Problem in kernel

Hi,

We are migrating to CE 6 and we can't finalize our OEMInit function because
of a DEBUGCHK falure in vm.c(in debug configuration of course). I activated
DebugMsg in the vm.c files. When we use the retail configuration it works but
fails later in a driver initialization (but this is another story). Here is
the output:

4294767296 PID:0 TID:0 Start x86InitRomChain.
4294767296 PID:0 TID:0 Looking for rom chain
4294767296 PID:0 TID:0 Rom chain NOT found
4294767296 PID:0 TID:0 End x86InitRomChain
4294767296 PID:0 TID:0 Start InitFPGA.
4294767296 PID:0 TID:0 +InitFPGA
4294767296 PID:0 TID:0 PCIGetBusDataByOffset Start
4294767296 PID:0 TID:0 PCIGetBusDataByOffset End
4294767296 PID:0 TID:0 PCIGetBusDataByOffset Start
4294767296 PID:0 TID:0 PCIGetBusDataByOffset end
4294767296 PID:0 TID:0 NkCreateStaticMapping EEE00000
4294767296 PID:0 TID:0 CreateStaticMapping: 00eee000 00004000
4294767296 PID:0 TID:0 DoVMReserve - process-id: 00000000, dwAddr =
00000000, cPages = 0x4, base = 00000000, fAllocType = 00000000
4294767296 PID:0 TID:0 DoVMReserve - searching from va = 00000000
4294767296 PID:0 TID:0 NK Kernel: DEBUGCHK failed in file
E:\IRIS\OS\WINCE600\private\winceos\COREOS\nk\kernel\nknormal\..\vm.c at line
550

Just after x86InitRomChain we call our FPGAInit function in which we call a
NKCreateStaticMapping to map PCI adress in virtual memory. Is it too early?
Is it wrong (it worked in CE 5)? It seams that the process 0 (nk.exe) has a
starting free VM address of 0 and that's why we receive a DEBUGCHK.

Any help would be welcome.

Thanks.

Ross
Thu Sep 27 16:35:24 CDT 2007

This is a known issue that we found after shipping CE6. If you would like to
see a fix (QFE), please open a support request either with whomever you
obtained CE from, or you can start a request here:
http://support.microsoft.com/oas/default.aspx?&c1=509&gprid=12015&

You can mention my name / this thread to help escalate the process.

-Ross


"Vinz" <Vinz@discussions.microsoft.com> wrote in message
news:25ED4D95-C419-495D-ACC1-050E6741826A@microsoft.com...
> Hi,
>
> We are migrating to CE 6 and we can't finalize our OEMInit function
> because
> of a DEBUGCHK falure in vm.c(in debug configuration of course). I
> activated
> DebugMsg in the vm.c files. When we use the retail configuration it works
> but
> fails later in a driver initialization (but this is another story). Here
> is
> the output:
>
> 4294767296 PID:0 TID:0 Start x86InitRomChain.
> 4294767296 PID:0 TID:0 Looking for rom chain
> 4294767296 PID:0 TID:0 Rom chain NOT found
> 4294767296 PID:0 TID:0 End x86InitRomChain
> 4294767296 PID:0 TID:0 Start InitFPGA.
> 4294767296 PID:0 TID:0 +InitFPGA
> 4294767296 PID:0 TID:0 PCIGetBusDataByOffset Start
> 4294767296 PID:0 TID:0 PCIGetBusDataByOffset End
> 4294767296 PID:0 TID:0 PCIGetBusDataByOffset Start
> 4294767296 PID:0 TID:0 PCIGetBusDataByOffset end
> 4294767296 PID:0 TID:0 NkCreateStaticMapping EEE00000
> 4294767296 PID:0 TID:0 CreateStaticMapping: 00eee000 00004000
> 4294767296 PID:0 TID:0 DoVMReserve - process-id: 00000000, dwAddr =
> 00000000, cPages = 0x4, base = 00000000, fAllocType = 00000000
> 4294767296 PID:0 TID:0 DoVMReserve - searching from va = 00000000
> 4294767296 PID:0 TID:0 NK Kernel: DEBUGCHK failed in file
> E:\IRIS\OS\WINCE600\private\winceos\COREOS\nk\kernel\nknormal\..\vm.c at
> line
> 550
>
> Just after x86InitRomChain we call our FPGAInit function in which we call
> a
> NKCreateStaticMapping to map PCI adress in virtual memory. Is it too
> early?
> Is it wrong (it worked in CE 5)? It seams that the process 0 (nk.exe) has
> a
> starting free VM address of 0 and that's why we receive a DEBUGCHK.
>
> Any help would be welcome.
>
> Thanks.

Vinz
Fri Sep 28 10:05:01 CDT 2007

OK...

I openned a case.

Meanwhile, can I just comment out the debugchk or can I do anything else to
make it continue the kernel initialisation?