Re: asm register constraint. Was: [PATCH v2 bpf-next 2/5] bpf: Introduce "volatile compare" macro

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Mon, Jan 15, 2024 at 8:33 AM Eduard Zingerman <eddyz87@xxxxxxxxx> wrote:
> [1] selftests
>     https://gist.github.com/eddyz87/276f1ecc51930017dcddbb56e37f57ad
> [2] Cilium
>     https://gist.github.com/eddyz87/4a485573556012ec730c2de0256a79db
>     Note: this is based upon branch 'libbpf-friendliness'
>           from https://github.com/anakryiko/cilium
> [3] Tetragon
>     https://gist.github.com/eddyz87/ca9a4b68007c72469307f2cce3f83bb1


The changes to all three make sense, but they might cause regressions
if they are not synchronized with new llvm.
cilium/tetragon can control the llvm version to some degree, but not selftests.
Should we add clang macro like __BPF_CPU_VERSION__ and ifdef
different asm style depending on that?
I suspect this "(short)" workaround will still be needed for quite
some time while people upgrade to the latest llvm.
something like __BPF_STRICT_ASM_CONSTRAINT__ ?
Maybe a flag too that can revert to old behavior without warnings?





[Index of Archives]     [Linux Samsung SoC]     [Linux Rockchip SoC]     [Linux Actions SoC]     [Linux for Synopsys ARC Processors]     [Linux NFS]     [Linux NILFS]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]


  Powered by Linux