From mboxrd@z Thu Jan 1 00:00:00 1970 From: ezequiel.garcia@free-electrons.com (Ezequiel Garcia) Date: Tue, 9 Apr 2013 17:49:47 -0300 Subject: [PATCH v6 1/5] drivers: memory: Introduce Marvell EBU Device Bus driver In-Reply-To: <20130409203855.GA27797@obsidianresearch.com> References: <1365539181-29242-1-git-send-email-ezequiel.garcia@free-electrons.com> <1365539181-29242-2-git-send-email-ezequiel.garcia@free-electrons.com> <20130409203855.GA27797@obsidianresearch.com> Message-ID: <20130409204946.GA2294@localhost> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Tue, Apr 09, 2013 at 02:38:55PM -0600, Jason Gunthorpe wrote: > On Tue, Apr 09, 2013 at 05:26:17PM -0300, Ezequiel Garcia wrote: > > + /* > > + * Since this is only part of the workaround, > > + * we can do this dirty 'ranges' property parsing. > > + * Of course, this will be removed once the address windows > > + * are declared in the device tree. > > + */ > > + err = of_property_read_u32_array(node, "ranges", addr_region, > > + ARRAY_SIZE(addr_region)); > > + if (err < 0) { > > + dev_err(dev, "%s has a malformed 'range' property\n", > > + node->full_name); > > + return err; > > + } > > > + /* > > + * Create an mbus address windows. > > + * FIXME: Remove this, together with the above code, once the > > + * address windows are declared in the device tree. > > + */ > > + err = mvebu_mbus_add_window(devbus_wins[cs], > > + addr_region[1], addr_region[2]); > > Oh, this is a dangerous way to parse ranges.. > > Try something like: > > const __be32 *prop; > u64 base; > u64 size; > > prop = of_get_property("ranges",&psize); > if (prop == NULL || psize != of_n_addr_cells(node) + of_n_size_cells(node)) > err; > base = of_translate_address(node,prop); > if (base == OF_BAD_ADDR) > err; > size = of_read_number(prop + of_n_addr_cells(node),of_n_size_cells(node)); > > err = mvebu_mbus_add_window(devbus_wins[cs],base,size); > No problem, I'll try something like the above. Mind explaining me why do you think it's dangerous or in what cases do you expect the other (dirty) solution to break? -- Ezequiel Garc?a, Free Electrons Embedded Linux, Kernel and Android Engineering http://free-electrons.com