Hi Bao Cheng,
On Sat, 30 Jul 2022 at 03:05, Su, Bao Cheng baocheng.su@siemens.com wrote:
Commit 71551055cbdb ("spl: fit: Load devicetree when a Linux payload is found") made a change to not report the spl_fit_append_fdt error at all if next-stage image is u-boot.
However for u-boot image without CONFIG_OF_EMBED, the error should be reported to uplevel caller. Otherwise, uplevel caller would think the fdt is already loaded which is obviously not true.
Signed-off-by: Baocheng Su baocheng.su@siemens.com
Changes in v2:
- Fix the wrong wrapping
common/spl/spl_fit.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-)
diff --git a/common/spl/spl_fit.c b/common/spl/spl_fit.c index a35be52965..00404935cb 100644 --- a/common/spl/spl_fit.c +++ b/common/spl/spl_fit.c @@ -770,8 +770,12 @@ int spl_load_simple_fit(struct spl_image_info *spl_image, */ if (os_takes_devicetree(spl_image->os)) { ret = spl_fit_append_fdt(spl_image, info, sector, &ctx);
if (ret < 0 && spl_image->os != IH_OS_U_BOOT)return ret;
if (ret < 0) {if (spl_image->os != IH_OS_U_BOOT)return ret;else if (!IS_ENABLED(CONFIG_OF_EMBED))return ret;
This is a pretty unpleasant condition. I think we would be better to report the error and let the caller figure it out.
There are no tests associated with this, so it is hard to know what is actually going on.
If we must have this workaround, I suggest adding a Kconfig so boards that need it can turn it on, and other boards can use normal operation, which is to report errors.
Regards, Simon