Michael Smith has posted comments on this change. ( http://gerrit.cloudera.org:8080/24940 )
Change subject: IMPALA-11917: Upgrade to GCC 15 and LLVM 22 ...................................................................... Patch Set 18: (1 comment) http://gerrit.cloudera.org:8080/#/c/24940/15/be/src/codegen/llvm-codegen.h File be/src/codegen/llvm-codegen.h: http://gerrit.cloudera.org:8080/#/c/24940/15/be/src/codegen/llvm-codegen.h@102 PS15, Line 102: pe == TYPE_DECIMAL && col_type.GetByte > I'll look into this more; presumably align(8) and align(1) result in produc Some related discussion: https://discourse.llvm.org/t/packed-structs-and-vectors-miscompilation-or-ub/63738. I could dig more into how this changes the overall generated code for some set of queries if that would help. There's no significant difference between align(8) and align(1) for our code generation on DecimalVal specifically, and I haven't run into problems with the default alignment otherwise. Unspecified alignment on a load/store uses architecture defaults. We could test out setting align(1) for everything that accesses slot memory; I'd want to do that in a follow-up commit since this seems to pass our tests. -- To view, visit http://gerrit.cloudera.org:8080/24940 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: Impala-ASF Gerrit-Branch: master Gerrit-MessageType: comment Gerrit-Change-Id: I7dda730fa98ebe3825969265627a560b0c3095f9 Gerrit-Change-Number: 24940 Gerrit-PatchSet: 18 Gerrit-Owner: Michael Smith <[email protected]> Gerrit-Reviewer: Balazs Hevele <[email protected]> Gerrit-Reviewer: Csaba Ringhofer <[email protected]> Gerrit-Reviewer: Impala Public Jenkins <[email protected]> Gerrit-Reviewer: Joe McDonnell <[email protected]> Gerrit-Reviewer: Laszlo Gaal <[email protected]> Gerrit-Reviewer: Michael Smith <[email protected]> Gerrit-Comment-Date: Mon, 05 Oct 2026 15:56:07 +0000 Gerrit-HasComments: Yes
