There're some platforms still doesn't support 1GB large page on CPU side.
Such as lakefield, TNT and EHL platforms on which have some silicon bug and
this case CPU don't support 1GB large page.
This patch tries to release this constrain to support more hardware platform.
Note this patch doesn't release the constrain on IOMMU side.
Tracked-On: #4550
Signed-off-by: Li Fei1 <fei1.li@intel.com>
In Commit 127c73c3, we remove the strict check for adding page table mapping. However,
we just replace the ASSERT of pr_fatal in add_pte. This is not enough. We still add
the virtual address by 4K if the page table mapping is exist and check the virtual
address is over the virtual address region for this mapping. Otherwise, The complain
will continue for 512 times at most.
Tracked-On: #3475
Signed-off-by: Li Fei1 <fei1.li@intel.com>
Deterministic is important for RTVM. The mitigation for MCE on
Page Size Change converts a large page to 4KB pages runtimely during
the vmexit triggered by the instruction fetch in the large page.
These vmexits increase nondeterminacy, which should be avoided for RTVM.
This patch builds 4KB page mapping in EPT for RTVM to avoid these vmexits.
Tracked-On: #4101
Signed-off-by: Binbin Wu <binbin.wu@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
Issue description:
-----------------
Machine Check Error on Page Size Change
Instruction fetch may cause machine check error if page size
and memory type was changed without invalidation on some
processors[1][2]. Malicious guest kernel could trigger this issue.
This issue applies to both primary page table and extended page
tables (EPT), however the primary page table is controlled by
hypervisor only. This patch mitigates the situation in EPT.
Mitigation details:
------------------
Implement non-execute huge pages in EPT.
This patch series clears the execute permission (bit 2) in the
EPT entries for large pages. When EPT violation is triggered by
guest instruction fetch, hypervisor converts the large page to
smaller 4 KB pages and restore the execute permission, and then
re-execute the guest instruction.
The current patch turns on the mitigation by default.
The follow-up patches will conditionally turn on/off the feature
per processor model.
[1] Refer to erratum KBL002 in "7th Generation Intel Processor
Family and 8th Generation Intel Processor Family for U Quad Core
Platforms Specification Update"
https://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/7th-gen-core-family-spec-update.pdf
[2] Refer to erratum SKL002 in "6th Generation Intel Processor
Family Specification Update"
https://www.intel.com/content/www/us/en/products/docs/processors/core/desktop-6th-gen-core-family-spec-update.html
Tracked-On: #4101
Signed-off-by: Binbin Wu <binbin.wu@intel.com>
Reviewed-by: Eddie Dong <eddie.dong@intel.com>
1. Print warning message instead of ASSERT when
the caller try to modify the attribute for
memory region that is not present.
2. To avoid above warning message for memory region
below 1M,its attribute may be updated by Service
VM when updating MTTR setting.
Tracked-On: #3992
Signed-off-by: Yonghua Huang <yonghua.huang@intel.com>
Reviewed-by: Fei Li <fei1.li@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
When we support PCI MSI-X table BAR remapping, we may re-delete the MSI-X table BAR
region. This patch removes strict check for deleting page table mapping.
Tracked-On: #3475
Signed-off-by: Li, Fei1 <fei1.li@intel.com>
The current implement only do "only add a page table mapping for a region when
it's not mapped" check when this page table entry is a PTE entry. However, it
need to do this check for PDPTE entry and PDE entry too.
Tracked-On: #3475
Signed-off-by: Li, Fei1 <fei1.li@intel.com>
EPT tables are shared by MMU and IOMMU.
Some IOMMUs don't support page-walk coherency, the cpu cache of EPT entires
should be flushed to memory after modifications, so that the modifications
are visible to the IOMMUs.
This patch adds a new interface to flush the cache of modified EPT entires.
There are different implementations for EPT/PPT entries:
- For PPT, there is no need to flush the cpu cache after update.
- For EPT, need to call iommu_flush_cache to make the modifications visible
to IOMMUs.
Tracked-On: #3607
Signed-off-by: Binbin Wu <binbin.wu@intel.com>
Reviewed-by: Anthony Xu <anthony.xu@intel.com>
Panic should only be used when system booting. Once the system boot done,
it could never be used. While ASSERT could be used in some situations, such
as, there are some pre-assumption for some code, using ASSERT here for debug.
Tracked-On: #861
Signed-off-by: Li, Fei1 <fei1.li@intel.com>
Acked-by: Eddie Dong <eddie.dong@inte.com>
1) Fix procedure has more than one exit point.
2) Add some brackets to make logical conjunctions more readable.
Tracked-On: #861
Signed-off-by: Li, Fei1 <fei1.li@intel.com>
Acked-by: Anthony Xu <anthony.xu@intel.com>
Replace dynamic memory allocation in memory management with static memory allocation.
Since the static memory allocation can guarantee the allocation never failed, so
split_large_page and construct_pgentry don't need to return a errno any more.
Besides, the destroy_ept don't need to free page any more. Instead, it would memset
the eptp to 0 to make sure we can't walk this paging table again.
Tracked-On: #861
Signed-off-by: Li, Fei1 <fei1.li@intel.com>
Fix violations for function whose parameter can be read-only.
Tracked-On: #861
Signed-off-by: Huihuang Shi <huihuang.shi@intel.com>
Acked-by: Anthony Xu <anthony.xu@intel.com>
- add missing brackets for 'if/else' statements based on MISRA-C
requirements
v1 -> v2:
* add brackets for each conditions in 'if' statements to improve
the readability
* modify 'ptdev_init' to make the logic clearer
Tracked-On: #861
Signed-off-by: Shiqing Gao <shiqing.gao@intel.com>
In the current hypervisor design, when HPA is not
found for the specified gpa by calling gpa2hpa or
local_gpa2hpa, 0 will be returned as a error code,
but 0 may be a valid HPA for vm0; error checking
is missed when invoking gpa2hpa or local_gpa2hpa;
when invoking lookup_address, the caller guarantees
that parameter pointer pml4_page and pointer pg_size
is not NULL.
If local_gpa2hpa/gpa2hpa returns a invalid HPA,
it means that this function fails to find the
HPA of the specified gpa of vm. If local_gpa2hpa/gpa2hpa
return value is a valid HPA, it means that this
function have found the HPA of the specified gpa of vm.
Each valid vm's EPTP is initialized during vm creating,
vm's EPTP is valid until this vm is destroyed. So the caller
can guarantee parameter pointer pml4_page is not NULL.
The caller uses a temporary variable to store page size.
So the caller can guarantee parameter pointer pg_size
is not NULL.
In this patch, define a invalid HPA for gpa2hpa and
local_gpa2hpa;add some error checking when invoking
local_gpa2hpa/gpa2hpa;add precondition for lookup_address
function and remove redundant error checking.
V1-->V2:
Define INVALID_HPA as a invalid HPA for gpa2hpa
and local_gpa2hpa;
Updated related error checking when invoking
gpa2hpa or local_gpa2hpa;
V2-->V3:
Add some debug information if specified gpa2hpa
mapping doesn't exit and ept_mr_del is called;
Update INVALID_HPA definition easier to be reviewed.
V3-->V4:
Add vm->id and gpa into pr_error;
Add precondition to ept_mr_del to cover [gpa,gpa+size)
unmapping case.
V4-->V5:
Update comments;
Update pr_error message.
Tracked-On: #1258
Signed-off-by: Xiangyang Wu <xiangyang.wu@linux.intel.com>
Reviewed-by: Li, Fei1 <fei1.li@intel.com>
Page table operation would not fault except:
1. the hypervisor it out of memory to allcate a page for page table operation
2. there is a bug with page table operation in hypervisor or devicemodle
While we assue that these would not happened in our platform when release, so
there is no need to check whether there is a fault with page table operation. However,
for debug version, we would panic the hypervisor if we can't meet the conditions really.
Tracked-On: #1124
Signed-off-by: Li, Fei1 <fei1.li@intel.com>
There are chances that names with leading underscore declared by
developers are conflict with the ones reserved for the compiler.
What this patch does:
- rename these functions/variables/macros starting with
underscore to avoid such unintentational mistakes.
- remove gpr.h without any contents
Tracked-On: #861
Signed-off-by: Shiqing Gao <shiqing.gao@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
- convert function like macros to inline functions based on MISRA-C
requirement
- remove some unused and duplicated macros
Tracked-On: #861
Signed-off-by: Shiqing Gao <shiqing.gao@intel.com>
Reviewed-by: Junjie Mao <junjie.mao@intel.com>
This patch fixes the following issues:
- Assignment operation in expression.
- For loop incrementation is not simple.
- No brackets to loop body.
- Use of comma operator.
v1 -> v2:
* Replace &x->y with &(x->y) based on our new coding rule
Tracked-On: #861
Signed-off-by: Shiqing Gao <shiqing.gao@intel.com>
Reviewed-by: Junjie Mao <junjie.mao@intel.com>
Add lookup_address to lokup the page table for a virtual address
of primary page table or a physical address of extended page table.
The remaining obtain_last_page_table_entry could be removed with their
called function later.
Signed-off-by: Li, Fei1 <fei1.li@intel.com>
The ept_mr_add still use the old map_mem. The old API will
keep for a peroid until the SOS the same refine be merged.
Signed-off-by: Li, Fei1 <fei1.li@intel.com>
Merge mmu_modify with mmu_del to mmu_modify_or_del(..., type).
While type is MR_MODIFY, the actual action is doing mmu_modify; while
type is MR_DEL, the actual action is doing mmu_del.
Signed-off-by: Li, Fei1 <fei1.li@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
If try to modify a memory region which will cross a large page and the first
page virtual address in it would not align to the large page size, it would
modify the whole large page attributes which is fatal wrong.
Signed-off-by: Li, Fei1 <fei1.li@intel.com>
1. fix some description for hv mmu_modify
2. add pml4_page input parameter for ept_mr_modify to keep align with
ept_mr_add and ept_mr_del which will support add or delete MR for trusty.
Signed-off-by: Li, Fei1 <fei1.li@intel.com>
Acked-by: Eddie Dong <eddie.dong@intel.com>
