For the record, we've removed the SRV record for keys.gentoo.org for now, to work around the problem. Without the SRV record, everything works as expected.

Ok, it seems that my reproducer isn't correct after all. The user just confirmed that the SRV lookup succeeds on their system, so it seems GPG hits some loop repeating that for no apparent reason.

In T4735#135315, @werner wrote:

Shall we backport this to 2.2 which is our LTS release?

The problem seems to have returned in 2.2.24.

I can confirm that the patch seems to resolve the issue for me.

Yes, I do have a signing key (that is distinct from the primary key, primary key I don't store on the smartcard).

No. I mean to have an option for the caller to provide apparent UID in context to the --verify option, and have that influence the result. This is important for automated software that can't rely on user rechecking the result.

Ping.

This can easily be solved by adding two more cases to handle_send_request_error(): for GPG_ERR_EADDRNOTAVAIL (that's IPv6 disabled via procfs) and GPG_ERR_EAFNOSUPPORT (that's missing kernel support). Normally I'd submit a patch but I don't care enough to jump through all the hoops just to get two-line change in.

It seems that my patch does actually break gcc... i'm not sure how to handle
this correctly anyway. It seems that the problem is actually unsolvable, and
there's no correct way to mix inline functions and external assembly definitions.

D154: 346_0001-Fix-use-of-inline-to-avoid-multiple-declarations.patch