Re: [PATCH] spi: enable spi_board_info to be registered after spi_master

[Feng Tang <feng.tang-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>]

On Wed, 28 Jul 2010 09:23:06 +0800
Feng Tang <feng.tang-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org> wrote:

> > > @@ -537,6 +557,10 @@ int spi_register_master(struct spi_master
> > > *master) dev_dbg(dev, "registered master %s%s\n",
> > > dev_name(&master->dev), dynamic ? " (dynamic)" : "");
> > >
> > > +       mutex_lock(&master_lock);
> > > +       list_add_tail(&master->list, &master_list);
> > > +       mutex_unlock(&master_lock);
> > > +
> > 
> > There seems to be a race condition here.  If a boardinfo is
> > registered after this list add, but before scan_boardinfo(master),
> > then a boardinfo can get parsed twice for a single master.  I think
> > the proper thing to do is to hold the mutex over the whole
> > operation; from registering the list or master, until after the
> > scan is complete.
> > 
> 
> Good point. Maybe moving the list_add(master) after scan_boardinfo() 
> can fix this race condition.

My solution still has problem, it was:

	scan_boardinfo(master);

        mutex_lock(&spi_master_lock);
        list_add_tail(&master->list, &spi_master_list);
        mutex_unlock(&spi_master_lock);

If another boardinfo registering just after scan_boardinfo() but before
list_add(), it will still get ignored.

I will just follow Grant's suggestion.

Thanks,
Feng

------------------------------------------------------------------------------
The Palm PDK Hot Apps Program offers developers who use the
Plug-In Development Kit to bring their C/C++ apps to Palm for a share 
of $1 Million in cash or HP Products. Visit us here for more details:
http://ad.doubleclick.net/clk;226879339;13503038;l?
http://clk.atdmt.com/CRS/go/247765532/direct/01/