Re: Panic on boot PPC64 / e1000 for 2.6.14-git10 & 2.6.14-mm1

[Benjamin Herrenschmidt <benh@kernel.crashing.org>]

On Mon, 2005-11-07 at 12:39 -0800, Andrew Morton wrote:
> "Martin J. Bligh" <mbligh@mbligh.org> wrote:

e1000 did a config space access which caused a hash miss (probably when
accessing the IO space) which blew up in the hashing
(.hash_page_do_lazy_icache called from __hash_page_4K).

I think the bug is a mistake on my side:

.hash_page_do_lazy_icache() used to begin with:

        if (!pfn_valid(pte_pfn(pte)))
                return pp;

And that got lost. It needs to go back in. I'll send a patch later today
(I'm at home without the necessary stuff at hand).

Ben.


> > Power 4 LPAR panics on boot. -git9 was fine.
> > 
> > Intel(R) PRO/1000 Network Driver - version 6.1.16-k2
> > Copyright (c) 1999-2005 Intel Corporation.
> > Oops: Kernel access of bad area, sig: 11 [#1]
> > SMP NR_CPUS=32 NUMA PSERIES LPAR 
> > Modules linked in:
> > NIP: C00000000002F4A4 LR: C00000000002F7C8 CTR: C00000000030FA08
> > REGS: c00000003ffcb0d0 TRAP: 0300   Not tainted  (2.6.14-git10-autokern1)
> > MSR: 8000000000001032 <ME,IR,DR>  CR: 48004088  XER: 20000000
> > DAR: C00000078002EEC8, DSISR: 0000000040010000
> > TASK = c00000077ffd07c0[1] 'swapper' THREAD: c00000003ffc8000 CPU: 1
> > GPR00: C00000000002F7C8 C00000003FFCB350 C0000000006518C0 00000000000001BC 
> > GPR04: 000000003FE48080 0000000000000300 C000000003DF4000 0000000000000300 
> > GPR08: C0000000006578A8 C00000077FFEF080 C000000000657928 000000000003FE48 
> > GPR12: C000000000650000 C000000000557800 0000000000000000 0000000000000000 
> > GPR16: 0000000000000000 0000000000000000 0000000000000000 0000000000000000 
> > GPR20: 0000000000230000 0000000003A10000 0000000000000060 0000000000000001 
> > GPR24: 0000000000000400 C00000003FFB5000 C000000003DF2830 C000000003DF2000 
> > GPR28: 0000000ADC7D4C2D 00000000000001BC 00007FC901000FB9 00007FC9010003B9 
> > NIP [C00000000002F4A4] .hash_page_do_lazy_icache+0x3c/0x108
> > LR [C00000000002F7C8] .__hash_page_4K+0x98/0xac
> > Call Trace:
> > [C00000003FFCB350] [C00000003FFCB540] 0xc00000003ffcb540 (unreliable)
> > [C00000003FFCB3E0] [C00000000002F7C8] .__hash_page_4K+0x98/0xac
> > [C00000003FFCB4E0] [C00000000002F444] .hash_page+0x158/0x17c
> > [C00000003FFCB560] [C000000000004734] .do_hash_page+0x34/0x40
> > --- Exception: 301 at .e1000_init_eeprom_params+0x40/0x314
> >     LR = .e1000_probe+0x550/0x9f0
> > [C00000003FFCB850] [C00000000028E078] .pci_bus_read_config_word+0x94/0xe0 (unreliable)
> > [C00000003FFCB900] [C0000000004FC9A4] .e1000_probe+0x550/0x9f0
> > [C00000003FFCB9D0] [C000000000291494] .pci_device_probe+0x11c/0x16c
> > [C00000003FFCBA70] [C0000000002DEE50] .driver_probe_device+0x80/0x160
> > [C00000003FFCBB00] [C0000000002DEFD8] .__driver_attach+0xa8/0xc4
> > [C00000003FFCBB90] [C0000000002DE50C] .bus_for_each_dev+0x7c/0xd0
> > [C00000003FFCBC40] [C0000000002DEC08] .driver_attach+0x28/0x40
> > [C00000003FFCBCC0] [C0000000002DDCAC] .bus_add_driver+0xd4/0x1ec
> > [C00000003FFCBD70] [C0000000002DF0D4] .driver_register+0x54/0x70
> > [C00000003FFCBE00] [C000000000291054] .pci_register_driver+0x80/0xc4
> > [C00000003FFCBE80] [C0000000004FC43C] .e1000_init_module+0x40/0x58
> > [C00000003FFCBF00] [C000000000009584] .init+0x1e8/0x44c
> > [C00000003FFCBF90] [C00000000000A358] .kernel_thread+0x4c/0x68
> > Instruction dump:
> > fb81ffe0 fba1ffe8 fbe1fff8 e902aef0 7c7d1b78 fbc1fff0 788ba302 f8010010 
> > f821ff71 60000000 60000000 e92a0000 <7c0958ae> 78001f24 7d28002a 38000038 
> >  <0>Kernel panic - not syncing: Attempted to kill init!
> 
> I'm going to need help understanding that trace.  Was the crash in
> e1000_init_eeprom_params() or in hash_page_do_lazy_icache()?  The former, I
> think?
> 
> git-netdev-all.patch touches e1000, but I don't see any substantial changes
> in there.