Re: [PATCH] blktap2: update connection handling to fix build with gcc5

[Olaf Hering <olaf@aepfle.de>]

On Mon, Jul 20, Wei Liu wrote:

> On Mon, Jul 20, 2015 at 11:16:34AM +0200, Olaf Hering wrote:
> > On Mon, Jul 20, Jan Beulich wrote:
> > 
> > > >>> On 19.07.15 at 11:33, <olaf@aepfle.de> wrote:
> > > > [  198s] block-log.c:549:23: error: array subscript is above array bounds [-Werror=array-bounds]
> > > > [  198s]      if (s->connections[i].id == id)
> > > > [  198s]                        ^
> > > 
> > > So what makes the compiler right with that complaint? I.e. how does
> > > it know i > 0 here? After all - afaict - s->connected can only be 0 or
> > 
> > It has to assume that ->connected can get any value because the input
> > comes from outside the unit.
> > 
> > To reduce the patch size "&& i < MAX_CONNECTIONS" could be added.
> > 
> 
> A smaller patch would be preferable at this stage.

I disagree with that.

What is the longterm goal of that binary?
Will it ever be able to handle more than one connection? If so the loops
have to handle holes in ->connections[], which increases patch size.

If it will ever handle only a single conection, please review and spot
possible errors in my patch.

Olaf