From: Laszlo Ersek <lersek@redhat.com>
To: Jian J Wang <jian.j.wang@intel.com>, edk2-devel@lists.01.org
Cc: Star Zeng <star.zeng@intel.com>, Eric Dong <eric.dong@intel.com>,
Jiewen Yao <jiewen.yao@intel.com>,
Michael Kinney <michael.d.kinney@intel.com>,
Jordan Justen <jordan.l.justen@intel.com>,
Ayellet Wolman <ayellet.wolman@intel.com>
Subject: Re: [PATCH v2 6/6] OvmfPkg/QemuVideoDxe: Bypass NULL pointer detection during VBE SHIM installing
Date: Fri, 22 Sep 2017 13:50:23 +0200 [thread overview]
Message-ID: <6ed8b38e-3387-e3a1-972f-5922470ed4c7@redhat.com> (raw)
In-Reply-To: <20170921052032.13652-7-jian.j.wang@intel.com>
This patch looks great to me, I would like to request a few small updates:
On 09/21/17 07:20, Jian J Wang wrote:
> QemuVideoDxe driver will install VBE SHIM into page 0. If NULL pointer
(1) please replace the word "install" with "link".
The VBE Shim is technically installed into the "real-mode" C segment,
only the int 0x10 vector lives in page 0.
> detection is enabled, this driver will fail to load. NULL pointer detection
> bypassing code is added to prevent such problem during boot.
>
> Please note that Windows 7 will try to access VBE SHIM during boot if it's
> installed, and then cause boot failure. This can be fixed by setting BIT7
> of PcdNullPointerDetectionPropertyMask to disable NULL pointer detection
> after EndOfDxe. As far as we know, there's no other OSs has such issue.
This is not a request, just a comment: I verified the default value in
the .dec, and I see it is 0. So there's no need to post an additional
patch for the OVMF DSC files, in order to set BIT7.
>
> Cc: Star Zeng <star.zeng@intel.com>
> Cc: Eric Dong <eric.dong@intel.com>
> Cc: Laszlo Ersek <lersek@redhat.com>
> Cc: Jiewen Yao <jiewen.yao@intel.com>
> Cc: Michael Kinney <michael.d.kinney@intel.com>
> Cc: Jordan Justen <jordan.l.justen@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>
> ---
> OvmfPkg/QemuVideoDxe/QemuVideoDxe.inf | 2 ++
> OvmfPkg/QemuVideoDxe/VbeShim.c | 22 ++++++++++++++++++++++
> 2 files changed, 24 insertions(+)
>
> diff --git a/OvmfPkg/QemuVideoDxe/QemuVideoDxe.inf b/OvmfPkg/QemuVideoDxe/QemuVideoDxe.inf
> index 577e07b0a8..8078232ded 100644
> --- a/OvmfPkg/QemuVideoDxe/QemuVideoDxe.inf
> +++ b/OvmfPkg/QemuVideoDxe/QemuVideoDxe.inf
> @@ -67,6 +67,7 @@
> UefiBootServicesTableLib
> UefiDriverEntryPoint
> UefiLib
> + DxeServicesTableLib
>
> [Protocols]
> gEfiDriverSupportedEfiVersionProtocolGuid # PROTOCOL ALWAYS_PRODUCED
> @@ -77,3 +78,4 @@
> [Pcd]
> gOptionRomPkgTokenSpaceGuid.PcdDriverSupportedEfiVersion
> gUefiOvmfPkgTokenSpaceGuid.PcdOvmfHostBridgePciDevId
> + gEfiMdeModulePkgTokenSpaceGuid.PcdNullPointerDetectionPropertyMask
> diff --git a/OvmfPkg/QemuVideoDxe/VbeShim.c b/OvmfPkg/QemuVideoDxe/VbeShim.c
> index e45a08e887..c3fb6d8d3c 100644
> --- a/OvmfPkg/QemuVideoDxe/VbeShim.c
> +++ b/OvmfPkg/QemuVideoDxe/VbeShim.c
> @@ -21,10 +21,13 @@
> WARRANTIES OR REPRESENTATIONS OF ANY KIND, EITHER EXPRESS OR IMPLIED.
> **/
>
> +#include <Pi/PiDxeCis.h>
(2) Question: what exactly is this necessary for?
(I would think that "DxeServicesTableLib.h" gave you everything you
needed, but I could be wrong.)
> #include <IndustryStandard/LegacyVgaBios.h>
> #include <Library/DebugLib.h>
> #include <Library/PciLib.h>
> #include <Library/PrintLib.h>
> +#include <Library/DxeServicesTableLib.h>
> +
> #include <OvmfPlatforms.h>
>
> #include "Qemu.h"
> @@ -74,11 +77,21 @@ InstallVbeShim (
> UINT8 *Ptr;
> UINTN Printed;
> VBE_MODE_INFO *VbeModeInfo;
> + EFI_STATUS Status;
>
> Segment0 = 0x00000;
> SegmentC = 0xC0000;
> SegmentF = 0xF0000;
>
> + //
> + // Disable NULL pointer detection temporarily. Otherwise the installation
> + // will fail due to the lack of memory access right.
> + //
> + if ((PcdGet8 (PcdNullPointerDetectionPropertyMask) & (BIT0|BIT7)) == BIT0) {
> + Status = gDS->SetMemorySpaceAttributes (0, EFI_PAGES_TO_SIZE (1), 0);
> + ASSERT_EFI_ERROR (Status);
> + }
> +
(3) Please hoist the
Segment0Pages = 1;
assignment from below, and use it in the SetMemorySpaceAttributes()
call, for the "Length" argument.
(4) Please use the variable "Segment0" as the "BaseAddress" argument.
(5) The Attributes=0 argument surprises me, and the end of this patch
seems to confirm that I'm right to be surprised :)
Instead of setting 0, can you please
- first get the original attributes with GetMemorySpaceDescriptor(),
- then clear only the attributes here that prevent read/write access,
- and at the end of the function, restore the original attributes?
I think this can be done without dynamic memory allocation, you just
need a local EFI_GCD_MEMORY_SPACE_DESCRIPTOR object.
> //
> // Attempt to cover the real mode IVT with an allocation. This is a UEFI
> // driver, hence the arch protocols have been installed previously. Among
> @@ -304,5 +317,14 @@ InstallVbeShim (
> Int0x10->Segment = (UINT16) ((UINT32)SegmentC >> 4);
> Int0x10->Offset = (UINT16) ((UINTN) (VbeModeInfo + 1) - SegmentC);
>
> + //
> + // Get NULL pointer detection back
> + //
> + if ((PcdGet8 (PcdNullPointerDetectionPropertyMask) & (BIT0|BIT7)) == BIT0) {
> + Status = gDS->SetMemorySpaceAttributes (0, EFI_PAGES_TO_SIZE(1),
> + EFI_MEMORY_RP);
> + ASSERT_EFI_ERROR (Status);
> + }
> +
> DEBUG ((EFI_D_INFO, "%a: VBE shim installed\n", __FUNCTION__));
> }
>
(6) The InstallVbeShim() function actually contains *two* earlier exits
than this. Please search the function for "return" statements.
I suggest the following:
- put the restoration of the original page attributes at the very end of
the function,
- put a label called "RestoreSegment0Attributes" between the DEBUG
message and the page attributes restoration,
- replace the "return" statements with "goto RestoreSegment0Attributes".
Thanks!
Laszlo
next prev parent reply other threads:[~2017-09-22 11:47 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <Add NULL pointer detection feature>
2017-09-21 5:20 ` [PATCH v2 0/6] Add NULL pointer detection feature Jian J Wang
2017-09-21 5:20 ` [PATCH v2 1/6] MdeModulePkg/MdeModulePkg.dec: Add NULL pointer detection PCD Jian J Wang
2017-09-25 8:01 ` Zeng, Star
2017-09-25 8:25 ` Wang, Jian J
2017-09-21 5:20 ` [PATCH v2 2/6] MdeModulePkg/DxeIpl: Implement NULL pointer detection Jian J Wang
2017-09-25 8:50 ` Yao, Jiewen
2017-09-26 0:54 ` Wang, Jian J
2017-09-21 5:20 ` [PATCH v2 3/6] UefiCpuPkg/PiSmmCpuDxeSmm: Implement NULL pointer detection for SMM code Jian J Wang
2017-09-21 5:20 ` [PATCH v2 4/6] MdeModulePkg/Core/Dxe: Add EndOfDxe workaround for NULL pointer detection Jian J Wang
2017-09-21 5:20 ` [PATCH v2 5/6] IntelFrameworkModulePkg/Csm: Add code to bypass " Jian J Wang
2017-09-21 5:20 ` [PATCH v2 6/6] OvmfPkg/QemuVideoDxe: Bypass NULL pointer detection during VBE SHIM installing Jian J Wang
2017-09-22 11:50 ` Laszlo Ersek [this message]
2017-09-22 15:29 ` Laszlo Ersek
2017-09-25 1:04 ` Wang, Jian J
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-list from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=6ed8b38e-3387-e3a1-972f-5922470ed4c7@redhat.com \
--to=devel@edk2.groups.io \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox