REF: https://bugzilla.tianocore.org/show_bug.cgi?id=1305
The patch reverts commit 1ed6498c4a
* UefiCpuPkg/CommonFeature: Skip locking when the feature is disabled
FEATURE_CONTROL.Lock bit is controlled by feature
CPU_FEATURE_LOCK_FEATURE_CONTROL_REGISTER. The commit 1ed649 fixes
a bug that when the feature is disabled, the Lock bit is cleared.
But it's a security hole if the bit is cleared when booting OS.
We can argue that platform needs to make sure the value
of PcdCpuFeaturesUserConfiguration should be set properly to make
sure feature CPU_FEATURE_LOCK_FEATURE_CONTROL_REGISTER is enabled.
But it's better to guarantee this in the generic core code.
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ruiyu Ni <ruiyu.ni@intel.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
Acked-by: Laszlo Ersek <lersek@redhat.com>
Cc: Andrew Fish <afish@apple.com>
Cc: Leif Lindholm <leif.lindholm@linaro.org>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
In current implementation, core and package level sync uses same semaphores.
Sharing the semaphore may cause wrong execution order.
For example:
1. Feature A has CPU_FEATURE_CORE_BEFORE dependency with Feature B.
2. Feature C has CPU_FEATURE_PACKAGE_AFTER dependency with Feature B.
The expected feature initialization order is A B C:
A ---- (Core Depends) ----> B ---- (Package Depends) ----> C
For a CPU has 1 package, 2 cores and 4 threads. The feature initialization
order may like below:
Thread#1 Thread#2 Thread#3 Thread#4
[A.Init] [A.Init] [A.Init]
Release(S1, S2) Release(S1, S2) Release(S3, S4)
Wait(S1) * 2 Wait(S2) * 2 <------------------------------- Core sync
[B.Init] [B.Init]
Release (S1,S2,S3,S4)
Wait (S1) * 4 <----------------------------------------------------- Package sync
Wait(S4 * 2) <- Core sync
[B.Init]
In above case, for thread#4, when it syncs in core level, Wait(S4) * 2 isn't
blocked and [B.Init] runs. But [A.Init] hasn't run in thread#3. It's wrong!
Thread#4 should execute [B.Init] after thread#3 executes [A.Init] because B
core level depends on A.
The reason of the wrong execution order is that S4 is released in thread#1
by calling Release (S1, S2, S3, S4) and in thread #4 by calling
Release (S3, S4).
To fix this issue, core level sync and package level sync should use separate
semaphores.
In above example, the S4 released in Release (S1, S2, S3, S4) should not be the
same semaphore as that in Release (S3, S4).
Related BZ: https://bugzilla.tianocore.org/show_bug.cgi?id=1311
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Acked-by: Laszlo Ersek <lersek@redhat.com>
In current implementation, core and package level sync uses same semaphores.
Sharing the semaphore may cause wrong execution order.
For example:
1. Feature A has CPU_FEATURE_CORE_BEFORE dependency with Feature B.
2. Feature C has CPU_FEATURE_PACKAGE_AFTER dependency with Feature B.
The expected feature initialization order is A B C:
A ---- (Core Depends) ----> B ---- (Package Depends) ----> C
For a CPU has 1 package, 2 cores and 4 threads. The feature initialization
order may like below:
Thread#1 Thread#2 Thread#3 Thread#4
[A.Init] [A.Init] [A.Init]
Release(S1, S2) Release(S1, S2) Release(S3, S4)
Wait(S1) * 2 Wait(S2) * 2 <------------------------------- Core sync
[B.Init] [B.Init]
Release (S1,S2,S3,S4)
Wait (S1) * 4 <----------------------------------------------------- Package sync
Wait(S4 * 2) <- Core sync
[B.Init]
In above case, for thread#4, when it syncs in core level, Wait(S4) * 2 isn't
blocked and [B.Init] runs. But [A.Init] hasn't run in thread#3. It's wrong!
Thread#4 should execute [B.Init] after thread#3 executes [A.Init] because B
core level depends on A.
The reason of the wrong execution order is that S4 is released in thread#1
by calling Release (S1, S2, S3, S4) and in thread #4 by calling
Release (S3, S4).
To fix this issue, core level sync and package level sync should use separate
semaphores.
In above example, the S4 released in Release (S1, S2, S3, S4) should not be the
same semaphore as that in Release (S3, S4).
Related BZ: https://bugzilla.tianocore.org/show_bug.cgi?id=1311
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Acked-by: Laszlo Ersek <lersek@redhat.com>
V2 changes:
V1 change has regression which caused by change feature order.
V2 changes logic to detect dependence not only for the
neighborhood features. It need to check all features in the list.
V1 Changes:
In current code logic, only adjust feature position if current
CPU feature position not follow the request order. Just like
Feature A need to be executed before feature B, but current
feature A registers after feature B. So code will adjust the
position for feature A, move it to just before feature B. If
the position already met the requirement, code will not adjust
the position.
This logic has issue when met all below cases:
1. feature A has core or package level dependence with feature B.
2. feature A is register before feature B.
3. Also exist other features exist between feature A and B.
Root cause is driver ignores the dependence for this case, so
threads may execute not follow the dependence order.
Fix this issue by change code logic to adjust feature position
for CPU features which has dependence relationship.
Related BZ: https://bugzilla.tianocore.org/show_bug.cgi?id=1311
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ruiyu Ni <Ruiyu.ni@intel.com>
When static paging is disabled, page table for below 4GB is created
and page table for above 4GB is created dynamically in page fault
handler.
Today's implementation only allow SMM access-out to below types of
memory address no matter static paging is enabled or not:
1. Reserved, run time and ACPI NVS type
2. MMIO
But certain platform feature like RAS may need to access other types
of memory from SMM. Today's code blocks these platforms.
This patch simplifies the policy to only block when static paging
is used so that the static paging can be disabled in these platforms
to meet their SMM access-out need.
Setting PcdCpuSmmStaticPageTable to FALSE can disable the static
paging.
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jiewen Yao <jiewen.yao@intel.com>
Signed-off-by: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
Acked-by: Laszlo Ersek <lersek@redhat.com>
REF: https://bugzilla.tianocore.org/show_bug.cgi?id=1305
Today's code unconditionally sets the IA32_FEATURE_CONTROL.Lock to 1
no matter the feature is enabled or not.
The patch fixes this issue by only setting the Lock bit to 1 when
the feature is enabled.
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ruiyu Ni <ruiyu.ni@intel.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
In some special cases, after BSP sends Init-sipi-sipi signal
AP needs more time to start the Ap procedure. In this case
BSP may think AP has finished its task but in fact AP hasn't began
yet.
Rollback former change to keep the status which only be used
when AP really finished task.
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Build UefiCpuPkg with below configuration:
Architecture(s) = IA32
Build target = NOOPT
Toolchain = VS2015x86
Below error info shows up:
DxeRegisterCpuFeaturesLib.lib(CpuFeaturesInitialize.obj) :
error LNK2001: unresolved external symbol __allmul
Valid mDependTypeStr type only have 5 items, use UINT32 type cast
to fix this error.
Cc: Dandan Bi <dandan.bi@intel.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Index is initialized to MAX_UINT16 as default failure value, which
is what the ASSERT is supposed to test for. The ASSERT condition
however can never return FALSE for INT16 != int, as due to
Integer Promotion[1], Index is converted to int, which can never
result in -1.
Furthermore, Index is used as a for loop index variable inbetween its
initialization and the ASSERT, so the value is unconditionally
overwritten too.
Fix the ASSERT check to compare Index to its upper boundary, which it
will be equal to if the loop was not broken out of on success.
[1] ISO/IEC 9899:2011, 6.5.9.4
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Marvin Haeuser <Marvin.Haeuser@outlook.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
Index is initialized to MAX_UINT16 as default failure value, which
is what the ASSERT is supposed to test for. The ASSERT condition
however can never return FALSE for INT16 != int, as due to
Integer Promotion[1], Index is converted to int, which can never
result in -1.
Furthermore, Index is used as a for loop index variable inbetween its
initialization and the ASSERT, so the value is unconditionally
overwritten too.
Fix the ASSERT check to compare Index to its upper boundary, which it
will be equal to if the loop was not broken out of on success.
[1] ISO/IEC 9899:2011, 6.5.9.4
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Marvin Haeuser <Marvin.Haeuser@outlook.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
Current code assume only one dependence (before or after) for one
feature. Enhance code logic to support feature has two dependence
(before and after) type.
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Code initialized in function can't be correctly detected by build tool.
Add code to clearly initialize the local variable before use it.
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Dandan Bi <dandan.bi@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Remove extra white space at the end of line.
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Dandan Bi <dandan.bi@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Changes include:
1. Remove extra white space at the end of line.
2. Add comments for the new add function parameter.
3. Update IN OUT tag for function parameter.
Cc: Dandan Bi <dandan.bi@intel.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Remove extra white space at the end of line.
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Dandan Bi <dandan.bi@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Code initialized in function can't be correctly detected by build tool.
Add code to clearly initialize the local variable before use it.
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Dandan Bi <dandan.bi@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
The freed-memory guard feature will cause a recursive calling
of InitializePageTablePool(). This is due to a fact that
AllocateAlignedPages() is used to allocate page table pool memory.
This function will most likely call gBS->FreePages to free unaligned
pages and then cause another round of page attributes change, like
below (freed pages will be always marked not-present if freed-memory
guard is enabled)
FreePages() <===============|
=> CpuSetMemoryAttributes() |
=> <if out of page table> |
=> InitializePageTablePool() |
=> AllocateAlignedPages() |
=> FreePages() ================|
The solution is add a global variable as a lock in page table pool
allocation function and fail any other requests if it has not been
done.
Since this issue will only happen if free-memory guard is enabled,
it won't affect CpuSetMemoryAttributes() in default build of a BIOS.
If free-memory guard is enabled, it only affect the pages
(failed to mark them as not-present) freed in AllocateAlignedPages().
Since those freed pages haven't been used yet (their addresses not
yet exposed to code outside AllocateAlignedPages), it won't compromise
the freed-memory guard feature.
This change will just fail the CpuSetMemoryAttributes() called from
FreePages() but it won't fail the FreePages(). So the error status
won't be propagated upper layer of code.
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Star Zeng <star.zeng@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jian J Wang <jian.j.wang@intel.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
Non-stop mode was introduced / explained in commit 8f2613628a
("MdeModulePkg/MdeModulePkg.dec: add new settings for PCDs",
2018-08-30).
The macro HEAP_GUARD_NONSTOP_MODE was added to CpuDxe in commit
dcc026217f ("UefiCpuPkg/CpuDxe: implement non-stop mode for uefi",
2018-08-30).
Another instance of the macro HEAP_GUARD_NONSTOP_MODE was added to
PiSmmCpuDxeSmm -- with BIT1|BIT0 replaced with BIT3|BIT2 -- in commit
09afd9a42a ("UefiCpuPkg/PiSmmCpuDxeSmm: implement non-stop mode for
SMM", 2018-08-30)
Since the freed-memory guard is for UEFI-only. This patch only updates
HEAP_GUARD_NONSTOP_MODE in "UefiCpuPkg/CpuDxe/CpuDxe.h" (add BIT4).
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Star Zeng <star.zeng@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jian J Wang <jian.j.wang@intel.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
Because MSR has scope attribute, driver has no needs to set
MSR for all APs if MSR scope is core or package type. This patch
updates code to base on the MSR scope value to add MSR to the register
table.
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
V4 changes:
1. Serial console log for different threads when program register table.
2. Check the AcpiCpuData before use it to avoid potential ASSERT.
V3 changes:
1. Use global variable instead of internal function to return string for register type
and dependence type.
2. Add comments for some complicated logic.
V1 changes:
Because this driver needs to set MSRs saved in normal boot phase, sync
semaphore logic from RegisterCpuFeaturesLib code which used for normal boot phase.
Detail see below change for RegisterCpuFeaturesLib:
UefiCpuPkg/RegisterCpuFeaturesLib: Add logic to support semaphore type.
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Acked-by: Laszlo Ersek <lersek@redhat.com>
Regression-tested-by: Laszlo Ersek <lersek@redhat.com>
V4 changes include:
1. Serial debug message for different threads when program the register table.
V3 changes include:
1. Use global variable instead of internal function to return string for register type
and dependence type.
2. Add comments for some complicated logic.
V2 changes include:
1. Add more description for the code part which need easy to understand.
2. Refine some code base on feedback for V1 changes.
V1 changes include:
In a system which has multiple cores, current set register value task costs huge times.
After investigation, current set MSR task costs most of the times. Current logic uses
SpinLock to let set MSR task as an single thread task for all cores. Because MSR has
scope attribute which may cause GP fault if multiple APs set MSR at the same time,
current logic use an easiest solution (use SpinLock) to avoid this issue, but it will
cost huge times.
In order to fix this performance issue, new solution will set MSRs base on their scope
attribute. After this, the SpinLock will not needed. Without SpinLock, new issue raised
which is caused by MSR dependence. For example, MSR A depends on MSR B which means MSR A
must been set after MSR B has been set. Also MSR B is package scope level and MSR A is
thread scope level. If system has multiple threads, Thread 1 needs to set the thread level
MSRs and thread 2 needs to set thread and package level MSRs. Set MSRs task for thread 1
and thread 2 like below:
Thread 1 Thread 2
MSR B N Y
MSR A Y Y
If driver don't control execute MSR order, for thread 1, it will execute MSR A first, but
at this time, MSR B not been executed yet by thread 2. system may trig exception at this
time.
In order to fix the above issue, driver introduces semaphore logic to control the MSR
execute sequence. For the above case, a semaphore will be add between MSR A and B for
all threads. Semaphore has scope info for it. The possible scope value is core or package.
For each thread, when it meets a semaphore during it set registers, it will 1) release
semaphore (+1) for each threads in this core or package(based on the scope info for this
semaphore) 2) acquire semaphore (-1) for all the threads in this core or package(based
on the scope info for this semaphore). With these two steps, driver can control MSR
sequence. Sample code logic like below:
//
// First increase semaphore count by 1 for processors in this package.
//
for (ProcessorIndex = 0; ProcessorIndex < PackageThreadsCount ; ProcessorIndex ++) {
LibReleaseSemaphore ((UINT32 *) &SemaphorePtr[PackageOffset + ProcessorIndex]);
}
//
// Second, check whether the count has reach the check number.
//
for (ProcessorIndex = 0; ProcessorIndex < ValidApCount; ProcessorIndex ++) {
LibWaitForSemaphore (&SemaphorePtr[ApOffset]);
}
Platform Requirement:
1. This change requires register MSR setting base on MSR scope info. If still register MSR
for all threads, exception may raised.
Known limitation:
1. Current CpuFeatures driver supports DXE instance and PEI instance. But semaphore logic
requires Aps execute in async mode which is not supported by PEI driver. So CpuFeature
PEI instance not works after this change. We plan to support async mode for PEI in phase
2 for this task.
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
v3 changes:
1. Move CPU_FEATURE_DEPENDENCE_TYPE definition here from RegisterCpuFeaturesLib.h file.
2. Add Invalid type for REGISTER_TYPE which will be used in code.
v2 changes:
1. Add more description about why we do this change.
2. Change structure field type from pointer to EFI_PHYSICAL_ADDRESS because it will
be share between PEI and DXE.
v1 Changes:
In order to support semaphore related logic, add new definition for it.
In a system which has multiple cores, current set register value task costs huge times.
After investigation, current set MSR task costs most of the times. Current logic uses
SpinLock to let set MSR task as an single thread task for all cores. Because MSR has
scope attribute which may cause GP fault if multiple APs set MSR at the same time,
current logic use an easiest solution (use SpinLock) to avoid this issue, but it will
cost huge times.
In order to fix this performance issue, new solution will set MSRs base on their scope
attribute. After this, the SpinLock will not needed. Without SpinLock, new issue raised
which is caused by MSR dependence. For example, MSR A depends on MSR B which means MSR A
must been set after MSR B has been set. Also MSR B is package scope level and MSR A is
thread scope level. If system has multiple threads, Thread 1 needs to set the thread level
MSRs and thread 2 needs to set thread and package level MSRs. Set MSRs task for thread 1
and thread 2 like below:
Thread 1 Thread 2
MSR B N Y
MSR A Y Y
If driver don't control execute MSR order, for thread 1, it will execute MSR A first, but
at this time, MSR B not been executed yet by thread 2. system may trig exception at this
time.
In order to fix the above issue, driver introduces semaphore logic to control the MSR
execute sequence. For the above case, a semaphore will be add between MSR A and B for
all threads. Semaphore has scope info for it. The possible scope value is core or package.
For each thread, when it meets a semaphore during it set registers, it will 1) release
semaphore (+1) for each threads in this core or package(based on the scope info for this
semaphore) 2) acquire semaphore (-1) for all the threads in this core or package(based
on the scope info for this semaphore). With these two steps, driver can control MSR
sequence. Sample code logic like below:
//
// First increase semaphore count by 1 for processors in this package.
//
for (ProcessorIndex = 0; ProcessorIndex < PackageThreadsCount ; ProcessorIndex ++) {
LibReleaseSemaphore ((UINT32 *) &SemaphorePtr[PackageOffset + ProcessorIndex]);
}
//
// Second, check whether the count has reach the check number.
//
for (ProcessorIndex = 0; ProcessorIndex < ValidApCount; ProcessorIndex ++) {
LibWaitForSemaphore (&SemaphorePtr[ApOffset]);
}
Platform Requirement:
1. This change requires register MSR setting base on MSR scope info. If still register MSR
for all threads, exception may raised.
Known limitation:
1. Current CpuFeatures driver supports DXE instance and PEI instance. But semaphore logic
requires Aps execute in async mode which is not supported by PEI driver. So CpuFeature
PEI instance not works after this change. We plan to support async mode for PEI in phase
2 for this task.
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Regression-tested-by: Laszlo Ersek <lersek@redhat.com>
REF: https://bugzilla.tianocore.org/show_bug.cgi?id=1237
Sometimes the memory will be contaminated by random data left in last
boot (warm reset). The code should not assume the allocated memory is
always filled with zero. This patch add code to clear data structure
used for stack switch to prevent such problem from happening.
Cc: Eric Dong <eric.dong@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jian J Wang <jian.j.wang@intel.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
V5:
1. Add ASSERT to indicate this assumption that environment is 32 bit mode.
2. Add description in INF about this driver's expected result
in different environment.
V4:
Only disable paging when it is enabled.
V3 changes:
No need to change inf file.
V2 changes:
Only disable paging in 32 bit mode, no matter it is enable or not.
V1 changes:
PEI Stack Guard needs to enable paging. This might cause #GP if code
trying to write CR3 register with PML4 page table while the processor
is enabled with PAE paging.
Simply disabling paging before updating CR3 can solve this conflict.
It's an regression caused by change: 0a0d5296e4
BZ: https://bugzilla.tianocore.org/show_bug.cgi?id=1232
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Jian J Wang <jian.j.wang@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Regression-tested-by: Laszlo Ersek <lersek@redhat.com>
REF:https://bugzilla.tianocore.org/show_bug.cgi?id=1194
Speculative execution is used by processor to avoid having to wait for
data to arrive from memory, or for previous operations to finish, the
processor may speculate as to what will be executed.
If the speculation is incorrect, the speculatively executed instructions
might leave hints such as which memory locations have been brought into
cache. Malicious actors can use the bounds check bypass method (code
gadgets with controlled external inputs) to infer data values that have
been used in speculative operations to reveal secrets which should not
otherwise be accessed.
It is possible for SMI handler(s) to call EFI_SMM_CPU_PROTOCOL service
ReadSaveState() and use the content in the 'CommBuffer' (controlled
external inputs) as the 'CpuIndex'. So this commit will insert AsmLfence
API to mitigate the bounds check bypass issue within SmmReadSaveState().
For SmmReadSaveState():
The 'CpuIndex' will be passed into function ReadSaveStateRegister(). And
then in to ReadSaveStateRegisterByIndex().
With the call:
ReadSaveStateRegisterByIndex (
CpuIndex,
SMM_SAVE_STATE_REGISTER_IOMISC_INDEX,
sizeof(IoMisc.Uint32),
&IoMisc.Uint32
);
The 'IoMisc' can be a cross boundary access during speculative execution.
Later, 'IoMisc' is used as the index to access buffers 'mSmmCpuIoWidth'
and 'mSmmCpuIoType'. One can observe which part of the content within
those buffers was brought into cache to possibly reveal the value of
'IoMisc'.
Hence, this commit adds a AsmLfence() after the check of 'CpuIndex'
within function SmmReadSaveState() to prevent the speculative execution.
A more detailed explanation of the purpose of commit is under the
'Bounds check bypass mitigation' section of the below link:
https://software.intel.com/security-software-guidance/insights/host-firmware-speculative-execution-side-channel-mitigation
And the document at:
https://software.intel.com/security-software-guidance/api-app/sites/default/files/337879-analyzing-potential-bounds-Check-bypass-vulnerabilities.pdf
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Hao Wu <hao.a.wu@intel.com>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
Acked-by: Laszlo Ersek <lersek@redhat.com>
Regression-tested-by: Laszlo Ersek <lersek@redhat.com>
The PCD below is unused, so it has been removed from inf.
gUefiCpuPkgTokenSpaceGuid.PcdCpuFeaturesSupport
Cc: Eric Dong <eric.dong@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: shenglei <shenglei.zhang@intel.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
https://bugzilla.tianocore.org/show_bug.cgi?id=967
Request to add a library function for GetAcpiTable() in order
to get ACPI table using signature as input.
After evaluation, we found there are many duplicated code to
find ACPI table by signature in different modules.
This patch updates PiSmmCpuDxeSmm to use new
EfiLocateFirstAcpiTable() and remove the duplicated code.
Cc: Younas khan <pmdyounaskhan786@gmail.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Cc: Liming Gao <liming.gao@intel.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Star Zeng <star.zeng@intel.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
V3 changes include:
1. Keep the ReservedX not change if bit info not changed for this field.
V2 changes include:
1. Use X in ReservedX fields from totally new value if MSR structure definition changed.
For example, if in current structure, the max reserved variable is Reserved2, in new
definition, reserved variable is begin with Reserved3.
V1 Changes:
Changes includes:
1. Update MSR structure definition, change some reserved fields to useful fields:
1. MSR_XEON_PHI_PKG_CST_CONFIG_CONTROL_REGISTER
2. MSR_XEON_PHI_SMM_MCA_CAP_REGISTER
2. For MSR_XEON_PHI_PMG_IO_CAPTURE_BASE_REGISTER structure, it expand the field range.
Old definition like below:
typedef union {
///
/// Individual bit fields
///
struct {
///
/// [Bits 15:0] LVL_2 Base Address (R/W).
///
UINT32 Lvl2Base:16;
///
/// [Bits 18:16] C-state Range (R/W) Specifies the encoding value of the
/// maximum C-State code name to be included when IO read to MWAIT
/// redirection is enabled by MSR_PKG_CST_CONFIG_CONTROL[bit10]: 100b - C4
/// is the max C-State to include 110b - C6 is the max C-State to include.
///
UINT32 CStateRange:3;
UINT32 Reserved1:13;
UINT32 Reserved2:32;
} Bits;
///
/// All bit fields as a 32-bit value
///
UINT32 Uint32;
///
/// All bit fields as a 64-bit value
///
UINT64 Uint64;
} MSR_XEON_PHI_PMG_IO_CAPTURE_BASE_REGISTER;
This patch make below changes for this data structure, it expand "CStateRange" field width.
old one:
UINT32 CStateRange:3;
UINT32 Reserved1:13;
new one:
UINT32 CStateRange:7;
UINT32 Reserved1:9;
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Acked-by: Laszlo Ersek <lersek@redhat.com>
V3 changes include:
1. Keep ReservedX not change if bit info not changed for this field.
V2 changes include:
1. Use X in ReservedX fields from totally new value if MSR structure definition changed.
For example, if in current structure, the max reserved variable is Reserved2, in new
definition, reserved variable is begin with Reserved3.
V1 Changes includes:
1. Change fields which is reserved in old version: MSR_IA32_RTIT_CTL_REGISTER
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Acked-by: Laszlo Ersek <lersek@redhat.com>
Changes includes:
1. Remove old MSR which not existed in 2018-05 version spec:
1. MSR_CORE_ROB_CR_BKUPTMPDR6
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Acked-by: Laszlo Ersek <lersek@redhat.com>
Changes includes:
1. Remove MSR which not existed in 2018-05 version spec: MSR_P6_ROB_CR_BKUPTMPDR6.
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Acked-by: Laszlo Ersek <lersek@redhat.com>
Changes includes:
1. Remove old MSR which not existed in 2018-05 version spec:
1. MSR_CORE2_BBL_CR_CTL3
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Acked-by: Laszlo Ersek <lersek@redhat.com>
Changes includes:
1. Add new MSR file which used for goldmont plus microarchitecture.
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Acked-by: Laszlo Ersek <lersek@redhat.com>
Latest SDM has moved MSR related content from volume 3 chapter 35 to volume 4
chapter 2. Current MSR's comments need to be updated to reference the new
chapter info.
Changes includes:
1. Update referenced chapter info from some MSRs.
2. Update referenced SDM version info.
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ruiyu Ni <ruiyu.ni@intel.com>
Acked-by: Laszlo Ersek <lersek@redhat.com>
REF: https://bugzilla.tianocore.org/show_bug.cgi?id=1187
The patch reverts 9c8c4478cf
"UefiCpuPkg/MtrrLib: Skip Base MSR access when the pair is invalid".
Microsoft Windows will report an error in event manager if MTRR
usage is different across hibernate even when the difference is
in an non valid MTRR pair. This seems like a bug in Windows but
for compatibility and servicing reasons we think a change in UEFI
would wise.
A Windows change has already been submitted for the next iteration
(2019 time frame).
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Ruiyu Ni <ruiyu.ni@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
Cc: Sean Brogan <sean.brogan@microsoft.com>
REF: https://bugzilla.tianocore.org/show_bug.cgi?id=1166
Visual Studio 2012 will complain uninitialized variable, StackBase,
in the CpuPaging.c. This patch adds code to init it to zero and
ASSERT check against 0. This is enough since uninit case will only
happen during retrieving stack memory via gEfiHobMemoryAllocStackGuid.
But this HOB will always be created in advance.
Cc: Dandan Bi <dandan.bi@intel.com>
Cc: Hao A Wu <hao.a.wu@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jian J Wang <jian.j.wang@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
BZ: https://bugzilla.tianocore.org/show_bug.cgi?id=1191
Before commit e21e355e2c, jmp _SmiHandler
is commented. And below code, ASM_PFX(CpuSmmDebugEntry) is moved into rax,
then call it. But, this code doesn't work in XCODE5 tool chain. Because XCODE5
doesn't generated the absolute address in the EFI image. So, rax stores the
relative address. Once this logic is moved to another place, it will not work.
; jmp _SmiHandler ; instruction is not needed
...
mov rax, ASM_PFX(CpuSmmDebugEntry)
call rax
Commit e21e355e2c is to support XCODE5.
One tricky way is selected to fix it. Although SmiEntry logic is copied to
another place and run, but here jmp _SmiHandler is enabled to jmp the original
code place, then call ASM_PFX(CpuSmmDebugEntry) with the relative address.
mov rax, strict qword 0 ; mov rax, _SmiHandler
_SmiHandlerAbsAddr:
jmp rax
...
call ASM_PFX(CpuSmmDebugEntry)
Now, BZ 1191 raises the issue that SmiHandler should run in the copied address,
can't run in the common address. So, jmp _SmiHandler is required to be removed,
the code is kept to run in copied address. And, the relative address is
requried to be fixed up to the absolute address. The necessary changes should
not affect the behavior of platforms that already consume PiSmmCpuDxeSmm.
OVMF SMM boot to shell with VS2017, GCC5 and XCODE5 tool chain has been verified.
...
mov rax, strict qword 0 ; call ASM_PFX(CpuSmmDebugEntry)
CpuSmmDebugEntryAbsAddr:
call rax
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Liming Gao <liming.gao@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Tested-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
Ruiyu Ni
Eric Dong
Jiewen Yao
Marvin H?user
Jian J Wang
Hao Wu
shenglei
Star Zeng
Liming Gao