Re: Re: Panic on cpu0

["Jan Beulich" <JBeulich@novell.com>]

>>> On 16.12.10 at 10:30, ZhouPeng <zpengxen@gmail.com> wrote:
> The err msg appears in
> void *map_domain_page(unsigned long mfn)
>       BUG_ON(idx >= MAPCACHE_ENTRIES);
> 
> printk idx is 1024 and BUG_ON is called.
> 
> I print  Xen call trace msg:
> 
> [<ffcd3c2>]  map_domain_page+0x1a2/0x230
> [<ff1cd718>] unmap_domain_page+0x148/0x160
>                    addr_to_dma_page_maddr+0x1f6/0x250
>                    __find_next_zero_bit+0x80/0x90
>                    intel_iommu_map_page+0x14a/0x250
>                    domain_context_mapping+0x14a/0x250
>                    rmrr_identity_mapping+0xb7/0x120
>                    intel_iommu_dom0_init+0x143/0x1eb
>                    construct_dom0+0x138f/0x16b0
>                    __start_xen+0x109a/0x1680
>                    e820nf+0x0/0x4
>                    __high_start+0x58/0x5a

Could you give the below patch a try?

Jan

--- a/xen/drivers/passthrough/vtd/iommu.c
+++ b/xen/drivers/passthrough/vtd/iommu.c
@@ -1331,6 +1331,7 @@ int domain_context_mapping_one(
     if ( context_set_domain_id(context, domain, iommu) )
     {
         spin_unlock(&iommu->lock);
+        unmap_vtd_domain_page(context_entries);
         return -EFAULT;
     }
 
@@ -1672,6 +1673,7 @@ static int intel_iommu_map_page(
     if ( old.val == new.val )
     {
         spin_unlock(&hd->mapping_lock);
+        unmap_vtd_domain_page(page);
         return 0;
     }
     *pte = new;