* [FFmpeg-devel] [PATCH] riscv/cpu: add V subset feature detection (PR #20873)
@ 2025-11-09 13:06 Rémi Denis-Courmont via ffmpeg-devel
0 siblings, 0 replies; only message in thread
From: Rémi Denis-Courmont via ffmpeg-devel @ 2025-11-09 13:06 UTC (permalink / raw)
To: ffmpeg-devel; +Cc: Rémi Denis-Courmont
PR #20873 opened by Rémi Denis-Courmont (Courmisch)
URL: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/20873
Patch URL: https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/20873.patch
This adds support for detecting Zve32x, Zve32f, Zve64x and Zve64d on
a system that does not support (or declare) the full V feature set.
Because the code assumes that vector detected at runtime are at least
128-bit long, we have to manually check that too (in theory, vectors
can be 32-bit or 64-bit, though that makes little sense in 2025).
From 3ea7c5449d2e35c01debdac8fb981675eb861644 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?R=C3=A9mi=20Denis-Courmont?= <remi@remlab.net>
Date: Sun, 9 Nov 2025 15:01:03 +0200
Subject: [PATCH] riscv/cpu: add V subset feature detection
This adds support for detecting Zve32x, Zve32f, Zve64x and Zve64d on
a system that does not support (or declare) the full V feature set.
Because the code assumes that vector detected at runtime are at least
128-bit long, we have to manually check that too (in theory, vectors
can be 32-bit or 64-bit, though that makes little sense in 2025).
---
libavutil/riscv/cpu.c | 16 ++++++++++++++++
1 file changed, 16 insertions(+)
diff --git a/libavutil/riscv/cpu.c b/libavutil/riscv/cpu.c
index 379703651a..8cc91625c7 100644
--- a/libavutil/riscv/cpu.c
+++ b/libavutil/riscv/cpu.c
@@ -22,6 +22,7 @@
# define _GNU_SOURCE
#endif
#include "libavutil/cpu.h"
+#include "libavutil/riscv/cpu.h"
#include "libavutil/cpu_internal.h"
#include "libavutil/macros.h"
#include "libavutil/log.h"
@@ -64,6 +65,21 @@ int ff_get_cpu_flags_riscv(void)
if (pairs[1].value & RISCV_HWPROBE_IMA_V)
ret |= AV_CPU_FLAG_RVV_I32 | AV_CPU_FLAG_RVV_I64
| AV_CPU_FLAG_RVV_F32 | AV_CPU_FLAG_RVV_F64;
+#ifdef RISCV_HWPROBE_EXT_ZVE32X
+ else if ((pairs[1].value & RISCV_HWPROBE_EXT_ZVE32X) &&
+ ff_get_rv_vlenb() >= 16) { // runtime detect assumes 128+ bits
+ ret |= AV_CPU_FLAG_RVV_I32;
+
+ if (pairs[1].value & RISCV_HWPROBE_EXT_ZVE32F)
+ ret |= AV_CPU_FLAG_RVV_F32;
+ if (pairs[1].value & RISCV_HWPROBE_EXT_ZVE64X) {
+ ret |= AV_CPU_FLAG_RVV_I64;
+
+ if (pairs[1].value & RISCV_HWPROBE_EXT_ZVE64D)
+ ret |= AV_CPU_FLAG_RVV_F64;
+ }
+ }
+#endif
#endif
#ifdef RISCV_HWPROBE_EXT_ZBB
if (pairs[1].value & RISCV_HWPROBE_EXT_ZBB)
--
2.49.1
_______________________________________________
ffmpeg-devel mailing list -- ffmpeg-devel@ffmpeg.org
To unsubscribe send an email to ffmpeg-devel-leave@ffmpeg.org
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2025-11-09 13:07 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2025-11-09 13:06 [FFmpeg-devel] [PATCH] riscv/cpu: add V subset feature detection (PR #20873) Rémi Denis-Courmont via ffmpeg-devel
Git Inbox Mirror of the ffmpeg-devel mailing list - see https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
This inbox may be cloned and mirrored by anyone:
git clone --mirror https://master.gitmailbox.com/ffmpegdev/0 ffmpegdev/git/0.git
# If you have public-inbox 1.1+ installed, you may
# initialize and index your mirror using the following commands:
public-inbox-init -V2 ffmpegdev ffmpegdev/ https://master.gitmailbox.com/ffmpegdev \
ffmpegdev@gitmailbox.com
public-inbox-index ffmpegdev
Example config snippet for mirrors.
AGPL code for this site: git clone https://public-inbox.org/public-inbox.git