Re: [PATCH v2 6/7] EmbeddedPkg: Fix DwEmmc SendCommand polling

["Loh, Tien Hock" <tien.hock.loh@intel.com>]

> -----Original Message-----
> From: Haojian Zhuang <haojian.zhuang@linaro.org>
> Sent: Thursday, May 30, 2019 3:06 PM
> To: Leif Lindholm <leif.lindholm@linaro.org>
> Cc: Loh, Tien Hock <tien.hock.loh@intel.com>; devel@edk2.groups.io;
> thloh85@gmail.com; Ard Biesheuvel <ard.biesheuvel@linaro.org>
> Subject: Re: [PATCH v2 6/7] EmbeddedPkg: Fix DwEmmc SendCommand
> polling
> 
> On Tue, May 28, 2019 at 07:04:09PM +0100, Leif Lindholm wrote:
> > +Haojian,
> >
> > Haojian - since you are the original author, can you comment on the
> > delays? Are these silicon bug workarounds (so we need to add a Pcd),
> > or does these changes work on your platforms too?
> 
> I'm not in the loop, so I missed the patch series.
> 
> The patch series can't work on my platform for the eMMC. Although a
> variable is created to identify whether it's a SD or eMMC device, it doesn't
> identify the eMMC device by the right way. So the eMMC device isn't
> initialized successfully on my platform.
> 
> 1. Since MMC framework could identify whether it's eMMC device or SD
> device, we need to make device driver gets this kind of information from the
> MMC framework. And we need to support multiple eMMC/SD instances in
> MMC framework.

Yeah my bad I didn't read through the SD/MMC specs on that. Now I check the specs and you're right, the information should be read from MMC framework.

> 
> 2. I sent a patch series to support both eMMC device and SD device before.
> https://edk2.groups.io/g/devel/message/28572
> &&
> https://edk2.groups.io/g/devel/message/28615
> Maybe it's missed. Could you help to review that patch series?

> 
> Best Regards
> Haojian
> 
> >
> > Regards,
> >
> > Leif
> >
> > On Mon, May 27, 2019 at 05:30:27PM +0800, tien.hock.loh@intel.com
> wrote:
> > > From: "Tien Hock, Loh" <tien.hock.loh@intel.com>
> > >
> > > Change SendCommand polling mode to remove unnecessary delay, and
> > > check for transfer done only when block data is to be read/write.
> > > This would also increase performance slightly.
> > >
> > > Signed-off-by: "Tien Hock, Loh" <tien.hock.loh@intel.com>
> > > Cc: Leif Lindholm <leif.lindholm@linaro.org>
> > > Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
> > > ---
> > >  EmbeddedPkg/Drivers/DwEmmcDxe/DwEmmcDxe.c | 43
> +++++++++++++++-----
> > >  1 file changed, 33 insertions(+), 10 deletions(-)
> > >
> > > diff --git a/EmbeddedPkg/Drivers/DwEmmcDxe/DwEmmcDxe.c
> > > b/EmbeddedPkg/Drivers/DwEmmcDxe/DwEmmcDxe.c
> > > index c6c8e04917..b57833458f 100644
> > > --- a/EmbeddedPkg/Drivers/DwEmmcDxe/DwEmmcDxe.c
> > > +++ b/EmbeddedPkg/Drivers/DwEmmcDxe/DwEmmcDxe.c
> > > @@ -286,16 +286,13 @@ SendCommand (
> > >              DWEMMC_INT_RCRC | DWEMMC_INT_RE;
> > >    ErrMask |= DWEMMC_INT_DCRC | DWEMMC_INT_DRT |
> DWEMMC_INT_SBE;
> > >    do {
> > > -    MicroSecondDelay(500);
> > >      Data = MmioRead32 (DWEMMC_RINTSTS);
> > > -
> > > -    if (Data & ErrMask) {
> > > -      return EFI_DEVICE_ERROR;
> > > -    }
> > > -    if (Data & DWEMMC_INT_DTO) {     // Transfer Done
> > > -      break;
> > > -    }
> > >    } while (!(Data & DWEMMC_INT_CMD_DONE));
> > > +
> > > +  if (Data & ErrMask) {
> > > +    return EFI_DEVICE_ERROR;
> > > +  }
> > > +
> > >    return EFI_SUCCESS;
> > >  }
> > >
> > > @@ -550,8 +547,9 @@ DwEmmcReadBlockData (
> > >    )
> > >  {
> > >    EFI_STATUS  Status;
> > > -  UINT32      DescPages, CountPerPage, Count;
> > > +  UINT32      DescPages, CountPerPage, Count, ErrMask;
> > >    EFI_TPL     Tpl;
> > > +  UINTN Rintsts = 0;
> > >
> > >    Tpl = gBS->RaiseTPL (TPL_NOTIFY);
> > >
> > > @@ -574,6 +572,18 @@ DwEmmcReadBlockData (
> > >      DEBUG ((DEBUG_ERROR, "Failed to read data,
> mDwEmmcCommand:%x, mDwEmmcArgument:%x, Status:%r\n",
> mDwEmmcCommand, mDwEmmcArgument, Status));
> > >      goto out;
> > >    }
> > > +
> > > +  while(!((MmioRead32(DWEMMC_RINTSTS) & (DWEMMC_INT_DTO))))
> {
> > > +    Rintsts = MmioRead32 (DWEMMC_RINTSTS);  }  ErrMask =
> > > + DWEMMC_INT_EBE | DWEMMC_INT_HLE | DWEMMC_INT_RTO |
> > > +            DWEMMC_INT_RCRC | DWEMMC_INT_RE |
> DWEMMC_INT_DCRC |
> > > +            DWEMMC_INT_DRT | DWEMMC_INT_SBE;
> > > +
> > > +  if (Rintsts & ErrMask) {
> > > +    Status = EFI_DEVICE_ERROR;
> > > +    goto out;
> > > +  }
> > >  out:
> > >    // Restore Tpl
> > >    gBS->RestoreTPL (Tpl);
> > > @@ -589,8 +599,9 @@ DwEmmcWriteBlockData (
> > >    )
> > >  {
> > >    EFI_STATUS  Status;
> > > -  UINT32      DescPages, CountPerPage, Count;
> > > +  UINT32      DescPages, CountPerPage, Count, ErrMask;
> > >    EFI_TPL     Tpl;
> > > +  UINTN Rintsts = 0;
> > >
> > >    Tpl = gBS->RaiseTPL (TPL_NOTIFY);
> > >
> > > @@ -613,6 +624,18 @@ DwEmmcWriteBlockData (
> > >      DEBUG ((DEBUG_ERROR, "Failed to write data,
> mDwEmmcCommand:%x, mDwEmmcArgument:%x, Status:%r\n",
> mDwEmmcCommand, mDwEmmcArgument, Status));
> > >      goto out;
> > >    }
> > > +
> > > +  while(!((MmioRead32(DWEMMC_RINTSTS) & (DWEMMC_INT_DTO))))
> {
> > > +    Rintsts = MmioRead32 (DWEMMC_RINTSTS);  }  ErrMask =
> > > + DWEMMC_INT_EBE | DWEMMC_INT_HLE | DWEMMC_INT_RTO |
> > > +            DWEMMC_INT_RCRC | DWEMMC_INT_RE |
> DWEMMC_INT_DCRC |
> > > +            DWEMMC_INT_DRT | DWEMMC_INT_SBE;
> > > +
> > > +  if (Rintsts & ErrMask) {
> > > +    Status = EFI_DEVICE_ERROR;
> > > +    goto out;
> > > +  }
> > >  out:
> > >    // Restore Tpl
> > >    gBS->RestoreTPL (Tpl);
> > > --
> > > 2.19.0
> > >