diff options
author | Ben Crocker <[email protected]> | 2017-02-10 18:08:07 -0500 |
---|---|---|
committer | Emil Velikov <[email protected]> | 2017-02-20 18:21:22 +0000 |
commit | b934aae36410e09892e772b26f61fed5d037fc69 (patch) | |
tree | c31473e334f257a26ff746b63261991cdfd8aba0 /src/gallium/auxiliary/gallivm | |
parent | a8e9c630f3dc656d107c4252b90c8e991438a3c4 (diff) |
gallivm: Override getHostCPUName() "generic" w/ "pwr8" (v4)
If llvm::sys::getHostCPUName() returns "generic", override
it with "pwr8" (on PPC64LE).
This is a work-around for a bug in LLVM: a table entry for "POWER8NVL"
is missing, resulting in (big-endian) "generic" being returned on
little-endian Power8NVL systems. The result is that code that
attempts to load the least significant 32 bits of a 64-bit quantity in
memory loads the wrong half.
This omission should be fixed in the next version of LLVM (4.0),
but this work-around should be left in place in case some
future version of POWER<n> also ends up unrepresented in LLVM's table.
This workaround fixes failures in the Piglit arb_gpu_shader_fp64 conversion
tests on POWER8NVL processors.
(V4: add similar comment in the code.)
Signed-off-by: Ben Crocker <[email protected]>
Cc: 12.0 13.0 17.0 <[email protected]>
Acked-by: Emil Velikov <[email protected]>
Diffstat (limited to 'src/gallium/auxiliary/gallivm')
-rw-r--r-- | src/gallium/auxiliary/gallivm/lp_bld_misc.cpp | 13 |
1 files changed, 13 insertions, 0 deletions
diff --git a/src/gallium/auxiliary/gallivm/lp_bld_misc.cpp b/src/gallium/auxiliary/gallivm/lp_bld_misc.cpp index 4115f1e9f40..6c64e4d2a6c 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_misc.cpp +++ b/src/gallium/auxiliary/gallivm/lp_bld_misc.cpp @@ -644,6 +644,19 @@ lp_build_create_jit_compiler_for_module(LLVMExecutionEngineRef *OutJIT, * when not using MCJIT so no instructions are generated which the old JIT * can't handle. Not entirely sure if we really need to do anything yet. */ +#if defined(PIPE_ARCH_LITTLE_ENDIAN) && defined(PIPE_ARCH_PPC_64) + /* + * Versions of LLVM prior to 4.0 lacked a table entry for "POWER8NVL", + * resulting in (big-endian) "generic" being returned on + * little-endian Power8NVL systems. The result was that code that + * attempted to load the least significant 32 bits of a 64-bit quantity + * from memory loaded the wrong half. This resulted in failures in some + * Piglit tests, e.g. + * .../arb_gpu_shader_fp64/execution/conversion/frag-conversion-explicit-double-uint + */ + if (MCPU == "generic") + MCPU = "pwr8"; +#endif builder.setMCPU(MCPU); if (gallivm_debug & (GALLIVM_DEBUG_IR | GALLIVM_DEBUG_ASM | GALLIVM_DEBUG_DUMP_BC)) { debug_printf("llc -mcpu option: %s\n", MCPU.str().c_str()); |