On Tue, May 23, 2023 at 09:55:24AM +0200, Uwe Kleine-König wrote:
> hello,
> 
> On Tue, May 23, 2023 at 09:16:21AM +0200, Ahmad Fatoum wrote:
> > On platforms that don't have a 2nd stage (in my case i.MX6 without
> > imx6_barebox_start_usb), it usually happens that the transfer limit for
> > the first (and only) upload is bigger than the actual file length.
> > Then the right thing to do is processing the complete image (minus its
> > header), but not more. This was broken by recent refactoring and fixed
> > for the transfer case with commit 3cf4bcd86419 ("imx-usb-loader: Don't
> > try to transfer more data than contained in the image").
> > 
> > The same bug persisted in the verification code though, breaking
> > imx-usb-loader -c:
> > 
> >     verifying file...
> >     mismatch at offset 0x000999c0. expected:
> >     [ hexdump of last bytes of barebox binary ]
> > 
> > A jump to the binary will then be skipped and subsequent imx-usb-loader
> > invocations will have their DCD writes unanswered leading to the
> > dreaded:
> > 
> >     main dcd length 328
> >     DCD write: sub dcd length: 0x0324, flags: 0x04w3 in err=-7, 
> > last_trans=0  00 00 00 00
> >     addr=0x021b001c, val=0x04088032               w4 in err=-7, 
> > last_trans=0  00 00 00 00
> >     !!perform_dcd returned -7
> >     4 in err=-7, last_trans=0  00 00 00 00
> > 
> > Applying the same fix as in 3cf4bcd86419 fixes this issue as well.
> > 
> > Fixes: 3367ebc55ebe ("scripts: imx-usb-loader: simplify code flow for file 
> > size calculations")
> > Cc: Uwe Kleine-König <[email protected]>
> > Reported-by: Roland Hieber <[email protected]>
> > Signed-off-by: Ahmad Fatoum <[email protected]>
> > ---
> >  scripts/imx/imx-usb-loader.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/scripts/imx/imx-usb-loader.c b/scripts/imx/imx-usb-loader.c
> > index 5f9c7ff3a458..676f077c2557 100644
> > --- a/scripts/imx/imx-usb-loader.c
> > +++ b/scripts/imx/imx-usb-loader.c
> > @@ -1415,7 +1415,7 @@ static int do_irom_download(struct usb_work *curr, 
> > int verify)
> >     if (verify) {
> >             printf("verifying file...\n");
> >  
> > -           ret = verify_memory(image, firststage_len, header_addr);
> > +           ret = verify_memory(image, min(fsize, firststage_len), 
> > header_addr);
> >             if (ret < 0) {
> >                     printf("verifying failed\n");
> >                     goto cleanup;
> 
> Ah, I see. Looks right.
> 
> Acked-by: Uwe Kleine-König <[email protected]>
> 
> (hmm, already too late.)

Tested-by: Roland Hieber <[email protected]>

(hmm, also too late. Thanks anyway!)

 - Roland

> 
> Best regards
> Uwe
> 
> -- 
> Pengutronix e.K.                           | Uwe Kleine-König            |
> Industrial Linux Solutions                 | https://www.pengutronix.de/ |



-- 
Roland Hieber, Pengutronix e.K.          | [email protected]     |
Steuerwalder Str. 21                     | https://www.pengutronix.de/ |
31137 Hildesheim, Germany                | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686         | Fax:   +49-5121-206917-5555 |

Reply via email to