Re: Add missing JIT inline pass for llvm>=17

Hi,

Good catch, thanks for the patch.

On 16/01/26 06:10, Anthonin Bonnefoy wrote:
> On Thu, Jan 15, 2026 at 2:51 PM Andres Freund <andres@anarazel.de> wrote:
>> I'm strongly against removing the always inline pass, I see absolutely no
>> reason for doing that. The whole point of always inline is that it happens
>> unconditionally. It's not an expensive pass either.
> 
> I've looked into more details on what was provided by 'default<O0>',
> and it turns out it includes an always-inline pass[0]. This is also
> visible when using debug-pass-manager:
> 
> llvm-as < /dev/null | opt -disable-output --passes='default<O0>'
> -debug-pass-manager
> Running analysis: InnerAnalysisManagerProxy<FunctionAnalysisManager,
> Module> on [module]
> Running pass: AlwaysInlinerPass on [module]
> Running analysis: ProfileSummaryAnalysis on [module]
> Running pass: CoroConditionalWrapper on [module]
> Running pass: VerifierPass on [module]
> Running analysis: VerifierAnalysis on [module]
> 
> With the pre-LLVM17 legacy pass manager, that doesn't seem to be the
> case[1] (despite the confusing comment? Inliner is only set by
> LLVMPassManagerBuilderUseInlinerWithThreshold).
> 
> So, with 'default<O0>,mem2reg', we replicate the same behaviour as
> pre-LLVM17 as it includes the always-inline pass.
> 
> I've updated the patch to only add the inline pass when PGJIT_INLINE
> is on. I've also added a comment to mention that always-inline is
> included in O0.
> 
> [0]: https://github.com/llvm/llvm-project/blob/701040d48f759369dce755f185a21aa6b92ba3ae/llvm/lib/Passes/PassBuilderPipelines.cpp#L2360-L2365
> [1]: https://github.com/llvm/llvm-project/blob/release/16.x/llvm/lib/Transforms/IPO/PassManagerBuilder.cpp#L290-L313

I've tested this patch with LLVM 22 using the patches from [1] and it
seems to me that it's working as expected.

On master with jit_inline_above_cost=0 and jit_inline_above_cost=10000 I
see the same bc:

    op.0.fetch:                                       ; preds = %b.op.0.start
      call void @deform_0_1(ptr %v_scanslot)
      br label %b.op.1.start

With v3 patch and with jit_inline_above_cost=0 it inline the deform
function as expected:

    op.0.fetch:                                       ; preds = %b.op.0.start
      %29 = getelementptr inbounds nuw %struct.TupleTableSlot, ptr %v_scanslot, i32 0, i32 5
      %tts_values.i = load ptr, ptr %29, align 8
      %30 = getelementptr inbounds nuw %struct.TupleTableSlot, ptr %v_scanslot, i32 0, i32 6
      %tts_ISNULL.i = load ptr, ptr %30, align 8
      %31 = getelementptr inbounds nuw %struct.TupleTableSlot, ptr %v_scanslot, i32 0, i32 1
      ...


[1] https://www.postgresql.org/message-id/flat/CA%2BhUKGJTumad75o8Zao-LFseEbt%3DenbUFCM7LZVV%3Dc8yg2i7dg%40mail.gmail.com

--
Matheus Alcantara
EDB: https://www.enterprisedb.com