This patch is to check BspIndex first before lock cmpxchg operation.
If BspIndex has not been set, then do the lock cmpxchg, otherwise,
the APs don't need to lock cmpxchg the BspIndex value since the BSP
election has been done. It's the optimization to lower the resource
contention caused by the atomic compare exchange operation, so as to
improve the SMI performance for BSP election.
Cc: Ray Ni <ray.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Zeng Star <star.zeng@intel.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Rahul Kumar <rahul1.kumar@intel.com>
Cc: Kinney Michael D <michael.d.kinney@intel.com>
Signed-off-by: Jiaxin Wu <jiaxin.wu@intel.com>
Reviewed-by: Ray Ni <ray.ni@intel.com>
Use MAX_UINT32 directly instead of typecasting from signed
to unsigned value.
Cc: Ray Ni <ray.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Zeng Star <star.zeng@intel.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Rahul Kumar <rahul1.kumar@intel.com>
Cc: Kinney Michael D <michael.d.kinney@intel.com>
Signed-off-by: Jiaxin Wu <jiaxin.wu@intel.com>
Reviewed-by: Ray Ni <ray.ni@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
After BSP returned from SmmCoreEntry, there are several rounds BSP
and AP sync in BSP handler:
1 .ReleaseAllAPs(); /// Notify all APs to exit.
if (SmmCpuFeaturesNeedConfigureMtrrs()) {
2. SmmCpuSyncWaitForAPs(); /// Wait for all APs to program MTRRs.
3. ReleaseAllAPs(); /// Signal APs to restore MTRRs.
}
4. SmmCpuSyncWaitForAPs(); /// Wait for all APs to complete pending
tasks including MTRR.
5. ReleaseAllAPs(); /// Signal APs to Reset states.
6. SmmCpuSyncWaitForAPs(); /// Gather APs to exit SMM synchronously.
Before step 6 and after step 5, BSP performs below items:
A. InitializeDebugAgent() /// Stop source level debug.
B. SmmCpuUpdate() /// Perform pending operations for hot-plug.
C. Present = FALSE; /// Clear the Present flag of BSP.
For InitializeDebugAgent(), BSP needs to wait all APs complete
pending tasks and then notify all APs to stop source level debug.
So, above step 4 & step 5 are required for InitializeDebugAgent().
For SmmCpuUpdate(), it's to perform pending operations for
hot-plug CPUs take effect in next SMI. Existing APs in SMI do not
reply on the CPU switch & hot-add & hot-remove operations. So, no
need step 4 and step 5 for additional one round BSP & AP sync.
Step 6 can make sure all APs are ready to exit SMM, then hot-plug
operation can take effect in next SMI.
For BSP "Present" flag, AP does not reply on it. No need step 4
and step 5 for additional one round BSP & AP sync.
Based on above analysis, step 4 and step 5 are only required if
need configure MTRR and support SMM source level debug. So, we can
reduce one round BSP and AP sync if both are unsupported. With
this change, SMI performance can be improved.
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Ray Ni <ray.ni@intel.com>
Cc: Zeng Star <star.zeng@intel.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Rahul Kumar <rahul1.kumar@intel.com>
Signed-off-by: Jiaxin Wu <jiaxin.wu@intel.com>
Reviewed-by: Ray Ni <ray.ni@Intel.com>
Existing BSP handler stops source level debug, then call ReleaseAllAPs
to tell all APs can reset the Present flag to FALSE:
InitializeDebugAgent (); /// Stop source level debug
ReleaseAllAPs (); /// Tell APs can reset "Present" flag.
This patch is to invert ReleaseAllAPs & InitializeDebugAgent:
ReleaseAllAPs (); /// Tell APs can reset "Present" flag.
InitializeDebugAgent (); /// Stop source level debug
After this change, there is no negative impact since SMM source level
debug feature doesn't depend on AP's "Present" flag, no impact to the
SMM source level debug capability.
Instead, the change will benefit the AP source level debug capability
to trace its "Present" flag change for SMI exit since the source
level debug feature will be stopped after each AP has the chance to
reset the state.
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Ray Ni <ray.ni@intel.com>
Cc: Zeng Star <star.zeng@intel.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Rahul Kumar <rahul1.kumar@intel.com>
Signed-off-by: Jiaxin Wu <jiaxin.wu@intel.com>
Reviewed-by: Ray Ni <ray.ni@Intel.com>
Below piece of code is the BSP and AP sync logic for SMI exit.
1. AP after finish the scheduled procedure:
if (SmmCpuFeaturesNeedConfigureMtrrs ()) {
SmmCpuSyncReleaseBsp ();
SmmCpuSyncWaitForBsp ();
...
}
SmmCpuSyncReleaseBsp ();
SmmCpuSyncWaitForBsp ();
SmmCpuSyncReleaseBsp ();
2. BSP after return from SmmCoreEntry:
SmmCpuSyncWaitForAPs ();
if (SmmCpuFeaturesNeedConfigureMtrrs ()) {
ReleaseAllAPs ();
...
SmmCpuSyncWaitForAPs ();
}
ReleaseAllAPs ();
SmmCpuSyncWaitForAPs();
This patch is to make BSP same as AP sync logic:
if (SmmCpuFeaturesNeedConfigureMtrrs ()) {
SmmCpuSyncWaitForAPs ();
ReleaseAllAPs ();
...
}
SmmCpuSyncWaitForAPs ();
ReleaseAllAPs ();
SmmCpuSyncWaitForAPs();
With the change, it will be easy to understand the sync flow as
below:
BSP: SmmCpuSyncWaitForAPs <-- AP: SmmCpuSyncReleaseBsp
BSP: ReleaseAllAPs --> AP: SmmCpuSyncWaitForBsp
This patch doesn't have function impact.
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Ray Ni <ray.ni@intel.com>
Cc: Zeng Star <star.zeng@intel.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Rahul Kumar <rahul1.kumar@intel.com>
Signed-off-by: Jiaxin Wu <jiaxin.wu@intel.com>
Reviewed-by: Ray Ni <ray.ni@Intel.com>
This patch is to check SMM Debug Agent support or not before
InitializeDebugAgent.
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Ray Ni <ray.ni@intel.com>
Cc: Zeng Star <star.zeng@intel.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Rahul Kumar <rahul1.kumar@intel.com>
Signed-off-by: Jiaxin Wu <jiaxin.wu@intel.com>
Reviewed-by: Ray Ni <ray.ni@Intel.com>
There is the SmmCpuSyncLib Library class define the SMM CPU sync
flow, which is aligned with existing SMM CPU driver sync behavior.
This patch is to consume SmmCpuSyncLib instance directly.
With this change, SMM CPU Sync flow/logic can be customized
with different implementation no matter for any purpose, e.g.
performance tuning, handle specific register, etc.
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Ray Ni <ray.ni@intel.com>
Cc: Zeng Star <star.zeng@intel.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Rahul Kumar <rahul1.kumar@intel.com>
Signed-off-by: Jiaxin Wu <jiaxin.wu@intel.com>
Reviewed-by: Ray Ni <ray.ni@intel.com>
To decrease the count of RunningApCount, InterlockedDecrement is
enough to achieve that.
This patch is to simplify RunningApCount decrement.
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Ray Ni <ray.ni@intel.com>
Cc: Zeng Star <star.zeng@intel.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Rahul Kumar <rahul1.kumar@intel.com>
Signed-off-by: Jiaxin Wu <jiaxin.wu@intel.com>
Reviewed-by: Ray Ni <ray.ni@intel.com>
This patch is to define 3 new functions (WaitForBsp & ReleaseBsp &
ReleaseOneAp) used for the semaphore sync between BSP & AP. With the
change, BSP and AP Sync flow will be easy understand as below:
BSP: ReleaseAllAPs or ReleaseOneAp --> AP: WaitForBsp
BSP: WaitForAllAPs <-- AP: ReleaseBsp
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Ray Ni <ray.ni@intel.com>
Cc: Zeng Star <star.zeng@intel.com>
Cc: Rahul Kumar <rahul1.kumar@intel.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Signed-off-by: Jiaxin Wu <jiaxin.wu@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Ray Ni <ray.ni@intel.com>
Currently, if BSP election is not enabled, will use Core0 as SMM BSP.
However, Core0 does not always have the highest performance.
So, we can used NonSmm BSP as default BSP.
This will take effect in normal SMM init flow and S3 boot flow.
In normal SMM flow, the code is executed before first SMI.
In S3 flow, the code is executed in Non-SMM BSP's first SMI, where the
gSmmCpuPrivate keeps the data from last boot.
Reviewed-by: Ray Ni <ray.ni@intel.com>
Cc: Rahul Kumar <rahul1.kumar@intel.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Zhiguang Liu <zhiguang.liu@intel.com>
Use GenSmmPageTable() to create both IA32 and X64 Smm S3
page table.
Signed-off-by: Dun Tan <dun.tan@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ray Ni <ray.ni@intel.com>
Cc: Rahul Kumar <rahul1.kumar@intel.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Extern mSmmShadowStackSize in PiSmmCpuDxeSmm.h and remove
extern for mSmmShadowStackSize in c files to simplify code.
Signed-off-by: Dun Tan <dun.tan@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ray Ni <ray.ni@intel.com>
Cc: Rahul Kumar <rahul1.kumar@intel.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
When a platform has lots of CPU cores/threads, perf-logging on every
AP produces lots of records. When this multiplies with number of SMIs
during post, the records are even more.
So, this patch adds a new PCD PcdSmmApPerfLogEnable (default TRUE)
to allow platform to turn off perf-logging on APs.
Signed-off-by: Ray Ni <ray.ni@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Rahul Kumar <rahul1.kumar@intel.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Jiaxin Wu <jiaxin.wu@intel.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
MP procedures are those procedures that run in every CPU thread.
The EDKII perf infra is not MP safe so it doesn't support to be called
from those MP procedures.
The patch adds SMM MP perf-logging support in SmmMpPerf.c.
The following procedures are perf-logged:
* SmmInitHandler
* SmmCpuFeaturesRendezvousEntry
* PlatformValidSmi
* SmmCpuFeaturesRendezvousExit
Signed-off-by: Ray Ni <ray.ni@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Rahul Kumar <rahul1.kumar@intel.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Jiaxin Wu <jiaxin.wu@intel.com>
Reviewed-by: Jiaxin Wu <jiaxin.wu@intel.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
REF: https://bugzilla.tianocore.org/show_bug.cgi?id=4337
Existing SMBASE Relocation is in the PiSmmCpuDxeSmm driver, which
will relocate the SMBASE of each processor by setting the SMBASE
field in the saved state map (at offset 7EF8h) to a new value.
The RSM instruction reloads the internal SMBASE register with the
value in SMBASE field when each time it exits SMM. All subsequent
SMI requests will use the new SMBASE to find the starting address
for the SMI handler (at SMBASE + 8000h).
Due to the default SMBASE for all x86 processors is 0x30000, the
APs' 1st SMI for rebase has to be executed one by one to avoid
the processors over-writing each other's SMM Save State Area (see
existing SmmRelocateBases() function), which means the next AP has
to wait for the previous AP to finish its 1st SMI, then it can call
into its 1st SMI for rebase via Smi Ipi command, thus leading the
existing SMBASE Relocation has to be running in series. Besides, it
needs very complex code to handle the AP exit semaphore
(mRebased[Index]), which will hook return address of SMM Save State
so that semaphore code can be executed immediately after AP exits
SMM for SMBASE relocation (see existing SemaphoreHook() function).
With SMM Base Hob support, PiSmmCpuDxeSmm does not need the RSM
instruction to do the SMBASE Relocation. SMBASE Register for each
processors have already been programmed and all SMBASE address have
recorded in SMM Base Hob. So the same default SMBASE Address
(0x30000) will not be used, thus the processors over-writing each
other's SMM Save State Area will not happen in PiSmmCpuDxeSmm driver.
This way makes the first SMI init can be executed in parallel and
save boot time on multi-core system. Besides, Semaphore Hook code
logic is also not required, which will greatly simplify the SMBASE
Relocation flow.
Mainly changes as below:
* Assume the biggest possibility of tile size is 8k.
* Combine 2 SMIs (gcSmmInitTemplate & gcSmiHandlerTemplate) into one
(gcSmiHandlerTemplate), the new SMI handler needs to run to 2 paths:
one to SmmCpuFeaturesInitializeProcessor(), the other to SMM Core
Entry Point.
* Issue SMI IPI (All Excluding Self SMM IPI + BSP SMM IPI) for first
SMI init before normal SMI sources happen.
* Call SmmCpuFeaturesInitializeProcessor() in parallel.
Cc: Eric Dong <eric.dong@intel.com>
Cc: Ray Ni <ray.ni@intel.com>
Cc: Zeng Star <star.zeng@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Acked-by: Gerd Hoffmann <kraxel@redhat.com>
Cc: Rahul Kumar <rahul1.kumar@intel.com>
Signed-off-by: Jiaxin Wu <jiaxin.wu@intel.com>
Reviewed-by: Ray Ni <ray.ni@intel.com>
REF: https://bugzilla.tianocore.org/show_bug.cgi?id=4173
Due to more core count increasement, it's hard to reflect all APs
state via AP bitvector support in the register. Actually, SMM CPU
driver doesn't need to check each AP state to know all CPUs in SMI
or not, one alternative method is to check the SMM Delayed & Blocked
AP Count number:
APs in SMI + Blocked Count + Disabled Count >= All supported Aps
(code comments explained why can be > All supported Aps)
With above change, the returned value of "SmmRegSmmEnable" &
"SmmRegSmmDelayed" & "SmmRegSmmBlocked" from SmmCpuFeaturesLib
should be the AP count number within the existing CPU package.
For register that return the bitvector state, require
SmmCpuFeaturesGetSmmRegister() returns count number of all bit per
logical processor within the same package.
For register that return the AP count, require
SmmCpuFeaturesGetSmmRegister() returns the register value directly.
v3:
- Refine the coding style
v2:
- Rename "mPackageBspInfo" to "mPackageFirstThreadIndex"
- Clarify the expected value of "SmmRegSmmEnable" & "SmmRegSmmDelayed" &
"SmmRegSmmBlocked" returned from SmmCpuFeaturesLib.
- Thread: https://edk2.groups.io/g/devel/message/96722
v1:
- Thread: https://edk2.groups.io/g/devel/message/96671
Cc: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ray Ni <ray.ni@intel.com>
Cc: Zeng Star <star.zeng@intel.com>
Signed-off-by: Jiaxin Wu <jiaxin.wu@intel.com>
REF: https://bugzilla.tianocore.org/show_bug.cgi?id=4083
In CPU relaxed mode, it doesn't reset the value of
mSmmMpSyncData->AllApArrivedWithException when BSP exit smm mode.
So this patch will reset this variable.
Cc: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ray Ni <ray.ni@intel.com>
Signed-off-by: Zhihao Li <zhihao.li@intel.com>
Reviewed-by: Abner Chang <abner.chang@amd.com>
REF? https://bugzilla.tianocore.org/show_bug.cgi?id=3815
This patch define a new Protocol with the new services
SmmWaitForAllProcessor(), which can be used by SMI handler
to optionally wait for other APs to complete SMM rendezvous in
relaxed AP mode.
A new library SmmCpuRendezvousLib is provided to abstract the service
into library API to simple SMI handler code.
Cc: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ray Ni <ray.ni@intel.com>
Cc: Rahul Kumar <rahul1.kumar@intel.com>
Cc: Siyuan Fu <siyuan.fu@intel.com>
Cc: Zhihao Li <zhihao.li@intel.com>
Signed-off-by: Zhihao Li <zhihao.li@intel.com>
REF: https://bugzilla.tianocore.org/show_bug.cgi?id=3737
Apply uncrustify changes to .c/.h files in the UefiCpuPkg package
Cc: Andrew Fish <afish@apple.com>
Cc: Leif Lindholm <leif@nuviainc.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Signed-off-by: Michael Kubacki <michael.kubacki@microsoft.com>
Reviewed-by: Ray Ni <ray.ni@intel.com>
REF: https://bugzilla.tianocore.org/show_bug.cgi?id=3739
Update all use of EFI_D_* defines in DEBUG() macros to DEBUG_* defines.
Cc: Andrew Fish <afish@apple.com>
Cc: Leif Lindholm <leif@nuviainc.com>
Cc: Michael Kubacki <michael.kubacki@microsoft.com>
Signed-off-by: Michael D Kinney <michael.d.kinney@intel.com>
Reviewed-by: Ray Ni <ray.ni@intel.com>
5-level paging can be enabled on CPU which supports up to 52 physical
address size. But when the feature was enabled, the 48 address size
limit was not removed and the 5-level paging testing didn't access
address >= 2^48. So the issue wasn't detected until recently an
address >= 2^48 is accessed.
Signed-off-by: Ray Ni <ray.ni@intel.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Rahul Kumar <rahul1.kumar@intel.com>
REF: https://bugzilla.tianocore.org/show_bug.cgi?id=3199
When Token points to mSmmStartupThisApToken, this routine is called
from SmmStartupThisAp() in non-blocking mode due to
PcdCpuSmmBlockStartupThisAp == FALSE.
In this case, caller wants to startup AP procedure in non-blocking
mode and cannot get the completion status from the Token because there
is no way to return the Token to caller from SmmStartupThisAp().
Caller needs to use its specific way to query the completion status.
There is no need to allocate a token for such case so the 3 overheads
can be avoided:
1. Call AllocateTokenBuffer() when there is no free token.
2. Get a free token from the token buffer.
3. Call ReleaseToken() in APHandler().
Signed-off-by: Ray Ni <ray.ni@intel.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Cc: Rahul Kumar <rahul1.kumar@intel.com>
In function InitGdt(), SmiPFHandler() and Gen4GPageTable(), it uses
CpuIndex * mSmmStackSize to get the SMM stack address offset for
multi processor. It misses the SMM Shadow Stack Size. Each processor
will use mSmmStackSize + mSmmShadowStackSize in the memory.
It should use CpuIndex * (mSmmStackSize + mSmmShadowStackSize) to get
this SMM stack address offset. If mSmmShadowStackSize > 0 and multi
processor enabled, it will get the wrong offset value.
CET shadow stack feature will set the value of mSmmShadowStackSize.
REF: https://bugzilla.tianocore.org/show_bug.cgi?id=3237
Signed-off-by: Sheng Wei <w.sheng@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Ray Ni <ray.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Rahul Kumar <rahul1.kumar@intel.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Roger Feng <roger.feng@intel.com>
Reviewed-by: Jiewen Yao <jiewen.yao@intel.com>
Reviewed-by: Ray Ni <ray.ni@intel.com>
Most busy waits (spinlocks) in "UefiCpuPkg/PiSmmCpuDxeSmm/MpService.c"
already call CpuPause() in their loop bodies; see SmmWaitForApArrival(),
APHandler(), and SmiRendezvous(). However, the "main wait" within
APHandler():
> //
> // Wait for something to happen
> //
> WaitForSemaphore (mSmmMpSyncData->CpuData[CpuIndex].Run);
doesn't do so, as WaitForSemaphore() keeps trying to acquire the semaphore
without pausing.
The performance impact is especially notable in QEMU/KVM + OVMF
virtualization with CPU overcommit (that is, when the guest has
significantly more VCPUs than the host has physical CPUs). The guest BSP
is working heavily in:
BSPHandler() [MpService.c]
PerformRemainingTasks() [PiSmmCpuDxeSmm.c]
SetUefiMemMapAttributes() [SmmCpuMemoryManagement.c]
while the many guest APs are spinning in the "Wait for something to
happen" semaphore acquisition, in APHandler(). The guest APs are
generating useless memory traffic and saturating host CPUs, hindering the
guest BSP's progress in SetUefiMemMapAttributes().
Rework the loop in WaitForSemaphore(): call CpuPause() in every iteration
after the first check fails. Due to Pause Loop Exiting (known as Pause
Filter on AMD), the host scheduler can favor the guest BSP over the guest
APs.
Running a 16 GB RAM + 512 VCPU guest on a 448 PCPU host, this patch
reduces OVMF boot time (counted until reaching grub) from 20-30 minutes to
less than 4 minutes.
The patch should benefit physical machines as well -- according to the
Intel SDM, PAUSE "Improves the performance of spin-wait loops". Adding
PAUSE to the generic WaitForSemaphore() function is considered a general
improvement.
Cc: Eric Dong <eric.dong@intel.com>
Cc: Philippe Mathieu-Daudé <philmd@redhat.com>
Cc: Rahul Kumar <rahul1.kumar@intel.com>
Cc: Ray Ni <ray.ni@intel.com>
Ref: https://bugzilla.redhat.com/show_bug.cgi?id=1861718
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Message-Id: <20200729185217.10084-1-lersek@redhat.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
REF: https://bugzilla.tianocore.org/show_bug.cgi?id=2388
After remove Used parameter, below code in ResetTokens can also be
removed:
1. The RunningApCount parameter will be reset in GetFreeToken.
2. The ReleaseSpinLock should be called in ReleaseToken function,
Code in this function seems like a later fix if ReleaseToken not
Release it. We should remove code here and fix the real issue if
existed.
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ray Ni <ray.ni@intel.com>
Cc: Star Zeng <star.zeng@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
REF: https://bugzilla.tianocore.org/show_bug.cgi?id=2388
After patch "UefiCpuPkg/PiSmmCpuDxeSmm: Improve the
performance of GetFreeToken()" which adds new parameter
FirstFreeToken, it's not need to use Uses parameter.
This patch used to remove this parameter.
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ray Ni <ray.ni@intel.com>
Cc: Star Zeng <star.zeng@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Today's GetFreeToken() runs at the algorithm complexity of O(n)
where n is the size of the token list.
The change introduces a new global variable FirstFreeToken and it
always points to the first free token. So the algorithm complexity
of GetFreeToken() decreases from O(n) to O(1).
The improvement matters when some SMI code uses StartupThisAP()
service for each of the AP such that the algorithm complexity
becomes O(n) * O(m) where m is the AP count.
As next steps,
1. PROCEDURE_TOKEN.Used field can be optimized out because
all tokens before FirstFreeToken should have "Used" set while all
after FirstFreeToken should have "Used" cleared.
2. ResetTokens() can be optimized to only reset tokens before
FirstFreeToken.
v2: add missing line in InitializeDataForMmMp.
v3: update copyright year to 2020.
Signed-off-by: Ray Ni <ray.ni@intel.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
REF: https://bugzilla.tianocore.org/show_bug.cgi?id=2434
Current code use below loops to enumerate the CPUs:
for (Index = mMaxNumberOfCpus; Index-- > 0;) {
it has no issue but not easy for the developers to read the code.
Update above code to below style,
for (Index = 0; Index < mMaxNumberOfCpus; Index++) {
It make the developers easy to read and consistent with other
similar cases in this driver.
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Cc: Ray Ni <ray.ni@intel.com>
Signed-off-by: Eric Dong <eric.dong@intel.com>
This reverts commit 123b720eeb.
The commit message for commit 123b720eeb is not correct.
Cc: Ray Ni <ray.ni@intel.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Eric Dong <eric.dong@intel.com>
REF: https://bugzilla.tianocore.org/show_bug.cgi?id=2388
Token is new introduced by MM MP Protocol. Current logic allocate Token
every time when need to use it. The logic caused SMI latency raised to
very high. Update logic to allocate Token buffer at driver's entry point.
Later use the token from the allocated token buffer. Only when all the
buffer have been used, then need to allocate new buffer.
Former change (9caaa79dd7) missed
PROCEDURE_TOKEN part, this change covers it.
Reviewed-by: Ray Ni <ray.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Eric Dong <eric.dong@intel.com>
The size for the array of mSmmMpSyncData->CpuData[] is 0 ~
mMaxNumberOfCpus -1. But current code may use
mSmmMpSyncData->CpuData[mMaxNumberOfCpus].
This patch fixed this issue.
Reviewed-by: Ray Ni <ray.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Eric Dong <eric.dong@intel.com>
REF: https://bugzilla.tianocore.org/show_bug.cgi?id=2268
In current implementation, when check whether APs called by StartUpAllAPs
or StartUpThisAp, it checks the Tokens value used by other APs. Also the AP
will update the Token value for itself if its task finished. In this
case, the potential race condition issues happens for the tokens.
Because of this, system may trig ASSERT during cycling test.
This change enhance the code logic, add new attributes for the token to
remove the reference for the tokens belongs to other APs.
Reviewed-by: Ray Ni <ray.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Eric Dong <eric.dong@intel.com>
REF: https://bugzilla.tianocore.org/show_bug.cgi?id=2388
Token is new introduced by MM MP Protocol. Current logic allocate Token
every time when need to use it. The logic caused SMI latency raised to
very high. Update logic to allocate Token buffer at driver's entry point.
Later use the token from the allocated token buffer. Only when all the
buffer have been used, then need to allocate new buffer.
Reviewed-by: Ray Ni <ray.ni@intel.com>
Signed-off-by: Eric Dong <eric.dong@intel.com>
Acked-by: Laszlo Ersek <lersek@redhat.com>
Due to needs a tackling the deficiency of the AP API, it's necessary to
ensure that in non-blocking mode previous AP executed command is
finished before starting new one.
To remedy above:
1) execute AcquireSpinLock instead AcquireSpinLockOrFail - this will
ensure time "window" to eliminate potential race condition between
BSP and AP spinLock release in non-blocking mode.
This also will eliminate possibility to start executing new AP
function before last is finished.
2) remove returns EFI_STATUS - EFI_NOT_READY - in new scenario returned
status is not necessary to caller.
Signed-off-by: Damian Nikodem <damian.nikodem@intel.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ray Ni <ray.ni@intel.com>
Cc: Benjamin You <benjamin.you@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Krzysztof Rusocki <krzysztof.rusocki@intel.com>
1. Update CPUStatus to CpuStatus in comments to align comments
with code.
2. Add "OUT" attribute for "ProcedureArguments" parameter in function
header.
Cc: Eric Dong <eric.dong@intel.com>
Cc: Ray Ni <ray.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Shenglei Zhang <shenglei.zhang@intel.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
1.Remove "out" attribute for " Buffer" parameter in function header.
2.Add "out" attribute for " Token" parameter in function header.
3.Update ProcedureArgument to ProcedureArguments.
Cc: Eric Dong <eric.dong@intel.com>
Cc: Ray Ni <ray.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Shenglei Zhang <shenglei.zhang@intel.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
REF: https://bugzilla.tianocore.org/show_bug.cgi?id=1937
Add MM Mp Protocol in PiSmmCpuDxeSmm driver.
Cc: Ray Ni <ray.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Ray Ni <ray.ni@intel.com>
Regression-tested-by: Laszlo Ersek <lersek@redhat.com>
BZ: https://bugzilla.tianocore.org/show_bug.cgi?id=1593
For every SMI occurrence, save and restore CR2 register only when SMM
on-demand paging support is enabled in 64 bit operation mode.
This is not a bug but to have better improvement of code.
Patch5 is updated with separate functions for Save and Restore of CR2
based on review feedback.
Patch6 - Removed Global Cr2 instead used function parameter.
Patch7 - Removed checking Cr2 with 0 as per feedback.
Patch8 and 9 - Aligned with EDK2 Coding style.
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Vanguput Narendra K <narendra.k.vanguput@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Ray Ni <ray.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Yao Jiewen <jiewen.yao@intel.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
Reviewed-by: Nate DeSimone <nathaniel.l.desimone@intel.com>
Reviewed-by: Ray Ni <ray.ni@intel.com>
Regression-tested-by: Laszlo Ersek <lersek@redhat.com>
REF: https://bugzilla.tianocore.org/show_bug.cgi?id=1521
We scan the SMM code with ROPgadget.
http://shell-storm.org/project/ROPgadget/https://github.com/JonathanSalwan/ROPgadget/tree/master
This tool reports the gadget in SMM driver.
This patch enabled CET ShadowStack for X86 SMM.
If CET is supported, SMM will enable CET ShadowStack.
SMM CET will save the OS CET context at SmmEntry and
restore OS CET context at SmmExit.
Test:
1) test Intel internal platform (x64 only, CET enabled/disabled)
Boot test:
CET supported or not supported CPU
on CET supported platform
CET enabled/disabled
PcdCpuSmmCetEnable enabled/disabled
Single core/Multiple core
PcdCpuSmmStackGuard enabled/disabled
PcdCpuSmmProfileEnable enabled/disabled
PcdCpuSmmStaticPageTable enabled/disabled
CET exception test:
#CF generated with PcdCpuSmmStackGuard enabled/disabled.
Other exception test:
#PF for normal stack overflow
#PF for NX protection
#PF for RO protection
CET env test:
Launch SMM in CET enabled/disabled environment (DXE) - no impact to DXE
The test case can be found at
https://github.com/jyao1/SecurityEx/tree/master/ControlFlowPkg
2) test ovmf (both IA32 and X64 SMM, CET disabled only)
test OvmfIa32/Ovmf3264, with -D SMM_REQUIRE.
qemu-system-x86_64.exe -machine q35,smm=on -smp 4
-serial file:serial.log
-drive if=pflash,format=raw,unit=0,file=OVMF_CODE.fd,readonly=on
-drive if=pflash,format=raw,unit=1,file=OVMF_VARS.fd
QEMU emulator version 3.1.0 (v3.1.0-11736-g7a30e7adb0-dirty)
3) not tested
IA32 CET enabled platform
Cc: Eric Dong <eric.dong@intel.com>
Cc: Ray Ni <ray.ni@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Yao Jiewen <jiewen.yao@intel.com>
Reviewed-by: Ray Ni <ray.ni@intel.com>
Regression-tested-by: Laszlo Ersek <lersek@redhat.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>
1. Do not use tab characters
2. No trailing white space in one line
3. All files must end with CRLF
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Liming Gao <liming.gao@intel.com>
The mechanism behind is the same as NULL pointer detection enabled in EDK-II
core. SMM has its own page table and we have to disable page 0 again in SMM
mode.
Cc: Star Zeng <star.zeng@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Michael Kinney <michael.d.kinney@intel.com>
Cc: Ayellet Wolman <ayellet.wolman@intel.com>
Suggested-by: Ayellet Wolman <ayellet.wolman@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Jian J Wang <jian.j.wang@intel.com>
Reviewed-by: Star Zeng <star.zeng@intel.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
https://bugzilla.tianocore.org/show_bug.cgi?id=674
Add CPUID check to see if the CPU supports the Machine
Check Architecture before accessing the Machine Check
Architecture related MSRs.
Cc: Eric Dong <eric.dong@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.1
Signed-off-by: Michael Kinney <michael.d.kinney@intel.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
Cc: Jeff Fan <jeff.fan@intel.com>
Cc: Ruiyu Ni <ruiyu.ni@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Eric Dong <eric.dong@intel.com>
Reviewed-by: Jeff Fan <jeff.fan@intel.com>
If PcdCpuHotPlugSupport is TRUE, gSmst->NumberOfCpus will be the
PcdCpuMaxLogicalProcessorNumber. If gSmst->SmmStartupThisAp() is invoked for
those un-existed processors, ASSERT() happened in ConfigSmmCodeAccessCheck().
This fix is to check if ProcessorId is valid before invoke
gSmst->SmmStartupThisAp() in ConfigSmmCodeAccessCheck() and to check if
ProcessorId is valid in InternalSmmStartupThisAp() to avoid unexpected DEBUG
error message displayed.
Cc: Jiewen Yao <jiewen.yao@intel.com>
Cc: Eric Dong <eric.dong@intel.com>
Contributed-under: TianoCore Contribution Agreement 1.0
Signed-off-by: Jeff Fan <jeff.fan@intel.com>
Reviewed-by: Eric Dong <eric.dong@intel.com>
Jiaxin Wu
Zhiguang Liu
Dun Tan
Ray Ni
Zhihao Li
Michael Kubacki
Michael D Kinney
Sheng Wei
Laszlo Ersek
Dong, Eric
Ray Ni
Damian Nikodem
Shenglei Zhang
Vanguput, Narendra K
Jiewen Yao
Liming Gao
Jian J Wang
Jeff Fan