On Mon, Dec 2, 2024 at 2:33 AM Andreas Schwab <sch...@suse.de> wrote: > > On Nov 21 2024, Heinrich Schuchardt wrote: > > > That ls works does not mean that the file system is not corrupted. > > This obviously has nothing to do with the file system: > > StarFive # load mmc 1 0x40200000 efi/boot/bootriscv64.efi > 167936 bytes read in 14 ms (11.4 MiB/s) > StarFive # load mmc 1 0x100000000 efi/boot/bootriscv64.efi > mmc fail to send stop cmd > Error reading cluster > Failed to load 'efi/boot/bootriscv64.efi' > StarFive # load mmc 1 0x40200000 efi/boot/bootriscv64.efi > ** fs_devread read error - block > ** fs_devread read error - block > Can't set block device > > Only a reset can revive the MMC interface. > > -- > Andreas Schwab, SUSE Labs, sch...@suse.de > GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE 1748 E4D4 88E3 0EEA B9D7 > "And now for something completely different."
Thank you for confirming this Andreas. Can you please test and review the patch today from Sughosh?: https://lore.kernel.org/u-boot/20241202070624.1295497-1-sughosh.g...@linaro.org/ ( Patchwork for same: https://patchwork.ozlabs.org/project/uboot/list/?series=434841 ) It can be applied on v2025.01-rc3 and with CONFIG_EFI_LOADER_BOUNCE_BUFFER=n CONFIG_BOUNCE_BUFFER=n I have been testing with CONFIG_BOOTSTD=n There should be now a success on 4GB+ DRAM VisionFive2 board loading some < 1GB data file from mmc to U-Boot address > 4GB. Please test and reply to Sughosh patch with remarks. Thank you! -E
