Re: [XEN PATCH for-4.14] tools/xen-ucode: fix error code propagation of microcode load operation

[Ian Jackson <ian.jackson@xxxxxxxxxx>]

Igor Druzhinin writes ("Re: [XEN PATCH for-4.14] tools/xen-ucode: fix error 
code propagation of microcode load operation"):
> On 12/06/2020 17:53, Ian Jackson wrote:
> > Unfortunately I don't think this is right.  errno might not fit into a
> > return value.
> 
> errno codes that Xen return are all lower than 127. It fits perfectly fine.

I don't think this is a stable aspect of Xen's ABI, is it ?  And of
course what you get from libxc is not a Xen errno in Xen encoding, but
a Xen errno in host errno encoding, whioch might be >=127 even if the
Xen number for the same EFOOBAR is <=127.

But anyway, if this is controversial I will drop it.

> Probably but that is identical to what was suggested.
> Cleaning resource is just a nice thing to do although unnecessary.
> Can change to just "return errno" if that's what you'd prefer.

Yes please.

Would you care to at least arrange for bad usage to exit nonzero ?
I will leave it up to you to resolve this quandry: your view is that
this program's exit status is a Xen errno value (in host encoding,
presumably, although you don't state this) but now you need to return
an error that didn't come from Xen.

Ian.