BeagleBone Black kernel panic

[Hartley Sweeten]

Hello,

Lately I have been seeing a LOT of kernel panics with brand new BeagleBone Black boards. All of them show this panic when first powered on:

[    0.997639] Unhandled fault: external abort on non-linefetch (0x1008) at 0xe085c000

[    1.005624] Internal error: : 1008 [#1] SMP THUMB2

[    1.010615] Modules linked in:

[    1.013803] CPU: 0    Not tainted  (3.8.13-bone84 #1)

[    1.019074] PC is at cpsw_probe+0x348/0x960

[    1.023444] LR is at ioremap_page_range+0x95/0xf8

[    1.028345] pc : [<c02a58e8>]    lr : [<c01d5d55>]    psr: a0000033

[    1.028345] sp : df071e18  ip : 00000000  fp : df0d5400

[    1.040311] r10: c07e2acc  r9 : de2d1000  r8 : de2d1590

[    1.045755] r7 : de2d1000  r6 : de2d1540  r5 : 00000000  r4 : e085c000

[    1.052554] r3 : 80000000  r2 : 00000000  r1 : e085d000  r0 : e085c000

[    1.059357] Flags: NzCv  IRQs on  FIQs on  Mode SVC_32  ISA Thumb  Segment kernel

[    1.067146] Control: 50c5387d  Table: 9e2e4019  DAC: 00000015

[    1.073143] Process swapper/0 (pid: 1, stack limit = 0xdf070240)

[    1.079400] Stack: (0xdf071e18 to 0xdf072000)

[    1.083938] 1e00:                                                       00000000 c00b6a49

[    1.092451] 1e20: c082fbb8 c082fbb8 de2d1540 de2d1298 df0d5410 de5fac08 00000000 df071e90

[    1.100970] 1e40: df071e90 c00ff173 22222222 22222222 00000000 c082fbb8 de5fac08 de2db7c0

[    1.109486] 1e60: de5fac08 c00ff0af de5fac08 00000000 df071e90 de5fac88 df0d9c48 c00ff927

[    1.118012] 1e80: df0494b8 c0049657 00000000 df0d5444 00000020 00000008 df0d5410 c08de0ec

[    1.126531] 1ea0: df0d5410 c084d9d0 00000000 c07c4e55 00000102 c07e2acc 00000000 c02484c9

[    1.135057] 1ec0: c02484b9 c024796b 00000000 df0d5410 c084d9d0 df0d5444 00000000 c0247acb

[    1.143581] 1ee0: c084d9d0 c0247a81 00000000 c0246a9f df049478 df0c6180 c084d9d0 c0842e20

[    1.152103] 1f00: de2db7c0 c024745f c06f3f5c c084d9d0 c084d9d0 df070000 c086ecc0 00000000

[    1.160622] 1f20: c07c4e55 c0247dcd 00000000 c07cf748 df070000 c086ecc0 00000000 c07c4e55

[    1.169146] 1f40: 00000102 c000867f 00000007 00000007 c0827c98 c07cf744 c07cf748 00000007

[    1.177668] 1f60: c07cf728 c086ecc0 c07a91c9 c07e2acc 00000000 c07a96a3 00000007 00000007

[    1.186183] 1f80: c07a91c9 c0cfbfc0 00000000 c044d1c1 00000000 00000000 00000000 00000000

[    1.194710] 1fa0: 00000000 c044d1c7 00000000 c000c8fd 00000000 00000000 00000000 00000000

[    1.203228] 1fc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000

[    1.211755] 1fe0: 00000000 00000000 00000000 00000000 00000013 00000000 00000000 00000000

[    1.220302] [<c02a58e8>] (cpsw_probe+0x348/0x960) from [<c02484c9>] (platform_drv_probe+0x11/0x14)

[    1.229645] [<c02484c9>] (platform_drv_probe+0x11/0x14) from [<c024796b>] (driver_probe_device+0x53/0x168)

[    1.239711] [<c024796b>] (driver_probe_device+0x53/0x168) from [<c0247acb>] (__driver_attach+0x4b/0x4c)

[    1.249495] [<c0247acb>] (__driver_attach+0x4b/0x4c) from [<c0246a9f>] (bus_for_each_dev+0x27/0x48)

[    1.258919] [<c0246a9f>] (bus_for_each_dev+0x27/0x48) from [<c024745f>] (bus_add_driver+0xe3/0x168)

[    1.268356] [<c024745f>] (bus_add_driver+0xe3/0x168) from [<c0247dcd>] (driver_register+0x3d/0xc4)

[    1.277692] [<c0247dcd>] (driver_register+0x3d/0xc4) from [<c000867f>] (do_one_initcall+0x1f/0xf4)

[    1.287045] [<c000867f>] (do_one_initcall+0x1f/0xf4) from [<c07a96a3>] (kernel_init_freeable+0xc3/0x158)

[    1.296933] [<c07a96a3>] (kernel_init_freeable+0xc3/0x158) from [<c044d1c7>] (kernel_init+0x7/0x98)

[    1.306363] [<c044d1c7>] (kernel_init+0x7/0x98) from [<c000c8fd>] (ret_from_fork+0x11/0x34)

[    1.315071] Code: f000 81d2 f8c7 05c0 (6803) f44f

[    1.320069] ---[ end trace d14620964e853f1d ]---

[    1.325064] Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b

They seem to work fine initially then this starts happening. Once it starts the boards seem to only boot about 1% of the time.

I have verified that the SD-Card image I am using works fine on some older BeagleBone Black boards.

All of the failing boards appear to have a serial number date code of 1826, 1827, or 1828 (i.e. 000C1828BBBG2011).

I have tried searching for a solution, or explanation, to the problem and it appears to be related to the PHY not resetting correctly but all the info I found is quite old.

Is this a known problem? Is there a real solution? Is this just a problem with a batch of boards?

I'm really stuck with this, any help would be appreciated.

Thanks,

Hartley

[Robert Nelson]

What driver/interface/feature is causing you to be stuck on 3.8.13-bone84?

Regards,

--
Robert Nelson
https://rcn-ee.com/