On Sat, Sep 17, 2011 at 05:27:01PM +0100, Russell King - ARM Linux wrote:
> On Sat, Sep 17, 2011 at 10:05:14AM -0600, Grant Likely wrote:
> > > +static struct omap_device *omap_device_alloc(struct platform_device *pdev,
> > > + struct omap_hwmod **ohs, int oh_cnt,
> > > + struct omap_device_pm_latency *pm_lats,
> > > + int pm_lats_cnt)
> > > +{
> > > + int ret = -ENOMEM;
> > > + struct omap_device *od;
> > > + struct resource *res = NULL;
> > > + int i, res_count;
> > > + struct omap_hwmod **hwmods;
> > > +
> > > + od = kzalloc(sizeof(struct omap_device), GFP_KERNEL);
> >
> > possible enhancement: devm_kzalloc() perhaps? Would simplify the cleanup
> > paths.
>
> Are you sure about that - have you thought about the lifetime issues?
> devm_*() lifetime is supposed to be from driver binding to driver
> unbinding.
>
> The lifetime of 'omap_device' appears to be from device creation to
> device destruction, which is different from what devm_*() gives you.
>
> So, using devm_*() will result in the 'omap_device' being destroyed
> when a driver is unbound - and presumably an oops if its attempted to
> be re-bound later (because the 'omap_device' structure will have been
> freed.)
Ugh, yes. You're absolutely right. devm_*() is doesn't at all work here.
g.
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
