Skip to content

[AArch64] Improve host feature detection. #160410

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
Open

[AArch64] Improve host feature detection. #160410

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
fc5d883
[AArch64] Improve host feature detection.
efriedma-quic Sep 23, 2025
2dc6d3e
Fix jscvt detection
efriedma-quic Sep 24, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
64 changes: 62 additions & 2 deletions llvm/lib/TargetParser/Host.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2260,20 +2260,80 @@ StringMap<bool> sys::getHostCPUFeatures() {
uint32_t Sha2 = CAP_SHA1 | CAP_SHA2;
Features["aes"] = (crypto & Aes) == Aes;
Features["sha2"] = (crypto & Sha2) == Sha2;

// SVE support is disabled in for cores which are identified as supporting
Copy link
Collaborator

@DavidSpickett DavidSpickett Oct 23, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This comment wording still needs fixing. "in for" does not sound right.

Copy link
Collaborator

@DavidSpickett DavidSpickett Oct 23, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Maybe you were thinking "disabled for cores that have sve disabled in them" and combined the two thoughts here.

// SVE; disable SVE if we don't detect support at runtime.
if (!Features.contains("sve"))
Features["sve"] = false;
#endif

return Features;
}
#elif defined(_WIN32) && (defined(__aarch64__) || defined(_M_ARM64) || \
defined(__arm64ec__) || defined(_M_ARM64EC))
#ifndef PF_ARM_V82_DP_INSTRUCTIONS_AVAILABLE
#define PF_ARM_V82_DP_INSTRUCTIONS_AVAILABLE 43
#endif
#ifndef PF_ARM_V83_JSCVT_INSTRUCTIONS_AVAILABLE
#define PF_ARM_V83_JSCVT_INSTRUCTIONS_AVAILABLE 44
#endif
#ifndef PF_ARM_V83_LRCPC_INSTRUCTIONS_AVAILABLE
#define PF_ARM_V83_LRCPC_INSTRUCTIONS_AVAILABLE 45
#endif
#ifndef PF_ARM_SVE_INSTRUCTIONS_AVAILABLE
#define PF_ARM_SVE_INSTRUCTIONS_AVAILABLE 46
#endif
#ifndef PF_ARM_SVE2_INSTRUCTIONS_AVAILABLE
#define PF_ARM_SVE2_INSTRUCTIONS_AVAILABLE 47
#endif
#ifndef PF_ARM_SVE_PMULL128_INSTRUCTIONS_AVAILABLE
#define PF_ARM_SVE_PMULL128_INSTRUCTIONS_AVAILABLE 50
#endif
#ifndef PF_ARM_SVE_SHA3_INSTRUCTIONS_AVAILABLE
#define PF_ARM_SVE_SHA3_INSTRUCTIONS_AVAILABLE 55
#endif
#ifndef PF_ARM_SVE_SM4_INSTRUCTIONS_AVAILABLE
#define PF_ARM_SVE_SM4_INSTRUCTIONS_AVAILABLE 56
#endif
#ifndef PF_ARM_SVE_I8MM_INSTRUCTIONS_AVAILABLE
#define PF_ARM_SVE_I8MM_INSTRUCTIONS_AVAILABLE 57
#endif
#ifndef PF_ARM_SVE_F32MM_INSTRUCTIONS_AVAILABLE
#define PF_ARM_SVE_F32MM_INSTRUCTIONS_AVAILABLE 58
#endif
#ifndef PF_ARM_SVE_F64MM_INSTRUCTIONS_AVAILABLE
#define PF_ARM_SVE_F64MM_INSTRUCTIONS_AVAILABLE 59
#endif
StringMap<bool> sys::getHostCPUFeatures() {
StringMap<bool> Features;

// If we're asking the OS at runtime, believe what the OS says
Features["neon"] =
IsProcessorFeaturePresent(PF_ARM_NEON_INSTRUCTIONS_AVAILABLE);
Features["crc"] =
IsProcessorFeaturePresent(PF_ARM_V8_CRC32_INSTRUCTIONS_AVAILABLE);
Features["lse"] =
IsProcessorFeaturePresent(PF_ARM_V81_ATOMIC_INSTRUCTIONS_AVAILABLE);
Features["dotprod"] =
IsProcessorFeaturePresent(PF_ARM_V82_DP_INSTRUCTIONS_AVAILABLE);
Features["jsconv"] =
IsProcessorFeaturePresent(PF_ARM_V83_JSCVT_INSTRUCTIONS_AVAILABLE);
Features["rcpc"] =
IsProcessorFeaturePresent(PF_ARM_V83_LRCPC_INSTRUCTIONS_AVAILABLE);
Features["sve"] =
IsProcessorFeaturePresent(PF_ARM_SVE_INSTRUCTIONS_AVAILABLE);
Features["sve2"] =
IsProcessorFeaturePresent(PF_ARM_SVE2_INSTRUCTIONS_AVAILABLE);
Features["sve-aes"] =
IsProcessorFeaturePresent(PF_ARM_SVE_PMULL128_INSTRUCTIONS_AVAILABLE);
Features["sve-sha3"] =
IsProcessorFeaturePresent(PF_ARM_SVE_SHA3_INSTRUCTIONS_AVAILABLE);
Features["sve-sm4"] =
IsProcessorFeaturePresent(PF_ARM_SVE_SM4_INSTRUCTIONS_AVAILABLE);
Features["f32mm"] =
IsProcessorFeaturePresent(PF_ARM_SVE_F32MM_INSTRUCTIONS_AVAILABLE);
Features["f64mm"] =
IsProcessorFeaturePresent(PF_ARM_SVE_F64MM_INSTRUCTIONS_AVAILABLE);
Features["i8mm"] =
IsProcessorFeaturePresent(PF_ARM_SVE_I8MM_INSTRUCTIONS_AVAILABLE);

// Avoid inferring "crypto" means more than the traditional AES + SHA2
bool TradCrypto =
Expand Down