Closed tmakatos closed 1 year ago
I really hate PRix64 and friends, they make code so unreadable. Can't we just use %llx and cast the arg instead?
Also, if we want to support these (I'm surprised i686 is still relevant), we need some CI to at least build + test them.
I really hate PRix64 and friends, they make code so unreadable. Can't we just use %llx and cast the arg instead?
Sure.
Also, if we want to support these (I'm surprised i686 is still relevant), we need some CI to at least build + test them.
Indeed.
BTW, someone pointed me to: https://lore.kernel.org/lkml/alpine.DEB.2.22.394.2208251435370.104368@digraph.polyomino.org.uk/
Looking forward to that!!
why do we need our own vfio_user_region_info?
why do we need our own vfio_user_region_info?
We don't, that was from the first version where I was working around the different typedef for u64
between x86_64 and i386.
Also, if we want to support these (I'm surprised i686 is still relevant), we need some CI to at least build + test them.
Indeed.
From a quick look I had at https://docs.github.com/en/actions/using-github-hosted-runners/about-github-hosted-runners, it doesn't look like GitHub provides a 32-bit VM. Not sure we can/want to provide our own-hosted 32-bit runner. Alternatively we'd have to enable cross-compilation somehow.
Next, I'll test on PPC and then will mark PR ready for review.
Cross-compilation doesn't help too much, since we couldn't run tests.
I've added a few more fixes
Signed-off-by: Thanos Makatos thanos.makatos@nutanix.com