public inbox for devel@edk2.groups.io
 help / color / mirror / Atom feed
From: "Omkar Anand Kulkarni" <omkar.kulkarni@arm.com>
To: "devel@edk2.groups.io" <devel@edk2.groups.io>,
	"huangming@linux.alibaba.com" <huangming@linux.alibaba.com>,
	Sami Mujawar <Sami.Mujawar@arm.com>,
	"ardb+tianocore@kernel.org" <ardb+tianocore@kernel.org>,
	"jiewen.yao@intel.com" <jiewen.yao@intel.com>,
	Supreeth Venkatesh <Supreeth.Venkatesh@arm.com>
Cc: "ming.huang-@outlook.com" <ming.huang-@outlook.com>
Subject: Re: [edk2-devel] [PATCH edk2 v1 3/3] StandaloneMmPkg: Fix check buffer address failed issue from TF-A
Date: Wed, 8 Dec 2021 17:46:13 +0000	[thread overview]
Message-ID: <DBAPR08MB578115A6383DF6DA1DCB18D7926F9@DBAPR08MB5781.eurprd08.prod.outlook.com> (raw)
In-Reply-To: <20211015090623.52511-4-huangming@linux.alibaba.com>

Hi Ming,

Thanks for this patch. This patch helps to resolve Standalone MM issue while exercising RAS use case.
Few comments mentioned inline.

- Omkar


 On 10/15/21 2:39 PM, Ming Huang via groups.io wrote:
> There are two scene communicate with StandaloneMm(MM):
> 1 edk2 -> TF-A -> MM, communicate MM use non-secure buffer which
>   specify by EFI_SECURE_PARTITION_BOOT_INFO.SpNsCommBufBase;
> 2 RAS scene: fiq -> TF-A -> MM, use secure buffer which
>   specify by EFI_SECURE_PARTITION_BOOT_INFO.SpShareBufBase;
>
> For now, the second scene will failed because check buffer address.
> This patch add CheckBufferAddr() to support check address for secure buffer.
>
> Signed-off-by: Ming Huang <huangming@linux.alibaba.com>
> ---
>  StandaloneMmPkg/Drivers/StandaloneMmCpu/EventHandle.c     | 70
> ++++++++++++++++----
>  StandaloneMmPkg/Drivers/StandaloneMmCpu/StandaloneMmCpu.c | 21
> ++++++
> StandaloneMmPkg/Drivers/StandaloneMmCpu/StandaloneMmCpu.h |  1 +
>  3 files changed, 79 insertions(+), 13 deletions(-)
>
> diff --git a/StandaloneMmPkg/Drivers/StandaloneMmCpu/EventHandle.c
> b/StandaloneMmPkg/Drivers/StandaloneMmCpu/EventHandle.c
> index 5dfaf9d751..63fab1bd78 100644
> --- a/StandaloneMmPkg/Drivers/StandaloneMmCpu/EventHandle.c
> +++ b/StandaloneMmPkg/Drivers/StandaloneMmCpu/EventHandle.c
> @@ -50,6 +50,7 @@ EFI_MM_COMMUNICATE_HEADER
> **PerCpuGuidedEventContext = NULL;
>
>  // Descriptor with whereabouts of memory used for communication with
> the normal world  EFI_MMRAM_DESCRIPTOR  mNsCommBuffer;
> +EFI_MMRAM_DESCRIPTOR  mSCommBuffer;
>
>  MP_INFORMATION_HOB_DATA *mMpInformationHobData;
>
> @@ -60,6 +61,58 @@ EFI_MM_CONFIGURATION_PROTOCOL mMmConfig =
> {
>
>  STATIC EFI_MM_ENTRY_POINT     mMmEntryPoint = NULL;
>
> +STATIC
> +EFI_STATUS
> +CheckBufferAddr (
> +  IN UINTN CommBufferAddr
> +  )
> +{
> +  UINTN      CommBufferSize;
> +  EFI_STATUS Status;
> +
> +  Status =  EFI_SUCCESS;
> +  if (CommBufferAddr < mNsCommBuffer.PhysicalStart) {
> +    Status = EFI_ACCESS_DENIED;
> +  }
> +
> +  if ((CommBufferAddr + sizeof (EFI_MM_COMMUNICATE_HEADER)) >=
> +      (mNsCommBuffer.PhysicalStart + mNsCommBuffer.PhysicalSize)) {
> +    Status =  EFI_INVALID_PARAMETER;

Single space after "Status = "

- Omkar


> +  }
> +
> +  // Find out the size of the buffer passed  CommBufferSize =
> + ((EFI_MM_COMMUNICATE_HEADER *) CommBufferAddr)->MessageLength
> +
> +    sizeof (EFI_MM_COMMUNICATE_HEADER);
> +
> +  // perform bounds check.
> +  if (CommBufferAddr + CommBufferSize >=
> +      mNsCommBuffer.PhysicalStart + mNsCommBuffer.PhysicalSize) {
> +    Status =  EFI_ACCESS_DENIED;

Single space after "Status = "

- Omkar

> +  }
> +
> +  if (!EFI_ERROR (Status)) {


In case of error this function call will not return from here. It will execute the code below comparing the MM Communicate buffer address with the Secure buffer address, which may cause wrong return type being returned. Can you check this, please?

- Omkar


> +    return EFI_SUCCESS;
> +  }
> +
> +  Status =  EFI_SUCCESS;
> +  if (CommBufferAddr < mSCommBuffer.PhysicalStart) {
> +    Status = EFI_ACCESS_DENIED;
> +  }
> +
> +  if ((CommBufferAddr + sizeof (EFI_MM_COMMUNICATE_HEADER)) >=
> +      (mSCommBuffer.PhysicalStart + mSCommBuffer.PhysicalSize)) {
> +    Status =  EFI_INVALID_PARAMETER;
> +  }
> +
> +  // perform bounds check.
> +  if (CommBufferAddr + CommBufferSize >=
> +      mSCommBuffer.PhysicalStart + mSCommBuffer.PhysicalSize) {
> +    Status =  EFI_ACCESS_DENIED;
> +  }
> +
> +  return Status;
> +}
> +


CheckBufferAddr() function performs validity and overflow checks on the Communication buffers. These checks are same for both the non-secure
MM communicate buffer and secure buffer shared between EL3 and S-EL0. Can this code be combined ( example below)? This will help mitigate the above mentioned return type issue as well.

STATIC
EFI_STATUS
CheckBufferAddr (
  IN UINTN CommBufferAddr
  )
{
  UINTN                CommBufferSize;
  EFI_STATUS           Status;
  EFI_MMRAM_DESCRIPTOR CommBuffer;

  if (CommBufferAddr < mNsCommBuffer.PhysicalStart ||
      CommBufferAddr > (mNsCommBuffer.PhysicalStart + mNsCommBuffer.PhysicalSize)) {
    CommBuffer = mSCommBuffer;
  } else {
    CommBuffer = mNsCommBuffer;
  }

  if (CommBufferAddr < CommBuffer.PhysicalStart) {
    Status = EFI_ACCESS_DENIED;
  }

  if ((CommBufferAddr + sizeof (EFI_MM_COMMUNICATE_HEADER)) >=
      (CommBuffer.PhysicalStart + CommBuffer.PhysicalSize)) {
    Status = EFI_INVALID_PARAMETER;
  }

  // Find out the size of the buffer passed
  CommBufferSize = ((EFI_MM_COMMUNICATE_HEADER *) CommBufferAddr)->MessageLength +
    sizeof (EFI_MM_COMMUNICATE_HEADER);

  // perform bounds check.
  if (CommBufferAddr + CommBufferSize >=
      CommBuffer.PhysicalStart + CommBuffer.PhysicalSize) {
    Status = EFI_ACCESS_DENIED;
  }

  return Status;
}

- Omkar


>  /**
>    The PI Standalone MM entry point for the TF-A CPU driver.
>
> @@ -104,25 +157,16 @@ PiMmStandaloneArmTfCpuDriverEntry (
>      return EFI_INVALID_PARAMETER;
>    }
>
> -  if (NsCommBufferAddr < mNsCommBuffer.PhysicalStart) {
> -    return EFI_ACCESS_DENIED;
> -  }
> -
> -  if ((NsCommBufferAddr + sizeof (EFI_MM_COMMUNICATE_HEADER)) >=
> -      (mNsCommBuffer.PhysicalStart + mNsCommBuffer.PhysicalSize)) {
> -    return EFI_INVALID_PARAMETER;
> +  Status = CheckBufferAddr (NsCommBufferAddr);  if (EFI_ERROR (Status))
> + {
> +    DEBUG ((DEBUG_ERROR, "Check Buffer failed: %r\n", Status));
> +    return Status;
>    }
>
>    // Find out the size of the buffer passed
>    NsCommBufferSize = ((EFI_MM_COMMUNICATE_HEADER *)
> NsCommBufferAddr)->MessageLength +
>      sizeof (EFI_MM_COMMUNICATE_HEADER);
>
> -  // perform bounds check.
> -  if (NsCommBufferAddr + NsCommBufferSize >=
> -      mNsCommBuffer.PhysicalStart + mNsCommBuffer.PhysicalSize) {
> -    return EFI_ACCESS_DENIED;
> -  }
> -
>    GuidedEventContext = NULL;
>    // Now that the secure world can see the normal world buffer, allocate
>    // memory to copy the communication buffer to the secure world.
> diff --git
> a/StandaloneMmPkg/Drivers/StandaloneMmCpu/StandaloneMmCpu.c
> b/StandaloneMmPkg/Drivers/StandaloneMmCpu/StandaloneMmCpu.c
> index fd9c59b4da..96dad20dd1 100644
> --- a/StandaloneMmPkg/Drivers/StandaloneMmCpu/StandaloneMmCpu.c
> +++ b/StandaloneMmPkg/Drivers/StandaloneMmCpu/StandaloneMmCpu.c
> @@ -107,6 +107,7 @@ StandaloneMmCpuInitialize (
>    UINTN                            Index;
>    UINTN                            ArraySize;
>    VOID                            *HobStart;
> +  EFI_MMRAM_HOB_DESCRIPTOR_BLOCK  *MmramRangesHob;
>
>    ASSERT (SystemTable != NULL);
>    mMmst = SystemTable;
> @@ -186,6 +187,26 @@ StandaloneMmCpuInitialize (
>    CopyMem (&mNsCommBuffer, NsCommBufMmramRange,
> sizeof(EFI_MMRAM_DESCRIPTOR));
>    DEBUG ((DEBUG_INFO, "mNsCommBuffer: 0x%016lx - 0x%lx\n",
> mNsCommBuffer.CpuStart, mNsCommBuffer.PhysicalSize));
>
> +  Status = GetGuidedHobData (
> +             HobStart,
> +             &gEfiMmPeiMmramMemoryReserveGuid,
> +             (VOID **) &MmramRangesHob
> +             );
> +  if (EFI_ERROR (Status)) {
> +    DEBUG ((DEBUG_ERROR, "MmramRangesHob data extraction failed -
> 0x%x\n", Status));
> +    return Status;
> +  }
> +
> +  //
> +  // As CreateHobListFromBootInfo(), the base and size of buffer shared
> + with  // privileged Secure world software is in second one.
> +  //
> +  CopyMem (
> +    &mSCommBuffer,
> +    &MmramRangesHob->Descriptor[0] + 1,

Can this be changed to
&MmramRangesHob->Descriptor[1],

- Omkar

> +    sizeof(EFI_MMRAM_DESCRIPTOR)
> +  );
> +
>    //
>    // Extract the MP information from the Hoblist
>    //
> diff --git
> a/StandaloneMmPkg/Drivers/StandaloneMmCpu/StandaloneMmCpu.h
> b/StandaloneMmPkg/Drivers/StandaloneMmCpu/StandaloneMmCpu.h
> index 2c96439c15..2e03b20d85 100644
> --- a/StandaloneMmPkg/Drivers/StandaloneMmCpu/StandaloneMmCpu.h
> +++ b/StandaloneMmPkg/Drivers/StandaloneMmCpu/StandaloneMmCpu.h
> @@ -30,6 +30,7 @@ extern EFI_MM_CPU_PROTOCOL mMmCpuState;  //
>  extern EFI_MM_COMMUNICATE_HEADER    **PerCpuGuidedEventContext;
>  extern EFI_MMRAM_DESCRIPTOR          mNsCommBuffer;
> +extern EFI_MMRAM_DESCRIPTOR          mSCommBuffer;
>  extern MP_INFORMATION_HOB_DATA       *mMpInformationHobData;
>  extern EFI_MM_CONFIGURATION_PROTOCOL mMmConfig;
>
> --
> 2.17.1
>
>
>
> 
>

IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.

  reply	other threads:[~2021-12-08 17:46 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-10-15  9:06 [PATCH edk2 v1 0/3] Fix several issues in StanaloneMmPkg Ming Huang
2021-10-15  9:06 ` [PATCH edk2 v1 1/3] StandaloneMmPkg: Fix issue about SpPcpuSharedBufSize field Ming Huang
2021-12-15  8:44   ` Ard Biesheuvel
2021-12-16  8:07     ` Masahisa Kojima
2021-12-16 10:05       ` [edk2-devel] " Jeff Fan
2021-12-16 11:09       ` Ard Biesheuvel
2021-10-15  9:06 ` [PATCH edk2 v1 2/3] StandaloneMmPkg: Replace DEBUG_INFO with DEBUG_ERROR Ming Huang
2021-10-15  9:06 ` [PATCH edk2 v1 3/3] StandaloneMmPkg: Fix check buffer address failed issue from TF-A Ming Huang
2021-12-08 17:46   ` Omkar Anand Kulkarni [this message]
2021-12-15 15:02     ` [edk2-devel] " Ming Huang
2021-12-21 14:59     ` Ming Huang
  -- strict thread matches above, loose matches on Subject: below --
2021-12-16  9:15 Marvin Häuser
2021-12-23 10:46 ` Ming Huang
2021-12-23 11:05   ` Marvin Häuser
2021-12-24  1:18     ` Ming Huang
2021-12-24 13:52       ` Marvin Häuser
2021-12-25  2:09         ` Ming Huang
2021-12-30 12:27           ` Marvin Häuser
2021-12-31 10:49             ` Ming Huang

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=DBAPR08MB578115A6383DF6DA1DCB18D7926F9@DBAPR08MB5781.eurprd08.prod.outlook.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