public inbox for devel@edk2.groups.io
 help / color / mirror / Atom feed
From: "Wu, Hao A" <hao.a.wu@intel.com>
To: Gaurav Jain <gaurav.jain@nxp.com>,
	"devel@edk2.groups.io" <devel@edk2.groups.io>,
	"Gao, Liming" <liming.gao@intel.com>,
	"afish@apple.com" <afish@apple.com>,
	"lersek@redhat.com" <lersek@redhat.com>,
	"leif@nuviainc.com" <leif@nuviainc.com>,
	"Kinney, Michael D" <michael.d.kinney@intel.com>
Cc: "Wang, Jian J" <jian.j.wang@intel.com>,
	"Ni, Ray" <ray.ni@intel.com>,
	"Ard Biesheuvel" <ard.biesheuvel@linaro.org>,
	Pankaj Bansal <pankaj.bansal@nxp.com>
Subject: Re: [EXT] RE: [edk2-stable202002][edk2-devel] [PATCH v2 1/1] MdeModulePkg/Pci: Fixed Asserts in SCT PCIIO Protocol Test.
Date: Mon, 24 Feb 2020 08:26:14 +0000	[thread overview]
Message-ID: <B80AF82E9BFB8E4FBD8C89DA810C6A093C9AAF08@SHSMSX104.ccr.corp.intel.com> (raw)
In-Reply-To: <DB7PR04MB4091E8DF36C3A35174445205E7EC0@DB7PR04MB4091.eurprd04.prod.outlook.com>

> -----Original Message-----
> From: Gaurav Jain [mailto:gaurav.jain@nxp.com]
> Sent: Monday, February 24, 2020 3:04 PM
> To: Wu, Hao A; devel@edk2.groups.io; Gao, Liming; afish@apple.com;
> lersek@redhat.com; leif@nuviainc.com; Kinney, Michael D
> Cc: Wang, Jian J; Ni, Ray; Ard Biesheuvel; Pankaj Bansal
> Subject: RE: [EXT] RE: [edk2-stable202002][edk2-devel] [PATCH v2 1/1]
> MdeModulePkg/Pci: Fixed Asserts in SCT PCIIO Protocol Test.
> 
> 
> > I think the above check for 'Attributes' can be dropped.
> > I found that the implementation of the PciIoGetBarAttributes() function
> does not
> > expose any configurable attributes. So the logic can fall through to the
> ASSERT
> > (for DEBUG images) and then returns EFI_UNSUPPORTED.
> 
> I agree that PciIoGetBarAttributes() function sets *Supports as 0.
> But In SCT Test for SetBarAttributes, there is a test case for Unsupported
> Attribute which expects EFI_UNSUPPORTED. If I drop this check, ASSERT will
> come, which is not expected.
> Can we keep check for 'Attributes'?


Oh, I forgot that.

I have one question, is there any special reason for you to pick the supported
bits specified by:
EFI_PCI_DEVICE_ENABLE | EFI_PCI_IO_ATTRIBUTE_DUAL_ADDRESS_CYCLE

Is it relating with the SCT test case?

Best Regards,
Hao Wu


> 
> > -----Original Message-----
> > From: Wu, Hao A <hao.a.wu@intel.com>
> > Sent: Friday, February 21, 2020 6:53 AM
> > To: devel@edk2.groups.io; Gaurav Jain <gaurav.jain@nxp.com>; Gao,
> Liming
> > <liming.gao@intel.com>; afish@apple.com; lersek@redhat.com;
> > leif@nuviainc.com; Kinney, Michael D <michael.d.kinney@intel.com>
> > Cc: Wang, Jian J <jian.j.wang@intel.com>; Ni, Ray <ray.ni@intel.com>; Ard
> > Biesheuvel <ard.biesheuvel@linaro.org>; Pankaj Bansal
> > <pankaj.bansal@nxp.com>
> > Subject: [EXT] RE: [edk2-stable202002][edk2-devel] [PATCH v2 1/1]
> > MdeModulePkg/Pci: Fixed Asserts in SCT PCIIO Protocol Test.
> >
> > Caution: EXT Email
> >
> > A couple of inline comments below. Please help to handle them in the next
> > version of patch.
> > With them addressed,
> > Reviewed-by: Hao A Wu <hao.a.wu@intel.com>
> >
> >
> > Hello Liming and Stewards,
> >
> > I would like to confirm with you for whether the patch should catch the
> > upcoming stable tag.
> >
> > My personal take is that the patch is more like a code refinement rather
> than a
> > bug fix.
> >
> > Could you help to make a final call for this one? Thanks in advance.
> >
> > Best Regards,
> > Hao Wu
> >
> >
> > > -----Original Message-----
> > > From: devel@edk2.groups.io [mailto:devel@edk2.groups.io] On Behalf
> Of
> > > Gaurav Jain
> > > Sent: Thursday, February 20, 2020 11:40 PM
> > > To: devel@edk2.groups.io
> > > Cc: Wang, Jian J; Wu, Hao A; Ni, Ray; Ard Biesheuvel; Pankaj Bansal;
> > > Gaurav Jain
> > > Subject: [edk2-devel] [PATCH v2 1/1] MdeModulePkg/Pci: Fixed Asserts
> > > in SCT PCIIO Protocol Test.
> > >
> > > ASSERT in PollMem_Conf, CopyMem_Conf, SetBarAttributes_Conf
> > > Conformance Test.
> > > SCT Test expect return as Invalid Parameter or Unsupported.
> > > Added Checks for Function Parameters.
> > > return Invalid or Unsupported if Check fails.
> > >
> > > Added Checks in PciIoPollIo(), PciIoIoRead()
> > > PciIoIoWrite()
> > >
> > > Signed-off-by: Gaurav Jain <gaurav.jain@nxp.com>
> > > ---
> > >
> > > Notes:
> > >     v2
> > >     - Reverted ASSERT(FALSE) code.
> > >     - Added Checks for Width, BarIndex, Buffer,
> > >       Address range in PciIoIoRead, PciIoIoWrite.
> > >     - Added Checks for Width, BarIndex, Result,
> > >       Address range in PciIoPollIo, PciIoPollMem,
> > >       PciIoCopyMem.
> > >     - Added Checks for Attributes, BarIndex,
> > >       Address range in PciIoSetBarAttributes.
> > >
> > >  .../NonDiscoverablePciDeviceIo.c              | 180 ++++++++++++++++++
> > >  1 file changed, 180 insertions(+)
> > >
> > > diff --git
> > >
> a/MdeModulePkg/Bus/Pci/NonDiscoverablePciDeviceDxe/NonDiscoverableP
> > > ciDeviceIo.c
> > >
> b/MdeModulePkg/Bus/Pci/NonDiscoverablePciDeviceDxe/NonDiscoverableP
> > > ciDeviceIo.c
> > > index 2d55c9699322..4dd804356021 100644
> > > ---
> > >
> a/MdeModulePkg/Bus/Pci/NonDiscoverablePciDeviceDxe/NonDiscoverableP
> > > ciDeviceIo.c
> > > +++
> > >
> b/MdeModulePkg/Bus/Pci/NonDiscoverablePciDeviceDxe/NonDiscoverableP
> > > ciDeviceIo.c
> > > @@ -93,6 +93,35 @@ PciIoPollMem (
> > >    OUT UINT64                      *Result
> > >    )
> > >  {
> > > +  NON_DISCOVERABLE_PCI_DEVICE         *Dev;
> > > +  EFI_ACPI_ADDRESS_SPACE_DESCRIPTOR   *Desc;
> > > +  UINTN                               Count;
> > > +  EFI_STATUS                          Status;
> > > +
> > > +  if ((UINT32)Width > EfiPciIoWidthUint64) {
> > > +    return EFI_INVALID_PARAMETER;
> > > +  }
> > > +
> > > +  if (BarIndex >= PCI_MAX_BAR) {
> > > +    return EFI_UNSUPPORTED;
> > > +  }
> > > +
> > > +  if (Result == NULL) {
> > > +    return EFI_INVALID_PARAMETER;
> > > +  }
> > > +
> > > +  Dev = NON_DISCOVERABLE_PCI_DEVICE_FROM_PCI_IO(This);
> > > +  Count = 1;
> > > +
> > > +  Status = GetBarResource (Dev, BarIndex, &Desc);  if (EFI_ERROR
> > > + (Status)) {
> > > +    return Status;
> > > +  }
> > > +
> > > +  if (Offset + (Count << (Width & 0x3)) > Desc->AddrLen) {
> > > +    return EFI_UNSUPPORTED;
> > > +  }
> > > +
> > >    ASSERT (FALSE);
> > >    return EFI_UNSUPPORTED;
> > >  }
> > > @@ -126,6 +155,35 @@ PciIoPollIo (
> > >    OUT UINT64                      *Result
> > >    )
> > >  {
> > > +  NON_DISCOVERABLE_PCI_DEVICE         *Dev;
> > > +  EFI_ACPI_ADDRESS_SPACE_DESCRIPTOR   *Desc;
> > > +  UINTN                               Count;
> > > +  EFI_STATUS                          Status;
> > > +
> > > +  if ((UINT32)Width > EfiPciIoWidthUint64) {
> > > +    return EFI_INVALID_PARAMETER;
> > > +  }
> > > +
> > > +  if (BarIndex >= PCI_MAX_BAR) {
> > > +    return EFI_UNSUPPORTED;
> > > +  }
> > > +
> > > +  if (Result == NULL) {
> > > +    return EFI_INVALID_PARAMETER;
> > > +  }
> > > +
> > > +  Dev = NON_DISCOVERABLE_PCI_DEVICE_FROM_PCI_IO(This);
> > > +  Count = 1;
> > > +
> > > +  Status = GetBarResource (Dev, BarIndex, &Desc);  if (EFI_ERROR
> > > + (Status)) {
> > > +    return Status;
> > > +  }
> > > +
> > > +  if (Offset + (Count << (Width & 0x3)) > Desc->AddrLen) {
> > > +    return EFI_UNSUPPORTED;
> > > +  }
> > > +
> > >    ASSERT (FALSE);
> > >    return EFI_UNSUPPORTED;
> > >  }
> > > @@ -396,6 +454,33 @@ PciIoIoRead (
> > >    IN OUT VOID                         *Buffer
> > >    )
> > >  {
> > > +  NON_DISCOVERABLE_PCI_DEVICE         *Dev;
> > > +  EFI_ACPI_ADDRESS_SPACE_DESCRIPTOR   *Desc;
> > > +  EFI_STATUS                          Status;
> > > +
> > > +  if ((UINT32)Width > EfiPciIoWidthUint64) {
> > > +    return EFI_INVALID_PARAMETER;
> > > +  }
> >
> >
> > For PciIoIoRead(), I think enum values smaller than EfiPciIoWidthMaximum
> are
> > all valid. The above check seems to strict.
> 
> Will address this in v3.
> >
> >
> > > +
> > > +  if (BarIndex >= PCI_MAX_BAR) {
> > > +    return EFI_UNSUPPORTED;
> > > +  }
> > > +
> > > +  if (Buffer == NULL) {
> > > +    return EFI_INVALID_PARAMETER;
> > > +  }
> > > +
> > > +  Dev = NON_DISCOVERABLE_PCI_DEVICE_FROM_PCI_IO(This);
> > > +
> > > +  Status = GetBarResource (Dev, BarIndex, &Desc);  if (EFI_ERROR
> > > + (Status)) {
> > > +    return Status;
> > > +  }
> > > +
> > > +  if (Offset + (Count << (Width & 0x3)) > Desc->AddrLen) {
> > > +    return EFI_UNSUPPORTED;
> > > +  }
> > > +
> > >    ASSERT (FALSE);
> > >    return EFI_UNSUPPORTED;
> > >  }
> > > @@ -425,6 +510,33 @@ PciIoIoWrite (
> > >    IN OUT VOID                         *Buffer
> > >    )
> > >  {
> > > +  NON_DISCOVERABLE_PCI_DEVICE         *Dev;
> > > +  EFI_ACPI_ADDRESS_SPACE_DESCRIPTOR   *Desc;
> > > +  EFI_STATUS                          Status;
> > > +
> > > +  if ((UINT32)Width > EfiPciIoWidthUint64) {
> > > +    return EFI_INVALID_PARAMETER;
> > > +  }
> >
> >
> > For PciIoIoWrite(), I think enum values smaller than EfiPciIoWidthMaximum
> are
> > all valid. The above check seems to strict.
> 
> Will address this in v3.
> >
> >
> > > +
> > > +  if (BarIndex >= PCI_MAX_BAR) {
> > > +    return EFI_UNSUPPORTED;
> > > +  }
> > > +
> > > +  if (Buffer == NULL) {
> > > +    return EFI_INVALID_PARAMETER;
> > > +  }
> > > +
> > > +  Dev = NON_DISCOVERABLE_PCI_DEVICE_FROM_PCI_IO(This);
> > > +
> > > +  Status = GetBarResource (Dev, BarIndex, &Desc);  if (EFI_ERROR
> > > + (Status)) {
> > > +    return Status;
> > > +  }
> > > +
> > > +  if (Offset + (Count << (Width & 0x3)) > Desc->AddrLen) {
> > > +    return EFI_UNSUPPORTED;
> > > +  }
> > > +
> > >    ASSERT (FALSE);
> > >    return EFI_UNSUPPORTED;
> > >  }
> > > @@ -556,6 +668,40 @@ PciIoCopyMem (
> > >    IN     UINTN                        Count
> > >    )
> > >  {
> > > +  NON_DISCOVERABLE_PCI_DEVICE         *Dev;
> > > +  EFI_ACPI_ADDRESS_SPACE_DESCRIPTOR   *DestDesc;
> > > +  EFI_ACPI_ADDRESS_SPACE_DESCRIPTOR   *SrcDesc;
> > > +  EFI_STATUS                          Status;
> > > +
> > > +  if ((UINT32)Width > EfiPciIoWidthUint64) {
> > > +    return EFI_INVALID_PARAMETER;
> > > +  }
> > > +
> > > +  if (DestBarIndex >= PCI_MAX_BAR ||
> > > +      SrcBarIndex >= PCI_MAX_BAR) {
> > > +    return EFI_UNSUPPORTED;
> > > +  }
> > > +
> > > +  Dev = NON_DISCOVERABLE_PCI_DEVICE_FROM_PCI_IO(This);
> > > +
> > > +  Status = GetBarResource (Dev, DestBarIndex, &DestDesc);  if
> > > + (EFI_ERROR (Status)) {
> > > +    return Status;
> > > +  }
> > > +
> > > +  if (DestOffset + (Count << (Width & 0x3)) > DestDesc->AddrLen) {
> > > +    return EFI_UNSUPPORTED;
> > > +  }
> > > +
> > > +  Status = GetBarResource (Dev, SrcBarIndex, &SrcDesc);  if
> > > + (EFI_ERROR (Status)) {
> > > +    return Status;
> > > +  }
> > > +
> > > +  if (SrcOffset + (Count << (Width & 0x3)) > SrcDesc->AddrLen) {
> > > +    return EFI_UNSUPPORTED;
> > > +  }
> > > +
> > >    ASSERT (FALSE);
> > >    return EFI_UNSUPPORTED;
> > >  }
> > > @@ -1414,6 +1560,40 @@ PciIoSetBarAttributes (
> > >    IN OUT UINT64                       *Length
> > >    )
> > >  {
> > > +  NON_DISCOVERABLE_PCI_DEVICE         *Dev;
> > > +  EFI_ACPI_ADDRESS_SPACE_DESCRIPTOR   *Desc;
> > > +  EFI_PCI_IO_PROTOCOL_WIDTH           Width;
> > > +  UINTN                               Count;
> > > +  EFI_STATUS                          Status;
> > > +
> > > +  #define DEV_SUPPORTED_ATTRIBUTES \
> > > +    (EFI_PCI_DEVICE_ENABLE |
> > > EFI_PCI_IO_ATTRIBUTE_DUAL_ADDRESS_CYCLE)
> > > +
> > > +  if ((Attributes & (~DEV_SUPPORTED_ATTRIBUTES)) != 0) {
> > > +    return EFI_UNSUPPORTED;
> > > +  }
> >
> >
> > I think the above check for 'Attributes' can be dropped.
> > I found that the implementation of the PciIoGetBarAttributes() function
> does not
> > expose any configurable attributes. So the logic can fall through to the
> ASSERT
> > (for DEBUG images) and then returns EFI_UNSUPPORTED.
> >
> > Best Regards,
> > HaoWu
> >
> >
> > > +
> > > +  if (BarIndex >= PCI_MAX_BAR) {
> > > +    return EFI_UNSUPPORTED;
> > > +  }
> > > +
> > > +  if (Offset == NULL || Length == NULL) {
> > > +    return EFI_INVALID_PARAMETER;
> > > +  }
> > > +
> > > +  Dev = NON_DISCOVERABLE_PCI_DEVICE_FROM_PCI_IO(This);
> > > +  Width = EfiPciIoWidthUint8;
> > > +  Count = (UINT32) *Length;
> > > +
> > > +  Status = GetBarResource(Dev, BarIndex, &Desc);  if (EFI_ERROR
> > > + (Status)) {
> > > +    return Status;
> > > +  }
> > > +
> > > +  if (*Offset + (Count << (Width & 0x3)) > Desc->AddrLen) {
> > > +    return EFI_UNSUPPORTED;
> > > +  }
> > > +
> > >    ASSERT (FALSE);
> > >    return EFI_UNSUPPORTED;
> > >  }
> > > --
> > > 2.17.1
> > >
> > >
> > > 


  reply	other threads:[~2020-02-24  8:26 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-02-21  1:23 [edk2-stable202002][edk2-devel] [PATCH v2 1/1] MdeModulePkg/Pci: Fixed Asserts in SCT PCIIO Protocol Test Wu, Hao A
2020-02-24  7:04 ` [EXT] " Gaurav Jain
2020-02-24  8:26   ` Wu, Hao A [this message]
2020-02-24  8:42     ` Gaurav Jain
2020-02-24 12:50       ` Wu, Hao A

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=B80AF82E9BFB8E4FBD8C89DA810C6A093C9AAF08@SHSMSX104.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