compiler_rt: add __clzdi2 and __clzti2

[matu3ba]

• structure derived from shift.zig
• rename clzsi2.zig to count0bits.zig
• test cases derived from clzsi2_test.zig

See #1290

[matu3ba]

Looks like the i386 can not handle word length of 128bit (i128). The manual only mentions quadword, which are 64 bits.

I will create a README.md and word.zig shortly explaining what word size etc means and why we need to do the word dance wrt endianess and maximum supported word size.

Personally I think having optimised versions by 1. (maximum) word size and 2. endianess would make sense, because zig may ship no LLVM and hence no peephole optimiser. However this complicates code and may introduce more maintenance, so its a tradeoff. Little endian algorithms are always simpler (+ faster), if architecture supports according word size.

Test [127/139] compiler_rt.clzti2_test.test "compiler-rt-i386-linux-none-Debug-bare-mult... Segmentation fault at address 0xc /home/vsts/work/1/s/lib/std/special/compiler_rt/clzti2.zig:25:5: 0x8b81b7 in compiler_rt.clzti2.__clzti2_generic (test) return n - @bitcast(i128, x); ^ /home/vsts/work/1/s/lib/std/special/compiler_rt/clzti2_test.zig:15:30: 0x9683d7 in compiler_rt.clzti2_test.test__clzti2 (test) var result = actualClzti2(x); ^ /home/vsts/work/1/s/lib/std/special/compiler_rt/clzti2_test.zig:20:21: 0x8e85f9 in compiler_rt.clzti2_test.test "compiler-rt-i386-linux-none-Debug-bare-multi clzti2" (test) try test__clzti2(0x00800000_00000000_00000000_00000000, 8); ^ /home/vsts/work/1/s/lib/std/special/test_runner.zig:76:28: 0x9430a5 in std.special.main (test) } else test_fn.func(); ^ /home/vsts/work/1/s/lib/std/start.zig:505:22: 0x93996a in std.start.callMain (test) root.main(); ^ /home/vsts/work/1/s/lib/std/start.zig:457:12: 0x920fa2 in std.start.callMainWithArgs (test) return @call(.{ .modifier = .always_inline }, callMain, .{}); ^ /home/vsts/work/1/s/lib/std/start.zig:371:17: 0x91d002 in std.start.posixCallMainAndExit (test) std.os.exit(@call(.{ .modifier = .always_inline }, callMainWithArgs, .{ argc, argv, envp })); ^ /home/vsts/work/1/s/lib/std/start.zig:284:5: 0x91cead in std.start._start (test) @call(.{ .modifier = .never_inline }, posixCallMainAndExit, .{}); ^ qemu: uncaught target signal 6 (Aborted) - core dumped error: the following test command crashed: qemu-i386 /home/vsts/work/1/s/zig-cache/o/043ef600021677d5b227f1f9a5e165be/test /home/vsts/work/1/s/build/release/bin/zig The following command exited with error code 1: /home/vsts/work/1/s/build/release/bin/zig test /home/vsts/work/1/s/lib/std/special/compiler_rt.zig --test-name-prefix compiler-rt-i386-linux-none-Debug-bare-multi --cache-dir /home/vsts/work/1/s/zig-cache --global-cache-dir /home/vsts/.cache/zig --name test -target i386-linux-none --test-cmd qemu-i386 --test-cmd-bin -I /home/vsts/work/1/s/test --zig-lib-dir /home/vsts/work/1/s/lib error: the following build command failed with exit code 1: /home/vsts/work/1/s/zig-cache/o/c7fc03ac1f14311b29096b9002efd9df/build /home/vsts/work/1/s/build/release/bin/zig /home/vsts/work/1/s /home/vsts/work/1/s/zig-cache /home/vsts/.cache/zig test-toolchain -Denable-qemu -Denable-wasmtime ##[error]Bash exited with code '1'.

[andrewrk]

Thanks @matu3ba - this looks ready to merge. However-

LLVM 12.x regressed on MIPS, so big endian can not be tested in CI and thus was omitted.

Today I pushed d0f0482 and fixed the CI tarballs with 75263e1. Mind rebasing against latest master with those MIPS test enabled?

[matu3ba]

@andrewrk Sweetly enough it works without any changes and the ugly big endian spaghetti code of LLVM. Just hope debug semantics match release ones.

[matu3ba]

I will change the text in a followup PR for parity or ctz.

[matu3ba]

This one comment should be changed, which I will do in a follow-up PR.