Re: [PATCH v2 01/15] RISC-V: Build Infastructure

[Palmer Dabbelt <palmer at dabbelt dot com>]

On Wed, 20 Dec 2017 08:43:18 PST (-0800), joseph@codesourcery.com wrote:
> Another observation:
>
> Since you're using ldbl-128, that presumably means _Float128 and _Float64x
> type names are supported with the same format as long double.  So you need
> to add RISC-V to the list in manual/math.texi of ports for which those
> type names and associated interfaces are supported.

That's our intent.  How does this look?

diff --git a/manual/math.texi b/manual/math.texi
index a9f2a9813832..d19a14b47dc3 100644
--- a/manual/math.texi
+++ b/manual/math.texi
@@ -70,7 +70,7 @@ provided for @code{_Float32}, @code{_Float64} and @code{_Float32x} on
all platforms.
It is also provided for @code{_Float128} and @code{_Float64x} on
powerpc64le (PowerPC 64-bits little-endian), x86_64, x86, ia64,
-aarch64, alpha, mips64, s390 and sparc.
+aarch64, alpha, mips64, riscv, s390 and sparc.

@menu
* Mathematical Constants::      Precise numeric values for often-used

> I don't think you strictly need to add it to the similar lists for
> _Float128 and _Float64x support in the NEWS section for glibc 2.27, since
> the whole port is new rather than this being a new feature for an existing
> port in the RISC-V case.  However, the patch series needs to include a
> patch adding a NEWS item about the port, and the port also needs to be
> added to the list in the README file of supported glibc configurations
> using the Linux kernel.

How does this look?

commit 462ad608e82fa7a4c340f05dc6c300610282d8d9
Author: Palmer Dabbelt <palmer@dabbelt.com>
Date:   Fri Dec 22 19:33:57 2017 -0800

   NEWS and README

diff --git a/NEWS b/NEWS
index a43ff26e83cf..a7476265ca9e 100644
--- a/NEWS
+++ b/NEWS
@@ -40,13 +40,13 @@ Major new features:
  process aborts as the result of assertion failures.

* On platforms where long double has the IEEE binary128 format (aarch64,
-  alpha, mips64, s390 and sparc), the math library now implements _Float128
-  interfaces for that type, as defined by ISO/IEC TS 18661-3:2015.  These
-  are the same interfaces added in version 2.26 for some platforms where
+  alpha, mips64, riscv, s390 and sparc), the math library now implements
+  _Float128 interfaces for that type, as defined by ISO/IEC TS 18661-3:2015.
+  These are the same interfaces added in version 2.26 for some platforms where
  this format is supported but is not the format of long double.

* On platforms with support for _Float64x (aarch64, alpha, i386, ia64,
-  mips64, powerpc64le, s390, sparc and x86_64), the math library now
+  mips64, powerpc64le, riscv, s390, sparc and x86_64), the math library now
  implements interfaces for that type, as defined by ISO/IEC TS
  18661-3:2015.  These are corresponding interfaces to those supported for
  _Float128.
@@ -67,6 +67,17 @@ Major new features:
  collation ordering.  Previous glibc versions used locale-specific
  ordering, the change might break systems that relied on that.

+* Support for the RISC-V ISA running on Linux has been added.  This port
+  requires at least binutils-2.30, gcc-7.3.0, and linux-4.15; and is supported
+  for the following ISA and ABI pairs:
+
+    - rv32imac ilp32
+    - rv32imafdc ilp32
+    - rv32imafdc ilp32d
+    - rv64imac lp64
+    - rv64imafdc lp64
+    - rv64imafdc lp64d
+
Deprecated and removed features, and other changes affecting compatibility:

* On GNU/Linux, the obsolete Linux constant PTRACE_SEIZE_DEVEL is no longer
diff --git a/README b/README
index c3d17d137887..4ab2bfc6f4c4 100644
--- a/README
+++ b/README
@@ -39,6 +39,8 @@ The GNU C Library supports these configurations for using Linux kernels:
	powerpc64*-*-linux-gnu	Big-endian and little-endian.
	s390-*-linux-gnu
	s390x-*-linux-gnu
+	riscv32-*-linux-gnu
+	riscv64-*-linux-gnu
	sh[34]-*-linux-gnu
	sparc*-*-linux-gnu
	sparc64*-*-linux-gnu