FS#1145 - ar71xx/mikrotik: sysupgrade on rb-nor devices broken after 4k sectors re-enabled #6119
Comments
|
mkresin: It should only happen if you are doing a sysupgrade with keeping your settings. And it should only happen once, if you are doing a sysupgrade from an image without CONFIG_MTD_SPI_NOR_USE_4K_SECTORS=y to an image with CONFIG_MTD_SPI_NOR_USE_4K_SECTORS=y. See FS#385 and more specific my comment from 16.01.2017 00:11 in that ticket. The question I have, was your config preserved successfully? |
|
joaochainho: Hi Mathias, the config was never preserved. |
|
nbd: Please test my staging tree at https://git.lede-project.org/?p=lede/nbd/staging.git;a=summary |
|
joaochainho: Hi Felix, let me thank you for your work. |
|
joaochainho: Hi Felix, I tested on a RB750UPr2 and it works fine. Unfortunately I don't have a RB951Ui-2nD at hand right now, but I think it will work fine as well. |
joaochainho:
Commit b7ea141 (ar71xx: re-enable 4k sectors for the mikrotik subtargets) breaks sysupgrade on Mikrotik RB-NOR devices, such as RB951Ui-2nD (hAP lite) and RB750UPr2 (hEX PoE lite).
NAND devices seem unaffected.
Here's some kernel log from a RB951Ui-2nD device right after sysupgrade.
... [ 24.702338] jffs2: Node at 0x00005044 with length 0x00001044 would run over the end of the erase block [ 24.712033] jffs2: Perhaps the file system was created with the wrong erase size? [ 24.848051] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x0000504c: 0xee2d instead [ 24.889528] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00005064: 0x59fd instead [ 25.014379] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00005068: 0x59fd instead [ 25.024182] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x0000506c: 0x59fd instead [ 25.078651] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00005080: 0x2134 instead [ 25.177825] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00005084: 0x5455 instead [ 25.214600] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00005088: 0x1f8b instead [ 25.242184] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00005090: 0x0203 instead [ 25.274445] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00005094: 0x6b77 instead [ 25.284246] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00005098: 0x72f9 instead [ 25.294061] jffs2: Further such events for this erase block will not be printed ... [ 26.091293] jffs2: Node at 0x00006088 with length 0x00001044 would run over the end of the erase block [ 26.100961] jffs2: Perhaps the file system was created with the wrong erase size? [ 26.178417] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00007000: 0xcc1d instead [ 26.283376] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00007004: 0x3d2b instead ... [ 27.784339] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00009020: 0x2c01 instead [ 27.794148] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00009024: 0xe20a instead [ 27.803960] jffs2: Further such events for this erase block will not be printed [ 27.914893] jffs2_scan_eraseblock(): End of filesystem marker found at 0x15000 [ 27.923162] jffs2_build_filesystem(): unlocking the mtd device... [ 27.929414] done. [ 27.931406] jffs2_build_filesystem(): erasing all blocks after the end marker... ...I did remove the option ** CONFIG_MTD_SPI_NOR_USE_4K_SECTORS=y** and everything looks good.The text was updated successfully, but these errors were encountered: