While indeed nothing prevents building dmidecode on all architectures, I can't see the point in doing that. dmidecode is only useful on systems which support the SMBIOS standard. Building dmidecode on ppc64le or s390x for example serves no purpose at all, it will only slow down the build service by building useless packages, and increase the media size by shipping these useless packages. To make things worse, if the dmidecode package is always available then other packages (such as rasdaemon) will depend on it unconditionally, so the useless packages will be forcibly installed on customer systems, wasting their disk space and network bandwidth.
Why would we want to inflict that to ourselves? The exclusive archs are listed for a purpose. If a new architecture gains support for SMBIOS then we can simply add it to the list, problem solved.
In particular I suspect if Arm is using this RiscV will likely as well
While indeed nothing prevents building dmidecode on all architectures, I can't see the point in doing that. dmidecode is only useful on systems which support the SMBIOS standard. Building dmidecode on ppc64le or s390x for example serves no purpose at all, it will only slow down the build service by building useless packages, and increase the media size by shipping these useless packages. To make things worse, if the dmidecode package is always available then other packages (such as rasdaemon) will depend on it unconditionally, so the useless packages will be forcibly installed on customer systems, wasting their disk space and network bandwidth. Why would we want to inflict that to ourselves? The exclusive archs are listed for a purpose. If a new architecture gains support for SMBIOS then we can simply add it to the list, problem solved.
@jdelvare: review reminder