Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Policy for VERIFY_CHECK and #ifdef VERIFY #1381

Closed
real-or-random opened this issue Jul 24, 2023 · 5 comments · Fixed by #1393
Closed

Policy for VERIFY_CHECK and #ifdef VERIFY #1381

real-or-random opened this issue Jul 24, 2023 · 5 comments · Fixed by #1393

Comments

@real-or-random
Copy link
Contributor

Current situation

The purpose of VERIFY_CHECK(cond) is to assert cond in test builds, i.e., when VERIFY is defined. While cond is supposed to have no side effects, VERIFY_CHECK will evaluate it even in production builds, to make sure that no side effects are masked in production builds.

In the case of simple cond, we can rely on the ability of the compiler to detect that cond has no side effects and to omit its evaluation entirely. However, many recently introduced cond are neither simple nor have negligible computation costs. In this case, we usually wrap VERIFY_CHECK(cond) in an #ifdef VERIFY block, which forces the evaluation to be omitted in production, at the cost of foregoing the aforementioned "side-effect safety" of VERIFY_CHECK. [1]

On top of this, we now have (full) functions like secp256k1_fe_verify, which internally perform VERIFY_CHECKs.

I don't think anything we currently do is wrong, but sometimes the borders are a bit arbitrary, and with a mix of #ifdef VERIFY and no #ifdef VERIFY, and a mix of uppercase and lowercase, also readability and consistency suffer.

Options

When it comes to forcing evaluations to be omitted, we can either keep the current convention and force omitting on a case-by-case basis. Alternatively, we could just always force omitting, e.g., by defining VERIFY_CHECK to be empty in production. That will keep things simple and save us the decision of whether to force omitting. The comment in #904 (comment) appears to agree with this.

When it comes to readability, I think that #ifdef blocks tend to clutter the code. If we redefine VERIFY_CHECK to be empty in production, we could omit many #ifdef lines. Even if we keep two types of VERIFY_CHECKs, I suggest introducing a separate macro instead to replace #ifdef.

One remaining purpose of the #ifdefs is to highlight the test code blocks, e.g., when we call secp256k1_fe_verify in field_impl.h. But then we could just make these function names uppercase. And if we want blocks/grouping for better readability, we can simply use empty lines instead of #ifdef lines.

So I suggest

  • redefining VERIFY_CHECK to be empty in production
  • renaming _check functions to uppercase.

And while we're touching this anyway, we could consider renaming "check" to "assert", which is a more precise term. (In fact, if we redefine VERIFY_CHECK to be empty in production, we have almost reimplemented assert.h...)

Does this make sense?


[1] One idea was to improve this situation was to introduce a C hack that will tell us whether a compiler is able to determine that a check has no side effects, but I agree with #904 (comment) that this is not the way to go. In addition to the points brought up by @sipa, relying on the compiler is just very brittle, with optimizations depending on compiler versions and flags.

@jonasnick
Copy link
Contributor

Adding to the current situation is that VERIFY_CHECK is not side-effect safe when built with --enable-coverage which lead to bugs in coverage mode in the past. In fact, there are again unused variable warnings right now when compiling master with --enable-coverage. So an additional advantage of your suggestion is that VERIFY_CHECK in coverage mode isn't special anymore which would hopefully reduce maintenance of that mode.

@sipa
Copy link
Contributor

sipa commented Jul 26, 2023

Agree, I think we should make VERIFY_CHECK compile to nothing in non-VERIFY mode. It was worthwhile to aim for side-effect freeness, but the way to codebase has evolved I believe makes the benefit not worth the cost. As pointed out in #904, the benefit is effectively restricted already to the simplest cases, where arguably it doesn't matter much anyway.

I wouldn't actually rename the verify_ functions to be upper case, but instead leave them as lowercase functions, but with an uppercase macro around them (maybe one that also compiles to nothing in non-VERIFY mode, so that not even a dummy verify_ function is needed in that case). This is just for style reasons - I think it's good that a reader can guess what is a function and what is a macro just on the name.

@real-or-random
Copy link
Contributor Author

I wouldn't actually rename the verify_ functions to be upper case, but instead leave them as lowercase functions, but with an uppercase macro around them [...]

Sounds good to me.

@theStack Are you interested to look into this by any chance?

@theStack
Copy link
Contributor

theStack commented Aug 3, 2023

I wouldn't actually rename the verify_ functions to be upper case, but instead leave them as lowercase functions, but with an uppercase macro around them [...]

Sounds good to me.

@theStack Are you interested to look into this by any chance?

Yes, planning to tackle this within the next days.

theStack added a commit to theStack/secp256k1 that referenced this issue Aug 4, 2023
As suggested in issue bitcoin-core#1381, this will make things simpler and
improve code readability, as we don't need to force omitting of
evaluations on a case-by-case basis anymore and hence can remove
lots of `#ifdef VERIFY`/`#endif` lines (see next commit). Plus,
VERIFY_CHECK behaves now identical in both non-VERIFY and coverage mode,
making the latter not special anymore and hopefully decreasing
maintenance burden. The idea of "side-effect safety" is given up.

Note that at two places in the ellswift module void-casts of return
values have to be inserted for non-VERIFY builds, in order to avoid
   "variable ... set but not used [-Wunused-but-set-variable]"
warnings.
theStack added a commit to theStack/secp256k1 that referenced this issue Aug 16, 2023
As suggested in issue bitcoin-core#1381, this will make things simpler and
improve code readability, as we don't need to force omitting of
evaluations on a case-by-case basis anymore and hence can remove
lots of `#ifdef VERIFY`/`#endif` lines (see next commit). Plus,
VERIFY_CHECK behaves now identical in both non-VERIFY and coverage mode,
making the latter not special anymore and hopefully decreasing
maintenance burden. The idea of "side-effect safety" is given up.

Note that at two places in the ellswift module void-casts of return
values have to be inserted for non-VERIFY builds, in order to avoid
   "variable ... set but not used [-Wunused-but-set-variable]"
warnings.
theStack added a commit to theStack/secp256k1 that referenced this issue Aug 18, 2023
As suggested in issue bitcoin-core#1381, this will make things simpler and
improve code readability, as we don't need to force omitting of
evaluations on a case-by-case basis anymore and hence can remove
lots of `#ifdef VERIFY`/`#endif` lines (see next commit). Plus,
VERIFY_CHECK behaves now identical in both non-VERIFY and coverage mode,
making the latter not special anymore and hopefully decreasing
maintenance burden. The idea of "side-effect safety" is given up.

Note that at two places in the ellswift module void-casts of return
values have to be inserted for non-VERIFY builds, in order to avoid
   "variable ... set but not used [-Wunused-but-set-variable]"
warnings.
theStack added a commit to theStack/secp256k1 that referenced this issue Sep 21, 2023
As suggested in issue bitcoin-core#1381, this will make things simpler and
improve code readability, as we don't need to force omitting of
evaluations on a case-by-case basis anymore and hence can remove
lots of `#ifdef VERIFY`/`#endif` lines (see next commit). Plus,
VERIFY_CHECK behaves now identical in both non-VERIFY and coverage mode,
making the latter not special anymore and hopefully decreasing
maintenance burden. The idea of "side-effect safety" is given up.

Note that at two places in the ellswift module void-casts of return
values have to be inserted for non-VERIFY builds, in order to avoid
   "variable ... set but not used [-Wunused-but-set-variable]"
warnings.
theStack added a commit to theStack/secp256k1 that referenced this issue Dec 1, 2023
As suggested in issue bitcoin-core#1381, this will make things simpler and
improve code readability, as we don't need to force omitting of
evaluations on a case-by-case basis anymore and hence can remove
lots of `#ifdef VERIFY`/`#endif` lines (see next commit). Plus,
VERIFY_CHECK behaves now identical in both non-VERIFY and coverage mode,
making the latter not special anymore and hopefully decreasing
maintenance burden. The idea of "side-effect safety" is given up.

Note that at two places in the ellswift module void-casts of return
values have to be inserted for non-VERIFY builds, in order to avoid
   "variable ... set but not used [-Wunused-but-set-variable]"
warnings.
real-or-random added a commit that referenced this issue Dec 1, 2023
…(issue #1381)

bb46723 remove VERIFY_SETUP define (Sebastian Falbesoner)
a3a3e11 remove unneeded VERIFY_SETUP uses in ECMULT_CONST_TABLE_GET_GE macro (Sebastian Falbesoner)
a0fb68a introduce and use SECP256K1_SCALAR_VERIFY macro (Sebastian Falbesoner)
cf25c86 introduce and use SECP256K1_{FE,GE,GEJ}_VERIFY macros (Sebastian Falbesoner)
5d89bc0 remove superfluous `#ifdef VERIFY`/`#endif` preprocessor conditions (Sebastian Falbesoner)
c2688f8 redefine VERIFY_CHECK to empty in production (non-VERIFY) mode (Sebastian Falbesoner)

Pull request description:

  As suggested in #1381, this PR reworks the policy for VERIFY_CHECK and when to use #ifdef VERIFY, by:
  - redefining VERIFY_CHECK to empty in production (non-VERIFY) mode
  - removing many then superflous #ifdef VERIFY blocks (if they exclusively contained VERIFY_CHECKs)
  - introducing uppercase macros around verify_ functions and using them for better readabiliy

  What is _not_ included yet is the proposed renaming from "_check" to "_assert":
  > And while we're touching this anyway, we could consider renaming "check" to "assert", which is a more precise term. (In fact, if we redefine VERIFY_CHECK to be empty in production, we have almost reimplemented assert.h...)

  This should be easy to achieve with simple search-and-replace (e.g. using sed), but I was hesitant as this would probably case annoying merge conflicts on some of the open PRs. Happy to add this if the rename if desired (#1381 didn't get any feedback about the renaming idea yet).

ACKs for top commit:
  stratospher:
    ACK bb46723.
  real-or-random:
    utACK bb46723

Tree-SHA512: 226ca609926dea638aa3bb537d29d4fac8b8302dcd9da35acf767ba9573e5221d2dae04ea26c15d80a50ed70af1ab0dca10642c21df7dbdda432fa237a5ef2cc
@real-or-random
Copy link
Contributor Author

real-or-random commented Dec 1, 2023

This is done, except for the potential renaming to "assert", which is now tracked in #1449.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging a pull request may close this issue.

4 participants