[RFD 0/5] Add support for MTE stack tagging.
Luis Machado
luis.machado@arm.com
Fri Dec 6 17:40:33 GMT 2024
Hi,
On 12/5/24 22:58, Indu Bhagat wrote:
> Hello,
>
> I am sending some WIP patches to follow up on some of the questions I
> have had with regard to Binutils support for MTE stack tagging.
> https://gcc.gnu.org/pipermail/gcc-patches/2024-November/670319.html
>
> Since sending of the above message to gcc-patches mailing list, I
> changed the command line options a bit. I also realised that some text
> in the gcc-patches message was a bit incorrect. So I am sending the
> patches, hopefully that conveys the picture better.
>
> The commit log of each commit has some of the above questions listed
> too. Most of the questions are around the implementation of the two
> dynamic tags (related to MTE stack tagging). Although the patchset
> addresses three different aspects (i.e., .cfi_mte_tagged_frame CFI
> directive, DT_AARCH64_MEMTAG_MODE and DT_AARCH64_MEMTAG_STACK), I am
> sending them together for discussion/coordination for next steps.
>
> (PS: I can get rid of the ChangeLog in the commit message later when the
> patches are more mature. Also, I have an erratic and sparse work
> schedule for December, my responses may be slow.)
>
> Thanks,
>
> Indu Bhagat (5):
> binutils: make read_cie aware of new augmentation char 'G'
> gas: aarch64: suppport CFI directive .cfi_mte_tagged_frame
> ld: aarch64: make EH Frame parsing aware of augmentation char 'G'
> ld/aarch64elf: add support for DT_AARCH64_MEMTAG_MODE dynamic tag
> ld/aarch64elf: add support for DT_AARCH64_MEMTAG_STACK dynamic tag
>
> bfd/elf-eh-frame.c | 2 +
> bfd/elfnn-aarch64.c | 21 ++++++++-
> bfd/elfxx-aarch64.h | 27 ++++++++++-
> binutils/dwarf.c | 2 +
> binutils/readelf.c | 3 ++
> gas/config/tc-aarch64.c | 9 ++++
> gas/config/tc-aarch64.h | 17 +++++--
> gas/doc/c-aarch64.texi | 10 ++++
> gas/testsuite/gas/aarch64/mte_tagged_stack.d | 47 +++++++++++++++++++
> gas/testsuite/gas/aarch64/mte_tagged_stack.s | 24 ++++++++++
> include/elf/aarch64.h | 2 +
> ld/emultempl/aarch64elf.em | 47 ++++++++++++++++++-
> ld/ld.texi | 14 ++++++
> ld/testsuite/ld-aarch64/aarch64-elf.exp | 4 ++
> ld/testsuite/ld-aarch64/dt-memtag-mode.d | 7 +++
> ld/testsuite/ld-aarch64/dt-memtag-stack.d | 7 +++
> ld/testsuite/ld-aarch64/dt-memtag.s | 7 +++
> .../ld-aarch64/mte-tagged-frame-bar.s | 17 +++++++
> .../ld-aarch64/mte-tagged-frame-foo.s | 16 +++++++
> ld/testsuite/ld-aarch64/mte-tagged-frame.d | 29 ++++++++++++
> 20 files changed, 303 insertions(+), 9 deletions(-)
> create mode 100644 gas/testsuite/gas/aarch64/mte_tagged_stack.d
> create mode 100644 gas/testsuite/gas/aarch64/mte_tagged_stack.s
> create mode 100644 ld/testsuite/ld-aarch64/dt-memtag-mode.d
> create mode 100644 ld/testsuite/ld-aarch64/dt-memtag-stack.d
> create mode 100644 ld/testsuite/ld-aarch64/dt-memtag.s
> create mode 100644 ld/testsuite/ld-aarch64/mte-tagged-frame-bar.s
> create mode 100644 ld/testsuite/ld-aarch64/mte-tagged-frame-foo.s
> create mode 100644 ld/testsuite/ld-aarch64/mte-tagged-frame.d
>
Do you envision gdb having to know any of this for debugging purposes? I know
gdb parses the augmentation character, but it is not clear gdb can do anything
useful with it.
More information about the Binutils
mailing list