NetBSD-Users archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
Re: RPi4 died SD card weirdness?
On Sun, 23 Aug 2026, Ramiro Aceves wrote:
Sorry, I placed commands not in the mount directory. This is the right test:
netbsd-acer# pwd
/root
netbsd-acer# mount /dev/dk2 disco_fat/
netbsd-acer# cd disco_fat/
netbsd-acer# echo "this is a nightmare" > test
netbsd-acer# ls
EFI boot.cfg fixup4cd.dat
LICENCE.broadcom bootcode.bin fixup_cd.dat
RPI4_UEFI_Firmware_v1.50.zip cmdline.txt netbsd.img
RPI_EFI.fd cmdline.txt-bak overlays
RPI_EFI.fd-bak config.txt start.elf
Readme.md config.txt-bak start4.elf
bcm2711-rpi-4-b.dtb dtb start4cd.elf
bcm2711-rpi-400.dtb firmware start_cd.elf
bcm2711-rpi-cm4.dtb fixup.dat test
boot fixup4.dat
netbsd-acer# cat test
this is a nightmare
netbsd-acer# cd ..
netbsd-acer# umount disco_fat/
netbsd-acer# pwd
/root
netbsd-acer# mount /dev/dk2 disco_fat/
netbsd-acer# cd disco_fat/
netbsd-acer# ls
EFI boot.cfg fixup4cd.dat
LICENCE.broadcom bootcode.bin fixup_cd.dat
RPI4_UEFI_Firmware_v1.50.zip cmdline.txt netbsd.img
RPI_EFI.fd cmdline.txt-bak overlays
RPI_EFI.fd-bak config.txt start.elf
Readme.md config.txt-bak start4.elf
bcm2711-rpi-4-b.dtb dtb start4cd.elf
bcm2711-rpi-400.dtb firmware start_cd.elf
bcm2711-rpi-cm4.dtb fixup.dat
boot fixup4.dat
netbsd-acer# cat test
cat: test: No such file or directory
netbsd-acer#
Yes, I've heard of this exact same error of some SD devices becoming
"read-only". Back in ~2011 or so, this was reported on xda-developers in
the ViewSonic G-Tablet forums. This tablet has with 16 GB of internal storage.
Most, I believe came with eMMC, but, a few others had an SDHC device (same
capacity). And, some SDHC ones exhibited this exact same failure mode.
Android would keep writing, and as long as all the I/O was taking place on
the FS cache, it all looked OK. But the kernel would panic when it noticed
the inconsistency between the in-cache data and the on-disk stuff.
I got in touch with Pierre Ossman (Linux SDcard drivers), and he hadn't seen
this failure either, but diagnosed it as some weird HW issue.
-RVP
PS. Try seeing if a `dd' with `oflag=direct' with a very large file (10GB)
provokes an error.
Home |
Main Index |
Thread Index |
Old Index