Re: [PATCH 1/1] gdb, gdbserver: Add Intel APX register support

Eli Zaretskii <[email protected]> Sat, 25 Jul 2026 14:07:41 +0300
Newsgroups gmane.comp.gdb.patches
Message-ID <[email protected]>
> From: Christina Joos <[email protected]>
> Cc: [email protected]
> Date: Sat, 25 Jul 2026 10:53:35 +0000
> 
> From: "H.J. Lu" <[email protected]>
> 
> Intel Advanced Performance Extensions (APX) doubles the number of
> general-purpose registers (GPRs) from 16 to 32 (r16-r31).
> 
> Those new 64-bit Extended GPRs (EGPRs) do not change the size of the
> XSAVE area, as they take up the space left behind by the deprecated
> MPX registers.
> 
> In contrast to the legacy GPRS, EGPRs are not enabled by default in 64-bit
> mode but are xcr0 enabled based on APX_F Intel APX state.  Define
> X86_XSTATE_APX_F to support this new configuration.
> 
> Add gdb and gdbserver registers support for those new 64-bit EGPRs.
> Also add byte, word and dword pseudo register support.
> 
> Co-Authored-By:  Christina Joos <[email protected]>
> ---
>  gdb/NEWS                                      |   4 +
>  gdb/amd64-linux-tdep.c                        |   2 +
>  gdb/amd64-tdep.c                              | 262 ++++++++++++++----
>  gdb/amd64-tdep.h                              |   2 +
>  gdb/arch/amd64.c                              |   4 +
>  gdb/arch/x86-linux-tdesc-features.c           |   1 +
>  gdb/doc/gdb.texinfo                           |   5 +
>  gdb/features/Makefile                         |   1 +
>  gdb/features/i386/64bit-apx.c                 |  29 ++
>  gdb/features/i386/64bit-apx.xml               |  26 ++
>  gdb/i386-tdep.c                               |  32 ++-
>  gdb/i386-tdep.h                               |  26 ++
>  gdb/i387-tdep.c                               | 105 ++++++-
>  gdb/i387-tdep.h                               |   6 +
>  gdb/nat/x86-xstate.c                          |   1 +
>  gdb/target-debug.h                            |   1 +
>  gdb/testsuite/gdb.arch/amd64-apx-core.c       |  88 ++++++
>  gdb/testsuite/gdb.arch/amd64-apx-corefile.c   |  71 +++++
>  gdb/testsuite/gdb.arch/amd64-apx-corefile.exp |  67 +++++
>  gdb/testsuite/gdb.arch/amd64-apx.c            |  66 +++++
>  gdb/testsuite/gdb.arch/amd64-apx.exp          |  77 +++++
>  gdb/testsuite/lib/gdb.exp                     |  58 ++++
>  gdbserver/i387-fp.cc                          |  43 +++
>  gdbserver/linux-x86-low.cc                    |   4 +-
>  gdbsupport/x86-xstate.h                       |  12 +-
>  25 files changed, 931 insertions(+), 62 deletions(-)
>  create mode 100644 gdb/features/i386/64bit-apx.c
>  create mode 100644 gdb/features/i386/64bit-apx.xml
>  create mode 100644 gdb/testsuite/gdb.arch/amd64-apx-core.c
>  create mode 100644 gdb/testsuite/gdb.arch/amd64-apx-corefile.c
>  create mode 100644 gdb/testsuite/gdb.arch/amd64-apx-corefile.exp
>  create mode 100644 gdb/testsuite/gdb.arch/amd64-apx.c
>  create mode 100644 gdb/testsuite/gdb.arch/amd64-apx.exp

Thanks.

> diff --git a/gdb/NEWS b/gdb/NEWS
> index 8f40ca5cb11..5b3ba152188 100644
> --- a/gdb/NEWS
> +++ b/gdb/NEWS
> @@ -3,6 +3,10 @@
>  
>  *** Changes since GDB 17
>  
> +* Support for Intel APX registers on amd64 GNU/Linux.
> +  Support displaying and modifying Intel APX registers (Extended GPRs
> +  $r16 - $r31) including byte, word and dword pseudo registers.
> +
>  * Support for the Common Trace Format (CTF) has been removed.  GDB now
>    saves trace information exclusively in its own "tfile" format.

This part is okay.

> diff --git a/gdb/doc/gdb.texinfo b/gdb/doc/gdb.texinfo
> index 0030698dcee..69a2d714299 100644
> --- a/gdb/doc/gdb.texinfo
> +++ b/gdb/doc/gdb.texinfo
> @@ -50753,6 +50753,11 @@ amd64, 32 bits on amd64 with 32-bit pointer size (X32) and 32 bits on i386.
>  Following the restriction of the Linux kernel, only @value{GDBN} for amd64
>  targets makes use of this feature for now.
>  
> +The @samp{org.gnu.gdb.i386.apx} feature is optional.  It should describe
> +16 Extended General Purpose registers (@acronym{EGPRs})
> +@samp{r16} - @samp{r31} with 64 bits each.  All @acronym{Intel APX}
> +registers are valid only on amd64.

This is also OK, but please use @acronym only for "APX", leaving the
"Intel" part outside of it.

Reviewed-By: Eli Zaretskii <[email protected]>