x86_64 and other targets finish_dynamic_symbol
Alan Modra
amodra@gmail.com
Thu Feb 15 10:59:51 GMT 2024
More information about the Binutils mailing list
Thu Feb 15 10:59:51 GMT 2024
- Previous message (by thread): x86_64 and other targets finish_dynamic_symbol
- Next message (by thread): [PATCH v3 0/1] Move callx (nee callr) to v1 ISA variant
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
On Thu, Feb 15, 2024 at 07:14:00PM +1030, Alan Modra wrote: > On Thu, Feb 15, 2024 at 11:56:49AM +1030, Alan Modra wrote: > > On Tue, Feb 13, 2024 at 04:03:52PM -0800, H.J. Lu wrote: > > > On Tue, Feb 13, 2024 at 3:53 PM Alan Modra <amodra@gmail.com> wrote: > > > > I propose changing all occurrences of "return false" in > > > > finish_dynamic_symbol functions where targets don't first emit an > > > > error message, to BFD_ASSERTs. Opinions? > > > > > > Sounds good to me. > > > > This is what I'm about to apply. > > I didn't examine ld testsuite logs properly after cf95b909e2c2. > Replacing one of the "return false" with BFD_ASSERT in > finish_dynamic_symbol was wrong as it causes segmentation faults on > testcases expected to fail. Revert those changes and instead make > a bfd_final_link failure noisy. And this. Making a bfd_final_link failure noisy requires testsuite changes. diff --git a/ld/testsuite/ld-arm/non-contiguous-arm7.err b/ld/testsuite/ld-arm/non-contiguous-arm7.err index 3185ef58ffc..7b3a3d8c3c9 100644 --- a/ld/testsuite/ld-arm/non-contiguous-arm7.err +++ b/ld/testsuite/ld-arm/non-contiguous-arm7.err @@ -2,3 +2,4 @@ .* may change behaviour for section .?\.bss\.MY_BUF.? from .* .* discards section .?\.bss\.MY_BUF.? from .* .* unresolvable R_ARM_ABS32 relocation against symbol .?MY_BUF.? +.* final link failed diff --git a/ld/testsuite/ld-mips-elf/n64-plt-2.ed b/ld/testsuite/ld-mips-elf/n64-plt-2.ed index 4a42ec1e341..effdb1b522a 100644 --- a/ld/testsuite/ld-mips-elf/n64-plt-2.ed +++ b/ld/testsuite/ld-mips-elf/n64-plt-2.ed @@ -1 +1,2 @@ [^\n]*: `\.got\.plt' entry VMA of 0x7fff8000 outside the 32-bit range supported; consider using `-Ttext-segment=\.\.\.' +.* final link failed diff --git a/ld/testsuite/ld-mips-elf/n64-plt-3.ed b/ld/testsuite/ld-mips-elf/n64-plt-3.ed index 29a346b274d..400d91871a8 100644 --- a/ld/testsuite/ld-mips-elf/n64-plt-3.ed +++ b/ld/testsuite/ld-mips-elf/n64-plt-3.ed @@ -1 +1,2 @@ [^\n]*: `\.got\.plt' start VMA of 0xffffffff7fff7ff0 outside the 32-bit range supported; consider using `-Ttext-segment=\.\.\.' +.* final link failed -- Alan Modra Australia Development Lab, IBM
- Previous message (by thread): x86_64 and other targets finish_dynamic_symbol
- Next message (by thread): [PATCH v3 0/1] Move callx (nee callr) to v1 ISA variant
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the Binutils mailing list