public inbox for devel@edk2.groups.io
 help / color / mirror / Atom feed
From: "Zeng, Star" <star.zeng@intel.com>
To: Ard Biesheuvel <ard.biesheuvel@linaro.org>,
	edk2-devel-01 <edk2-devel@lists.01.org>,
	"Tian, Feng" <feng.tian@intel.com>,
	"Gao, Liming" <liming.gao@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>,
	Leif Lindholm <leif.lindholm@linaro.org>,
	"Ni, Ruiyu" <ruiyu.ni@intel.com>
Subject: Re: [PATCH 0/7] MdeModulePkg ArmVirtPkg: fixes for 64-bit PCI DMA
Date: Tue, 6 Sep 2016 08:04:44 +0000	[thread overview]
Message-ID: <0C09AFA07DD0434D9E2A0C6AEB048310394B7649@shsmsx102.ccr.corp.intel.com> (raw)
In-Reply-To: <CAKv+Gu9bWrycpUXRyuTNmmWpdSd-6zqFExfKAwpSFS9WhTS9SA@mail.gmail.com>

Also cc Ruiyu to see if any comments.

Thanks,
Star
-----Original Message-----
From: Ard Biesheuvel [mailto:ard.biesheuvel@linaro.org] 
Sent: Tuesday, September 6, 2016 3:48 PM
To: edk2-devel-01 <edk2-devel@lists.01.org>; Tian, Feng <feng.tian@intel.com>; Zeng, Star <star.zeng@intel.com>; Gao, Liming <liming.gao@intel.com>
Cc: Laszlo Ersek <lersek@redhat.com>; Leif Lindholm <leif.lindholm@linaro.org>; Ard Biesheuvel <ard.biesheuvel@linaro.org>
Subject: Re: [PATCH 0/7] MdeModulePkg ArmVirtPkg: fixes for 64-bit PCI DMA

Feng, Star: do you have any feedback on these patches? Thanks.

On 5 September 2016 at 10:17, Ard Biesheuvel <ard.biesheuvel@linaro.org> wrote:
> After moving ArmVirtQemu to the generic PciHostBridgeDxe, we noticed 
> that setting DmaAbove4G resulted in problems with the emulated EHCI 
> USB host controller, which were caused by the fact that the PCI layer 
> was providing DMA buffers allocated above 4 GB while the emulated EHCI 
> controller in QEMU does not indicate support for 64-bit addressing.
>
> As it turns out, the PCI drivers in MdeModulePkg *completely* ignore 
> the EFI_PCI_ATTRIBUTE_DUAL_ADDRESS_CYCLE attribute, and simply assume 
> that no PCI root bridge driver will produce mappings above 4 GB. On 
> ARM, this is problematic, since not all platforms have memory below 4 
> GB, and so having full support for DMA above 4 GB is indispensable.
>
> So first, make the various drivers under MdeModulePkg/Pci/Bus set the 
> EFI_PCI_IO_ATTRIBUTE_DUAL_ADDRESS_CYCLE attributes for devices that 
> can support 64-bit DMA addressing (patches #1 - #5). Then, we can 
> update the host bridge driver to actually take these attributes into 
> account, and only create mappings above 4 GB for devices that have indicated support for it.
>
> Finally, in patch #7 we can remove the 4 GB DMA limit from ArmVirtPkg.
>
> Branch can be found here:
> https://git.linaro.org/people/ard.biesheuvel/uefi-next.git/shortlog/re
> fs/heads/pci-64bit-dma-fixes
>
> Ard Biesheuvel (7):
>   MdeModulePkg/AtaAtapiPassThru: enable 64-bit PCI DMA
>   MdeModulePkg/EhciDxe: enable 64-bit PCI DMA
>   MdeModulePkg/NvmExpressDxe: enable 64-bit PCI DMA
>   MdeModulePkg/SdMmcPciHcDxe: enable 64-bit PCI DMA
>   MdeModulePkg/XhciDxe: enable 64-bit PCI DMA
>   MdeModulePkg/PciHostBridgeDxe: restrict 64-bit DMA to devices that
>     support it
>   ArmVirtPkg/FdtPciHostBridgeLib: enable 64-bit PCI DMA
>
>  ArmVirtPkg/Library/FdtPciHostBridgeLib/FdtPciHostBridgeLib.c |  2 +-
>  MdeModulePkg/Bus/Ata/AtaAtapiPassThru/AhciMode.c             | 20 +++++++++++++++++-
>  MdeModulePkg/Bus/Pci/EhciDxe/Ehci.c                          | 22 +++++++++++++++++++-
>  MdeModulePkg/Bus/Pci/EhciDxe/Ehci.h                          |  2 ++
>  MdeModulePkg/Bus/Pci/EhciDxe/EhciSched.c                     |  2 +-
>  MdeModulePkg/Bus/Pci/NvmExpressDxe/NvmExpressHci.c           | 13 ++++++++++++
>  MdeModulePkg/Bus/Pci/PciHostBridgeDxe/PciRootBridgeIo.c      | 14 +++++++++----
>  MdeModulePkg/Bus/Pci/SdMmcPciHcDxe/SdMmcPciHcDxe.c           | 20 ++++++++++++++++++
>  MdeModulePkg/Bus/Pci/XhciDxe/Xhci.c                          | 22 +++++++++++++++++++-
>  MdeModulePkg/Bus/Pci/XhciDxe/Xhci.h                          |  2 ++
>  10 files changed, 110 insertions(+), 9 deletions(-)
>
> --
> 2.7.4
>

  reply	other threads:[~2016-09-06  8:04 UTC|newest]

Thread overview: 20+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-09-05  9:17 [PATCH 0/7] MdeModulePkg ArmVirtPkg: fixes for 64-bit PCI DMA Ard Biesheuvel
2016-09-05  9:17 ` [PATCH 1/7] MdeModulePkg/AtaAtapiPassThru: enable " Ard Biesheuvel
2016-09-05  9:17 ` [PATCH 2/7] MdeModulePkg/EhciDxe: " Ard Biesheuvel
2016-09-05 12:19   ` Laszlo Ersek
2016-09-05 12:44     ` Ard Biesheuvel
2016-09-05 13:06       ` Laszlo Ersek
2016-09-05  9:17 ` [PATCH 3/7] MdeModulePkg/NvmExpressDxe: " Ard Biesheuvel
2016-09-05  9:17 ` [PATCH 4/7] MdeModulePkg/SdMmcPciHcDxe: " Ard Biesheuvel
2016-09-05  9:17 ` [PATCH 5/7] MdeModulePkg/XhciDxe: " Ard Biesheuvel
2016-09-05 12:14   ` Laszlo Ersek
2016-09-05  9:17 ` [PATCH 6/7] MdeModulePkg/PciHostBridgeDxe: restrict 64-bit DMA to devices that support it Ard Biesheuvel
2016-09-05 12:04   ` Laszlo Ersek
2016-09-05  9:17 ` [PATCH 7/7] ArmVirtPkg/FdtPciHostBridgeLib: enable 64-bit PCI DMA Ard Biesheuvel
2016-09-05 12:04   ` Laszlo Ersek
2016-09-06  7:48 ` [PATCH 0/7] MdeModulePkg ArmVirtPkg: fixes for " Ard Biesheuvel
2016-09-06  8:04   ` Zeng, Star [this message]
2016-09-06  8:54   ` Ni, Ruiyu
2016-09-06 10:36     ` Ard Biesheuvel
2016-09-06 12:41       ` Tian, Feng
2016-09-06 14:48         ` Ard Biesheuvel

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=0C09AFA07DD0434D9E2A0C6AEB048310394B7649@shsmsx102.ccr.corp.intel.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