[llvm] AMDGPU: Fix using unaligned vgprs in mimg error test (PR #157001)
Matt Arsenault via llvm-commits
llvm-commits at lists.llvm.org
Thu Sep 4 19:27:55 PDT 2025
https://github.com/arsenm created https://github.com/llvm/llvm-project/pull/157001
These instructions really have 2 errors, from the unsupported
image base instruction plus the unaligned vgpr usage. This
test intends to test the base instruction error, so fix the
secondary error to avoid changing the diagnostic in a future
patch.
>From 04a71d441b65da6b764a8e89df9af2f176fab952 Mon Sep 17 00:00:00 2001
From: Matt Arsenault <Matthew.Arsenault at amd.com>
Date: Fri, 5 Sep 2025 11:21:52 +0900
Subject: [PATCH] AMDGPU: Fix using unaligned vgprs in mimg error test
These instructions really have 2 errors, from the unsupported
image base instruction plus the unaligned vgpr usage. This
test intends to test the base instruction error, so fix the
secondary error to avoid changing the diagnostic in a future
patch.
---
llvm/test/MC/AMDGPU/mimg-err.s | 14 +++++++-------
1 file changed, 7 insertions(+), 7 deletions(-)
diff --git a/llvm/test/MC/AMDGPU/mimg-err.s b/llvm/test/MC/AMDGPU/mimg-err.s
index bec33bab984ab..63a5757f2a1b5 100644
--- a/llvm/test/MC/AMDGPU/mimg-err.s
+++ b/llvm/test/MC/AMDGPU/mimg-err.s
@@ -12,22 +12,22 @@ image_load v[4:6], v[237:240], s[28:35] dmask:0x7 tfe
// NOGFX9: error: image data size does not match dmask, d16 and tfe
// NOGFX90A: :[[@LINE-3]]:{{[0-9]+}}: error: invalid operand for instruction
-image_load v[4:5], v[237:240], s[28:35] dmask:0x7
+image_load v[4:5], v[236:239], s[28:35] dmask:0x7
// NOGCN: error: image data size does not match dmask and tfe
// NOGFX9: error: image data size does not match dmask, d16 and tfe
// NOGFX90A: :[[@LINE-3]]:{{[0-9]+}}: error: image data size does not match dmask and d16
-image_store v[4:7], v[237:240], s[28:35] dmask:0x7
+image_store v[4:7], v[236:239], s[28:35] dmask:0x7
// NOGCN: error: image data size does not match dmask and tfe
// NOGFX9: error: image data size does not match dmask, d16 and tfe
// NOGFX90A: :[[@LINE-3]]:{{[0-9]+}}: error: image data size does not match dmask and d16
-image_store v[4:7], v[237:240], s[28:35] dmask:0xe
+image_store v[4:7], v[236:239], s[28:35] dmask:0xe
// NOGCN: error: image data size does not match dmask and tfe
// NOGFX9: error: image data size does not match dmask, d16 and tfe
// NOGFX90A: :[[@LINE-3]]:{{[0-9]+}}: error: image data size does not match dmask and d16
-image_load v4, v[237:240], s[28:35] tfe
+image_load v4, v[236:239], s[28:35] tfe
// NOGCN: error: image data size does not match dmask and tfe
// NOGFX9: error: image data size does not match dmask, d16 and tfe
// NOGFX90A: :[[@LINE-3]]:{{[0-9]+}}: error: invalid operand for instruction
@@ -36,17 +36,17 @@ image_load v4, v[237:240], s[28:35] tfe
// Image Sample
//===----------------------------------------------------------------------===//
-image_sample v[193:195], v[237:240], s[28:35], s[4:7] dmask:0x7 tfe
+image_sample v[192:194], v[236:239], s[28:35], s[4:7] dmask:0x7 tfe
// NOGCN: error: image data size does not match dmask and tfe
// NOGFX9: error: image data size does not match dmask, d16 and tfe
// NOGFX90A: :[[@LINE-3]]:{{[0-9]+}}: error: invalid operand for instruction
-image_sample v[193:195], v[237:240], s[28:35], s[4:7] dmask:0x3
+image_sample v[192:194], v[236:239], s[28:35], s[4:7] dmask:0x3
// NOGCN: error: image data size does not match dmask and tfe
// NOGFX9: error: image data size does not match dmask, d16 and tfe
// NOGFX90A: :[[@LINE-3]]:{{[0-9]+}}: error: image data size does not match dmask and d16
-image_sample v[193:195], v[237:240], s[28:35], s[4:7] dmask:0xf
+image_sample v[192:194], v[236:239], s[28:35], s[4:7] dmask:0xf
// NOGCN: error: image data size does not match dmask and tfe
// NOGFX9: error: image data size does not match dmask, d16 and tfe
// NOGFX90A: :[[@LINE-3]]:{{[0-9]+}}: error: image data size does not match dmask and d16
More information about the llvm-commits
mailing list