Re: [PATCH v3 13/16] [gdb/generic] corefile/bug: Fixup (gcore) core file target description reading order

[Thiago Jung Bauermann <thiago.bauermann@linaro.org>]

Luis Machado via Gdb-patches <gdb-patches@sourceware.org> writes:

> Due to the nature of the AArch64 SVE/SME extensions in GDB, each thread
> can potentially have distinct target descriptions/gdbarches.
>
> When loading a gcore-generated core file, at the moment GDB gives priority
> to the target description dumped to NT_GDB_TDESC.  Though technically correct
> for most target, it doesn't work correctly for AArch64 with SVE or SME
> support.
>
> The correct approach for AArch64/Linux is to rely on the
> gdbarch_core_read_description hook, so it can figure out the proper target
> description for a given thread based on the various available register notes.
>
> I think this should work for other architectures as well. If not, we may
> need to adjust things so all architectures get the information that they
> need for discovering the target description of the core file.

The code to load the target description from the core file was added in
commit 95ce627aeb9d "gdb: write target description into core file", by
Andrew Burgess. My understanding of that commit message is that it was
added before the gdbarch hook on purpose, and my reading of:

 “My primary motivation for adding this feature is that, in a future
  commit, I will be adding support for bare metal core dumps on some
  targets.  For RISC-V specifically, I want to be able to dump all the
  available control status registers.  As different targets will present
  different sets of register in their target description, including
  registers that are possibly not otherwise known to GDB I wanted a way
  to capture these registers in the core dump.”

is that it doesn't work if reading from the target description comes
last, though I'm not sure about that. I copied him in this email to see
if he would like to chime in.

-- 
Thiago