Lakshmi Ramasubramanian <nramas@xxxxxxxxxxxxxxxxxxx> writes: > @@ -63,7 +29,22 @@ void remove_ima_buffer(void *fdt, int chosen_node) > if (!prop) > return; > > - ret = do_get_kexec_buffer(prop, len, &addr, &size); > + ret = fdt_address_cells(fdt, chosen_node); This change was already present in the previous version of the patch but it was just today that I noticed a problem: there's no #address-cells property in /chosen. This code will still work though because if there's no property this function returns 2 which is the correct value for ppc64. But it's conceptually wrong. You need to pass the node offset for / so that it gets the #address-cells property from there. > + if (ret < 0) > + return; > + addr_cells = ret; > + > + ret = fdt_size_cells(fdt, chosen_node); Here we're not so lucky. The default value returned when no #size-cells property is present is 1, which is wrong for ppc64 so this change introduces a bug. You also need to pass the node offset for / here. > + if (ret < 0) > + return; > + size_cells = ret; > + > + if (len < 4 * (addr_cells + size_cells)) > + return; > + > + addr = of_read_number(prop, addr_cells); > + size = of_read_number(prop + 4 * addr_cells, size_cells); > + > fdt_delprop(fdt, chosen_node, FDT_PROP_IMA_KEXEC_BUFFER); > if (ret) > return; > @@ -129,9 +110,15 @@ int setup_ima_buffer(const struct kimage *image, void *fdt, int chosen_node) > if (!image->arch.ima_buffer_size) > return 0; > > - ret = get_addr_size_cells(&addr_cells, &size_cells); > - if (ret) > + ret = fdt_address_cells(fdt, chosen_node); > + if (ret < 0) > + return ret; > + addr_cells = ret; > + > + ret = fdt_size_cells(fdt, chosen_node); > + if (ret < 0) > return ret; > + size_cells = ret; > > entry_size = 4 * (addr_cells + size_cells); Ditto here. -- Thiago Jung Bauermann IBM Linux Technology Center
