Received-SPF: pass (google.com: domain of linux-kernel+bounces-59682-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.199.223 as permitted sender) client-ip=147.75.199.223;
From: Mark Brown <broonie@kernel.org>
Date: Fri, 09 Feb 2024 16:53:36 +0000
Subject: [PATCH v2] arm64/sve: Document that __SVE_VQ_MAX is much larger
 than needed
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: 7bit
Message-Id: <20240209-arm64-sve-vl-max-comment-v2-1-111b283469ee@kernel.org>
To: Catalin Marinas <catalin.marinas@arm.com>, 
 Will Deacon <will@kernel.org>
Cc: Dave Martin <Dave.Martin@arm.com>, linux-arm-kernel@lists.infradead.org, 
 linux-kernel@vger.kernel.org, Mark Brown <broonie@kernel.org>

__SVE_VQ_MAX is defined without comment as 512 but the actual
architectural maximum is 16, a substantial difference which might not
be obvious to readers especially given the several different units used
for specifying vector sizes in various contexts and the fact that it's
often used via macros.  In an effort to minimise surprises for users who
might assume the value is the architectural maximum and use it to do
things like size allocations add a comment noting the difference, and
add a note for SVE_VQ_MAX to aid discoverability.

Signed-off-by: Mark Brown <broonie@kernel.org>
---
Changes in v2:
- Use new wording mostly provided by Dave Martin.
- Reference __SVE_VQ_MAX comment near SVE_VQ_MAX define.
- Link to v1: https://lore.kernel.org/r/20240206-arm64-sve-vl-max-comment-v1-1-dddf16414412@kernel.org
---
 arch/arm64/include/uapi/asm/sve_context.h | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/arch/arm64/include/uapi/asm/sve_context.h b/arch/arm64/include/uapi/asm/sve_context.h
index 754ab751b523..72aefc081061 100644
--- a/arch/arm64/include/uapi/asm/sve_context.h
+++ b/arch/arm64/include/uapi/asm/sve_context.h
@@ -13,6 +13,17 @@
 
 #define __SVE_VQ_BYTES		16	/* number of bytes per quadword */
 
+/*
+ * Yes, __SVE_VQ_MAX is 512 QUADWORDS.
+ *
+ * To help ensure forward portability, this is much larger than the
+ * current maximum value defined by the SVE architecture.  While arrays
+ * or static allocations can be sized based on this value, watch out!
+ * It will waste a surprisingly large amount of memory.
+ *
+ * Dynamic sizing based on the actual runtime vector length is likely to
+ * be preferable for most purposes.
+ */
 #define __SVE_VQ_MIN		1
 #define __SVE_VQ_MAX		512
 

---
base-commit: 54be6c6c5ae8e0d93a6c4641cb7528eb0b6ba478
change-id: 20240206-arm64-sve-vl-max-comment-64efa3f03625

Best regards,
-- 
Mark Brown <broonie@kernel.org>