Re: [PATCH v4 6/6] x86/microcode/intel: Print when early microcode loading fails

[Ashok Raj <ashok.raj@intel.com>]

On Tue, Jan 17, 2023 at 08:35:33AM -0800, Dave Hansen wrote:
> On 1/9/23 07:35, Ashok Raj wrote:
> > -static void print_ucode(int old_rev, int new_rev, int date)
> > +static void print_ucode(bool failed, int old_rev, int new_rev, int date)
> ...
> >  	if (rev != mc->hdr.rev)
> > -		return -1;
> > +		retval = -1;
> >  
> >  	uci->cpu_sig.rev = rev;
> >  
> >  	if (early)
> > -		print_ucode(old_rev, uci->cpu_sig.rev, mc->hdr.date);
> > +		print_ucode(retval, old_rev, mc->hdr.rev, mc->hdr.date);
> >  	else
> > -		print_ucode_info(old_rev, uci->cpu_sig.rev, mc->hdr.date);
> > +		print_ucode_info(retval, old_rev, uci->cpu_sig.rev, mc->hdr.date);
> >  
> > -	return 0;
> > +	return retval;
> >  }
> 
> I'm generally not a _huge_ fan of having an 'int' implicitly cast to a
> bool.  The:
> 
> 	print_ucode_info(retval, ...
> 
> Line could be right or wrong based on what the retval is logically.
> This, on the other hand:
> 
> 	bool failed = false;
> 	...
> 	if (rev != mc->hdr.rev) {
> 		retval = -1;
> 		failed = true;
> 	}
> 	...
> 	print_ucode_info(failed, old_rev, uci->cpu_sig.rev, ...
> 
> *Clearly* and unambiguously matches up with:
> 
> 	static void print_ucode(bool failed, int old_rev, ...

Yes, it makes good sense.. I'll fix up next update including the commit log
that you called out.