Fix __SOFTFP__ case for arm getcontext

[am11]

Upstream PR: libunwind/libunwind#400.

[ghost]

I couldn't figure out the best area label to add to this PR. If you have write-permissions please help me learn by adding exactly one area label.

[ghost]

Tagging subscribers to this area: @hoyosjs
See info in area-owners.md if you want to be subscribed.

Issue Details

---

Upstream PR: libunwind/libunwind#400.

Author:	am11
Assignees:	-
Labels:	area-Infrastructure-coreclr, community-contribution
Milestone:	-

[am11]

Linux arm checked failure is #73247. I have double checked that the leg is actually building with hard fp support (and #ifdef __SOFTFP__ is false).

We use __SOFTFP__ macro in debugger support e.g.

runtime/src/coreclr/debug/createdump/dumpwriterelf.cpp

Line 410 in 73cf30d

#if defined(__arm__) && defined(__VFP_FP__) && !defined(__SOFTFP__)

so it is a known macro for both clang and gcc arm soft float toolchains.

[janvorli]

LGTM, thank you!

[AaronRobinsonMSFT]

@am11 Any concerns with waiting until upstream takes this?

[am11]

No concerns (with a but) :)
Just wanted to note that it might be useful for .NET 7 as linux-armv6 is new platform in 7.0 and CLR_CMAKE_HOST_ARCH_ARMV6 is used in libunwind.cmake; as corehost requires coreclr-pal (even for RuntimeFlavor=mono). Though I am not sure if we support (or have CI coverage for) softfp variant with armv6 as the compilation should had failed in the CI. git grep '\-mfpu=' suggests we don't but in places without -mfpu option set, it defaults to toolchain's baseline, so it is unclear if we actually support softfp/-mfpu=none.

[am11]

Upstream PR is merged.

[AaronRobinsonMSFT]

Issues are #73247 and #73474.