Pushed via:
PR - https://github.com/tianocore/edk2/pull/4841
Commit -
https://github.com/tianocore/edk2/commit/4fb69c2bee3006b08ca2abba58843f7573e0c5c6
Best Regards,
Hao Wu
> -Original Message-
> From: Wu, Hao A
> Sent: Thursday, August 10, 2023 10:39 AM
> To: Ranbir Singh ;
Reviewed-by: Hao A Wu
Best Regards,
Hao Wu
> -Original Message-
> From: Ranbir Singh
> Sent: Wednesday, August 16, 2023 1:38 PM
> To: devel@edk2.groups.io; rsi...@ventanamicro.com
> Cc: Wu, Hao A ; Ni, Ray ; Veeresh
> Sangolli
> Subject: [PATCH v2 1/2] MdeModulePkg/Bus/Pci/UhciDxe:
Hi All,
Could you please help review this patch. Very appreciated!
Thanks,
Xianglei
-Original Message-
From: devel@edk2.groups.io On Behalf Of Xianglei Cai
Sent: Monday, September 18, 2023 2:57 PM
To: devel@edk2.groups.io
Cc: Cai, Xianglei ; Wu, Hao A ; Ni,
Ray ; Wang, Jian J ; Gao,
*Reminder: TianoCore Bug Triage - APAC / NAMO*
*When:*
Tuesday, September 19, 2023
6:30pm to 7:30pm
(UTC-07:00) America/Los Angeles
*Where:*
Now that the transition to use SetMemoryProtectionsLib and
GetMemoryProtectionsLib is complete, delete the memory protection PCDs
to avoid confusing the interface. All memory protection settings
will now be set and consumed via the libraries.
Signed-off-by: Taylor Beebe
Cc: Jian J Wang
Cc:
Now that the transition to use SetMemoryProtectionsLib and
GetMemoryProtectionsLib is complete, delete the memory protection PCDs
to avoid confusing the interface. All memory protection settings
will now be set and consumed via the libraries.
Signed-off-by: Taylor Beebe
Cc: Ard Biesheuvel
Cc:
Replace references to the memory protection PCDs to instead
check the platform protections via GetMemoryProtectionsLib.
Because the protection profile is equivalent to the PCD settings,
this updated does not cause a torn state.
Signed-off-by: Taylor Beebe
Cc: Jian J Wang
Cc: Liming Gao
Cc:
Now that the transition to use SetMemoryProtectionsLib and
GetMemoryProtectionsLib is complete, delete the memory protection PCDs
to avoid confusing the interface. All memory protection settings
will now be set and consumed via the libraries.
Signed-off-by: Taylor Beebe
Cc: Ard Biesheuvel
Cc:
Now that the EDK2 tree uses GetMemoryProtectionsLib to query
the platform memory protection settings, OvmfPkg can be updated
to use QemuCfg to set the entire memory protection profile instead
of just SetNxForStack.
For example, the following will set the DXE memory protection to
the RELEASE
Set the memory protections on Arm virtual platforms. Because
the QemuFg parser is not currently available in ArmVirtPkg, use
the RELEASE profile by default.
Signed-off-by: Taylor Beebe
Cc: Ard Biesheuvel
Cc: Leif Lindholm
Cc: Sami Mujawar
Cc: Gerd Hoffmann
---
Now that the transition to use SetMemoryProtectionsLib and
GetMemoryProtectionsLib is complete, delete the PCD profile
to avoid confusing the interface. All memory protection settings
will now be set and consumed via the libraries.
Signed-off-by: Taylor Beebe
Cc: Jian J Wang
Cc: Liming Gao
---
MemoryProtectionConfigLib enables parsing the fw_cfg for the
memory protection profile.
Signed-off-by: Taylor Beebe
Cc: Ard Biesheuvel
Cc: Jiewen Yao
Cc: Jordan Justen
Cc: Gerd Hoffmann
---
OvmfPkg/Library/MemoryProtectionConfigLib/MemoryProtectionConfigLib.c | 118
QemuFwCfgParseString allows reading strings from the fw_cfg file.
This function is needed for parsing the memory protection profile
name.
Signed-off-by: Taylor Beebe
Cc: Ard Biesheuvel
Cc: Jiewen Yao
Cc: Jordan Justen
Cc: Gerd Hoffmann
---
Now that the EDK2 tree uses GetMemoryProtectionsLib to query
the platform memory protection settings, we can add additional
profiles to SetMemoryProtectionsLib to give plaforms more options
for setting memory protections.
Signed-off-by: Taylor Beebe
Cc: Jian J Wang
Cc: Liming Gao
---
Replace references to the memory protection PCDs to instead
check the platform protections via GetMemoryProtectionsLib.
Because the protection profile is equivalent to the PCD settings,
this updated does not cause a torn state.
Signed-off-by: Taylor Beebe
Cc: Eric Dong
Cc: Ray Ni
Cc: Rahul
Replace references to the memory protection PCDs to instead
check the platform protections via GetMemoryProtectionsLib.
Because the protection profile is equivalent to the PCD settings,
this updated does not cause a torn state.
Signed-off-by: Taylor Beebe
Cc: Andrew Fish
Cc: Ray Ni
---
Replace references to the memory protection PCDs to instead
check the platform protections via GetMemoryProtectionsLib.
Because the protection profile is equivalent to the PCD settings,
this updated does not cause a torn state.
Signed-off-by: Taylor Beebe
Cc: Ard Biesheuvel
Cc: Jiewen Yao
Cc:
Update the DXE handoff logic in MdeModulePkg to use
SetMemoryProtectionsLib to fetch the platform memory protection
settings and reference them when creating the page tables.
Because the protection profile is equivalent to the PCD settings
even when the platform does not explicitly set a profile,
Replace references to the memory protection PCDs to instead
check the platform protections via GetMemoryProtectionsLib.
Signed-off-by: Taylor Beebe
Cc: Leif Lindholm
Cc: Ard Biesheuvel
Cc: Sami Mujawar
---
ArmPkg/Drivers/CpuDxe/CpuDxe.c | 5 ++---
ArmPkg/ArmPkg.dsc| 1 +
Update the DXE handoff logic in UefiPayloadPkg to use
SetMemoryProtectionsLib to fetch the platform memory protection
settings and reference them when creating the page tables.
Because the protection profile is equivalent to the PCD settings
even when the platform does not explicitly set a
Use SetMemoryProtectionsLib to set the memory protections for
the platform in both normal and PEI-less boot. The protections
set are equivalent to the PCD settings and the ability to set
NxForStack via QemuCfg is preserved. Once the transition to use
SetMemoryProtectionsLib and
This patch updates the DXE Handoff in PEI-less Startup to use
SetMemoryProtectionsLib to get the platform memory protection settings
and build the page tables based on the applied protections.
Signed-off-by: Taylor Beebe
Cc: Ard Biesheuvel
Cc: Jiewen Yao
Cc: Jordan Justen
Cc: Gerd Hoffmann
Add library classes for SetMemoryProtectionsLib and
GetMemoryProtectionsLib to ArmVirtPkg.
Signed-off-by: Taylor Beebe
Cc: Ard Biesheuvel
Cc: Leif Lindholm
Cc: Sami Mujawar
Cc: Gerd Hoffmann
Cc: Julien Grall
---
ArmVirtPkg/ArmVirt.dsc.inc | 6 ++
1 file changed, 6 insertions(+)
diff
Add library classes for SetMemoryProtectionsLib and
GetMemoryProtectionsLib to OvmfPkg platfomrs.
Signed-off-by: Taylor Beebe
Cc: Ard Biesheuvel
Cc: Jiewen Yao
Cc: Jordan Justen
Cc: Gerd Hoffmann
Cc: Rebecca Cran
Cc: Peter Grehan
Cc: Corvin Köhne
Cc: Jianyong Wu
Cc: Anatol Belski
Cc:
Memory protection is not set in PEI and ingested during and
after DXE handoff. This paradigm means that the platform cannot
reliably query the stack guard setting during MpInit. Because the
execution path of PEI consistent and no third party
code is executed, setting the stack guard in MpInit on
When page/pool protections are active, the GCD sync process takes
quite a bit longer than normal. This behavior is primarily due to
a function which dumps the GCD memory map to the console. This
dump function runs only on DEBUG builds but will iterate through
the GCD memory map dozens of times
The SetMemoryProtectionsLib implementation has functionality for
setting protections based on a preset profile or a custom DXE/MM
profile passed in by the caller. The implementation also supports
locking the protections (tracked via an extra boolean stored
in the HOB entry) which prevents the
Because the platform memory protection settings will be stored
in the HOB, the HOB list should be marked read-only and non-executable
as soon as possible in boot.
This patch page-aligns the allocated HOB list in DXE and marks
it RO/NX during memory protection initialization.
Signed-off-by:
The NULL instances for GetMemoryProtectionsLib and
SetMemoryProtectionsLib just zero out the memory protections
structure effectively disabling memory protections.
Signed-off-by: Taylor Beebe
Cc: Jian J Wang
Cc: Liming Gao
---
HOB memory should not be written to in DXE phase. This patch
copies the PCD database from PEI into a new buffer so updates
to dynamic PCDs don't write to HOB memory.
Signed-off-by: Taylor Beebe
Cc: Jian J Wang
Cc: Liming Gao
Cc: Dandan Bi
---
MdeModulePkg/Universal/PCD/Dxe/Service.c | 6
v4:
-Update the memory protection profiles to align the allocated pools to the
tail guard by default (patch 20).
- Add a patch to create MemoryProtectionConfigLib which consolidates code
for parsing the fw_cfg for the memory protection profile strings (patch 22).
-Move the update to add
These headers provide settings definitions for memory protections,
settings profiles for easily enabling memory protections,
and the GUIDs used for producing the memory protection HOB entry.
The settings options are functionally 1:1 with the existing
PCD bitfield definitions. Instead of setting a
SetMemoryProtectionsLib is a PEIM which allows platforms to
apply memory protection settings to the current boot.
GetMemoryProtectionsLib has DXE and MM implementations to allow
platforms to query the current memory protection settings via a
global variable populated by the library
I do not think anyone has done the complete analysis since UEFI 2.7. Some
updates from newer specs
have been added as needed by different community members.
Unless there are dedicated tasks to review and verify that every spec change
has been added to edk2, the version can not be updated.
On Tue, Sep 19, 2023 at 14:18:01 +0100, Peter Maydell wrote:
> On Tue, 19 Sept 2023 at 12:20, Leif Lindholm
> wrote:
> >
> > An ASSERT trips when we try to add the NS-EL2 virtual timer to qemu
> > mach-virt.
> >
> > Add a new Pcd for the new private peripheral interrupt id,
> >
Pushed: https://github.com/tianocore/edk2-platforms/commit/cf8a716
-Original Message-
From: Huang, Yanbo
Sent: Tuesday, September 5, 2023 7:19 PM
To: devel@edk2.groups.io
Cc: Huang, Yanbo ; Isaac Oram ;
Desimone, Nathaniel L ; Gao, Liming
Subject: [PATCH v2] IpmiFeaturePkg: Refine
Reviewed-by: Nate DeSimone
-Original Message-
From: Huang, Yanbo
Sent: Tuesday, September 5, 2023 7:19 PM
To: devel@edk2.groups.io
Cc: Huang, Yanbo ; Isaac Oram ;
Desimone, Nathaniel L ; Gao, Liming
Subject: [PATCH v2] IpmiFeaturePkg: Refine code to avoid warning report
On Tue, 19 Sept 2023 at 16:19, Leif Lindholm wrote:
>
> On 2023-09-19 14:18, Peter Maydell wrote:
> > On Tue, 19 Sept 2023 at 12:20, Leif Lindholm
> > wrote:
> >>
> >> An ASSERT trips when we try to add the NS-EL2 virtual timer to qemu
> >> mach-virt.
> >>
> >> Add a new Pcd for the new private
Hi All,
Can you please share your thoughts if any idea in EDK2 compliance with UEFI 2.8
Support ?
From UEFI sources and latest edk2 master branch is still at 2.7. Same for
latest stable tag edk2-stable202308.
https://github.com/tianocore/edk2.git
#define EFI_SYSTEM_TABLE_REVISION
Gua may be able to help.
What error do you get trying to use stuart commands?
Mike
From: CrossedCarpet
Sent: Tuesday, September 19, 2023 9:45 AM
To: Kinney, Michael D ; devel@edk2.groups.io
Subject: Re: [edk2-devel] gTest code coverage
Thank you for your reply!
Unfortunately, I was aware of
In latest tianocore edk2 ( https://github.com/tianocore/edk2 ), we can see
System table is still pointing to 2.70 Version(EFI_2_70_SYSTEM_TABLE_REVISION)
https://github.com/tianocore/edk2/blob/master/MdePkg/Include/Uefi/UefiSpec.h
//
// EFI Runtime Services Table
//
#define
System Tbale is still pointing to Uefi 2.7 version.
Is the latest edk2 still not compliance with UEFI 2.8 ?
-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#108852): https://edk2.groups.io/g/devel/message/108852
Mute This Topic:
Thank you for your reply!
Unfortunately, I was aware of that documentation, I quoted it in my original
email when asking for help in its interpretation.
I have been using the build system instead of the stuart abstraction. If I try
using it in my current setup I get a bug saying my HostTestPkg
Thanks Leif, this unfortunate inclusion has already been remedied in the v2.
What happens is that random files in the tree get changed by virtue of just
doing a build, so it is very easy to accidentally pull it in. And it's a single
line that I missed manually looking over the patch.
Wrt
On 2023-09-19 14:18, Peter Maydell wrote:
On Tue, 19 Sept 2023 at 12:20, Leif Lindholm wrote:
An ASSERT trips when we try to add the NS-EL2 virtual timer to qemu
mach-virt.
Add a new Pcd for the new private peripheral interrupt id,
PcdArmArchTimerHypVirtIntrNum.
Update
My understanding is single patch is deemed accepted by one of the following
responses:
1) Maintainer provides Reviewed-by
2) Non-Maintainer provides Reviewed-by, and later a Maintainer provides
Acked-by to indicate they accept the review performed by someone else.
An Acked-by alone by
On 2023-09-19 15:46, Ard Biesheuvel wrote:
On Mon, 18 Sept 2023 at 22:35, Michael Kubacki
wrote:
Thanks Ard. I still need a Reviewed-by on this patch to complete the series.
Actually, you don't but I suppose there is some difference of
interpretation here.
In the Linux community (where
On Mon, 18 Sept 2023 at 22:35, Michael Kubacki
wrote:
>
> Thanks Ard. I still need a Reviewed-by on this patch to complete the series.
>
Actually, you don't but I suppose there is some difference of
interpretation here.
In the Linux community (where these tags originated), an ack from a
On 2023-09-19 15:31, Ard Biesheuvel wrote:
Hello Andrei,
On Tue, 19 Sept 2023 at 04:43, Andrei Warkentin
wrote:
InternalLongJump was not returning the 2nd parameter passed
to LongJmp (Value) as the return value from SetJmp.
Seen with code compiled with -Os, where an LongJmp (Buffer, -1)
From: Michael Kubacki
Adds a new module (dynamic shell command) to ShellPkg that lists
variable policy information for all UEFI variables on the system.
Some other UEFI variable related functionality is also included to
give a greater sense of platform UEFI variable state. This command
is
From: Michael Kubacki
Introduces two new APIs to EDKII_VARIABLE_POLICY_PROTOCOL:
1. GetVariablePolicyInfo()
2. GetLockOnVariableStateVariablePolicyInfo()
These allow a caller to retrieve policy information associated with
a UEFI variable given the variable name and vendor GUID.
Hello Andrei,
On Tue, 19 Sept 2023 at 04:43, Andrei Warkentin
wrote:
>
> InternalLongJump was not returning the 2nd parameter passed
> to LongJmp (Value) as the return value from SetJmp.
>
> Seen with code compiled with -Os, where an LongJmp (Buffer, -1)
> somehow translated to SetJmp returning
On 2023-09-19 13:03, Marcin Juszkiewicz wrote:
W dniu 18.09.2023 o 13:35, Leif Lindholm pisze:
On Mon, Sep 18, 2023 at 13:03:12 +0200, Marcin Juszkiewicz wrote:
There are some changes in progress which make BL1 bigger than 0x8000
which EDK2 uses.
TF-A defines BL1 size to be 0x12000 one. So
On Tue, 19 Sept 2023 at 13:20, Leif Lindholm wrote:
>
> An ASSERT trips when we try to add the NS-EL2 virtual timer to qemu
> mach-virt.
>
> Add a new Pcd for the new private peripheral interrupt id,
> PcdArmArchTimerHypVirtIntrNum.
>
> Update ArmVirtTimerFdtClientLib to:
> - Only assert on
Thanks, Mike! I will add your acked-by to commit message and merge it.
Regards,
Nickle
> -Original Message-
> From: Mike Maslenkin
> Sent: Tuesday, September 19, 2023 9:54 PM
> To: Nickle Wang
> Cc: devel@edk2.groups.io; Abner Chang ; Igor
> Kulchytskyy ; Nick Ramirez
> Subject: Re:
Dear All,
Can some help in clarifying the what is the current UEFI Spec version support
in latest tianocore edk2
In latest tianocore edk2 (https://github.com/tianocore/edk2 ), we can see
System table is pointing to 2.70 Version(EFI_2_70_SYSTEM_TABLE_REVISION)
On Tue, 19 Sept 2023 at 12:20, Leif Lindholm wrote:
>
> An ASSERT trips when we try to add the NS-EL2 virtual timer to qemu
> mach-virt.
>
> Add a new Pcd for the new private peripheral interrupt id,
> PcdArmArchTimerHypVirtIntrNum.
>
> Update ArmVirtTimerFdtClientLib to:
> - Only assert on
Hi all,
There is a test for prebuild Rust library link with UEFI C code in GCC build as
below. But I will encounter some build problems.
Case 1: prebuild Rust library with aarch64-unknown-uefi target command, it will
encounter "error adding symbols: file format not recognized".
Case 2: I tried
As written in the error message I'm sending the call stack.
(Python 3.8.10 on linux) Traceback (most recent call last):
File
"/home/pliops/development/edk2.git/edk2/BaseTools/BinWrappers/PosixLike/../../Source/Python/build/build.py",
line 2692, in Main
MyBuild = Build(Target, Workspace,
Hi Nickle,
patch looks good to me.
You can add my acked-by/reviewed-by if you wish.
Thanks,
Mike.
On Mon, Sep 18, 2023 at 4:15 PM Nickle Wang wrote:
>
> Hi Mike,
>
> Sorry for my late response. Version2 is here:
> https://edk2.groups.io/g/devel/message/108778 Please help me to review it
>
On 9/19/23 02:59, Gerd Hoffmann wrote:
On Fri, Jul 14, 2023 at 03:28:26PM -0500, Tom Lendacky wrote:
Commit 63c50d3ff2854a76432b752af4f2a76f33ff1974 changed the check that is
used to determine if SEV-ES is active. Originally, a CMP instruction with
a supporting JZ instruction was used for the
On Mon, 18 Sept 2023 at 15:06, Marcin Juszkiewicz
wrote:
>
> W dniu 18.09.2023 o 14:33, Ard Biesheuvel pisze:
> > On Mon, 18 Sept 2023 at 13:35, Leif Lindholm
> > wrote:
>
> >> Note for the interested:
> >> https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/23417
> >> TL;DR: the fip
W dniu 18.09.2023 o 13:35, Leif Lindholm pisze:
On Mon, Sep 18, 2023 at 13:03:12 +0200, Marcin Juszkiewicz wrote:
There are some changes in progress which make BL1 bigger than 0x8000
which EDK2 uses.
TF-A defines BL1 size to be 0x12000 one. So let follow it.
Signed-off-by: Marcin Juszkiewicz
On Tue, Sep 19, 2023 at 03:16:59AM -0500, Andrei Warkentin wrote:
> InternalLongJump was not returning the 2nd parameter passed
> to LongJmp (Value) as the return value from SetJmp.
>
> Seen with code compiled with -Os, where an LongJmp (Buffer, -1)
> somehow translated to SetJmp returning 0...
>
FEAT_VHE, introduced in ARMv8.1, adds a virtual EL2 timer.
However, this library verifies that exactly 3 or 4 12-byte timer
interrupts are provided in input DT, ASSERTing when the new timer
is added.
Change the assert to >= 36.
Extend the current logic, also initializing
PcdArmArchTimerHypVirtIntrNum, defaults to 28 as defined in Arm's
Base System Architecture (current version DEN0094C 1.0C).
Relevant for systems implementing FEAT_VHE, added in ARMv8.1.
Signed-off-by: Leif Lindholm
Cc: Ard Biesheuvel
Cc: Sami Mujawar
Signed-off-by: Leif Lindholm
---
An ASSERT trips when we try to add the NS-EL2 virtual timer to qemu
mach-virt.
Add a new Pcd for the new private peripheral interrupt id,
PcdArmArchTimerHypVirtIntrNum.
Update ArmVirtTimerFdtClientLib to:
- Only assert on receiving less information that required through DT.
- Set
InternalLongJump was not returning the 2nd parameter passed
to LongJmp (Value) as the return value from SetJmp.
Seen with code compiled with -Os, where an LongJmp (Buffer, -1)
somehow translated to SetJmp returning 0...
Cc: Yong Li
Cc: Sunil V L
Cc: Tuan Phan
Cc: Daniel Schaefer
I was playing around with building code with -Os and ran into
weird crashes, that I ended up chasing down to an obviously
incorrect InternalLongJump implementation, which has never
been correctly passing its 2nd parameter down as the return
value of SetJump.
You can find the pull at
My mistake. Didn't see that pulled in... Will rework.
От: Yao, Jiewen
Отправлено: Tuesday, September 19, 2023 3:01:34 AM
Кому: devel@edk2.groups.io ; Warkentin, Andrei
Копия: Li, Yong ; Sunil V L ; Tuan
Phan ; Daniel Schaefer
Тема: RE: [edk2-devel] [PATCH v1
I am OK for the RISC-V change.
Would you please let me know why we need openssl submodule ?
> -Original Message-
> From: devel@edk2.groups.io On Behalf Of Andrei
> Warkentin
> Sent: Tuesday, September 19, 2023 12:43 PM
> To: devel@edk2.groups.io
> Cc: Warkentin, Andrei ; Li, Yong
> ;
On Fri, Jul 14, 2023 at 03:28:26PM -0500, Tom Lendacky wrote:
> Commit 63c50d3ff2854a76432b752af4f2a76f33ff1974 changed the check that is
> used to determine if SEV-ES is active. Originally, a CMP instruction with
> a supporting JZ instruction was used for the check. It was changed to use
> the BT
72 matches
Mail list logo
