kexec: arm: possible overwrite of initrd

kexec: arm: possible overwrite of initrd

[Corentin Labbe]

Hello

Following https://lkml.org/lkml/2020/4/6/96 I was able to boot my Cubieboard4 via kexec reliabily.

But now I have started to use kernelCI builds, I got problems.
All sunxi_defconfig kernel works but not multi_v7_defconfig which got:
[    1.896540] Trying to unpack rootfs image as initramfs...
[    1.896947] rootfs image is not initramfs (invalid magic at start of compressed archive); looks like an initrd
Then:
[    3.927732] RAMDISK: Couldn't find valid RAM disk image starting at 0.
[    3.934489] VFS: Cannot open root device \"(null)\" or unknown-block(0,0): error -6

I have tryed to disable all related RD/RAMFS/compression CONFIGs without change.
Only the size of the kernel seems to matter which let me think that the initrd is overwritten by the kernel.

I use kexec-tools master
This is the output of my kexec run
run kexec with --debug --kexec-syscall --force --initrd /tmp/ramdisk --dtb /tmp/dtb --command-line='console=ttyS0,115200n8 root=/dev/ram0 earlycon=uart,mmio32,0x7000000 ip=dhcp'
Try gzip decompression.
kernel: 0xb65c0008 kernel_size: 0x853200
MEMORY RANGES
0000000020000000-000000009fffffff (0)
zImage header: 0x016f2818 0x00000000 0x00853200
zImage size 0x853200, file size 0x853200
zImage requires 0x00864200 bytes
  offset 0x0000b810 tag 0x5a534c4b size 8
Decompressed kernel sizes:
 text+data 0x01563f54 bss 0x0005ca84 total 0x015c09d8
Resulting kernel space: 0x01dc8154
Kernel: address=0x20008000 size=0x01dc8154
Initrd: address=0x21dd1000 size=0x01c64369
DT    : address=0x23a36000 size=0x000060bb
kexec_load: entry = 0x20008000 flags = 0x280000
nr_segments = 3
segment[0].buf   = 0xb65c0008
segment[0].bufsz = 0x853204
segment[0].mem   = 0x20008000
segment[0].memsz = 0x854000
segment[1].buf   = 0xb495b000
segment[1].bufsz = 0x1c64369
segment[1].mem   = 0x21dd1000
segment[1].memsz = 0x1c65000
segment[2].buf   = 0x4f030
segment[2].bufsz = 0x60bb
segment[2[   39.693411] sun7i-dwmac 830000.ethernet eth0: Link is Down
].mem   = 0x23a36000
segment[2].memsz = 0x7000
[   39.709586] kexec_core: Starting new kernel
[   40.120408] Bye!
[    0.000000] Booting Linux on physical CPU 0x0
[    0.000000] Linux version 5.6.11-dirty (compile@Red) (gcc version 9.2.0 (Gentoo 9.2.0-r2 p3)) #43 SMP Fri May 15 15:31:20 CEST 2020
[    0.000000] CPU: ARMv7 Processor [410fc075] revision 5 (ARMv7), cr=10c5387d
[    0.000000] CPU: div instructions available: patching division code
[    0.000000] CPU: PIPT / VIPT nonaliasing data cache, VIPT aliasing instruction cache
[    0.000000] OF: fdt: Machine model: Cubietech Cubieboard4
[    0.000000] Memory policy: Data cache writealloc
[    0.000000] efi: Getting EFI parameters from FDT:
[    0.000000] efi: UEFI not found.
[    0.000000] Ignoring RAM at 0x50000000-0xa0000000
[    0.000000] Consider using a HIGHMEM enabled kernel.
[    0.000000] cma: Reserved 64 MiB at 0x4c000000
[    0.000000] percpu: Embedded 20 pages/cpu s49228 r8192 d24500 u81920
[    0.000000] Built 1 zonelists, mobility grouping on.  Total pages: 195072
[    0.000000] Kernel command line: 'console=ttyS0,115200n8
[    0.000000] Dentry cache hash table entries: 131072 (order: 7, 524288 bytes, linear)
[    0.000000] Inode-cache hash table entries: 65536 (order: 6, 262144 bytes, linear)
[    0.000000] mem auto-init: stack:off, heap alloc:off, heap free:off
[    0.000000] Memory: 662712K/786432K available (12288K kernel code, 1455K rwdata, 4788K rodata, 2048K init, 370K bss, 58184K reserved, 65536K cma-reserved)
[    0.000000] SLUB: HWalign=64, Order=0-3, MinObjects=0, CPUs=8, Nodes=1
[    0.000000] rcu: Hierarchical RCU implementation.
[    0.000000] rcu: 	RCU event tracing is enabled.
[    0.000000] rcu: 	RCU restricting CPUs from NR_CPUS=16 to nr_cpu_ids=8.
[    0.000000] rcu: RCU calculated value of scheduler-enlistment delay is 10 jiffies.
[    0.000000] rcu: Adjusting geometry for rcu_fanout_leaf=16, nr_cpu_ids=8
[    0.000000] NR_IRQS: 16, nr_irqs: 16, preallocated irqs: 16
[    0.000000] random: get_random_bytes called from start_kernel+0x2fc/0x494 with crng_init=0
[    0.000000] arch_timer: cp15 timer(s) running at 24.00MHz (phys).
[    0.000000] clocksource: arch_sys_counter: mask: 0xffffffffffffff max_cycles: 0x588fe9dc0, max_idle_ns: 440795202592 ns
[    0.000006] sched_clock: 56 bits at 24MHz, resolution 41ns, wraps every 4398046511097ns
[    0.000018] Switching to timer-based delay loop, resolution 41ns
[    0.001476] clocksource: timer: mask: 0xffffffff max_cycles: 0xffffffff, max_idle_ns: 79635851949 ns
[    0.002630] Console: colour dummy device 80x30
[    0.002953] printk: console [tty0] enabled
[    0.002997] Calibrating delay loop (skipped), value calculated using timer frequency.. 48.00 BogoMIPS (lpj=240000)
[    0.003026] pid_max: default: 32768 minimum: 301
[    0.003201] Mount-cache hash table entries: 2048 (order: 1, 8192 bytes, linear)
[    0.003237] Mountpoint-cache hash table entries: 2048 (order: 1, 8192 bytes, linear)
[    0.003933] CPU: Testing write buffer coherency: ok
[    0.004409] CPU0: update cpu_capacity 523
[    0.004433] CPU0: thread -1, cpu 0, socket 0, mpidr 80000000
[    0.005032] Setting up static identity map for 0x20300000 - 0x203000ac
[    0.005475] ARM CCI driver probed
[    0.005783] sunxi multi cluster SMP support installed
[    0.006198] rcu: Hierarchical SRCU implementation.
[    0.008162] EFI services will not be available.
[    0.008500] smp: Bringing up secondary CPUs ...
[    0.009318] CPU1: update cpu_capacity 523
[    0.009324] CPU1: thread -1, cpu 1, socket 0, mpidr 80000001
[    0.010326] CPU2: update cpu_capacity 523
[    0.010333] CPU2: thread -1, cpu 2, socket 0, mpidr 80000002
[    0.011197] CPU3: update cpu_capacity 523
[    0.011203] CPU3: thread -1, cpu 3, socket 0, mpidr 80000003
[    0.012178] CPU4: update cpu_capacity 1538
[    0.012192] CPU4: thread -1, cpu 0, socket 1, mpidr 80000100
[    0.012208] CPU4: Spectre v2: using ICIALLU workaround
[    0.013422] CPU5: update cpu_capacity 1538
[    0.013435] CPU5: thread -1, cpu 1, socket 1, mpidr 80000101
[    0.013449] CPU5: Spectre v2: using ICIALLU workaround
[    0.014582] CPU6: update cpu_capacity 1538
[    0.014596] CPU6: thread -1, cpu 2, socket 1, mpidr 80000102
[    0.014611] CPU6: Spectre v2: using ICIALLU workaround
[    0.015757] CPU7: update cpu_capacity 1538
[    0.015771] CPU7: thread -1, cpu 3, socket 1, mpidr 80000103
[    0.015787] CPU7: Spectre v2: using ICIALLU workaround
[    0.016070] smp: Brought up 1 node, 8 CPUs
[    0.016110] SMP: Total of 8 processors activated (384.00 BogoMIPS).
[    0.016125] CPU: All CPU(s) started in SVC mode.
[    0.017054] devtmpfs: initialized
[    0.027537] clocksource: jiffies: mask: 0xffffffff max_cycles: 0xffffffff, max_idle_ns: 19112604462750000 ns
[    0.027597] futex hash table entries: 2048 (order: 5, 131072 bytes, linear)
[    0.029778] pinctrl core: initialized pinctrl subsystem
[    0.032100] thermal_sys: Registered thermal governor 'step_wise'
[    0.032374] DMI not present or invalid.
[    0.032934] NET: Registered protocol family 16
[    0.035272] DMA: preallocated 256 KiB pool for atomic coherent allocations
[    0.038735] cpuidle: using governor menu
[    0.038866] No ATAGs?
[    0.039281] hw-breakpoint: found 5 (+1 reserved) breakpoint and 4 watchpoint registers.
[    0.039335] hw-breakpoint: maximum watchpoint size is 8 bytes.
[    0.041953] Serial: AMBA PL011 UART driver
[    0.044873] sram 20000.sram: can't request region for resource [mem 0x00020000-0x0005ffff]
[    0.044925] sram 20000.sram: could not map SRAM registers
[    0.044976] sram: probe of 20000.sram failed with error -16
[    0.104030] AT91: Could not find identification node
[    0.105987] iommu: Default domain type: Translated 
[    0.106465] vgaarb: loaded
[    0.107861] SCSI subsystem initialized
[    0.108599] usbcore: registered new interface driver usbfs
[    0.108705] usbcore: registered new interface driver hub
[    0.108893] usbcore: registered new device driver usb
[    0.110629] pps_core: LinuxPPS API ver. 1 registered
[    0.110663] pps_core: Software ver. 5.3.6 - Copyright 2005-2007 Rodolfo Giometti <giometti@linux.it>
[    0.110732] PTP clock support registered
[    0.111020] EDAC MC: Ver: 3.0.0
[    0.114979] clocksource: Switched to clocksource arch_sys_counter
[    1.892786] NET: Registered protocol family 2
[    1.893828] tcp_listen_portaddr_hash hash table entries: 512 (order: 0, 6144 bytes, linear)
[    1.893912] TCP established hash table entries: 8192 (order: 3, 32768 bytes, linear)
[    1.894090] TCP bind hash table entries: 8192 (order: 4, 65536 bytes, linear)
[    1.894411] TCP: Hash tables configured (established 8192 bind 8192)
[    1.894622] UDP hash table entries: 512 (order: 2, 16384 bytes, linear)
[    1.894725] UDP-Lite hash table entries: 512 (order: 2, 16384 bytes, linear)
[    1.895155] NET: Registered protocol family 1
[    1.896009] RPC: Registered named UNIX socket transport module.
[    1.896047] RPC: Registered udp transport module.
[    1.896077] RPC: Registered tcp transport module.
[    1.896106] RPC: Registered tcp NFSv4.1 backchannel transport module.
[    1.896144] PCI: CLS 0 bytes, default 64
[    1.896540] Trying to unpack rootfs image as initramfs...
[    1.896947] rootfs image is not initramfs (invalid magic at start of compressed archive); looks like an initrd
[    2.040365] Freeing initrd memory: 29076K

Regards

Re: kexec: arm: possible overwrite of initrd

[Russell King - ARM Linux admin]

On Fri, May 15, 2020 at 03:57:12PM +0200, Corentin Labbe wrote:
> Hello
> 
> Following https://lkml.org/lkml/2020/4/6/96 I was able to boot my Cubieboard4 via kexec reliabily.

You can try increasing the kernel size that kexec thinks the kernel
needs, but it should be extremely accurate with modern kexec.

--image-size $((0x01dc8154 + 0x10000))

will add 64k on top of what you currently have.  Note where the first
figure comes from (you'll find it in the debug output, see
"Resulting kernel space").

The best I can say is try playing around with that - but, kexec's
calculations should be spot on to stop the booting kernel from
overwriting the initrd.

The only way to debug that is to get the booted kernel to hexdump the
initrd so it's possible to see what happened to it.

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line in suburbia: sync at 10.2Mbps down 587kbps up