add -fthinlto-index= option to clang-cl

[inglorion]

Summary

This adds a -fthinlto-index= option to clang-cl, which allows it to
be used to drive ThinLTO backend passes. This allows clang-cl to be
used for distributed ThinLTO.

[inglorion]

This is similar to r254927 for the clang driver, and reuses the tests from that commit for the backend part. On the frontend, it differs in that clang-cl does not support -x (per the comment in clang/lib/Driver/Driver.cpp line 2068: "// No driver mode exposes -x and /TC or /TP; we don't support mixing them.") Instead of requiring -x ir, the behavior added to clang-cl here is to expect that what would be detected as a native object file based on file extension is really a bitcode file. If it isn't, the backend will cause us to say "error: Invalid bitcode signature", which I think is adequate as a diagnostic.

[pcc]

Do we really need to support clang-cl syntax here? Can't the user of -fthinlto-index= use the regular clang driver?

[rnk]

In D64458#1577435, @pcc wrote:

Do we really need to support clang-cl syntax here? Can't the user of -fthinlto-index= use the regular clang driver?

I think it depends on how many compiler flags we have to pass to backend compilation actions. I don't have a good sense of how many of those actually matter. Some flags affect the object file and are not encoded in bitcode, like -ffunction-sections. It's potentially onerous for a build system to have to work out for itself that cl -O2 implicitly enables -Gy, which is the MSVC spelling of -ffunction-sections, so therefore we need this other set of backend action flags. It seems nicer to for the build system to just take the first stage command line and re-run it with a new input and this new flag.

If we want to make the behavior consistent, I would say we should port this new -x ir behavior over to clang so it's the same as clang-cl.

[inglorion]

Do we really need to support clang-cl syntax here? Can't the user of -fthinlto-index= use the regular clang driver?

We could, of course, require using the clang driver for this feature. But I would argue that clang-cl exists to make it easy for projects that use MSVC-style command lines to switch to Clang and get access to the benefits it provides, including features for which there is no equivalent MSVC flag. So I think this is reasonable to add.

[tejohnson]

LGTM, but please wait for @pcc to make sure he is satisfied.
I think it would be good to port the file type deduction to clang (possibly as a follow on), to make these consistent.

[inglorion]

Simplified after rebasing on top of r366127.

[rnk]

lgtm too, since now this is just whitelisting a clang flag for clang-cl.

[pcc]

I think it's a little unfortunate that we're continuing to go down the road of letting users pass more flags to the ThinLTO backend action, but I won't stand in the way here.

cfe/trunk/include/clang/Driver/Options.td

 def fno_lto : Flag<["-"], "fno-lto">, Group<f_Group>,
   HelpText<"Disable LTO mode (default)">;
 def flto_jobs_EQ : Joined<["-"], "flto-jobs=">,
   Flags<[CC1Option]>, Group<f_Group>,
   HelpText<"Controls the backend parallelism of -flto=thin (default "
            "of 0 means the number of threads will be derived from "
            "the number of CPUs detected)">;
 def fthinlto_index_EQ : Joined<["-"], "fthinlto-index=">,
-  Flags<[CC1Option]>, Group<f_Group>,
+  Flags<[CoreOption, CC1Option]>, Group<f_Group>,
   HelpText<"Perform ThinLTO importing using provided function summary index">;
 def fmacro_backtrace_limit_EQ : Joined<["-"], "fmacro-backtrace-limit=">,
                                 Group<f_Group>, Flags<[DriverOption, CoreOption]>;
 def fmerge_all_constants : Flag<["-"], "fmerge-all-constants">, Group<f_Group>,
   Flags<[CC1Option, CoreOption]>, HelpText<"Allow merging of constants">;
 def fmessage_length_EQ : Joined<["-"], "fmessage-length=">, Group<f_Group>;
 def fms_extensions : Flag<["-"], "fms-extensions">, Group<f_Group>, Flags<[CC1Option, CoreOption]>,
   HelpText<"Accept some non-standard constructs supported by the Microsoft compiler">;

cfe/trunk/test/Driver/cl-thinlto-backend.c

+// RUN: %clang_cl -c -flto=thin -Fo%t.obj %s
+// RUN: llvm-lto2 run -thinlto-distributed-indexes -o %t.exe %t.obj
+
+// -fthinlto_index should be passed to cc1
+// RUN: %clang_cl -### -c -fthinlto-index=%t.thinlto.bc -Fo%t1.obj \
+// RUN:     %t.obj 2>&1 | FileCheck %s
+
+// CHECK: -fthinlto-index=
+// CHECK: "-x" "ir"