From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Chen, Tiejun" Subject: Re: [v2][PATCH 1/1] xen:vtd: missing RMRR mapping while share EPT Date: Thu, 24 Jul 2014 17:03:27 +0800 Message-ID: <53D0CBDF.4010606@intel.com> References: <1406191841-5258-1-git-send-email-tiejun.chen@intel.com> <53D0CA69.6090500@citrix.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; Format="flowed" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <53D0CA69.6090500@citrix.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xen.org Errors-To: xen-devel-bounces@lists.xen.org To: Andrew Cooper , JBeulich@suse.com, yang.z.zhang@intel.com, kevin.tian@intel.com Cc: xen-devel@lists.xen.org List-Id: xen-devel@lists.xenproject.org On 2014/7/24 16:57, Andrew Cooper wrote: > On 24/07/14 09:50, Tiejun Chen wrote: >> intel_iommu_map_page() does nothing if VT-d shares EPT page table. >> So rmrr_identity_mapping() never create RMRR mapping but in some >> cases like some GFX drivers it still need to access RMRR. >> >> Here we will create those RMRR mappings even in shared EPT case. >> >> Signed-off-by: Tiejun Chen >> --- >> xen/drivers/passthrough/vtd/iommu.c | 18 ++++++++++++++++-- >> 1 file changed, 16 insertions(+), 2 deletions(-) >> >> v2: >> >> * Fix coding style. >> * Still need to abide intel_iommu_map_page(), so we should do nothing >> if dom0 and iommu supports pass thru. >> >> diff --git a/xen/drivers/passthrough/vtd/iommu.c b/xen/drivers/passthrough/vtd/iommu.c >> index 042b882..aca79db 100644 >> --- a/xen/drivers/passthrough/vtd/iommu.c >> +++ b/xen/drivers/passthrough/vtd/iommu.c >> @@ -41,6 +41,7 @@ >> #include "extern.h" >> #include "vtd.h" >> #include "../ats.h" >> +#include "../../../arch/x86/mm/mm-locks.h" > > iommu.c:38:29: fatal error: asm/mm/mm-locks.h: No such file or directory #include ^ compilation terminated. Tiejun > > ~Andrew > >> >> struct mapped_rmrr { >> struct list_head list; >> @@ -1842,6 +1843,7 @@ static int rmrr_identity_mapping(struct domain *d, >> unsigned long base_pfn, end_pfn; >> struct mapped_rmrr *mrmrr; >> struct hvm_iommu *hd = domain_hvm_iommu(d); >> + struct p2m_domain *p2m = p2m_get_hostp2m(d); >> >> ASSERT(spin_is_locked(&pcidevs_lock)); >> ASSERT(rmrr->base_address < rmrr->end_address); >> @@ -1867,8 +1869,20 @@ static int rmrr_identity_mapping(struct domain *d, >> >> while ( base_pfn < end_pfn ) >> { >> - if ( intel_iommu_map_page(d, base_pfn, base_pfn, >> - IOMMUF_readable|IOMMUF_writable) ) >> + if ( iommu_use_hap_pt(d) && (!iommu_passthrough || >> + !is_hardware_domain(d)) ) >> + { >> + p2m_lock(p2m); >> + if ( p2m_set_entry(p2m, base_pfn, _mfn(base_pfn), PAGE_ORDER_4K, >> + p2m_mmio_direct, p2m_access_rw) ) >> + { >> + p2m_unlock(p2m); >> + return -1; >> + } >> + p2m_unlock(p2m); >> + } >> + else if ( intel_iommu_map_page(d, base_pfn, base_pfn, >> + IOMMUF_readable|IOMMUF_writable) ) >> return -1; >> base_pfn++; >> } > > >