Fix FTBFS on armhf

[Danilo Egea Gondolfo]

bpftrace build is failing, on Ubuntu and Debian, on armhf. CMake is not detecting the architecture correctly due to the cross-compilation. Also, apparently due to a bug on gcc, the build will fail due to the use of the __attribute__(target("arm")) in src/triggers.h.

PPA: https://launchpad.net/~danilogondolfo/+archive/ubuntu/bpftrace/

I tested it on a bare metal armhf system and bpftrace is working fine as far as I can see.

[Danilo Egea Gondolfo (danilogondolfo)]

Patch also submitted to Debian: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1037185

[Ubuntu Foundations Team Bug Bot (crichton)]

The attachment "Fix FTBFS on armhf" seems to be a debdiff. The ubuntu-sponsors team has been subscribed to the bug report so that they can review and hopefully sponsor the debdiff. If the attachment isn't a patch, please remove the "patch" flag from the attachment, remove the "patch" tag, and if you are member of the ~ubuntu-sponsors, unsubscribe the team.

[This is an automated message performed by a Launchpad user owned by ~brian-murray, for any issue please contact him.]

[Lucas Kanashiro (lucaskanashiro)]

Thanks for the patch Danilo. Could you also reference this bug in the changelog? So it can be closed once the package lands in the release pocket.

[Sergio Durigan Junior (sergiodj)]

There seems to be some some character encoding problem with the d/changelog entry as well. I'd suggest trimming the string and removing any special characters.

[Michael Hudson-Doyle (mwhudson)]

I think the CMAKE_SYSTEM_PROCESSOR thing should be taken upstream: I am not sure there are _any_ systems where uname -m can come out as armv7-a. I think values that should be interpreted as 32 bit arm are "arm", "armv7l", "armv8l".

[Michael Hudson-Doyle (mwhudson)]

I filed an upstream report while the issue was still in my head https://github.com/iovisor/bpftrace/issues/2654.

[Andreas Hasenack (ahasenack)]

I added some logs to https://github.com/iovisor/bpftrace/issues/2654, I think we should wait a bit since the armhf maintainer was pinged in that bug just 10h ago.

[Julian Andres Klode (juliank)]

Do you actually need bpftrace on armhf? I'd be inclined to close this as Won't Fix on our side as introducing a delta against Debian for armhf of all things seems out of proportion for the amount of work it will incur.

Debian has precisely the same issue.

We generally also do not cross-compile packages, neither does Debian - The packages are compiled in a native armhf toolchain, but they may run on arm64-capable hardware, possibly with arm64 kernels.

[Benjamin Drung (bdrung)]

I agree with Michael's opinion that CMAKE_SYSTEM_PROCESSOR should be taken upstream. So instead of extending debian/rules, better patch the cmake configuration and submit that change upstream. I am unsubscribing ~ubuntu-sponsors. Please resubscribe ~ubuntu-sponsors once you have addressed the previous comments.

[Danilo Egea Gondolfo (danilogondolfo)]

Both problems were fixed upstream https://github.com/iovisor/bpftrace/issues/2654