RFR: 8338197: [ubsan] ad_x86.hpp:6417:11: runtime error: shift exponent 100 is too large for 32-bit type 'unsigned int' [v3]

[Boris Ulasevich]

On Tue, 26 Aug 2025 10:51:43 GMT, Andrew Dinn <adinn at openjdk.org> wrote:

> Yes, I think we should fix both, output_h.cpp and fixed_latency(100) on all platforms, then we can get rid of the workarounds and arm32-specific logic.

Thanks, @dean-long

Along with capping the possible max shift in the generated ad files, I also changed the hard-coded artificially large latency value from 100 to 30 — it’s still sufficiently large but stays within the bit-width range. Do you think this form is OK for integration, or should we refine it further?

-------------

PR Comment: https://git.openjdk.org/jdk/pull/26890#issuecomment-3334057450