binutils-gdb
2
0
Fork 0
mirror of https://sourceware.org/git/binutils-gdb.git synced 2025-01-12 12:16:04 +08:00
Code Issues Packages Projects Releases Wiki Activity
69a284867c
binutils-gdb/ld/testsuite/ld-elf/pr26580-2.sd

5 lines
79 B
Plaintext
Raw Normal View History

PR26580, Size and alignment of commons vs as-needed shared lib Two pieces to this puzzle: 1) Revert HJ's fix for PR13250 so that size and alignment isn't sticky, instead attack the real underlying problem that _bfd_generic_link_add_one_symbol does the wrong thing in making a common section in a shared library bfd. 2) Save and restore common u.c.p fields, which hold the section and alignment. A better fix for (2) would be to throw away all of that horrible code saving and restoring the hash table when loading as-needed library symbols, and instead do a scan over as-needed library symbols before adding anything. bfd/ PR 13250 PR 26580 * elflink.c (_bfd_elf_merge_symbol): Make "override" a bfd**. Return oldbfd in override when old common should override new common. (_bfd_elf_add_default_symbol): Adjust to suit. (elf_link_add_object_symbols): Likewise. Pass "override" to _bfd_generic_link_add_one_symbol. Save and restore common u.c.p field for --as-needed shared libraries. Revert pr13250 changes. ld/ * testsuite/ld-elf/pr26580-a.s, * testsuite/ld-elf/pr26580-b.s, * testsuite/ld-elf/pr26580-1.sd, * testsuite/ld-elf/pr26580-2.sd: New tests * testsuite/ld-elf/comm-data.exp: Run new tests. * testsuite/ld-elf/pr26580-a.c, * testsuite/ld-elf/pr26580-b.c, * testsuite/ld-elf/pr26580-3.out, * testsuite/ld-elf/pr26580-4.out: New tests. * testsuite/ld-elf/shared.exp: Run new tests.
2020-09-08 11:32:31 +08:00
#...
Re: PR26580, Size and alignment of commons vs as-needed shared lib Some MIPS targets, for reasons I didn't analyse, use the larger common symbol in a shared lib rather than a smaller common in an executable. That doesn't seem unreasonable, so allow that to pass for pr26580-2. bfin-elf complains about not supporting copy relocs, but it's quite silly to want a copy reloc for common symbols, so leave the fail there. mn10300-elf and score-elf both fail the test with "PHDR segment not covered by LOAD segment". Other tests fail similarly so one more doesn't hurt. The failure is a consequence of supporting dynamic objects but setting EMBEDDED in ld scripts. PR 26580 * testsuite/ld-elf/pr26580-2.sd: Accept undefined symbol.
2020-09-10 17:16:14 +08:00
.* (8 OBJECT GLOBAL DEFAULT .*|0 OBJECT GLOBAL DEFAULT UND) one
PR26580, Size and alignment of commons vs as-needed shared lib Two pieces to this puzzle: 1) Revert HJ's fix for PR13250 so that size and alignment isn't sticky, instead attack the real underlying problem that _bfd_generic_link_add_one_symbol does the wrong thing in making a common section in a shared library bfd. 2) Save and restore common u.c.p fields, which hold the section and alignment. A better fix for (2) would be to throw away all of that horrible code saving and restoring the hash table when loading as-needed library symbols, and instead do a scan over as-needed library symbols before adding anything. bfd/ PR 13250 PR 26580 * elflink.c (_bfd_elf_merge_symbol): Make "override" a bfd**. Return oldbfd in override when old common should override new common. (_bfd_elf_add_default_symbol): Adjust to suit. (elf_link_add_object_symbols): Likewise. Pass "override" to _bfd_generic_link_add_one_symbol. Save and restore common u.c.p field for --as-needed shared libraries. Revert pr13250 changes. ld/ * testsuite/ld-elf/pr26580-a.s, * testsuite/ld-elf/pr26580-b.s, * testsuite/ld-elf/pr26580-1.sd, * testsuite/ld-elf/pr26580-2.sd: New tests * testsuite/ld-elf/comm-data.exp: Run new tests. * testsuite/ld-elf/pr26580-a.c, * testsuite/ld-elf/pr26580-b.c, * testsuite/ld-elf/pr26580-3.out, * testsuite/ld-elf/pr26580-4.out: New tests. * testsuite/ld-elf/shared.exp: Run new tests.
2020-09-08 11:32:31 +08:00
#pass
Reference in New Issue Copy Permalink