Re: [PATCH v2 2/3] hw/usb/hcd-xhci-pci: Abort if setting link property failed

[Peter Maydell]

On Tue, 24 Aug 2021 at 14:58, Eduardo Habkost <ehabkost@redhat.com> wrote:
>
> On Tue, Aug 24, 2021 at 01:16:40PM +0100, Peter Maydell wrote:
> > On Tue, 24 Aug 2021 at 13:05, Markus Armbruster <armbru@redhat.com> wrote:
> > > When you know that all callers handle errors like &error_fatal does, use
> > > of &error_fatal doesn't produce wrong behavior.  It's still kind of
> > > wrong, because relying on such a non-local argument without a genuine
> > > need is.
> >
> > Not using error_fatal results in quite a bit of extra boilerplate
> > for all those extra explicit "check for failure, print the error
> > message and exit" points.
>
> I don't get it.  There's no need for extra boilerplate if the
> caller is using &error_fatal.

Yes, that is what I mean: if you do not use error_fatal,
then there is extra boilerplate. Markus is suggesting that
we should avoid using error_fatal, and my response is "the
cost of that would be that we add boilerplate".

> > Right now we have nearly a thousand instances of error_fatal
> > in the codebase, incidentally.
>
> It looks like 73 of them are in functions that take an Error**
> argument.

Those could probably be fairly easily adjusted to pass the error
back instead. Still leaves nearly 900 to go :-)

-- PMM