Re: [PATCH v3 2/2] fastboot: bail if update handler couldn't be found for bbu-partition

[Sascha Hauer <sha@pengutronix.de>]

On Fri, Jun 10, 2022 at 08:36:35AM +0200, Michael Olbrich wrote:
> On Thu, Jun 09, 2022 at 04:12:39PM +0200, Ahmad Fatoum wrote:
> > On 09.06.22 16:10, Sascha Hauer wrote:
> > > On Thu, Jun 09, 2022 at 03:09:36PM +0200, Ahmad Fatoum wrote:
> > >> Fastboot would fall back to a raw copy even for bbu- partitions if
> > >> no barebox_update handler was found. Prevent this by bailing out
> > >> with an error code.
> > >>
> > >> Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
> > >> ---
> > >> v2 -> v3:
> > >>   - bail out instead of only printing message and falling back
> > >>     to raw copy (Sascha)
> > >> v1 -> v2:
> > >>   - print message in case barebox_update handler is not found
> > >> ---
> > >>  common/fastboot.c | 9 +++++++--
> > >>  1 file changed, 7 insertions(+), 2 deletions(-)
> > >>
> > >> diff --git a/common/fastboot.c b/common/fastboot.c
> > >> index 330a06f5a32f..a5cf04b39ecd 100644
> > >> --- a/common/fastboot.c
> > >> +++ b/common/fastboot.c
> > >> @@ -683,8 +683,13 @@ static void cb_flash(struct fastboot *fb, const char *cmd)
> > >>  		};
> > >>  
> > >>  		handler = bbu_find_handler_by_device(data.devicefile);
> > >> -		if (!handler)
> > >> -			goto copy;
> > >> +		if (!handler) {
> > >> +			fastboot_tx_print(fb, FASTBOOT_MSG_FAIL,
> > >> +					  "No barebox update handler registered for %s",
> > >> +					  data.devicefile);
> > >> +			ret = -ENOENT;
> > >> +			goto out;
> > >> +		}
> > > 
> > > I didn't verify that, but I believe a full raw bootable disk image
> > > generated for i.MX will be detected as barebox image. With this patch we
> > > wouldn't be able to flash that anymore.
> > 
> > Can we just apply v2 and see how often we see the message?
> 
> Or maybe the 'is a barebox image' detection is not strict enough? We know
> the image size, right? Is there a way to determine the size of the actual
> barebox image? If the two differ then it's not a barebox image after all.

filetype_is_barebox_image() looks like this:

bool filetype_is_barebox_image(enum filetype ft)
{
        switch (ft) {
        case filetype_arm_barebox:
        case filetype_mips_barebox:
        case filetype_ch_image:
        case filetype_ch_image_be:
        case filetype_layerscape_image:
        case filetype_layerscape_qspi_image:
        case filetype_stm32_image_fsbl_v1:
        case filetype_fip:
                return true;
        default:
                return false;
        }
}

There's likely a way to determine the size for some of the images, but
probably not for all.

We could check the size of the image. If it's too big for a barebox
image then it is none. I don't have a very good feeling about such a
heuristic though.

Sascha

-- 
Pengutronix e.K.                           |                             |
Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox