Re: [PATCH 00 of 10] x86: unify asm/pgtable.h

[Jeremy Fitzhardinge <jeremy@goop.org>]

Ingo Molnar wrote:
> right now i'm trying to bisect this crash on 64-bit:
>
>  Calling initcall 0xffffffff80395877: pci_init+0x0/0x2b()
>  ------------[ cut here ]------------
>  kernel BUG at arch/x86/mm/pageattr_64.c:176!
>  [...]
>  Call Trace:
>   [<ffffffff8021db68>] change_page_attr_addr+0x9e/0x119
>   [<ffffffff8021d44f>] ioremap_change_attr+0x49/0x58
>   [<ffffffff8021d626>] iounmap+0xbe/0xe0
>   [<ffffffff804dfc95>] quirk_usb_early_handoff+0x396/0x3a5
>   [<ffffffff80396ea0>] pci_fixup_device+0x90/0x9e
>   [<ffffffff80395889>] pci_init+0x12/0x2b
>   [<ffffffff808c9545>] kernel_init+0xd0/0x23c
>   [<ffffffff8020bec8>] child_rip+0xa/0x12
>   [<ffffffff808c9475>] kernel_init+0x0/0x23c
>   [<ffffffff8020bebe>] child_rip+0x0/0x12
>
> config and bootlog attached.
>
> it's one of these patches:
>
>  Subject: x86: move all asm/pgtable constants into one place
>  Subject: x86: avoid name conflict for Voyager leave_mm
>  Subject: x86/pgtable: unify pagetable accessors
>   

Hm, you got the same crash the with my last attempt at this.   I'll take 
a closer look at that first patch.  The global flag differences between 
32 and 64 bit was a bit weird.

    J