REF:https://bugzilla.tianocore.org/show_bug.cgi?id=2457
This commit fixes an offset calculation that is used to write the
VarErrorFlag UEFI variable to the UEFI variable runtime cache.
Currently a physical address is used instead of an offset. This
commit changes the offset to zero with a length of the entire
non-volatile variable store so the entire non-volatile variable
store buffer in SMRAM (with the variable update modification) is
copied to the runtime variable cache. This follows the same pattern
used in other SynchronizeRuntimeVariableCache () calls for
consistency.
* Observable symptom: An exception in SMM will most likely occur
due to the invalid memory reference when the VarErrorFlag variable
is written. The variable is most commonly written when the UEFI
variable store is full.
* The issue only occurs when the variable runtime cache is enabled
by the following PCD being set to TRUE:
gEfiMdeModulePkgTokenSpaceGuid.PcdEnableVariableRuntimeCache
Fixes: aab3b9b9a1
Cc: Liming Gao <liming.gao@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Cc: Michael Turner <michael.turner@microsoft.com>
Cc: Jian J Wang <jian.j.wang@intel.com>
Cc: Hao A Wu <hao.a.wu@intel.com>
Signed-off-by: Michael Kubacki <michael.a.kubacki@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
Reviewed-by: Jian J Wang <jian.j.wang@intel.com>
The LoadFile protocol can report such a large buffer size that we cannot
allocate enough reserved pages for. This particularly affects HTTP(S)
Boot, if the remote file is very large (for example, an ISO image).
While the TianoCore wiki mentions this at
<https://github.com/tianocore/tianocore.github.io/wiki/HTTP-Boot#ram-disk-image-size>:
> The maximum RAM disk image size depends on how much continuous reserved
> memory block the platform could provide.
it's hard to remember; so log a DEBUG_ERROR message when the allocation
fails.
This patch produces error messages such as:
> UiApp:BmExpandLoadFile: failed to allocate reserved pages:
> BufferSize=4501536768
> LoadFile="PciRoot(0x0)/Pci(0x3,0x0)/MAC(5254001B103E,0x1)/
> IPv4(0.0.0.0,TCP,DHCP,192.168.124.106,192.168.124.1,255.255.255.0)/
> Dns(192.168.124.1)/
> Uri(https://ipv4-server/RHEL-7.7-20190723.1-Server-x86_64-dvd1.iso)"
> FilePath=""
(Manually rewrapped here for keeping PatchCheck.py happy.)
Cc: Hao A Wu <hao.a.wu@intel.com>
Cc: Jian J Wang <jian.j.wang@intel.com>
Cc: Ray Ni <ray.ni@intel.com>
Cc: Zhichao Gao <zhichao.gao@intel.com>
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Philippe Mathieu-Daude <philmd@redhat.com>
Reviewed-by: Siyuan Fu <siyuan.fu@intel.com>
Acked-by: Hao A Wu <hao.a.wu@intel.com>
Changed @rtval to @retval in SdMmcHcStartSdClock
function description.
Signed-off-by: Mateusz Albecki <mateusz.albecki@intel.com>
Reviewed-by: Hao A Wu <hao.a.wu@intel.com>
EnterS3WithImmediateWake () no longer has any callers, so remove it
from ResetSystemLib.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Ray Ni <ray.ni@intel.com>
Add edk2 platform boot manager protocol which would have platform
specific refreshes to the auto enumerated as well as NV boot options
for the platform.
Signed-off-by: Ashish Singhal <ashishsingha@nvidia.com>
Reviewed-by: Ray Ni <ray.ni@intel.com>
The endpoint selected by the driver needs to not
only be an interrupt type, but have direction IN
as required to set up an asynchronous interrupt transfer.
Currently, the driver assumes that the first INT endpoint
will be of type IN, but that is not true of all devices,
and will silently fail on devices which have the OUT endpoint
before the IN. Adjust the endpoint selection loop to explictly
check for direction IN.
Signed-off-by: Matt DeVillier <matt.devillier@gmail.com>
Reviewed-by: Hao A Wu <hao.a.wu@intel.com>
The endpoint selected by the driver needs to not
only be an interrupt type, but have direction IN
as required to set up an asynchronous interrupt transfer.
Currently, the driver assumes that the first INT endpoint
will be of type IN, but that is not true of all devices,
and will silently fail on devices which have the OUT endpoint
before the IN. Adjust the endpoint selection loop to explictly
check for direction IN.
Signed-off-by: Matt DeVillier <matt.devillier@gmail.com>
Reviewed-by: GuoMinJ <newexplorerj@gmail.com>
Reviewed-by: Hao A Wu <hao.a.wu@intel.com>
The endpoint selected by the driver needs to not
only be an interrupt type, but have direction IN
as required to set up an asynchronous interrupt transfer.
Currently, the driver assumes that the first INT endpoint
will be of type IN, but that is not true of all devices,
and will silently fail on devices which have the OUT endpoint
before the IN. Adjust the endpoint selection loop to explictly
check for direction IN.
Test: detachable keyboard on Google Pixel Slate now works.
Signed-off-by: Matt DeVillier <matt.devillier@gmail.com>
Reviewed-by: GuoMinJ <newexplorerj@gmail.com>
Reviewed-by: Hao A Wu <hao.a.wu@intel.com>
For eMMC modules we used to notify the platform about frequency
change only after sending CMD13 which meant that platform
might not get a chance to apply required post frequency
change fixes to get the clock stable. To fix this
notification has been moved to SdMmcHcClockSupply function
just after we start the SD clock. During first time setup
the notification won't be sent to avoid changing old behavior.
Cc: Hao A Wu <hao.a.wu@intel.com>
Cc: Marcin Wojtas <mw@semihalf.com>
Cc: Zhichao Gao <zhichao.gao@intel.com>
Cc: Liming Gao <liming.gao@intel.com>
Signed-off-by: Mateusz Albecki <mateusz.albecki@intel.com>
Tested-by: Marcin Wojtas <mw@semihalf.com>
Tested-by: Hao A Wu <hao.a.wu@intel.com>
Reviewed-by: Hao A Wu <hao.a.wu@intel.com>
REF:https://bugzilla.tianocore.org/show_bug.cgi?id=2118
When a packet is queued/completed for the asynchronous IO queue, the logic
to roll over to the front of the queue doesn't account for actual size of
the IO Submission/Completion queue.
This causes a device to hang due to doorbell being outside of visible
queue. An example would be if an NVMe drive only supported a queue size of
128 while the driver supports 256.
Cc: Jian J Wang <jian.j.wang@intel.com>
Signed-off-by: Sean Brogan <sean.brogan@microsoft.com>
Signed-off-by: Hao A Wu <hao.a.wu@intel.com>
Acked-by: Ray Ni <ray.ni@intel.com>
REF:https://bugzilla.tianocore.org/show_bug.cgi?id=2333
During a SetVariable () invocation, UpdateVariable () is called.
UpdateVariable () contains logic to determine whether a volatile or
non-volatile UEFI variable was set so the corresponding runtime
cache can be updated to reflect the change. The current logic simply
evaluates Variable->Volatile to determine which runtime cache should
be updated.
The problem is Variable->Volatile does not always reflect whether a
volatile variable is being set. Variable->Volatile is set to TRUE
only in the case a pre-existing variable is found in the volatile
variable store. Therefore, the value is FALSE when a new volatile
variable is written.
This change updates the logic to take this into account. If a new
variable is written successfully, the Attributes will accurately
reflect whether the variable is non-volatile. If a pre-existing
variable is modified, the Volatile field will reflect the type of
variable (Attributes are not reliable; e.g. 0x0 indicates deletion).
* Observable symptom: A volatile variable that was set successfully
might return EFI_NOT_FOUND when the variable should be found.
* The issue is a regression introduced to the variable services only
when the variable runtime cache is enabled by the following PCD
being set to TRUE:
gEfiMdeModulePkgTokenSpaceGuid.PcdEnableVariableRuntimeCache
* The issue was implemented in commit aab3b9b9a1 but the PCD was not
set to TRUE by default enabling the issue until commit e07b7d024a.
Fixes: aab3b9b9a1
Cc: Liming Gao <liming.gao@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Cc: Jian J Wang <jian.j.wang@intel.com>
Cc: Hao A Wu <hao.a.wu@intel.com>
Signed-off-by: Michael Kubacki <michael.a.kubacki@intel.com>
Reviewed-by: Liming Gao <liming.gao@intel.com>
Reviewed-by: Jian J Wang <jian.j.wang@intel.com>
REF: https://bugzilla.tianocore.org/show_bug.cgi?id=2303
Whenever a PCI device is discovered, PCI bus calls the
EDKII_DEVICE_SECURITY_PROTOCOL to authenticate it.
If the function returns success, the PCI bus allocates
the resource and installs the PCI_IO for the device.
If the function returns fail, the PCI bus skips the device.
It is similar to EFI_SECURITY_ARCH_PROTOCOL, which
is used to verify an EFI image.
Cc: Jian J Wang <jian.j.wang@intel.com>
Cc: Hao A Wu <hao.a.wu@intel.com>
Cc: Ray Ni <ray.ni@intel.com>
Cc: Yun Lou <yun.lou@intel.com>
Signed-off-by: Jiewen Yao <jiewen.yao@intel.com>
Reviewed-by: Ray Ni <ray.ni@intel.com>
When a boot loader examines the memory map, it can see that location 0
is available memory. If it chooses to use that memory, and
PcdNullPointerDetectionPropertyMask is enabled, use of memory in page 0
will cause an exception. This does occur when running the memtest86
program.
Leaving page 0 available is for legacy support purpose. Since we have
deprecated the support of legacy, the solution is just reserving it so
that it cannot be allocated for other uses.
Ref: https://bugzilla.tianocore.org/show_bug.cgi?id=1885
Cc: Dandan Bi <dandan.bi@intel.com>
Cc: Liming Gao <liming.gao@intel.com>
Cc: Ray Ni <ray.ni@intel.com>
Cc: Hao A Wu <hao.a.wu@intel.com>
Cc: Sean Brogan <sean.brogan@microsoft.com>
Signed-off-by: Jian J Wang <jian.j.wang@intel.com>
Reviewed-by: Ray Ni <ray.ni@intel.com>
REF:https://bugzilla.tianocore.org/show_bug.cgi?id=2220
This change reduces SMIs for GetVariable () by maintaining a
UEFI variable cache in Runtime DXE in addition to the pre-
existing cache in SMRAM. When the Runtime Service GetVariable()
is invoked, a Runtime DXE cache is used instead of triggering an
SMI to VariableSmm. This can improve overall system performance
by servicing variable read requests without rendezvousing all
cores into SMM.
The runtime cache can be disabled with by setting the FeaturePCD
gEfiMdeModulePkgTokenSpaceGuid.PcdEnableVariableRuntimeCache
to FALSE. If the PCD is set to FALSE, the runtime cache will not be
used and an SMI will be triggered for Runtime Service
GetVariable () and GetNextVariableName () invocations.
The following are important points regarding the behavior of the
variable drivers when the variable runtime cache is enabled.
1. All of the non-volatile storage contents are loaded into the
cache upon driver load. This one time load operation from storage
is preferred as opposed to building the cache on demand. An on-
demand cache would require a fallback SMI to load data into the
cache as variables are requested.
2. SetVariable () requests will continue to always trigger an SMI.
This occurs regardless of whether the variable is volatile or
non-volatile.
3. Both volatile and non-volatile variables are cached in a runtime
buffer. As is the case in the current EDK II variable driver, they
continue to be cached in separate buffers.
4. The cache in Runtime DXE and SMM are intended to be exact copies
of one another. All SMM variable accesses only return data from the
SMM cache. The runtime caches are only updated after the variable I/O
operation is successful in SMM. The runtime caches are only updated
from SMM.
5. Synchronization mechanisms are in place to ensure the runtime cache
content integrity with the SMM cache. These may result in updates to
runtime cache that are the same in content but different in offset and
size from updates to the SMM cache.
When using SMM variables with runtime cache enabled, two caches will now
be present.
1. "Runtime Cache" - Maintained in VariableSmmRuntimeDxe. Used to service
Runtime Services GetVariable () and GetNextVariableName () callers.
2. "SMM Cache" - Maintained in VariableSmm to service SMM GetVariable ()
and GetNextVariableName () callers.
a. This cache is retained so SMM modules do not operate on data outside
SMRAM.
Because a race condition can occur if an SMI occurs during the execution
of runtime code reading from the runtime cache, a runtime cache read lock
is introduced that explicitly moves pending updates from SMM to the runtime
cache if an SMM update occurs while the runtime cache is locked. Note that
it is not expected a Runtime services call will interrupt SMM processing
since all CPU cores rendezvous in SMM.
It is possible to view UEFI variable read and write statistics by setting
the gEfiMdeModulePkgTokenSpaceGuid.PcdVariableCollectStatistics FeaturePcd
to TRUE and using the VariableInfo UEFI application in MdeModulePkg to dump
variable statistics to the console. By doing so, a user can view the number
of GetVariable () hits from the Runtime DXE variable driver (Runtime Cache
hits) and the SMM variable driver (SMM Cache hits). SMM Cache hits for
GetVariable () will occur when SMM modules invoke GetVariable ().
Cc: Dandan Bi <dandan.bi@intel.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Liming Gao <liming.gao@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Cc: Ray Ni <ray.ni@intel.com>
Cc: Jian J Wang <jian.j.wang@intel.com>
Cc: Hao A Wu <hao.a.wu@intel.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Signed-off-by: Michael Kubacki <michael.a.kubacki@intel.com>
Reviewed-by: Jian J Wang <jian.j.wang@intel.com>
REF:https://bugzilla.tianocore.org/show_bug.cgi?id=2220
The current VariableInfo application only checks for variable
statistics from SMM if the variable information entries are
not present in the UEFI System Configuration table as published
by the DXE UEFI variable driver (VariableRuntimeDxe).
This change first checks for variable information entries in the
UEFI System Configuration but always checks for entries in SMM
as well. If the SMM variable driver is not present, an instance of
EFI_SMM_VARIABLE_PROTOCOL will not be found and the search for
SMM variable statistics will be aborted (an SW SMI to get variable
statistics will not be triggered).
In the case variable statistics are provided by both a Runtime DXE
driver (e.g. VariableSmmRuntimeDxe) and a SMM driver (VariableSmm),
this change will clearly identify statistics from each respective
driver.
Cc: Dandan Bi <dandan.bi@intel.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Liming Gao <liming.gao@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Cc: Ray Ni <ray.ni@intel.com>
Cc: Jian J Wang <jian.j.wang@intel.com>
Cc: Hao A Wu <hao.a.wu@intel.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Signed-off-by: Michael Kubacki <michael.a.kubacki@intel.com>
Reviewed-by: Jian J Wang <jian.j.wang@intel.com>
Acked-by: Hao A Wu <hao.a.wu@intel.com>
This change moves the following functions into a dedicated file
so they may be used in other variable files as needed. These are
commonly needed for basic variable data structure parsing
operations. The functions are grouped together in VariableParsing.c
to support cohesiveness for these operations in the file.
Furthermore, it reduces the overall size of the common Variable.c
file.
* DataSizeOfVariable ()
* FindVariableEx ()
* GetEndPointer ()
* GetNextVariablePtr ()
* GetStartPointer ()
* GetVariableDataOffset ()
* GetVariableDataPtr ()
* GetVariableHeaderSize ()
* GetVariableNamePtr ()
* GetVariableStoreStatus ()
* GetVendorGuidPtr ()
* IsValidVariableHeader ()
* NameSizeOfVariable ()
* SetDataSizeOfVariable ()
* SetNameSizeOfVariable ()
* UpdateVariableInfo ()
* VariableCompareTimeStampInternal ()
* VariableServiceGetNextVariableInternal ()
Cc: Dandan Bi <dandan.bi@intel.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Eric Dong <eric.dong@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>
Cc: Liming Gao <liming.gao@intel.com>
Cc: Michael D Kinney <michael.d.kinney@intel.com>
Cc: Ray Ni <ray.ni@intel.com>
Cc: Jian J Wang <jian.j.wang@intel.com>
Cc: Hao A Wu <hao.a.wu@intel.com>
Cc: Jiewen Yao <jiewen.yao@intel.com>
Signed-off-by: Michael Kubacki <michael.a.kubacki@intel.com>
Reviewed-by: Jian J Wang <jian.j.wang@intel.com>
Reviewed-by: Hao A Wu <hao.a.wu@intel.com>
Before called by GetBufferForValue(), Value has already been called
function IsTypeInBuffer to make sure the value must be buffer type.
So GetBufferForValue can not return NULL.
This commit adds ASSERT to assume (GetBufferForValue (&Value) is not
NULL.
Cc: Jian J Wang <jian.j.wang@intel.com>
Cc: Hao A Wu <hao.a.wu@intel.com>
Signed-off-by: Shenglei Zhang <shenglei.zhang@intel.com>
Reviewed-by: Dandan Bi <dandan.bi@intel.com>
