Re: File system is not working in Retail Version in WinCE 5.0

TFAT and FAL work fine in DEBUG and RETAIL version, you should check your
FMD, and mainly the timing on your chip select of something like that.
I would suggest you modify your fmd_init so that at the end, it performs an
erase of the whole flash, writes different patterns in each sector (and
spares), reads them back and compares to what is expected.

HTH

--
--
--
----------------------------------------------------------------
Anthony Pellerin (eMVP)
ADENEO (ADESET)
Windows Embedded Consultant
<apellerin AT adeneo DOT adetelgroup DOT com>
http://www.adeneo.adetelgroup.com
Tél : +33 (0)4.72.18.57.77
Fax : +33 (0)4.72.18.57.78
----------------------------------------------------------------
"prabu" <prabu@xxxxxxxxxxxxxxxxxxxxxxxxx> a écrit dans le message de news:
3B511CD4-9EE2-43A8-B95E-583A64CF89F4@xxxxxxxxxxxxxxxx
Hi all,

I am trying to implement the tfat filesystem for NAND and NOR flash in my
device on WINDOWSCE 5.0.
Both Nor and Nand Flash filesystem is working in the Debug version. But it
is not working in the retail version.

During the format operation,

FMD_WriteSector() is called by the FAL,after that FMD_ReadSector() is
called
by the FAL,then again the FMD_WriteSector() and FMD_ReadSector() is called
for each Sector Starting from 0th Sector.

But the same operation is not taken place in retail version for both NAND
and NOR filesystem. Only FMD_ReadSector() is called for some time and it
says
the following error on the Message box.

"A System Cannot Write to the Specified Device".

Is there any Code Optimisation Issue During the Compilation process?
Because
both the NAND and NOR FMD drivers are working correctly in the Debug
Version.The Same issue is occured in both the file system.

Also I am suspecting the issue in FAL. Because the FMD_ReadSector() and
FMD_WriteSector() is called by that Layer.


Advanced Thanks,
Prabu Kumar.K




.

Relevant Pages

  • Re: Capturing debug messages using CeLog on a release terminal
    ... Don't set the shipbuild flag in your retail kernel (because it will #define RETAILMSG and DEBUGMSG into nirvana) and don't set Enable CE Target control and don't set Enable Kernel Debugger. ... I'm not entirely sure if DEBUGMSGs (as opposed to RETAILMSGs) still work when replacing dll's in a retail build with their debug counterparts. ...
    (microsoft.public.windowsce.platbuilder)
  • Re: KITL over USB RNDIS
    ... have tried to compile the KITL library as retail version and kernel as Debug ... USB use a huge amount of stack in DEBUG builds. ...
    (microsoft.public.windowsce.platbuilder)
  • Re: SMBServer problem
    ... And, as expected, the RETAIL build also works properly when using the ... The QFE is bad. ... Paul T. ... the SMB_PACKET structure, in the case of a DEBUG build, ...
    (microsoft.public.windowsce.platbuilder)
  • Re: DEBUGMSGs in retail build + USB enumeration problem
    ... If you just build your debug kernel, then copy the debug versions of the drivers you want debug info on into your retail flatreldir, and do a makeimg, you should get debugmsg info out of the serial port in your retail image. ... device at a time plugged into the single USB port. ...
    (microsoft.public.windowsce.platbuilder)
  • Re: DEBUGMSGs in retail build + USB enumeration problem
    ... Still, those debug messages can be output because of many things that can go wrong with USB, you'll just have to try and see whether the QFE will solve it. ... USB1.1 hub; when I unplug it the following debugging messages are ... makeimg (for the retail image), you should get debugmsg info out of the ...
    (microsoft.public.windowsce.platbuilder)
Loading