Add 'Weight' for default parameter consistency.

added tests and did a small refactor.

In D147297#4235771, @wlei wrote:

Good catch, thank you! For the test case, perhaps we can reuse the test https://github.com/llvm/llvm-project/blob/main/llvm/test/tools/llvm-profdata/Inputs/sample-flatten-profile.proftext#L6 and add some dummy call target to the inlinees. like

   3: bar:10
    1: 10
+   2: 10 baz:10

add test case (and minor change to use 'try_emplace' that does nothing if key presents)

In D146754#4224991, @dmgreen wrote:

LGTM. You can land tests like this without review.

arc land <feature-branch> detects this pre-commit test case is not reviewed so decided to send this out as well. Thanks!

thanks for reviews! Going to submit this patch and the pre-commit test (D146754) together.

add RUN line for new test

remove empty lines

Run auto-updater on one of existing tests

The change to handle function alias for single-impl-devirt LGTM. The current way of bailing-out-upon-function-alias where function is assumed (e.g., virtual const propagation) sounds reasonable (since function-aliases are not handled correctly).

In rGc847e22db33e4625b1ad58b631f57d7d1e1944ca#1161057, @MsThang3677 wrote:

In rGc847e22db33e4625b1ad58b631f57d7d1e1944ca#1161036, @mingmingl wrote:

In rGc847e22db33e4625b1ad58b631f57d7d1e1944ca#1161033, @MsThang3677 wrote:

Hello, In case anyone is unaware "lol I'm the most unaware" I do not understand how this works. I know more than I started with.
What anyone reading this may not know is that I lose user accounts and invaluable memories everytime the titles and information of the materials taken from my devices is modified. I would appreciate it if someone could explain how this workers and how installing custom builds without my knowledge is even legal and would kindly ask you to stay out. If I have something of yours, please communicate that to me. I have copied pasted and stored data only for the sake of evidence. I have no intention of using anyone else's work. It's been close to two years of every device. User account and pretty much every second of my life being hacked. LLVM please stay off my devices. If there are any ideas on how I can restore my user accounts, particularly allys.fambam which held videos of my granddaughters birth and countless pictures if my kids over the years.

Hello there, sorry to hear that your account is hacked. I'm not sure if I follow; is the concern about this patch (that guards an intrinsic) causing a regression or build breakage, or is it a question about how this patch works? Or other concerns with this commit?

I suppose this iis a two part question. This began for me in 2021 July. I knew nothing about programming. I was hacked and this has been my story every since. So yes, I need to understand how this works over all. Why does every thing I find on my devices lead to these few projects and vise versa?
Everything about me in the metadata is erased and modified and so Any attempt to learn is blocked. Why am I being erased? Most importantly, if the hacking was not a result of any analytic data, is it possible that my modified metadata is your missing piece? .I have seen in the manifest files of multiple apps that my metadata was paused and the metadata of my "remote controller" is being swap in to replace mine. If that can be sorted out, I believe your build could be fixed

In rGc847e22db33e4625b1ad58b631f57d7d1e1944ca#1161033, @MsThang3677 wrote:

Hello, In case anyone is unaware "lol I'm the most unaware" I do not understand how this works. I know more than I started with.
What anyone reading this may not know is that I lose user accounts and invaluable memories everytime the titles and information of the materials taken from my devices is modified. I would appreciate it if someone could explain how this workers and how installing custom builds without my knowledge is even legal and would kindly ask you to stay out. If I have something of yours, please communicate that to me. I have copied pasted and stored data only for the sake of evidence. I have no intention of using anyone else's work. It's been close to two years of every device. User account and pretty much every second of my life being hacked. LLVM please stay off my devices. If there are any ideas on how I can restore my user accounts, particularly allys.fambam which held videos of my granddaughters birth and countless pictures if my kids over the years.

In D140432#4010143, @dmgreen wrote:

Thanks for putting together the patch. It may be worth mentioning in the commit message that aes currently includes both FEAT_AES and FEAT_PMULL.
Otherwise LGTM.

update commit message to mention target guard 'aes' includes both FEAT_AES and FEAT_PMULL currently.

indent intrinsic definition in arm_neon.td

In D138904#3970087, @bcl5980 wrote:

So how about this patch？

Submitted (as a part of D137689) in this commit

thanks for reviews! Diff3 and diff4 (in "History" tab of "Revision Contents") have no diff (meaning rebase doesn't introduce unintended diff), going to submit.

rebase after D137797

thanks! Going to submit (the x86 failure window looks irrelevant from error message, and ninja check-llvm passed).

resolve comments.

For the affected test case test_nonuseful_bits, one BFM and one ORR is generated now (https://godbolt.org/z/a3c68f7dE) with this commit. Going to abandon this patch (rather than rebase it for other BFI improvements). Thanks for the discussions around BFM/ORR.

run git clang-format HEAD~1

fix comments for the 80 column limit.

mostly rebase

run a post-approval 'clang-format'

thanks! run a post-approval clang-format and going to submit this.

In D137467#3911238, @davidxl wrote:

I see. The DOT CFG graph dump uses the layout order instead of the MBB number. Another way to fix this is to use the following in the node label: BBName[MBBNum][layout-order], but making the MBB num and layout number consistent under an option is fine to me.

In D135843#3862662, @qiongsiwu1 wrote:

Ah I just saw 5d3139aef113881fff9430c96069faf96630a5d8. Thanks @MaskRay !

In D137467#3909358, @davidxl wrote:

Does MIR text output show MBB number in the dump?

thanks for reviews! Going to submit and implement the FIXME (for SRL) in follow-up patches.

rebase after D137296

Thanks!

Resolve comments.

This update runs git clang-format HEAD~1 only, no functional change --> without this, pre-merge checks fails due to ERROR git-clang-format returned an non-zero exit code 1

Thanks for reviews! PTAL.

resolve comments.

In D135102#3899360, @dmgreen wrote:

Hmm. I had not considered am ORR with shift to be cheaper than a BFM before. From what I can tell it doesn't seem to be universal across all cpus, but does look like it will be faster or equal.

fix a bug around width computation (updated commit message and patch summary as well), and adjust whitespace in a few test cases (mainly those not updated by utils/update_llc_test_checks.py) to remove irrelevant diff.

In order to generate orr with shifted operand (not bfi) when orr is better, this patch does a comparison inside tryBitfieldInsertOpFromOr (i.e., after dag-combiner, alongside cpp-based instruction selection).

added tests for {i32,i64}x{bfi,bfxil}