public inbox for devel@edk2.groups.io
 help / color / mirror / Atom feed
From: Andrew Fish <afish@apple.com>
To: Rafael Machado <rafaelrodrigues.machado@gmail.com>
Cc: "edk2-devel@lists.01.org" <edk2-devel@lists.01.org>
Subject: Re: DMA Buffer write operation not persisted
Date: Thu, 25 Jan 2018 11:04:31 -0800	[thread overview]
Message-ID: <EB2CED4C-0EAC-42E3-BFBF-2E6F2EA96D62@apple.com> (raw)
In-Reply-To: <CACgnt79qcR1L1KpA6sX7=Yi_t4_=11Cb3-ATJns0w7P13mWfeA@mail.gmail.com>

Rafeal,

There are some good summaries in the UEFI Spec that really help. 

DMA Bus Master Common Buffer Operation
• Call AllocateBuffer() to allocate a common buffer.
• Call Map() for EfiPciOperationBusMasterCommonBuffer or EfiPciOperationBusMasterCommonBuffer64.
• Program the DMA Bus Master with the DeviceAddress returned by Map().
• The common buffer can now be accessed equally by the processor and the DMA bus master.
• Call Unmap().
• Call FreeBuffer().

Did you miss the PciIo->AllocateBuffer() call?

For x86 it can abstract things like DMA only supported < 4GB. 
For ARM it may need to change the cacheability of the region etc. 

Thanks,

Andrew Fish

> On Jan 25, 2018, at 10:53 AM, Rafael Machado <rafaelrodrigues.machado@gmail.com> wrote:
> 
> Hi everyone.
> 
> I'm currently work on a task, and I need to write some data at a DMA buffer.
> At the UEFI Driver Writer's guide, at page 359 there is a sample code of
> how to do that.
> 
> Considering that code and adapting to my scenario I got the following
> function (some debug prints are present for clarification):
> 
> EFI_STATUS
> EFIAPI
> DoBusMasterWrite (
>  IN EFI_PCI_ROOT_BRIDGE_IO_PROTOCOL *PciIo,
>  IN UINT8 *HostAddress,
>  IN UINTN *Length,
>  IN UINT32 Value
> )
> {
>  EFI_STATUS Status;
>  UINTN NumberOfBytes;
>  EFI_PHYSICAL_ADDRESS DeviceAddress;
>  VOID *Mapping;
>  UINT64 ReadValue;
> 
>  //
>  // Call Map() to retrieve the DeviceAddress to use for the bus
>  // master write operation. The Map() function may not support
>  // performing a DMA operation for the entire length, so it may
>  // be broken up into smaller DMA operations.
>  //
>  NumberOfBytes = *Length;
>  Status = PciIo->Map (PciIo, // This
> 
> EfiPciIoOperationBusMasterCommonBuffer, // Operation
>                                     (VOID *)HostAddress, // HostAddress
>                                     &NumberOfBytes, // NumberOfBytes
>                                     &DeviceAddress, //DeviceAddress
>                                     &Mapping //Mapping);
> 
>  if (EFI_ERROR (Status)) {
>    return Status;
>  }
> 
>  //
>  // Write the data to the desired address
>  // This write operation also starts the DMA transaction
>  //
>  Status = PciIo->Mem.Write (PciIo, // This
>                                               EfiPciIoWidthUint32, // Width
>                                               *HostAddress,
>                                               1, // Count
>                                               &Value // Buffer
>                                               );
> 
>  Print(L"NumberOfBytes: %d\r\n", NumberOfBytes);
>  Print(L"address: 0x%x\r\n", HostAddress);
>  Print(L"Value: 0x%x\r\n", Value);
>  Print(L"Status: %r\r\n", Status);
> 
>  if (EFI_ERROR (Status)) {
> return Status;
>  }
> 
>  //
>  // The operations performed by PollMem() also flush all posted
>  // writes from the PCI bus master and through PCI-to-PCI bridges.
>  //
>  Status = PciIo->PollMem (PciIo, // This
>                                               EfiPciIoWidthUint32, // Width
>                                               *HostAddress, // Offset
>                                               0xFFFFFFFF,// Mask
>                                               Value,// Value
>                                               EFI_TIMER_PERIOD_SECONDS
> (1), // Timeout
>                                               &ReadValue // Result
>                                               );
>  Print(L"Status2: %r\r\n", Status);
> 
>  if (EFI_ERROR (Status)) {
> return Status;
>  }
> 
>  //
>  // Call Flush() to flush all write transactions to system memory
>  //
>  Status = PciIo->Flush (PciIo);
>  Print(L"Status3: %r\r\n", Status);
> 
>  if (EFI_ERROR (Status)) {
> return Status;
>  }
> 
>  //
>  // Call Unmap() to complete the bus master write operation
>  //
>  Status = PciIo->Unmap (PciIo, Mapping);
>  Print(L"Status4: %r\r\n", Status);
> 
>  if (EFI_ERROR (Status)) {
> return Status;
>  }
>  return Status;
> }
> 
> The output of this function is this:
>  NumberOfBytes: 4
>  address: 0xCCCAC000
>  Value: 0xAAAAA
>  Status: Success
>  Status2: Success
>  Status3: Success
>  Status4: Success
> 
> The problem is that when I try to read this memory content using the dmem
> command at the efiShell the value 0xAAAAA cannot be found. Seems something
> is locking the DMA trasaction.
> Can someone give me some light?
> 
> Thanks and Regard
> Rafael R. Machado
> _______________________________________________
> edk2-devel mailing list
> edk2-devel@lists.01.org
> https://lists.01.org/mailman/listinfo/edk2-devel



  reply	other threads:[~2018-01-25 18:59 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-01-25 18:53 DMA Buffer write operation not persisted Rafael Machado
2018-01-25 19:04 ` Andrew Fish [this message]
2018-01-25 19:15   ` Rafael Machado
2018-01-25 19:52 ` Paulo Alcantara

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=EB2CED4C-0EAC-42E3-BFBF-2E6F2EA96D62@apple.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