Re: [LinuxBIOS] flashrom support for AMD CS5536

[Alan Alexander]

I really should do more reading before bothering you guys.
It would seem that the answer to my questions is CMOS non-volatile RAM,
accessed via IO ports. How very odd.

Thanks again,

Alan Alexander          mailto:[e-mail]
Software Toolchain Engineering
Icera Inc, 2520 The Quadrant, Aztec west, Bristol BS32 4AQ, UK
Tel. [phone] 
-----Original Message-----
From: [e-mail]
[mailto:[e-mail]] On Behalf Of Alan Alexander
Sent: 27 November 2007 11:09
To: bari
Cc: [e-mail]
Subject: Re: [LinuxBIOS] flashrom support for AMD CS5536

Many thanks to all of the help regarding flashing of the Pm49fl004
device through the AMD CS5536 chipset on my IEI Nano LX board.
I am now able to erase and write the entire 512KByte region as desired.

Unfortunately, it would appear that I have overlooked something in my
quest for an automatic way to modify the legacy bios settings in our
embedded Linux devices. Although I can now alter the BIOS at my hearts
content, it seems that the actual BIOS settings themselves are stored
elsewhere. For example:

0) BIOS settings are set to configuration 'a'
1) I make a change to the Legacy BIOS settings (using the BIOS boot
menu) - configuration now 'b' (not equal to 'a') 
2) I save the BIOS image (with configuration 'b') from the Pm49fl004
device via flashrom -r.
3) I revert the BIOS settings back to configuration 'a' using the BIOS
boot menu.
4) I re-apply the saved image ('b') using flashrom -w and verify it with
flashrom -v 
5) On reboot the settings change applied in configuration 'b' are
missing.

So, my conclusion is that the actual BIOS settings must be stored
elsewhere. I suspect that this is probably obvious to you guys.

Could someone please tell me if this is the usual setup (i.e. BIOS flash
+ some other storage device for settings). If so, could someone also
point me at a likely interface/device/address range for this 'other'
storage device. Is it likely to be a SPI flash or something?

Any help greatly appreciated.

Best regards,

Alan Alexander          mailto:[e-mail]
Software Toolchain Engineering
Icera Inc, 2520 The Quadrant, Aztec west, Bristol BS32 4AQ, UK
Tel. [phone] 
-----Original Message-----
From: bari [mailto:[e-mail]] 
Sent: 23 November 2007 18:35
To: Alan Alexander
Cc: [e-mail]
Subject: Re: [LinuxBIOS] flashrom support for AMD CS5536

Alan Alexander wrote:
> Many thanks for the information. Unfortunately I don't have access to
> the board schematics and getting technical support from the board
> supplier is proving rather difficult.
>
> I have confirmed via an oscilloscope that both the WP# and TBL# pins

are
> permanently pulled high - indicating that no hardware write protection
> is enabled (so it looks like I may have side stepped the GPIO locking
> issue).

The pins WP# and TBL# are active low and would be tied high even in the 
case of having write protection. Have you seen the WP# or TBL# go low on

the scope during any of your write attempts?

I have also observed activity on the LFRAME#/FWH4 pin when
> flashrom is attempting to write to the flash. Since I don't have

access
> to X-ray equipment I think my best shot is to take a look at waveforms
> produced during a byte program cycle. I just need to get hold of that
> Logic Analyzer.......

That will tell you how much you have working during an attempt at a byte

program cycle. Are you able now to write to the memory area below the 
top-boot-block of the Flash? If so, all you have left to do is find out 
how they control the TBL#.

-Bari