linux-kernel - Re: [PATCH 001/141] afs: Fix fall-through warnings for Clang

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <276a532e28d290abaaf9cd5969cbacecd51cabc9.camel@perches.com>
Date:   Mon, 23 Nov 2020 08:51:14 -0800
From:   Joe Perches <joe@...ches.com>
To:     David Howells <dhowells@...hat.com>
Cc:     "Gustavo A. R. Silva" <gustavoars@...nel.org>,
        linux-afs@...ts.infradead.org, linux-kernel@...r.kernel.org,
        linux-hardening@...r.kernel.org
Subject: Re: [PATCH 001/141] afs: Fix fall-through warnings for Clang

On Mon, 2020-11-23 at 16:10 +0000, David Howells wrote:
> Joe Perches <joe@...ches.com> wrote:
> 
> > >  		call->unmarshall++;
> > > +
> > > +		fallthrough;
> > 
> > My preference would be to change these to break and not fallthrough;
> > 
> > >  	case 5:
> > >  		break;
> > >  	}
> 
> My preference would be to fall through.  The case number is the state machine
> state, as indexed by call->unmarshall.

Then ideally the state machine states should be enums and not numbers
and the compiler should use a default block for unhandled states right?

Is code like call->marshall++ a common style for kernel state machines?
Perhaps not.

Does it work?
Sure.

Is it obvious what the transitions are?
No.

> All the other cases in the switch fall through.
> 
> I would also generally prefer that the fallthrough statement occur before the
> blank line, not after it, since it belongs with the previous clause, and not
> between a comment about a case statement and its associated case statement,
> i.e.:
> 
> 		afs_extract_to_tmp(call);
> 		call->unmarshall++;
> 
> 		/* extract the callback array and its count in two steps */
> 		fallthrough;
> 	case 3:
> 
> would be better written as:
> 
> 		afs_extract_to_tmp(call);
> 		call->unmarshall++;
> 		fallthrough;
> 
> 		/* extract the callback array and its count in two steps */
> 	case 3:

I agree completely.