Re: linux-next: scsi tree boot filure

[James Bottomley]

linux-scsi added to cc

On Sun, 2009-09-27 at 16:43 +1000, Stephen Rothwell wrote:
> Hi James,
> 
> next-20090926 does not boot on some of my PowerPC partitions:
> 
> calling  .ibmvscsi_module_init+0x0/0xb8 @ 1
> ibmvscsi 30000028: SRP_VERSION: 16.a
> scsi0 : IBM POWER Virtual SCSI Adapter 1.5.8
> ibmvscsi 30000028: partner initialization complete
> ibmvscsi 30000028: host srp version: 16.a, host partition 1-Didgo-VIOS (1), OS 3, max io 1048576
> ibmvscsi 30000028: Client reserve enabled
> ibmvscsi 30000028: sent SRP login
> ibmvscsi 30000028: SRP_LOGIN succeeded
> Unable to handle kernel paging request for data at address 0x00000058
> Faulting instruction address: 0xc0000000003a6280
> Oops: Kernel access of bad area, sig: 11 [#1]
> SMP NR_CPUS=128 NUMA pSeries
> Modules linked in:
> NIP: c0000000003a6280 LR: c0000000003a63b4 CTR: 0000000000000000
> REGS: c00000007c3f3020 TRAP: 0300   Not tainted  (2.6.31-autokern1)
> MSR: 8000000000009032 <EE,ME,IR,DR>  CR: 24002042  XER: 00000001
> DAR: 0000000000000058, DSISR: 0000000040000000
> TASK = c00000007c3e8000[1] 'swapper' THREAD: c00000007c3f0000 CPU: 3
> GPR00: 0000000000000000 c00000007c3f32a0 c000000000bc5390 c000000000a76420 
> GPR04: c000000000b97818 c0000000015abc70 0000000000000000 c00000007c81c918 
> GPR08: c00000007c81c888 0000000002000000 0000000000000002 c0000000014ecbcc 
> GPR12: 0000000024000042 c000000000c1ea80 0000000003500000 c00000000074af10 
> GPR16: c000000000749588 0000000000000000 0000000000000000 0000000000000000 
> GPR20: c00000007c3f3600 c000000079074c00 c00000007c81c000 0000000002f1f8e0 
> GPR24: 0000000000000000 0000000000000000 0000000000000000 c000000079074c28 
> GPR28: c00000007c81c000 0000000000000000 c000000000b353f0 c000000000b97818 
> NIP [c0000000003a6280] .__scsi_alloc_queue+0x2c/0x13c
> LR [c0000000003a63b4] .scsi_alloc_queue+0x24/0x84
> Call Trace:
> [c00000007c3f32a0] [c00000007c3f3330] 0xc00000007c3f3330 (unreliable)
> [c00000007c3f3330] [c0000000003a63b4] .scsi_alloc_queue+0x24/0x84
> [c00000007c3f33b0] [c0000000003a8f78] .scsi_alloc_sdev+0x198/0x2ac
> [c00000007c3f3470] [c0000000003a9450] .scsi_probe_and_add_lun+0x130/0xaac
> [c00000007c3f3580] [c0000000003aa20c] .__scsi_scan_target+0xf4/0x5fc
> [c00000007c3f36a0] [c0000000003aa768] .scsi_scan_channel+0x54/0xd0
> [c00000007c3f3740] [c0000000003aa8b0] .scsi_scan_host_selected+0xcc/0x144
> [c00000007c3f37f0] [c0000000003d5264] .ibmvscsi_probe+0x590/0x6e4
> [c00000007c3f38c0] [c000000000021e88] .vio_bus_probe+0x84/0xb0
> [c00000007c3f3960] [c00000000037cbac] .driver_probe_device+0xfc/0x1c0
> [c00000007c3f39f0] [c00000000037cd04] .__driver_attach+0x94/0xd8
> [c00000007c3f3a80] [c00000000037b9f8] .bus_for_each_dev+0x84/0xdc
> [c00000007c3f3b30] [c00000000037c954] .driver_attach+0x28/0x40
> [c00000007c3f3bb0] [c00000000037c290] .bus_add_driver+0x148/0x314
> [c00000007c3f3c60] [c00000000037d1b0] .driver_register+0xd4/0x1a8
> [c00000007c3f3d10] [c000000000021cbc] .vio_register_driver+0x40/0x5c
> [c00000007c3f3da0] [c00000000084f418] .ibmvscsi_module_init+0x80/0xb8
> [c00000007c3f3e30] [c0000000000094c8] .do_one_initcall+0x9c/0x1cc
> [c00000007c3f3ee0] [c000000000822cc0] .kernel_init+0x21c/0x298
> [c00000007c3f3f90] [c000000000026cb8] .kernel_thread+0x54/0x70
> Instruction dump:
> 4e800020 7c0802a6 fb81ffe0 fbe1fff8 fba1ffe8 7c7c1b78 f8010010 f821ff71 
> 7c9f2378 eba302a0 48000008 ebbd0000 <e81d0058> 7fa3eb78 2fa00000 419efff0 
> ---[ end trace 18604a042ee6e0ba ]---
> Kernel panic - not syncing: Attempted to kill init!
> Call Trace:
> [c00000007c3f2c80] [c00000000001024c] .show_stack+0x70/0x184 (unreliable)
> [c00000007c3f2d30] [c00000000006a410] .panic+0x80/0x1b4
> [c00000007c3f2dd0] [c00000000006eca4] .do_exit+0x84/0x728
> [c00000007c3f2e90] [c000000000024d2c] .die+0x24c/0x27c
> [c00000007c3f2f30] [c0000000000330c8] .bad_page_fault+0xb8/0xd4
> [c00000007c3f2fb0] [c0000000000051dc] handle_page_fault+0x3c/0x74
> --- Exception: 300 at .__scsi_alloc_queue+0x2c/0x13c
>     LR = .scsi_alloc_queue+0x24/0x84
> [c00000007c3f32a0] [c00000007c3f3330] 0xc00000007c3f3330 (unreliable)
> [c00000007c3f3330] [c0000000003a63b4] .scsi_alloc_queue+0x24/0x84
> [c00000007c3f33b0] [c0000000003a8f78] .scsi_alloc_sdev+0x198/0x2ac
> [c00000007c3f3470] [c0000000003a9450] .scsi_probe_and_add_lun+0x130/0xaac
> [c00000007c3f3580] [c0000000003aa20c] .__scsi_scan_target+0xf4/0x5fc
> [c00000007c3f36a0] [c0000000003aa768] .scsi_scan_channel+0x54/0xd0
> [c00000007c3f3740] [c0000000003aa8b0] .scsi_scan_host_selected+0xcc/0x144
> [c00000007c3f37f0] [c0000000003d5264] .ibmvscsi_probe+0x590/0x6e4
> [c00000007c3f38c0] [c000000000021e88] .vio_bus_probe+0x84/0xb0
> [c00000007c3f3960] [c00000000037cbac] .driver_probe_device+0xfc/0x1c0
> [c00000007c3f39f0] [c00000000037cd04] .__driver_attach+0x94/0xd8
> [c00000007c3f3a80] [c00000000037b9f8] .bus_for_each_dev+0x84/0xdc
> [c00000007c3f3b30] [c00000000037c954] .driver_attach+0x28/0x40
> [c00000007c3f3bb0] [c00000000037c290] .bus_add_driver+0x148/0x314
> [c00000007c3f3c60] [c00000000037d1b0] .driver_register+0xd4/0x1a8
> [c00000007c3f3d10] [c000000000021cbc] .vio_register_driver+0x40/0x5c
> [c00000007c3f3da0] [c00000000084f418] .ibmvscsi_module_init+0x80/0xb8
> [c00000007c3f3e30] [c0000000000094c8] .do_one_initcall+0x9c/0x1cc
> [c00000007c3f3ee0] [c000000000822cc0] .kernel_init+0x21c/0x298
> [c00000007c3f3f90] [c000000000026cb8] .kernel_thread+0x54/0x70
> Rebooting in 180 seconds..
> 
> I have bisected this down to commit
> 4acd10521ee002137b5d6791e234d7110033c782 ("[SCSI] scsi_lib_dma.c : fix
> bug /w dma maps on virtual vc ports") which was added between
> next-20090925 and next-20090926.
> 
> Reverting that single commit from next-20090926 allows it to boot.

OK, so my strongest suspicion is that the SCSI device is parented to
some IBM specific device that has no type.  This is causing SCSI to
wander up the tree until it hits a NULL device and panics on the deref.

Does this incremental diff fix it?

James

---

diff --git a/include/scsi/scsi_host.h b/include/scsi/scsi_host.h
index 2977806..9d5bfdc 100644
--- a/include/scsi/scsi_host.h
+++ b/include/scsi/scsi_host.h
@@ -718,7 +718,7 @@ static inline struct Scsi_Host *dev_to_shost(struct device *dev)
  */
 static inline struct device *dev_to_nonscsi_dev(struct device *dev)
 {
-	while (dev->type == NULL || scsi_is_host_device(dev))
+	while (dev->parent && (dev->type == NULL || scsi_is_host_device(dev)))
 		dev = dev->parent;
 	return dev;
 }


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/