[MD-2811-032] DoCMilPlus32 missrecognized on pxa250.

[MD-2811-032] DoCMilPlus32 missrecognized on pxa250.

[Balint Cristian]

Hello  !

 I write in a last epuization of ideas,a call on list, because i hit a strange problem
and want to call a little bit advice from authors.

 I have a PDA with an PXA250 and an MD-2811-032 on it, but my chip is missrecognized
by the kernel:

"Flash chip found: Manufacturer ID: 98, Chip ID: A5 (Toshiba: NAND 2GiB 1,8V)"

 Uhmm 2GiB ??!? it is 32M.

 I have checked DoC registers it is 16 bit wired,it can be found in CS0 area at 0x00000000 of PXA.
 Also played with MRC0 regs to enable 32/16 bit of course driver worked, only if i used 8bit or 16 bit I/O  access mode in doc2000.h:

 32bit mmio for __arm__  < --- this not worked
 16bit mmio for __ppc__   < --- worked .
 readb/writeb else.            <---- worked.

 In 32bit [wich was original destinated for arm acording doc2000.h] it even dont recognise the chip ! 
The strangest part in all is that in 32 bit if i printf a little bit what happenning, the ChipID value for eg is 0xD3, value 
wich curiosly in reality is not at 0x00001000 is I expected to read from, this 0xD3 value is only present at 0x00000000  
[as i see via JTAG] !!
 Is it possible to read from another address in 32 bit mode of driver ??!?

 As specs, I used CVS version [a week old since now] of mtd with heandhealds org-2.6.8.1 kernel from an CVS tarball,
and i used old /mtd/drivers instead mtd/nand implementation because nand implementation said that not support this chip 
or something like this.


 Guys, any advice is welcome, please correct me if somewhere my suppositions are horrible,any idea is welcome
 i can send more printf more debugs give access to device, my C language skills are not the best but this really give headbang to me.

 As final words, i started adapt jtag-0.5.1 tool from openwince and combined with mtd-CVS drivers, it works at last fine
but recognise the chip as 2GiB Toshiba too :)) like linux kernel.
 Anyone wants to take a look in these lines of jtag ? 
Would be nice  to have a JTAG tool to handle these devices.


Thanks in advance,
~cristian

PS:
 excuse but english levels are poor too, it may sound in strange accords :)

Re: [MD-2811-032] DoCMilPlus32 missrecognized on pxa250.

[Thomas Gleixner]

Please fix your mail client to use standard line breaks !

On Thu, 2004-10-14 at 09:27, Balint Cristian wrote:
> "Flash chip found: Manufacturer ID: 98, Chip ID: A5 
> (Toshiba: NAND 2GiB 1,8V)"

The doc is configured for 16 bit access, the driver uses 8-bit access.
That's a known issue.

Search the mail-archive for a thread with the subject "[PATCH]Probing at
0x0"

The solution is proposed in the thread

tglx

Re: [MD-2811-032] DoCMilPlus32 missrecognized on pxa250.

[Balint Cristian]

On Thursday 14 October 2004 10:50, Thomas Gleixner wrote:
Hi Thomas !

 Thanks for you response,
> Please fix your mail client to use standard line breaks !

 I set my kmail to "colum warp at 78" now, hope now is ok for the 
requirements.

>
> On Thu, 2004-10-14 at 09:27, Balint Cristian wrote:
> > "Flash chip found: Manufacturer ID: 98, Chip ID: A5
> > (Toshiba: NAND 2GiB 1,8V)"
>
> The doc is configured for 16 bit access, the driver uses 8-bit access.
> That's a known issue.

 Not yet a fix for this ?
 To try hardwire chip in 8 bit mode ? What do you think ?
 I can use an unused GPIO of pxa to enable via software to control the DoC 
8/16 bit enabler pin if i want to have control over this.

>
> Search the mail-archive for a thread with the subject "[PATCH]Probing at
> 0x0"

  but, the probing issue is solved for me by hardcoding in docprobe.c to probe 
address "#elseif _arm_ 0x00000000" in docprobe.c, but of course the patch is 
a more elegant solution.

>
> The solution is proposed in the thread

I see only patch for the 0x0000000 probing issue [already hardcoded for me], 
but my problem is that it see a 2GiB instead of 32MiB, is this still becouse
you tell DoC is in 16bit and driver try 8bit access ? What that mean ?
What can i do to figure out this ?

 Please, let me clarify more a little bit.

PS
 As hardware knowleger, can tell someone if there exist another more popular 
flash device [i am somehow getting sick about DoC and M-Systems] but to be 
pin-pin compatible with this device MD-2811-032 and to work nice with mtd 
drivers ?

>
> tglx

Re: [MD-2811-032] DoCMilPlus32 missrecognized on pxa250.

[Thomas Gleixner]

On Thu, 2004-10-14 at 10:23, Balint Cristian wrote:
>  Not yet a fix for this ?

Volunteers welcome :)

>  To try hardwire chip in 8 bit mode ? What do you think ?
>  I can use an unused GPIO of pxa to enable via software to control the DoC 
> 8/16 bit enabler pin if i want to have control over this.

8 bit mode should work

> I see only patch for the 0x0000000 probing issue [already hardcoded for me], 
> but my problem is that it see a 2GiB instead of 32MiB, is this still becouse
> you tell DoC is in 16bit and driver try 8bit access ? What that mean ?
> What can i do to figure out this ?

Read the whole thread !
The 16 bit todo's are discussed there !

Read also the "Issues with a Doc Milplus" thread an follow ups

tglx