search for: shyisc6hoqci

David Blaikie <dblaikie at gmail.com> 於 2020年1月6日 週一 下午2:23寫道： > If this is something that can vary per file in a compilation and resolve > correctly when one object file is built with one ABI and another object > file is built with a different ABI (that seems to be antithetical to the > concept of "ABI" Though) - then it should be a subtarget feature. > > ABI is

Hi all. There are two steps in LTO codegen so the problem is how to pass ABI info into LTO code generator. The easier way is pass -target-abi via option to LTO codegen, but there is linking issue when linking two bitcodes generated by different -mabi option. (see https://reviews.llvm.org/D71387#1792169) Usually the ABI info for a file is derived from target triple, mcpu or -mabi, but in RISC-V,

> On Jan 6, 2020, at 14:29, David Blaikie via llvm-dev <llvm-dev at lists.llvm.org> wrote: > > > > On Mon, Jan 6, 2020 at 5:58 AM Zakk <zakk0610 at gmail.com <mailto:zakk0610 at gmail.com>> wrote: > > > David Blaikie <dblaikie at gmail.com <mailto:dblaikie at gmail.com>> 於 2020年1月6日 週一 下午2:23寫道： > If this is something that can vary per

> On Jan 7, 2020, at 13:57, David Blaikie <dblaikie at gmail.com> wrote: > > > > On Mon, Jan 6, 2020 at 6:05 PM Daniel Sanders <daniel_l_sanders at apple.com <mailto:daniel_l_sanders at apple.com>> wrote: > > >> On Jan 6, 2020, at 14:29, David Blaikie via llvm-dev <llvm-dev at lists.llvm.org <mailto:llvm-dev at lists.llvm.org>> wrote:

David Blaikie via llvm-dev <llvm-dev at lists.llvm.org> 於 2020年1月11日 週六 上午2:03寫道： > Ah, OK - thanks for walking me through that. > > Fair enough, I think I understand the issue/tradeoff now - and that the > other module level metadata don't currently influence the target > configuration at this level? > > I'm not sure, I only know that the target-abi is decided

David Blaikie <dblaikie at gmail.com> 於 2020年1月14日 週二 上午2:15寫道： > > > On Mon, Jan 13, 2020 at 6:12 AM Zakk <zakk0610 at gmail.com> wrote: > >> >> >> David Blaikie via llvm-dev <llvm-dev at lists.llvm.org> 於 2020年1月11日 週六 >> 上午2:03寫道： >> >>> Ah, OK - thanks for walking me through that. >>> >>> Fair enough, I

On Mon, Jan 20, 2020 at 8:05 AM Sam Elliott via llvm-dev < llvm-dev at lists.llvm.org> wrote: > To follow up on this issue: > > Our plan is still to encode `target-abi` into the module flags for RISC-V > LLVM IR modules. As was pointed out earlier in this thread, the function > lowering in Clang is slightly different for the ABIs which support hardware > floating point.

I also work on the RISC-V backend, and have been doing a little work on the ELF psABI document for RISC-V. I agree that, conceptually, the psABI choice should be in the module metadata. Zakk, however, has discovered a phase ordering issue within LLVM that relates to this approach. The phase ordering problem is that the LTO backend is currently setup without interrogating the current module for

On Mon, Jan 27, 2020 at 3:04 PM Eric Christopher <echristo at gmail.com> wrote: > > > On Mon, Jan 27, 2020 at 2:56 PM David Blaikie via llvm-dev < > llvm-dev at lists.llvm.org> wrote: > >> >> >> On Mon, Jan 20, 2020 at 8:05 AM Sam Elliott via llvm-dev < >> llvm-dev at lists.llvm.org> wrote: >> >>> To follow up on this issue:

On Tue, Jan 7, 2020 at 5:27 PM Eric Christopher <echristo at gmail.com> wrote: > > > On Tue, Jan 7, 2020 at 3:18 PM Daniel Sanders via llvm-dev < > llvm-dev at lists.llvm.org> wrote: > >> >> >> On Jan 7, 2020, at 13:57, David Blaikie <dblaikie at gmail.com> wrote: >> >> >> >> On Mon, Jan 6, 2020 at 6:05 PM Daniel Sanders

Right. I think that's what we ended up doing rather than a more general attribute on the module itself. *shrugs* Probably ok? I'd probably prefer not to have to have target code to do the evaluation if possible, but everything is weird and an edge case - mips abis more than some :) -eric On Wed, Jan 8, 2020 at 8:58 AM David Blaikie <dblaikie at gmail.com> wrote: > Oh, I should