On Wed, Nov 17, 2010 at 02:36:47PM -0800, Erik Gilling wrote: > On Tue, Nov 16, 2010 at 2:41 PM, Robert Morell <rmorell@xxxxxxxxxx> wrote: > *snip* > > @@ -106,7 +106,7 @@ static int nvhost_channelopen(struct inode *inode, struct file *filp) > > gather_size = sizeof(struct nvhost_op_pair) * NVHOST_MAX_GATHERS; > > priv->gather_mem = nvmap_alloc(ch->dev->nvmap, gather_size, 32, > > NVMAP_HANDLE_CACHEABLE); > > - if (IS_ERR(priv->gather_mem)) > > + if (IS_ERR_OR_NULL(priv->gather_mem)) > > From my reading of nvmap_alloc, it will never return NULL. Have you > seen otherwise? Not in practice, and I agree that it doesn't look like it will. However, this seems like more of an implementation detail in the current nvmap_alloc than a guarantee. Looking at other calls to nvmap_alloc(), 2 of the 3 use IS_ERR_OR_NULL rather than IS_ERR. I guess we should pick one and make all of them consistent. I think the maintenance benefit outweighs the few extra branches, but if you think otherwise I'll remove that hunk from this patch and change the rest to just IS_ERR separately. Thanks, Robert > -Erik > -- To unsubscribe from this list: send the line "unsubscribe linux-tegra" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html