[PowerPC] Mask constant operands in ValueBit tracking #67653
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
ecnelises
requested review from
bzEq,
chenzheng1030,
orcguru,
nemanjai,
amy-kwan,
EsmeYi,
lei137 and
stefanp-ibm
ecnelises
commented
Sep 28, 2023
| ; RUN: llc -mtriple=powerpc64le-unknown-linux-gnu -mcpu=pwr7 < %s | FileCheck %s --check-prefix=LE64 | ||
| ; RUN: llc -mtriple=powerpcle-unknown-linux-gnu -mcpu=pwr7 < %s | FileCheck %s --check-prefix=LE32 | ||
| ; RUN: llc -mtriple=powerpc64-ibm-aix -mcpu=pwr7 < %s | FileCheck %s --check-prefix=BE64 | ||
| ; RUN: llc -mtriple=powerpc-ibm-aix -mcpu=pwr7 < %s | FileCheck %s --check-prefix=BE32 |
There was a problem hiding this comment.
-mcpu=pwr8 touches another codepath (such instructions are not generated, so only crash on -mcpu=pwr7).
The crash comes from constant shift amounts of srl_parts, which only presents after splitting v2i128 into v1i128 (from the subtraction parts).
|
I was going to report the following bug, but since it is already fixed by this change, maybe this even simpler repro would be useful to add as a test since it doesn't involve any vectors: target triple = "powerpc"
define i64 @entry(ptr) {
store i6 0, ptr %0
%2 = load i8, ptr %0
%3 = and i8 %2, 63
%4 = zext i8 %3 to i64
%5 = shl i64 1, %4
ret i64 %5
} |
|
Thank you @bzEq for the quick response. Would it be possible to backport this into 18? |
In IR or C code, left/right shift amount larger than value size is undefined behavior. But in practise, backend lowering for srl_parts/sra_parts/shl_parts produces add/sub of shift amounts, thus constant shift amounts might be negative or larger than value size. And the lowering depends on behavior in ISA. PowerPC ISA says, the lowest 7 bits (6 bits if in 32-bit instruction) will be taken, and if the highest among them is 1, result will be zero, otherwise the low 6 bits (or 5 on 32-bit) are used as shift amount. This patch emulates the behavior and avoids array overflow in bit permutation's value bits calculator.
ecnelises
force-pushed
the
shiftamt_mask
branch
from
13aeaba
to
6759869
Compare
|
/cherry-pick 292d9e8 |
|
Failed to create pull request for issue67653 https://github.com/llvm/llvm-project/actions/runs/7967856697 |
|
/cherry-pick 292d9e8 |
llvmbot
pushed a commit
to llvmbot/llvm-project
that referenced
this pull request
Feb 20, 2024
In IR or C code, shift amount larger than value size is undefined behavior. But in practice, backend lowering for shift_parts produces add/sub of shift amounts, thus constant shift amounts might be negative or larger than value size, which depends on ISA definition. PowerPC ISA says, the lowest 7 bits (6 bits for 32-bit instruction) will be taken, and if the highest among them is 1, result will be zero, otherwise the low 6 bits (or 5 on 32-bit) are used as shift amount. This commit emulates the behavior and avoids array overflow in bit permutation's value bits calculator. (cherry picked from commit 292d9e8)
|
/pull-request #82301 |
llvmbot
pushed a commit
to llvmbot/llvm-project
that referenced
this pull request
Feb 20, 2024
In IR or C code, shift amount larger than value size is undefined behavior. But in practice, backend lowering for shift_parts produces add/sub of shift amounts, thus constant shift amounts might be negative or larger than value size, which depends on ISA definition. PowerPC ISA says, the lowest 7 bits (6 bits for 32-bit instruction) will be taken, and if the highest among them is 1, result will be zero, otherwise the low 6 bits (or 5 on 32-bit) are used as shift amount. This commit emulates the behavior and avoids array overflow in bit permutation's value bits calculator. (cherry picked from commit 292d9e8)
Closed
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
In IR or C code, shift amount larger than value size is undefined
behavior. But in practice, backend lowering for shift_parts produces
add/sub of shift amounts, thus constant shift amounts might be
negative or larger than value size, which depends on ISA definition.
PowerPC ISA says, the lowest 7 bits (6 bits for 32-bit instruction)
will be taken, and if the highest among them is 1, result will be
zero, otherwise the low 6 bits (or 5 on 32-bit) are used as shift
amount.
This commit emulates the behavior and avoids array overflow in bit
permutation's value bits calculator.
Fixes #59074