New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
"UBI init error 22" on some 128MB SPI NOR chips #18
Comments
I flashed a 128MB image on the alt bmc chip on both systems and verified the checksum, and got the same result where r217 couldn't attach the ubi volume.
|
We should make sure that U-Boot (aspeed flash driver) and Linux (spi-nor driver) have support for this new flash chip and build flash images with the required changes, if needed. Regarding QEMU, we should check that the m25p80 model has the same support and I think we should add a new aspeed witherspoon-bmc machine with the definition of this new flash chip. That would be cleaner. QEMU defines :
Is that the correct chip ? but OpenBMC u-boot defines :
May be we have typo ? If this is the datasheet : it would nice to refer to it in the commit logs. |
Unfortunately, this is not the problem. |
The issue can be reproduced with QEMU. Include this patch to define a new machine :
|
I pushed a patch for this new witherspoon machine under : |
@anoo1 what you can start by doing is activate debug under u-boot. I think UBI has its own debug macros. Check that the partitions and the UBIFS are correctly read. To generate a custom flash image, you only have to
|
this is fixing the issue on QEMU :
Can you give it a try on a real system ? |
This patch solves the issue. I've flashed it on 2 systems that were previously unable to boot. Thanks @legoater ! |
Closing issue. Patch: https://patchwork.ozlabs.org/patch/1100654/ |
We generally close the issue when the patch lands. |
Cédric Le Goater (1): aspeed/flash: fix definition of the MT25QL01GB chip (From meta-aspeed rev: 37b73caedf44bc51e1bf798fe9d566f8646353b2) Resolves openbmc/u-boot#18 Change-Id: If4082230d50c8da4a5a731fc14d938edd018ea10 Signed-off-by: Joel Stanley <joel@jms.id.au> Signed-off-by: Brad Bishop <bradleyb@fuzziesquirrel.com>
Witherspoon-128 is a relatively new OpenBMC system that is the same as witherspoon but with 128MB NOR chips (instead of 32MB). We've been flashing a few systems with the latest master but lately with a new batch of chips which are the same part number as the first ones (Micron MT25QL01GBBB8ESF), u-boot is failing to read the UBI volume where the kernel and rootfs reside:
These systems can be netbooted, and the kernel and user space are able to mount and read the UBI volume (will attach dmesg logs below), so seems just u-boot is having an issue.
Collected data for comparison, "w81" is a working machine, "r217" is a failing machine so it's booted using netboot.
w81:
r217:
I added some debug traces to u-boot and seems that in drivers/mtd/ubi/attach.c the good system reads UBI_IO_FF and moves on to the next sector but the failing system goes off on a different path into ubi_add_to_av() and ubi_compare_lebs().
Files:
The hexdump output shows the content of the ubi volume only differs in the image_seq value which is expected to be a different number every time.
r217-dmesg.txt
r217-mtd7-hexdump.txt
w81-dmesg.txt
w81-mtd7-hexdump.txt
The text was updated successfully, but these errors were encountered: