Re: [Xen-ia64-devel][PATCH][VTD] add hypercalls needed for VTD

[Isaku Yamahata <yamahata@xxxxxxxxxxxxx>]

Applied, thanks.

On Wed, Oct 22, 2008 at 03:12:29PM +0800, Xu, Anthony wrote:
> The new one,
> I applies manually #Cset 18649 -- 18653 in xen-devel tree on #Cset 18637.
> Tested by booting rhel5.1 guest.
> 
> Please pull xen-devel tree and apply this patch.
> I'll send out other patches soon.
> 
> Thanks
> Anthony
> 
> 
> 
> 
> 
> Isaku Yamahata wrote:
> > Hi Anthony.
> > With the following 3 patches applied, xen panics.
> >
> > 0 A hypercall_for_vtd2.patch
> > 1 A ia64-px-xen-setpxinfo-update.patch
> > 2 A add_head_file_for_VTD2.patch
> >
> > dom0->arch.pdev_list doesn't seem to be initialized.
> > diversity_vtd.patch has the corresponding line, but the hunk should
> > be included in add_head_file_for_VTD2.patch. The patch should be
> > complete itself.
> >
> > ip corresponds to the line
> > int pci_add_device(u8 bus, u8 devfn)
> > ...
> >         list_add(&pdev->domain_list, &dom0->arch.pdev_list);
> >
> >
> > Please test patches before sending.
> >
> > thanks,
> >
> > (XEN) *** xen_handle_domain_access: exception table lookup failed,
> > iip=0xf40000000406e340, addr=0x8, spinning... (XEN) $$$$$ PANIC in
> > domain 0 (k6=0xf0000002f65d0000): *** xen_handle_domain_access:
> > exception table lookup failed, iip=0xf40000000406e340, addr=0x8,
> > spinning... (XEN) d 0xf000000004770080 domid 0 (XEN) vcpu
> > 0xf0000002f65d0000 vcpu 0 (XEN) (XEN) CPU 5
> > (XEN) psr : 0000121008226018 ifs : 800000000000050e ip  :
> > [<f40000000406e341>] (XEN) ip is at pci_add_device+0x1b1/0x250
> > (XEN) unat: 0000000000000000 pfs : 000000000000050e rsc :
> > 0000000000000003 (XEN) rnat: 0000000000000000 bsps: 0000000000000000
> > pr  : 000000000000a559 (XEN) ldrs: 0000000000000000 ccv :
> > 0000000000000000 fpsr: 0009804c0270033f (XEN) csd : 0000000000000000
> > ssd : 0000000000000000 (XEN) b0  : f40000000406e2e0 b6  :
> > f400000004099eb0 b7  : f400000004002e30 (XEN) f6  :
> > 000000000000000000000 f7  : 0ffe3d7c0000000000000 (XEN) f8  :
> > 1000bd7c0000000000000 f9  : 10000c000000000000000 (XEN) f10 :
> > 1000a8fd5555554c58000 f11 : 1003e00000000000008fd (XEN) r1  :
> > f4000000043f89f0 r2  : f0000002f65d7dd0 r3  : 000000000000050e (XEN)
> > r8  : f00000000436be30 r9  : 0000000000000000 r10 : 2000000000000000
> > (XEN) r11 : 0000000000000008 r12 : f0000002f65d7dd0 r13 :
> > f0000002f65d0000 (XEN) r14 : f000000004770740 r15 : 0000000000000000
> > r16 : 0000000000000008 (XEN) r17 : f00000000436be38 r18 :
> > f00000000436be50 r19 : f00000000436be70 (XEN) r20 : f00000000436be60
> > r21 : f00000000436be48 r22 : f00000000436be68 (XEN) r23 :
> > 0000000000000002 r24 : 00000000000000e8 r25 : 0000000000000000 (XEN)
> > r26 : 00000000000c0300 r27 : 0000000000000000 r28 : 0000000000000000
> > (XEN) r29 : 0000000080000000 r30 : 0000000000000000 r31 :
> > f00000000436be60 (XEN) (XEN) Call Trace: (XEN)  [<f4000000040fb4c0>]
> > show_stack+0x90/0xb0 (XEN)
> > sp=f0000002f65d7880 bsp=f0000002f65d18d8 (XEN)  [<f4000000040fbf80>]
> > show_registers+0xaa0/0xac0 (XEN)
> > sp=f0000002f65d7a50 bsp=f0000002f65d1898 (XEN)  [<f4000000040c1440>]
> > panic_domain+0x120/0x170 (XEN)
> > sp=f0000002f65d7a50 bsp=f0000002f65d1828 (XEN)  [<f4000000040b2220>]
> > ia64_do_page_fault+0x710/0x720 (XEN)
> > sp=f0000002f65d7b90 bsp=f0000002f65d1788 (XEN)  [<f4000000040f2ca0>]
> > ia64_leave_kernel+0x0/0x300 (XEN)
> > sp=f0000002f65d7bd0 bsp=f0000002f65d1788 (XEN)  [<f40000000406e340>]
> > pci_add_device+0x1b0/0x250 (XEN)
> > sp=f0000002f65d7dd0 bsp=f0000002f65d1718 (XEN)  [<f400000004099f40>]
> > do_physdev_op+0x5b0/0x8e0 (XEN)
> > sp=f0000002f65d7dd0 bsp=f0000002f65d16e8 (XEN)  [<f400000004002e60>]
> > fast_hypercall+0x170/0x310 (XEN)
> > sp=f0000002f65d7e00 bsp=f0000002f65d16e8 (XEN) domain_crash_sync
> > called from xenmisc.c:141 (XEN) Domain 0 (vcpu#0) crashed on cpu#5:
> > (XEN) d 0xf000000004770080 domid 0
> > (XEN) vcpu 0xf0000002f65d0000 vcpu 0
> > (XEN)
> >
> >
> > On Tue, Oct 21, 2008 at 06:36:35PM +0800, Xu, Anthony wrote:
> >> Updated one per your comments
> >>
> >> Thanks,
> >> Anthony
> >>
> >> Isaku Yamahata wrote:
> >>> On Tue, Oct 21, 2008 at 05:05:51PM +0800, Xu, Anthony wrote:
> >>>> Add hypercalls needed for VTD
> >>>>
> >>>> Signed-off-by; Anthony Xu < anthony.xu@xxxxxxxxx >
> >>>
> >>>
> >>>> diff -r 3b1ff707d130 -r 55251ad3e262 xen/arch/ia64/xen/hypercall.c
> >>>> --- a/xen/arch/ia64/xen/hypercall.c   Tue Oct 21 10:41:54 2008
> >>>> +0800 +++ b/xen/arch/ia64/xen/hypercall.c   Tue Oct 21 14:12:29
> >>>>  2008 +0800 @@ -35,6 +35,7 @@ #include
> >>>>  <public/arch-ia64/debug_op.h>  #include <asm/sioemu.h> #include
> >>>> <public/arch-ia64/sioemu.h> +#include <xen/pci.h>
> >>>>
> >>>>  static IA64FAULT
> >>>>  xen_hypercall (struct pt_regs *regs)
> >>>> @@ -313,6 +314,22 @@
> >>>>  iosapic_guest_write(
> >>>>      unsigned long physbase, unsigned int reg, u32 pval);
> >>>>
> >>>> +
> >>>> +/*
> >>>> + * XXX We don't support MSI for PCI passthrough, so just return
> >>>> success + */ +static int physdev_map_pirq(struct physdev_map_pirq
> >>>> *map) +{ +     return 0; +}
> >>>> +
> >>>> +static int physdev_unmap_pirq(struct physdev_unmap_pirq *unmap)
> >>>> +{ +     return 0; +}
> >>>> +
> >>>> +
> >>>> +
> >>>>  long do_physdev_op(int cmd, XEN_GUEST_HANDLE(void) arg)  {
> >>>>      int irq;
> >>>
> >>> -ENOSYS?
> >>> Or is other code confused with -ENOSYS?
> >>>
> >>>
> >>>> diff -r 3b1ff707d130 -r 55251ad3e262 xen/arch/ia64/xen/mm.c
> >>>> --- a/xen/arch/ia64/xen/mm.c  Tue Oct 21 10:41:54 2008 +0800
> >>>> +++ b/xen/arch/ia64/xen/mm.c  Tue Oct 21 14:12:29 2008 +0800 @@
> >>>>          -1456,12 +1456,38 @@ BUG_ON(mfn != pte_pfn(ret_pte));
> >>>> }
> >>>>
> >>>> +    perfc_incr(zap_domain_page_one);
> >>>> +    if(!mfn_valid(mfn))
> >>>> +        return;
> >>>> +
> >>>>      page = mfn_to_page(mfn);
> >>>>      BUG_ON((page->count_info & PGC_count_mask) == 0);
> >>>>
> >>>>      BUG_ON(clear_PGC_allocate && (page_get_owner(page) == NULL));
> >>>>      domain_put_page(d, mpaddr, pte, old_pte, clear_PGC_allocate);
> >>>> -    perfc_incr(zap_domain_page_one);
> >>>> +}
> >>>
> >>> The BUG_ON() above this hunk should be updated too. something like
> >>> -        BUG_ON(page_get_owner(mfn_to_page(mfn)) != d &&
> >>> -               page_get_owner(mfn_to_page(mfn)) != NULL);
> >>> +        BUG_ON(mfn_valid(mfn) &&
> >>> +               page_get_owner(mfn_to_page(mfn)) != d &&
> >>> +               page_get_owner(mfn_to_page(mfn)) != NULL);
> >>>
> >>>
> >>> thanks,
> >
> >
> >> _______________________________________________
> >> Xen-ia64-devel mailing list
> >> Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
> >> http://lists.xensource.com/xen-ia64-devel


> _______________________________________________
> Xen-ia64-devel mailing list
> Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
> http://lists.xensource.com/xen-ia64-devel

-- 
yamahata

_______________________________________________
Xen-ia64-devel mailing list
Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-ia64-devel