Skip to content

Conversation

@Alexandr-Konovalov
Copy link
Contributor

@Alexandr-Konovalov Alexandr-Konovalov commented May 5, 2025

Move SubmissionInfo to versioned public header and allocate it completely on stack.

Move SubmissionInfo to versioned public header and allocate it
complitely on stack.

Signed-off-by: Alexandr Konovalov <[email protected]>
Signed-off-by: Alexandr Konovalov <[email protected]>
Signed-off-by: Alexandr Konovalov <[email protected]>
Signed-off-by: Alexandr Konovalov <[email protected]>
@Alexandr-Konovalov Alexandr-Konovalov marked this pull request as ready for review May 7, 2025 08:45
@Alexandr-Konovalov Alexandr-Konovalov requested a review from a team as a code owner May 7, 2025 08:45
@steffenlarsen steffenlarsen changed the title [SYCL] Allocate SubmissionInfo complitely on stack [SYCL] Allocate SubmissionInfo completely on stack May 7, 2025
Signed-off-by: Alexandr Konovalov <[email protected]>
Signed-off-by: Alexandr Konovalov <[email protected]>
Copy link
Contributor

@steffenlarsen steffenlarsen left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM!

Signed-off-by: Alexandr Konovalov <[email protected]>
@vinser52 vinser52 requested a review from sergey-semenov May 8, 2025 15:22
@Alexandr-Konovalov
Copy link
Contributor Author

ThreadSanitizer/check_device_global.cpp failure seems unrelated to the PR, see #18349 .

@vinser52
Copy link
Contributor

vinser52 commented May 8, 2025

@intel/llvm-gatekeepers the PR is ready to merge

@sarnex sarnex merged commit 75aa60c into intel:sycl May 8, 2025
18 of 19 checks passed
Copy link
Contributor

@aelovikov-intel aelovikov-intel left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I also doubt that this approach would work well for non-NFC ABI-breaking changes (e.g. adding new fields or removing old ones). Providing old versions without those new fields could be troublesome, unless all of them are actually optional.

};
#endif

namespace v1 {
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This way one particular data structure claims entire sycl::detail::vN namespace, so that we can't use the same approach for other APIs (that's the same reason why I think our sycl::_v1 turned out to be rather useless). IMO, it would be better to have detail::submission_info::_v1::SubmissionInfo or something like that. Maybe make submission_info inline itself.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

6 participants