Fix return on suword() failure.

Further address @kib's feedback.

Address @kib's comments. Looks nicer now.

Abandoning in favor of another way.

I wasn't sure about adding new functions to libraries in stable/ branches, particularly to notify ports of the existence. I see now that stable/13 has the FBSD_1.7 in the top level Symbol.map anyway, so no issue.

If the syscall bitset is limited to just cpuset, the syscall boundary would be better, yeah. I couldn't tell up front if it was limited to this, but on further inspection it looks like it is restricted to just cpuset, so I'll go that route instead.

It was probably not included already because there's little need for wifi in powerpc64 machines.

It's not required in the !AIM case. This function won't even get called in the !AIM case. I think at some point we'll need to clean up that file and split the AIM and !AIM into their respective files.

Yeah, just move it under the BOOKE_E500 block.

It could be removed, and added back later. The goal was to be able to handle some machine checks as needed, but none have been needed yet.

The mfmsr is required, because if you look at line 763/764 it's doing a mtmsr of (msr | PSL_POW). It *could* be replaced with the mfmsr() in each instance of the mtmsr (inline function and the asm), but that's redundant.

Anything new on this?

Thanks for updating this!

Thanks for adding this!

In D30554#734742, @ehem_freebsd_m5p.com wrote:

In D30554#734639, @mhorne wrote:

Looks like it fails to attach since 16781571 > (1 << 16). @ehem_freebsd_m5p.com please relax the rman boundary constraints.

That I cannot dispute, so seems I have to let things be exactly the way they were before despite really not liking that.

Sparse interrupt numbering? That certainly explains why the PowerPC interrupt code is written the way it was.

Don't forget to update the commit message, with the latest changes you made.

Glad to see you solve the last piece of this puzzle!

Looks fine to me.

In D30793#692367, @imp wrote:

That is, I'd rather see ZFS64

include GENERIC64
options ZFS

etc

Address @alfredo's feedback

Is this still needed? We've had ifuncs for a year now, with no problems.

Address @bdragon's feedback.