Conversation
paulmckrcu

Paul E. McKenney

Some good progress on lifetime-end pointer zap and out-of-thin-air values in C++29! https://people.kernel.org/paulmck/c-pointer-zap-and-oota-progress
1
2
5
uecker@mastodon.social

Martin Uecker

Reply to @paulmckrcu

@paulmckrcu Have the semantic issues with angelic non-determinism been resolved?

1
0
0
paulmckrcu

Paul E. McKenney

Reply to @uecker@mastodon.social
@uecker If you are asking how this maps to the C-language ISO/IEC TS 6010, that is yet to be determined. I long ago resigned myself to different approaches for C++, C, and Rust.
1
0
0
uecker@mastodon.social

Martin Uecker

Reply to @paulmckrcu

@paulmckrcu No I mean whether it has clearly been defined what angelic non-determinism means semantically.

1
0
0
paulmckrcu

Paul E. McKenney

Reply to @uecker@mastodon.social
@uecker Define "clearly defined". ;-)

I would say "yes", but it is probably better to point you to this paper and let you decide: https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2025/p2434r4.html

I would expect that this would not satisfy you, but perhaps the good news is that I expect that things will be more clear in the context of ISO/IEC TS 6010. It would still be the case that a pointer exposed concurrently with an access would use that pointer, though it might be possible to provide at least some limits on this.
1
0
0
uecker@mastodon.social

Martin Uecker

Reply to @paulmckrcu

@paulmckrcu You are right, I do not find this satisfying.

1
0
0
paulmckrcu

Paul E. McKenney

Reply to @uecker@mastodon.social
@uecker If it makes you feel any better, I do understand that a different approach to lifetime-end pointer zap will be required for the C language. Presumably something based on ISO/IEC TS 6010.
1
0
1
uecker@mastodon.social

Martin Uecker

Reply to @paulmckrcu

@paulmckrcu The problem is that C++ messing up their semantics usually also ends up hurting us in C, e.g. as for the time-travel UB which we never had but we still had to clarify this. But I guess I should be happy that C++ also moved away from this. But why can't they just simply take TS 6010 but instead they again try to be smarter than what is actually good for them.

1
0
0
paulmckrcu

Paul E. McKenney

Reply to @uecker@mastodon.social
@uecker Here is hoping that in the case of C, starting with TS 6010 will make things easier and more clearly defined.
1
0
0
uecker@mastodon.social

Martin Uecker

Reply to @paulmckrcu

@paulmckrcu I hope so, because I do not think TS 6010 being "overly charitable to the programmer" is a valid criticism. And while demons and angels playing game might be entertaining, I do not think it is good basis for any kind of useful technical specification - at least as long as the rules for the game remain ambiguous.

1
0
0
paulmckrcu

Paul E. McKenney

Reply to @uecker@mastodon.social
@uecker I apologize, but I really cannot resist suggesting that providing a way of cleanly supporting ABA-tolerant concurrent algorithms based on TS 6010 is a necessary (though probably not sufficient) step towards convincing WG21 of the value of this TS. ;-)
1
0
0
uecker@mastodon.social

Martin Uecker

Reply to @paulmckrcu

@paulmckrcu Why wouldn't the same solution work for TS 6010?

1
0
0
paulmckrcu

Paul E. McKenney

Reply to @uecker@mastodon.social
@uecker You lost me on this one. The same pointer-zap solution as adopted by WG21? If so, I am reviewing N3226 (a working draft that led to TS 6010) to attempt to produce an initial proposal. I could present that here, but I suspect that Mastodon isn't the best communications medium for this sort of thing. So left to myself, I would post to the SC22WG14. But is there a better venue?
0
0
0
About social.kernel.org

Terms of service

  1. Please do not use this service in violation of the Linux Kernel Code of Conduct. Doing so will result in your account suspension with the referral of the matter to the CoC committee.

  2. "Repeating"/"boosting" someone else's status on this platform will be treated as endorsement and will fall under rule #1.

  3. You are encouraged to use this platform to promote your work on the Linux Kernel, but there is no restriction on permitted topics (with the exception of anything covered by #1 above).

  4. There is no requirement to post in English, but it should be considered the primary language of communication on this platform.

Privacy notice

The admins of this service have access to all posted statuses. They aren't looking, but if it's something they shouldn't know about, then you should not post it on this platform.

Please see the Linux Foundation Privacy Policy, which applies to this platform as well.

Getting your own account

If you would like an account on this instance, please check that the following applies to you:

  • You are listed in MAINTAINERS or CREDITS

  • OR: You have a kernel.org account or email address

  • OR: You have a long and established history of involvement with the Linux Kernel

If the above is true and you agree with the Terms of Service and Privacy Notice listed above, please use these instructions to request an account: