2. 技术支持
  3. AR# 7111:
  4. AR# 71114: 2017.1-2018.2 Zynq UltraScale+ MPSoC: Linux kernel boot failed while mounting a JFFS2 filesystem in QSPI boot mode

AR# 71114

|

2017.1-2018.2 Zynq UltraScale+ MPSoC: Linux kernel boot failed while mounting a JFFS2 filesystem in QSPI boot mode

描述

When Linux is booted with a JFFS2 filesystem on an UltraScale+ MPSoC evaluation board, the below errors (or similar errors like read_inode failure) are observed while mounting the filesystem in PetaLinux 2017.x and 2018.1/2.

2017.x boot log: 

When Linux is booted in QSPI boot mode along with a JFFS2 filesystem, the below error messages appears. 

The Full boot log is attached.

[    3.802595] ALSA device list:
[    3.805506]   #0: DisplayPort monitor
[    3.810220] jffs2: Node at 0x00001cec with length 0x00001044 would run over the end of the erase block
[    3.819531] jffs2: Perhaps the file system was created with the wrong erase size?
[    3.827014] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00001cf0: 0x1044 instead
[    3.836501] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00001cf4: 0x586d instead
[    3.845979] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00001cf8: 0x0012 instead
[    3.855467] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00001cfc: 0x0003 instead
[    3.864948] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00001d00: 0x81ed instead
[    3.874425] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00001d08: 0x3008 instead
[    3.883904] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00001d18: 0x2000 instead
[    3.893384] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00001d1c: 0x1000 instead
[    3.902863] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00001d20: 0x1000 instead
[    3.912344] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00001d28: 0x700c instead
[    3.921820] jffs2: Further such events for this erase block will not be printed
[    3.929294] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00002000: 0xeb45 instead
[    3.938779] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00002004: 0x7deb instead
[    3.948257] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00002008: 0xb31e instead
[    3.959251] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x0000200c: 0x9c13 instead
[    3.970237] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00002010: 0x0586 instead
[    3.981221] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00002014: 0xed13 instead
[    3.992215] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00002018: 0x8422 instead
[    4.003164] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x0000201c: 0xc654 instead
[    4.014065] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00002020: 0xca11 instead
[    4.024937] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00002024: 0x4ec6 instead
[    4.035808] jffs2: Further such events for this erase block will not be printed
[    4.045321] jffs2: Node at 0x00003c2c with length 0x00000691 would run over the end of the erase block
[    4.056143] jffs2: Perhaps the file system was created with the wrong erase size?
[    4.065364] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00004000: 0xc18a instead
[    4.076473] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00004004: 0x3a61 instead
[    4.087588] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00004008: 0xcd5f instead
[    4.098710] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x0000400c: 0x944f instead
[    4.109787] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00004010: 0xebcf instead
[    4.120818] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00004014: 0x8cff instead
[    4.131832] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00004018: 0x8f39 instead
[    4.142845] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x0000401c: 0x73a1 instead
[    4.153836] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00004020: 0xc358 instead
[    4.164787] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00004024: 0x0679 instead
[    4.175714] jffs2: Further such events for this erase block will not be printed
[    4.185348] jffs2: Node at 0x00005d08 with length 0x000006a6 would run over the end of the erase block
[    4.196119] jffs2: Perhaps the file system was created with the wrong erase size?
[    4.205361] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00006000: 0x153c instead
[    4.216361] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00006004: 0x0f31 instead
[    4.227323] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00006008: 0x4fb0 instead
[    4.238242] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x0000600c: 0xcfe2 instead
[    4.249125] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00006010: 0x3c88 instead
[    4.259955] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00006014: 0x8f83 instead
[    4.270719] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00006018: 0x0644 instead
[    4.281471] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x0000601c: 0x06ae instead
[    4.292208] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00006020: 0x3c41 instead
[    4.302886] jffs2: jffs2_scan_eraseblock(): Magic bitmask 0x1985 not found at 0x00006024: 0xbee9 instead
[    4.313509] jffs2: Further such events for this erase block will not be printed
 

2018.1/2 boot log: 

When the Kernel is booted in QSPI boot mode using a JFFS2 root filesystem, you will see filesystem mounting related issues as shown below:


[    4.237179] ALSA device list:
[    4.242208]   #0: DisplayPort monitor
[    4.387362] random: crng init done
[    5.109881] jffs2: notice: (1) jffs2_build_xattr_subsystem: complete building xattr subsystem, 0 of xdatum (0 unchecked, 0 orphan) and 0 of xref (0 dead, 0 orphan) found.
[    5.129424] VFS: Mounted root (jffs2 filesystem) on device 31:3.
[    5.137842] jffs2: error: (1) jffs2_do_read_inode_internal: CRC failed for read_inode of inode 4 at physical location 0x10c
[    5.151269] jffs2: iget() failed for ino #4
[    5.157703] devtmpfs: error mounting -5
[    5.163888] Freeing unused kernel memory: 512K
[    5.179677] jffs2: error: (1) jffs2_do_read_inode_internal: CRC failed for read_inode of inode 12 at physical location 0x490
[    5.193259] jffs2: iget() failed for ino #12
[    5.199889] Starting init: /sbin/init exists but couldn't execute it (error -5)
[    5.210270] jffs2: error: (1829) jffs2_do_read_inode_internal: CRC failed for read_inode of inode 128 at physical location 0x109408
[    5.216080] jffs2: error: (1) jffs2_do_read_inode_internal: CRC failed for read_inode of inode 5 at physical location 0x17c
[    5.216098] Unable to handle kernel paging request at virtual address ffffffc97a03c57e
[    5.216099] Mem abort info:
[    5.216102]   Exception class = DABT (current EL), IL = 32 bits
[    5.216103]   SET = 0, FnV = 0
[    5.216104]   EA = 0, S1PTW = 0
[    5.216105] Data abort info:
[    5.216107]   ISV = 0, ISS = 0x00000045
[    5.216108]   CM = 0, WnR = 1
[    5.216112] swapper pgtable: 4k pages, 39-bit VAs, pgd = ffffff8009062000
[    5.216114] [ffffffc97a03c57e] *pgd=0000000000000000, *pud=0000000000000000
[    5.216119] Internal error: Oops: 96000045 [#1] SMP
[    5.216121] Modules linked in:
[    5.216127] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 4.14.0-xilinx-v2018.1 #1
[    5.216128] Hardware name: ZynqMP ZCU102 Rev1.0 (DT)
[    5.216130] task: ffffffc87b846d00 task.stack: ffffff8008038000
[    5.216139] PC is at free_block.constprop.20+0xf0/0x1b0
[    5.216141] LR is at 0x1
[    5.216143] pc : [<ffffff8008189698>] lr : [<0000000000000001>] pstate: 400000c5
[    5.216145] sp : ffffff800803b7d0
[    5.216146] x29: ffffff800803b7d0 x28: ffffff800803b850
[    5.216150] x27: ffffffc87b824000 x26: ffffffc87b801080
[    5.216153] x25: ffffffc87ff75110 x24: 000000000000003c
[    5.216157] x23: ffffffc87a5f8b80 x22: 00000048771dc000
[    5.216160] x21: ffffffc87b800000 x20: ffffffc87ff75100
[    5.216163] x19: ffffff8008d99100 x18: 0000000000000001
[    5.216167] x17: 0000000000000006 x16: ffffffc87b801088
[    5.216170] x15: ffffffc87b8010a8 x14: ffffffbf00000000
[    5.216174] x13: dead000000000100 x12: dead000000000200
[    5.216177] x11: ffffffc87ff752f0 x10: ffffffbf1dab0d40
[    5.216180] x9 : ffffffc87a03c580 x8 : ffffffc87a03c580
[    5.216184] x7 : ffffffc87b801080 x6 : 00000000fffffffe
[    5.216187] x5 : 0000000000000000 x4 : 0000000000f4078b
[    5.216190] x3 : ffffff800803b850 x2 : ffffffbf1dab0d20
[    5.216193] x1 : ffffffc87ff75198 x0 : ffffffc87b800000
[    5.216198] Process swapper/0 (pid: 1, stack limit = 0xffffff8008038000)
[    5.216199] Call trace:
[    5.216202] Exception stack(0xffffff800803b690 to 0xffffff800803b7d0)
[    5.216205] b680:                                   ffffffc87b800000 ffffffc87ff75198
[    5.216209] b6a0: ffffffbf1dab0d20 ffffff800803b850 0000000000f4078b 0000000000000000
[    5.216212] b6c0: 00000000fffffffe ffffffc87b801080 ffffffc87a03c580 ffffffc87a03c580
[    5.216216] b6e0: ffffffbf1dab0d40 ffffffc87ff752f0 dead000000000200 dead000000000100
[    5.216219] b700: ffffffbf00000000 ffffffc87b8010a8 ffffffc87b801088 0000000000000006
[    5.216223] b720: 0000000000000001 ffffff8008d99100 ffffffc87ff75100 ffffffc87b800000
[    5.216227] b740: 00000048771dc000 ffffffc87a5f8b80 000000000000003c ffffffc87ff75110
[    5.216231] b760: ffffffc87b801080 ffffffc87b824000 ffffff800803b850 ffffff800803b7d0
[    5.216234] b780: 0000000000000001 ffffff800803b7d0 ffffff8008189698 00000000400000c5
[    5.216238] b7a0: ffffff800803b800 0000000000000003 ffffffffffffffff 00000000ffffffc8
[    5.216240] b7c0: ffffff800803b7d0 ffffff8008189698
[    5.216245] [<ffffff8008189698>] free_block.constprop.20+0xf0/0x1b0
[    5.216250] [<ffffff800818a498>] ___cache_free+0xc8/0x1b8
[    5.216254] [<ffffff800818acdc>] kfree+0x54/0x70
[    5.216260] [<ffffff80082e6944>] jffs2_free_full_dirent+0xc/0x18
[    5.216264] [<ffffff80082e9e94>] jffs2_do_clear_inode+0xbc/0x130
[    5.216268] [<ffffff80082f04e4>] jffs2_evict_inode+0x34/0x40
[    5.216273] [<ffffff80081b8224>] evict+0xa4/0x190
[    5.216276] [<ffffff80081b8628>] iput+0x148/0x1d0
[    5.216279] [<ffffff80081ba234>] iget_failed+0x24/0x30
[    5.216282] [<ffffff80082f0684>] jffs2_iget+0x194/0x320
[    5.216286] [<ffffff80082e503c>] jffs2_lookup+0x104/0x188
[    5.216291] [<ffffff80081a5758>] lookup_slow+0x90/0x160
[    5.216295] [<ffffff80081a88d4>] walk_component+0x1fc/0x308
[    5.216299] [<ffffff80081a8ca0>] link_path_walk+0x2c0/0x4e0
[    5.216303] [<ffffff80081a98ac>] path_openat+0x84/0x11f8
[    5.216307] [<ffffff80081abb08>] do_filp_open+0x60/0xd8
[    5.216310] [<ffffff80081a0b8c>] do_open_execat+0x5c/0x1c8
[    5.216314] [<ffffff80081a26e8>] do_execveat_common.isra.14+0x198/0x618
[    5.216318] [<ffffff80081a2b90>] do_execve+0x28/0x30
[    5.216323] [<ffffff8008083850>] run_init_process+0x28/0x38
[    5.216326] [<ffffff8008083874>] try_to_run_init_process+0x14/0x50
[    5.216332] [<ffffff8008a1c60c>] kernel_init+0xbc/0x100
[    5.216336] [<ffffff8008084a90>] ret_from_fork+0x10/0x18
[    5.216341] Code: 1ad22484 0b050084 1ad12484 b40005a9 (38264924)
[    5.216343] ---[ end trace 128c3b2905dc9d7e ]---
[    5.216368] Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
[    5.216368]
[    5.216372] SMP: stopping secondary CPUs
[    5.224555] Kernel Offset: disabled
[    5.224558] CPU features: 0x002004
[    5.224559] Memory Limit: none
[    5.884350] ---[ end Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
[    5.884350]

解决方案

Work-around for PetaLinux 2017.x:

The issue occurs because the JFFS2 filesystem is configured with an incorrect erase block size.

To work around this issue, create the JFFS2 filesystem with the steps below::

1) Install mkfs tools in your Linux PC and Create the JFFS2 filesystem manually using the following commands:

    $ mkfs.jffs2 --root=<filesystem directory> --eraseblock=<erase block size of your flash> -p -o rootfs.jffs2
    $ mkfs.jffs2 --root=<filesystem directory> --eraseblock=0x2000 -p -o rootfs.jffs2


Note: Block Erase size is 8KB(0x2000) for a Zynq UltraScale MPSoC evaluation board.

2) Build the PetaLinux project. 

Include the filesystem into the boot image using the petalinux-package command:

  $ petalinux-package --boot --u-boot --kernel image.ub --add rootfs.jffs2 --offset <offset-of-jffs2-partition>


Work-around for PetaLinux 2018.1/2:

This is a known issue in kernel 4.14 for DMA. 

You will need to turn off the DMA in the SPI controller by enabling the "has-io-mode" device-tree property in the QSPI's device-tree node.

Add the below content in <plnx-proj-root>/project-spec/meta-user/recipes-bsp/device-tree/files/system-user.dtsi

    &qspi {
        has-io-mode = <1>;
    };

附件

文件名 文件大小 File Type
2017-x-QSPI-boot-error-log.txt 73 KB TXT
2018-1-or-2-QSPI-boot-error-log.txt 41 KB TXT

链接问答记录

相关答复记录

Answer Number 问答标题 问题版本 已解决问题的版本
71439 2017.1-2018.2: Zynq UltraScale+ MPSoC: Linux kernel panic for JFFS2 filesystem on POR or reboot N/A N/A
AR# 71114
日期 08/21/2018
状态 Active
Type 综合文章
器件
Tools More Less
Boards & Kits
People Also Viewed