On Mon, Oct 23, 2017 at 09:20:36AM +0200, Sascha Hauer wrote: > On Wed, Oct 18, 2017 at 03:41:16PM +0200, Uwe Kleine-König wrote: > > From: Uwe Kleine-König <u.kleine-koenig@xxxxxxxxxxxxxx> > > > > This handler updates the non-active MMC boot partition and after a > > successful update makes the updated partition the active one. This way > > the machine should continue to be bootable when the update fails. > > Adding something like this as a comment above > imx6_bbu_internal_mmcboot_register_handler() would be nice. ok. > > +static int imx_bbu_internal_v2_mmcboot_update(struct bbu_handler *handler, > > + struct bbu_data *data) > > +{ > > + struct imx_internal_bbu_handler *imx_handler = > > + container_of(handler, struct imx_internal_bbu_handler, handler); > > + int ret; > > + uint32_t *barker; > > + char *bootpartvar; > > + const char *bootpart; > > + char *devicefile; > > + > > + barker = data->image + imx_handler->flash_header_offset; > > + > > + if (*barker != IVT_BARKER) { > > + printf("Board does not provide DCD data and this image is no imximage\n"); > > + return -EINVAL; > > + } > > + > > + ret = asprintf(&bootpartvar, "%s.boot", data->devicefile); > > + if (ret < 0) { > > + printf("Failed to allocate string for boot variable\n"); > > I think messages for failed memory allocations which eat more space than > the actual allocation should be avoided. Just return -ENOMEM without a > message. In general not giving an error message makes the problem harder to locate. But maybe if I'm unable to allocate 20 Bytes the problem is obvious. Best regards Uwe -- Pengutronix e.K. | Uwe Kleine-König | Industrial Linux Solutions | http://www.pengutronix.de/ | _______________________________________________ barebox mailing list barebox@xxxxxxxxxxxxxxxxxxx http://lists.infradead.org/mailman/listinfo/barebox