[FMV][AArch64] Remove features which expose non exploitable runtime behavior.

[labrinea]

Feature rpres allows an increase in the precision of the single-precision
floating-point reciprocal estimate and reciprocal square root estimate
from an 8-bit mantissa to a 12-bit mantissa depending on the value FPCR.AH
of the Floating-point Control Register.

Similarly, ebf16 allows existing floating-point instructions (BFDOT, BFMMLA,
BFMOPA, BFMOPS, and BFVDOT) to change numeric behaviour depending on the
value FPCR.EBF of the Floating-point Control Register.

Feature memtag3 allows Tag Check Faults to change behaviour depending on
the value SCTLR_ELx.{TCF, TCF0} of the System Control Register.

The runtime detection in FMV does not examine the content of control
registers, therefore runtime dispatch cannot be based on that. One may
argue there is value in altering the control register from a version,
for example "msr dit, #1" if the feature is available on hardware. The
registers FPCR and SCTLR_ELx can be accessed in the absence of rpres,
ebf16, and memtag3, making it hard to justify adding them to the compiler.

---

name: Pull request
about: Technical issues, document format problems, bugs in scripts or feature proposal.

---

Thank you for submitting a pull request!

If this PR is about a bugfix:

Please use the bugfix label and make sure to go through the checklist below.

If this PR is about a proposal:

We are looking forward to evaluate your proposal, and if possible to
make it part of the Arm C Language Extension (ACLE) specifications.

We would like to encourage you reading through the contribution
guidelines, in particular the section on submitting
a proposal.

Please use the proposal label.

As for any pull request, please make sure to go through the below
checklist.

Checklist: (mark with X those which apply)

• If an issue reporting the bug exists, I have mentioned it in the
  PR (do not bother creating the issue if all you want to do is
  fixing the bug yourself).
• I have added/updated the SPDX-FileCopyrightText lines on top
  of any file I have edited. Format is SPDX-FileCopyrightText: Copyright {year} {entity or name} <{contact informations}>
  (Please update existing copyright lines if applicable. You can
  specify year ranges with hyphen , as in 2017-2019, and use
  commas to separate gaps, as in 2018-2020, 2022).
• I have updated the Copyright section of the sources of the
  specification I have edited (this will show up in the text
  rendered in the PDF and other output format supported). The
  format is the same described in the previous item.
• I have run the CI scripts (if applicable, as they might be
  tricky to set up on non-*nix machines). The sequence can be
  found in the contribution
  guidelines. Don't
  worry if you cannot run these scripts on your machine, your
  patch will be automatically checked in the Actions of the pull
  request.
• I have added an item that describes the changes I have
  introduced in this PR in the section Changes for next
  release of the section Change Control/Document history
  of the document. Create Changes for next release if it does
  not exist. Notice that changes that are not modifying the
  content and rendering of the specifications (both HTML and PDF)
  do not need to be listed.
• When modifying content and/or its rendering, I have checked the
  correctness of the result in the PDF output (please refer to the
  instructions on how to build the PDFs
  locally).
• The variable draftversion is set to true in the YAML header
  of the sources of the specifications I have modified.
• Please DO NOT add my GitHub profile to the list of contributors
  in the README page of the project.

[labrinea]

@andrewcarlotti @DanielKristofKiss @tmatheson-arm @Wilco1

[andrewcarlotti]

LGTM

[andrewcarlotti]

The same justification applies to dit - the behaviour is determined by a control bit, and the architecture feature merely allows this bit to be switched on.

[labrinea]

Added feature dit to this list with justification.

[Wilco1]

Looks good to me.

[jroelofs]

I don't agree with the justification for removal of dit: you still need the ISA feature presence check around msr dit, x0 instructions that toggle the feature, and prior to this change, FMV could be used for that.

[labrinea]

I don't agree with the justification for removal of dit: you still need the ISA feature presence check around msr dit, x0 instructions that toggle the feature, and prior to this change, FMV could be used for that.

I kinda see your point, but the value of CPSR.DIT cannot be detected at runtime in order to choose a function version based on that. Even if it was the case I still can't think of a meaningful example. Perhaps something like the following snippet?

__attribute__((target_version("dit"))) bool enableDIT(void) { asm volatile ("msr dit, x0"); return true; }
__attribute__((ttarget_version("default"))) bool enableDIT(void) { return false; }

if (enableDIT())
  // do something
else
  // do something else

I am not sure what kind of decisions can be made in the if-else logic based on this.

[jroelofs]

I feel this should be sufficient justification to keep it:

__attribute__((target_version("dit"))) void enableDITIfExists(void) { asm volatile ("msr dit, #1"); }
__attribute__((target_version("default"))) void enableDITIfExists(void) { }

[jroelofs]

This solves a very real problem, by the way: https://developer.apple.com/documentation/xcode/writing-arm64-code-for-apple-platforms#Enable-DIT-for-constant-time-cryptographic-operations

[jroelofs]

I don't agree with the justification for removal of dit: you still need the ISA feature presence check around msr dit, x0 instructions that toggle the feature, and prior to this change, FMV could be used for that.

I kinda see your point, but the value of CPSR.DIT cannot be detected at runtime in order to choose a function version based on that

By "ISA feature presence check" I mean whether or not the cpu supports DIT at all, not runtime detection of whether or not DIT is enabled. Those two axes are almost orthogonal, but FMV is only useful for the former, so I agree that it is not useful for the latter.

[vhscampos]

Thanks. Merging now