Bug 251564
| Summary: | modify build of devel/libffi for riscv64 | ||||||
|---|---|---|---|---|---|---|---|
| Product: | Ports & Packages | Reporter: | Mark Linimon <linimon> | ||||
| Component: | Individual Port(s) | Assignee: | Niclas Zeising <zeising> | ||||
| Status: | Closed FIXED | ||||||
| Severity: | Affects Only Me | CC: | kp, mhorne | ||||
| Priority: | --- | Flags: | zeising:
maintainer-feedback+
|
||||
| Version: | Latest | ||||||
| Hardware: | Any | ||||||
| OS: | Any | ||||||
| See Also: | https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=250968 | ||||||
| Attachments: |
|
||||||
This papers over the real issue, which is that we lack a proper definition of __builtin___clear_cache on FreeBSD/RISC-V. This is something I'm investigating that doesn't have a clear long-term solution yet. Given the number of ports that depend on python to build, I think this patch has value, even as a temporary stopgap. I have no problems with this, approved! A commit references this bug: Author: kp Date: Tue Dec 8 16:04:37 UTC 2020 New revision: 557291 URL: https://svnweb.freebsd.org/changeset/ports/557291 Log: devel/libffi: Fix abort() on FreeBSD riscv related to __clear_case() Like on mips, disable the __builtin___clear_cache() call because it triggers aborts. PR: 251564 Approved by: zeising Changes: head/devel/libffi/files/patch-src_riscv_ffi.c Seems like this can be closed now, yes? |
Created attachment 220221 [details] new file devel/libffi/files/patch-src_riscv_ffi.c Adding this patch to devel/libffi/files fixes the build of dependent port devel/py-setuptools on riscv64, which itself blocks more than 5868 ports. This is similar to the patch used on mips* .