[flashrom] Tyan flash S2915 fails under linux - chip detection problem....?

Carl-Daniel Hailfinger c-d.hailfinger.devel.2006 at gmx.net
Mon Feb 1 01:00:15 CET 2010

Hi everyone,

one possible explanation of the failing chip probes is the amount of RAM
in the tested machines. The common pattern of the failing machines seems
to be that they have more than 3 GB (or more than 2 GB) of RAM.
Depending on BIOS and kernel versions, this might result in cached
mapping of the flash chip and that will cause probing to fail.

So if those who have problems detecting the flash chip could physically
remove some RAM to get down to 2 GB (or less) and run flashrom again in
verbose mode (please use unpatched latest flashrom), that might give us
a hint about what is failing here.

Those who have working flash detection: Please tell us how much RAM you use.

Besides that, it would be great if you could:
- tell us the exact BIOS version
- tell us the exact kernel version
- send in the first 200 lines of your dmesg (preferably directly after boot)
- send the contents of /proc/mtrr
- send the contents of /sys/kernel/debug/x86/pat_memtype_list (may need
mounted debugfs).

Please do this for your current config and for the config with less RAM
(and tell us which is which).

I expect some very interesting revelations.

On 28.01.2010 21:05, Phil wrote:
> Michael Karcher wrote:
>> We are still having inconsistent results. As a summary of mails on the
>> flashrom list about the S2915(-E):
>> 21.10.2009, Alessandro Polverini (r736): No chip found
>> 03.12.2009, Alessandro Polverini (r791+GPIO4): No chip found
>> 19.01.2010, Phil (r710): No chip found
>> 28.01.2010, Allesandro Polverini, other board (r884): Chip found
>> (SST49LF080A)
>> 21.01.2010, Nicolas Aveline (r877;HP board): No chip found
>> 08.12.2009, Joel Robertson (r710): Chip found (SST49LF080A)

Thanks for your cooperation.


Developer quote of the year:
"We are juggling too many chainsaws and flaming arrows and tigers."

More information about the flashrom mailing list