kexec on kirkwood kernels?

Has anyone successfully used kexec to boot one kirkwood kernel from another? I've been trying to do it on a Seagate DockStar, but all I see is this:

    # kexec -l --append="console=ttyS0,115200" zImage
    # kexec -e
    [   35.553632] Starting new kernel
    Uncompressing Linux... done, booting the kernel.

and then nothing. I've tried with both 2.6.35 and 2.6.38-rc1 kernels. I've tried modifying the kernel image to set the machine ID explicitly, but that makes no difference. I wonder if the boot parameters are being passed correctly, but I don't know how to check without any console output. I do have a JTAG setup, but so far I've only used it cookbook-fashion to re-flash the u-boot NAND partition. I'm hoping someone can help me avoid having to use gdb over JTAG to resolve this. Thanks.

Eric Cooper             e c c @ c m u . e d u

Hallo,

Try enabling DEBUG_LL and EARLY_PRINTK in the kernel to boot and pass

earlyprintk

as kernel parameter.

Best regards
Uwe

Pengutronix e.K.                           | Uwe Kleine-K?nig            |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |

Thanks, that helped quite a bit. After inserting various print statements, I found that the system is hanging in the readl call in orion_pcie_dev_id. The call chain is:

start_kernel
time_init
kirkwood_timer_init
kirkwood_find_tclk
kirkwood_pcie_id
orion_pcie_dev_id
readl

The readl is attempting to access the address
KIRKWOOD_REGS_VIRT_BASE + 0x40000.

This same code gets executed successfully when the kernel is started from u-boot; I'm not sure what's different in the post-kexec environment. Any suggestions on what to try next would be appreciated.

Eric Cooper             e c c @ c m u . e d u

This worked ok on dove processor with several fixes. What is the output of the failed bootup? If kirkwood comes with L2, please check L2 is flushed and cache disabled before booting into the 2nd kernel.

- Eric

# kexec -e
[   57.233847] Starting new kernel
Uncompressing Linux... done, booting the kernel.
[    0.000000] Linux version 2.6.38-rc1+ (ecc at stratocaster) (gcc version 4.4.1 (Sourcery G++ Lite 2010q1-188) ) #5 Sat Jan 29 14:52:49 EST 2011
[    0.000000] CPU: Feroceon 88FR131 [56251311] revision 1 (ARMv5TE), cr=00053177
[    0.000000] CPU: VIVT data cache, VIVT instruction cache
[    0.000000] Machine: Seagate FreeAgent DockStar
[    0.000000] bootconsole [earlycon0] enabled
[    0.000000] Memory policy: ECC disabled, Data cache writeback
[    0.000000] Built 1 zonelists in Zone order, mobility grouping on.  Total pages: 32512
[    0.000000] Kernel command line: earlyprintk console=ttyS0,115200
[    0.000000] PID hash table entries: 512 (order: -1, 2048 bytes)
[    0.000000] Dentry cache hash table entries: 16384 (order: 4, 65536 bytes)
[    0.000000] Inode-cache hash table entries: 8192 (order: 3, 32768 bytes)
[    0.000000] Memory: 128MB = 128MB total
[    0.000000] Memory: 126836k/126836k available, 4236k reserved, 0K highmem
[    0.000000] Virtual kernel memory layout:
[    0.000000]     vector  : 0xffff0000 - 0xffff1000   (   4 kB)
[    0.000000]     fixmap  : 0xfff00000 - 0xfffe0000   ( 896 kB)
[    0.000000]     DMA     : 0xffc00000 - 0xffe00000   (   2 MB)
[    0.000000]     vmalloc : 0xc8800000 - 0xfe800000   ( 864 MB)
[    0.000000]     lowmem  : 0xc0000000 - 0xc8000000   ( 128 MB)
[    0.000000]     modules : 0xbf000000 - 0xc0000000   (  16 MB)
[    0.000000]       .init : 0xc0008000 - 0xc0023000   ( 108 kB)
[    0.000000]       .text : 0xc0023000 - 0xc02b4518   (2630 kB)
[    0.000000]       .data : 0xc02b6000 - 0xc02d32e0   ( 117 kB)
[    0.000000] SLUB: Genslabs=13, HWalign=32, Order=0-3, MinObjects=0, CPUs=1, Nodes=1
[    0.000000] NR_IRQS:114

This output is identical to a correctly booting system as far as it goes, but the system is hanging in a read to the Kirkwood register virtual address space in "kirkwood_pcie_id". If I comment this operation out, the system hangs the next time it reads another Kirkwood register, and so on.

Clearly something isn't in the same state after kexec that it was after a hardware reset. I'll try looking into the L2 cache issue, thanks.

Eric Cooper             e c c @ c m u . e d u