Re: [platforms: PATCH 1/6] Marvell/Drivers: MvSpiFlash: Improve ReadId

[Marcin Wojtas <mw@semihalf.com>]

2017-11-01 4:14 GMT+01:00 Leif Lindholm <leif.lindholm@linaro.org>:
> On Tue, Oct 31, 2017 at 10:22:50AM +0100, Marcin Wojtas wrote:
>> 2017-10-31 10:07 GMT+01:00 Leif Lindholm <leif.lindholm@linaro.org>:
>> > On Tue, Oct 31, 2017 at 04:59:30AM +0100, Marcin Wojtas wrote:
>> >> Fix the ReadId routine by using master's ReadWrite callback
>> >> instead of the raw Transfer - no longer swapping and byte
>> >> shifting is needed. Simplify code by using local array
>> >> instead of dynamic allocation. Moreover store the FlashId
>> >> in an UINT8 array PCD instead of the concatenated UINT32
>> >> format - this way less overhead in the driver is needed
>> >> for comparing the buffers.
>> >>
>> >> The new handling allowed for cleaning Fupdate and Sf
>> >> shell commands FlashProbe routines.
>> >>
>> >> Contributed-under: TianoCore Contribution Agreement 1.1
>> >> Signed-off-by: Marcin Wojtas <mw@semihalf.com>
>> >> ---
>> >>  Platform/Marvell/Applications/FirmwareUpdate/FUpdate.c | 22 +++--------
>> >>  Platform/Marvell/Applications/SpiTool/SpiFlashCmd.c    | 37 ++++++------------
>> >>  Platform/Marvell/Armada/Armada70x0.dsc                 |  2 +-
>> >>  Platform/Marvell/Drivers/Spi/Devices/MvSpiFlash.c      | 41 ++++++++++++--------
>> >>  Platform/Marvell/Drivers/Spi/Devices/MvSpiFlash.h      |  2 +
>> >>  Platform/Marvell/Include/Protocol/SpiFlash.h           |  3 ++
>> >>  Platform/Marvell/Marvell.dec                           |  2 +-
>> >>  7 files changed, 48 insertions(+), 61 deletions(-)
>> >>
>> >> diff --git a/Platform/Marvell/Applications/FirmwareUpdate/FUpdate.c b/Platform/Marvell/Applications/FirmwareUpdate/FUpdate.c
>> >> index 664411a..d70645d 100644
>> >> --- a/Platform/Marvell/Applications/FirmwareUpdate/FUpdate.c
>> >> +++ b/Platform/Marvell/Applications/FirmwareUpdate/FUpdate.c
>> >> @@ -94,28 +94,16 @@ SpiFlashProbe (
>> >>    )
>> >>  {
>> >>    EFI_STATUS       Status;
>> >> -  UINT32           IdBuffer, Id, RefId;
>> >> +  UINT8           *FlashId;
>> >>
>> >> -  Id = PcdGet32 (PcdSpiFlashId);
>> >> -
>> >> -  IdBuffer = CMD_READ_ID & 0xff;
>> >> +  FlashId = (UINT8 *)PcdGetPtr (PcdSpiFlashId);
>> >>
>> >>    // Read SPI flash ID
>> >> -  SpiFlashProtocol->ReadId (Slave, sizeof (UINT32), (UINT8 *)&IdBuffer);
>> >> -
>> >> -  // Swap and extract 3 bytes of the ID
>> >> -  RefId = SwapBytes32 (IdBuffer) >> 8;
>> >> -
>> >> -  if (RefId == 0) {
>> >> -    Print (L"%s: No SPI flash detected");
>> >> -    return EFI_DEVICE_ERROR;
>> >> -  } else if (RefId != Id) {
>> >> -    Print (L"%s: Unsupported SPI flash detected with ID=%2x\n", CMD_NAME_STRING, RefId);
>> >> -    return EFI_DEVICE_ERROR;
>> >> +  Status = SpiFlashProtocol->ReadId (Slave, NOR_FLASH_ID_DEFAULT_LEN, FlashId);
>> >
>> > Is the length not possible to calculate somehow?
>> > Having a MAX_LEN defined and then using a DEFAULT_LEN or explicitly
>> > extracting 3 bytes from somewhere feels suboptimal.
>> >
>>
>> I know. It is however a change that was somewhat artificially
>> extracted, so that to make the next patch more readable
>> (NOR_FLASH_ID_DEFAULT_LEN is removed there). I will substitute it with
>> PcdGetSize (PcdSpiFlashId).
>
> Right, OK, that sort of thing is useful to mention in the cover
> letter.
>
> But there's also MAX_ID_LEN, which is added here only to allocate
> buffers that are hard-coded to always contain value of ID_DEFAULT_LEN.
> Surely this could just be left out?
>

The id can be either 3 or 5 bytes in the extended variant, which is
used by some flash devices. In theory we can define such pcd and
PcdGetSize ensures flexibility. However with the next commit and
NorFlashInfoLib table, MAX_ID_LEN is suitable and used with new
implementation. In the new version of commit I already removed
ID_DEFAULT_LEN.

Marcin