[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20241017-jesse_unaligned_vector-v10-6-5b33500160f8@rivosinc.com>
Date: Thu, 17 Oct 2024 12:00:23 -0700
From: Charlie Jenkins <charlie@...osinc.com>
To: Paul Walmsley <paul.walmsley@...ive.com>,
Palmer Dabbelt <palmer@...belt.com>, Albert Ou <aou@...s.berkeley.edu>,
Clément Léger <cleger@...osinc.com>,
Evan Green <evan@...osinc.com>, Jonathan Corbet <corbet@....net>
Cc: Palmer Dabbelt <palmer@...osinc.com>, linux-riscv@...ts.infradead.org,
linux-kernel@...r.kernel.org, linux-doc@...r.kernel.org,
Charlie Jenkins <charlie@...osinc.com>, Jesse Taube <jesse@...osinc.com>
Subject: [PATCH v10 6/6] RISC-V: hwprobe: Document unaligned vector perf
key
From: Jesse Taube <jesse@...osinc.com>
Document key for reporting the speed of unaligned vector accesses.
The descriptions are the same as the scalar equivalent values.
Signed-off-by: Jesse Taube <jesse@...osinc.com>
Reviewed-by: Charlie Jenkins <charlie@...osinc.com>
---
Documentation/arch/riscv/hwprobe.rst | 16 ++++++++++++++++
1 file changed, 16 insertions(+)
diff --git a/Documentation/arch/riscv/hwprobe.rst b/Documentation/arch/riscv/hwprobe.rst
index 85b709257918..ea4e0b9c73e7 100644
--- a/Documentation/arch/riscv/hwprobe.rst
+++ b/Documentation/arch/riscv/hwprobe.rst
@@ -274,3 +274,19 @@ The following keys are defined:
represent the highest userspace virtual address usable.
* :c:macro:`RISCV_HWPROBE_KEY_TIME_CSR_FREQ`: Frequency (in Hz) of `time CSR`.
+
+* :c:macro:`RISCV_HWPROBE_KEY_MISALIGNED_VECTOR_PERF`: An enum value describing the
+ performance of misaligned vector accesses on the selected set of processors.
+
+ * :c:macro:`RISCV_HWPROBE_MISALIGNED_VECTOR_UNKNOWN`: The performance of misaligned
+ vector accesses is unknown.
+
+ * :c:macro:`RISCV_HWPROBE_MISALIGNED_VECTOR_SLOW`: 32-bit misaligned accesses using vector
+ registers are slower than the equivalent quantity of byte accesses via vector registers.
+ Misaligned accesses may be supported directly in hardware, or trapped and emulated by software.
+
+ * :c:macro:`RISCV_HWPROBE_MISALIGNED_VECTOR_FAST`: 32-bit misaligned accesses using vector
+ registers are faster than the equivalent quantity of byte accesses via vector registers.
+
+ * :c:macro:`RISCV_HWPROBE_MISALIGNED_VECTOR_UNSUPPORTED`: Misaligned vector accesses are
+ not supported at all and will generate a misaligned address fault.
--
2.45.0
Powered by blists - more mailing lists