expat is required by cross for OpenOCD support, but not provided for native
builds, so replace BINUTILS_GDB_TARGET_FLAGS with BINUTILS_GDB_NATIVE_FLAGS.
248124ac2e RISC-V: Add missing privileged spec registers.
bee792f83a RISC-V: Add compressed instruction hints, and a few misc cleanups.
7d895b6f4c Add missing RISC-V fsrmi and fsflagsi instructions.
04d33e4a83 Objcopy interleave fails if section address not multiple of interleave.
71a8fbb5aa Really fix riscv shared library __global_pointer$ problem.
355497b4a7 Riscv shared libraries should not export __global_pointer$.
bfd4dc4363 Fix for texinfo 4.8.
caf839a4a4 Update and clean up RISC-V gas documentation.
df726395c9 Give Palmer co-credit for last patch.
0362ab46db Fix riscv malloc error on small alignment after norvc.
3adcc16f76 Merge pull request #126 from riscv/native-gdb
be4ba40749 Make native gdb builds work.
* Makefile.in (relsrcdir): New.
(stamps/build-gcc-linux-stage1, stamps/build-gcc-linux-stage2)
(stamps/build-gcc-newlib-stage1, stamps/build-gcc-newlib/stage2): Pass
--src to configure set from relsrcdir.
This might be an actaul newlib bug: it's not providing a range error
when taking log1p of -1e80, but it appears that this triggered a glibc
bug from a while ago that may have not been fixed in newlib. I'm not
sure why the bug is just manifesting now (the newlib sources haven't
changed for a long time), but the glibc version has some macros that
force the evaluation of particular floating-point expressions (one of
which produces the range error) that don't exist in newlib.
I'm OK whitelisting the error for the next release, despite it
technically being a regression. log1p isn't a standard function, and I
doubt embedded users are going to notice. That said, we should still
fix it.
* Makefile.in (NEWLIB_NANO_TARGET_BOARDS): New.
(build-libc): Change build-newlib to stamps/build-newlib. Add
stamps/build-newlib-nano.
(check-newlib-nano, check-gcc-newlib-nano, check-binutils-newlib-nano)
(check-gdb-newlib-nano, report-newlib-nano): New.
(check-dhrystone-linux, check-dhrystone-newlib): Delete first version
of these rules.
(build-gcc-newlib-stage1): Use --with-sysroot.
(build-newlib-nano, merge-newlib-nano): New.
(build-gcc-newlib-stage2): Depend on merge-newlib-nano. Use
--with-sysroot instead of --with-headers. Also use
--with-native-system-header-dir.
(build-dejagnu): Install it.
(check-gcc-newlib, check-gcc-linux, check-dhrystone-newlib-%)
(check-dhrystone-linux-%, check-binutils-newlib, check-binutils-linux)
(check-gdb-newlib, check-gdb-linux): Drop
DEJAGNULIBS environment variable set.
(check-gcc-newlib-nano, check-dhrystone-newlib-nano)
(check-dhrystone-newlib-nano-%, check-binutils-newlib-nano)
(check-gdb-newlib-nano, report-gcc-newlib-nano)
(report-dhrystone-newlib-nano, report-binutils-newlib-nano)
(report-gdb-newlib-nano): New.
(Makefile): Add rule to rebuild when Makefile.in changes.
(test/benchmarks/dhrystone/check): Add -specs= option. Use it in
$cc commands.
This has the same commits as the previous hash, it's just on the proper
branch now (so it's round-tripped through upstream's trunk). I archived
the old commit.
b66926e93524 RISC-V: Emit "i" suffix for instructions with immediate operands
a2fc54542b66 RISC-V: If -m[no-]strict-align is not passed, assume its value from -mtune
f34a83e82258 RISC-V: Set SLOW_BYTE_ACCESS=1
7dde69e2c5f7 RISC-V: Handle non-legitimate address in riscv_legitimize_move
1751fbe7b9e8 RISC-V: Use "@minus{}2 GB" instead of "-2 GB" in invoke.texi
6d1f1f891869 RISC-V: Document the medlow and medany code models