Message ID | 1585645160-108428-1-git-send-email-ye.li@nxp.com |
---|---|
State | Superseded |
Delegated to: | Tom Rini |
Headers | show |
Series | Revert "common: spl_fit: Default to IH_OS_U_BOOT if FIT_IMAGE_TINY enabled" | expand |
Hi Ye Li, On Tue, Mar 31, 2020 at 6:00 AM Ye Li <ye.li@nxp.com> wrote: > > The patch in commit cf8dcc5d02c32173b74bf1b7600dd2b990a90b13 is not correct, 1. "Patch in commit" is not correct. Please refer it to "commit" instead. 2. Please use the same kernel style to refer to commit: 12 SHA digit and the subject between (" xxx ") So this would become: Commit cf8dcc5d02c3 ("common: spl_fit: Default to IH_OS_U_BOOT if FIT_IMAGE_TINY enabled") is not correct... > it will append fdt to each loadable image. > Actually when using TINY FIT, the first loadable image is thought as u-boot > and already have fdt appended. > > Signed-off-by: Ye Li <ye.li@nxp.com> Tested on a imx6q and imx6qp sabresd: Tested-by: Fabio Estevam <festevam@gmail.com> Thanks
On Tue, Mar 31, 2020 at 06:40:28AM -0300, Fabio Estevam wrote: > Hi Ye Li, > > On Tue, Mar 31, 2020 at 6:00 AM Ye Li <ye.li@nxp.com> wrote: > > > > The patch in commit cf8dcc5d02c32173b74bf1b7600dd2b990a90b13 is not correct, > > 1. "Patch in commit" is not correct. Please refer it to "commit" instead. > > 2. Please use the same kernel style to refer to commit: 12 SHA digit > and the subject between (" xxx ") > > So this would become: > > Commit cf8dcc5d02c3 ("common: spl_fit: Default to IH_OS_U_BOOT if > FIT_IMAGE_TINY enabled") is not correct... > > > it will append fdt to each loadable image. > > Actually when using TINY FIT, the first loadable image is thought as u-boot > > and already have fdt appended. > > > > Signed-off-by: Ye Li <ye.li@nxp.com> > > Tested on a imx6q and imx6qp sabresd: > > Tested-by: Fabio Estevam <festevam@gmail.com> You should also mention that you're reverting the previous commit and I would really like Abel to confirm the problem they saw isn't reproducible. Thanks!
On Tue, Mar 31, 2020 at 01:59:20AM -0700, Ye Li wrote: > The patch in commit cf8dcc5d02c32173b74bf1b7600dd2b990a90b13 is not correct, > it will append fdt to each loadable image. > Actually when using TINY FIT, the first loadable image is thought as u-boot > and already have fdt appended. > > Signed-off-by: Ye Li <ye.li@nxp.com> Hi, I just rebased an i.mx8mm board from v2020.04-rc4 to v2020.04-rc5 that I'm preparing for submission upstream and noticed that this commit broke start of U-Boot. Reverting this or disabling FIT_IMAGE_TINY fixes it. -Matt > --- > common/spl/spl_fit.c | 4 ---- > 1 file changed, 4 deletions(-) > > diff --git a/common/spl/spl_fit.c b/common/spl/spl_fit.c > index aef1dbd..69dabd2 100644 > --- a/common/spl/spl_fit.c > +++ b/common/spl/spl_fit.c > @@ -646,10 +646,6 @@ int spl_load_simple_fit(struct spl_image_info *spl_image, > > if (!spl_fit_image_get_os(fit, node, &os_type)) > debug("Loadable is %s\n", genimg_get_os_name(os_type)); > -#if CONFIG_IS_ENABLED(FIT_IMAGE_TINY) > - else > - os_type = IH_OS_U_BOOT; > -#endif > > if (os_type == IH_OS_U_BOOT) { > spl_fit_append_fdt(&image_info, info, sector, > -- > 2.7.4 >
Hi Matt, > -----Original Message----- > From: Matt Porter <mporter@konsulko.com> > Sent: 2020年4月8日 5:38 > To: Ye Li <ye.li@nxp.com> > Cc: trini@konsulko.com; u-boot@lists.denx.de; Fabio Estevam > <fabio.estevam@nxp.com>; Peng Fan <peng.fan@nxp.com>; Abel Vesa > <abel.vesa@nxp.com>; dl-uboot-imx <uboot-imx@nxp.com> > Subject: [EXT] Re: [PATCH] Revert "common: spl_fit: Default to > IH_OS_U_BOOT if FIT_IMAGE_TINY enabled" > > Caution: EXT Email > > On Tue, Mar 31, 2020 at 01:59:20AM -0700, Ye Li wrote: > > The patch in commit cf8dcc5d02c32173b74bf1b7600dd2b990a90b13 is not > > correct, it will append fdt to each loadable image. > > Actually when using TINY FIT, the first loadable image is thought as > > u-boot and already have fdt appended. > > > > Signed-off-by: Ye Li <ye.li@nxp.com> > > Hi, > > I just rebased an i.mx8mm board from v2020.04-rc4 to v2020.04-rc5 that I'm > preparing for submission upstream and noticed that this commit broke start > of U-Boot. Reverting this or disabling FIT_IMAGE_TINY fixes it. > > -Matt The mkimage_fit_atf.sh has a bug. It reverses the u-boot and atf: atf is set as firmware, but u-boot as loadable. So when the commit is reverted, fdt is appended to atf only. Previously fdt was appended to both. I will send a patch to fix the bug. Best regards, Ye Li > > > --- > > common/spl/spl_fit.c | 4 ---- > > 1 file changed, 4 deletions(-) > > > > diff --git a/common/spl/spl_fit.c b/common/spl/spl_fit.c index > > aef1dbd..69dabd2 100644 > > --- a/common/spl/spl_fit.c > > +++ b/common/spl/spl_fit.c > > @@ -646,10 +646,6 @@ int spl_load_simple_fit(struct spl_image_info > > *spl_image, > > > > if (!spl_fit_image_get_os(fit, node, &os_type)) > > debug("Loadable is %s\n", > > genimg_get_os_name(os_type)); -#if > CONFIG_IS_ENABLED(FIT_IMAGE_TINY) > > - else > > - os_type = IH_OS_U_BOOT; > > -#endif > > > > if (os_type == IH_OS_U_BOOT) { > > spl_fit_append_fdt(&image_info, info, sector, > > -- > > 2.7.4 > >
diff --git a/common/spl/spl_fit.c b/common/spl/spl_fit.c index aef1dbd..69dabd2 100644 --- a/common/spl/spl_fit.c +++ b/common/spl/spl_fit.c @@ -646,10 +646,6 @@ int spl_load_simple_fit(struct spl_image_info *spl_image, if (!spl_fit_image_get_os(fit, node, &os_type)) debug("Loadable is %s\n", genimg_get_os_name(os_type)); -#if CONFIG_IS_ENABLED(FIT_IMAGE_TINY) - else - os_type = IH_OS_U_BOOT; -#endif if (os_type == IH_OS_U_BOOT) { spl_fit_append_fdt(&image_info, info, sector,
The patch in commit cf8dcc5d02c32173b74bf1b7600dd2b990a90b13 is not correct, it will append fdt to each loadable image. Actually when using TINY FIT, the first loadable image is thought as u-boot and already have fdt appended. Signed-off-by: Ye Li <ye.li@nxp.com> --- common/spl/spl_fit.c | 4 ---- 1 file changed, 4 deletions(-)