On Fri, 19 Apr 2013 12:32:15 +0800 Chi Zhang zhangchi866@gmail.com wrote:
在 2013年4月2日 星期二 15:19:25，Stefan Tauner 写道：
Thank you very much for this information. I had to guess the device IDs because they are not mentioned in the datasheet. The AT25F1024A has 0x60, the AT25F512A has 0x65, the AT25F025B has 0x6500... but apparently assuming at least some sanity has failed to reflect the truth.
So, according to your verbose log your AT25F512 has the same ID as the AT25F1024A (according to its datasheet). *sigh* Before I'll make some changes to the code I'll try to find someone with a AT25F1024 chip to test.
Bought an AT25F1024(without A) recently and tested. http://paste.flashrom.org/view.php?id=1620
Found a PCN from Atmel, stating that AT25F512's device code is 0x60, and the new AT25F512A's is 0x65: http://www.baite.com.hk/uploadFiles/service/SB_serial_mod_PCN_25F512.pdf
Oh that's very interesting. Thanks a lot for digging this up!
Another interesting thing is that my AT25F512 can actually hold 128KBytes of data. I generated the test data from /dev/urandom and confirmed that it contains no repetitive patterns with lzma -9 (a fake 128KB "random" data file made up with two 64KB ones with same content can be compressed into ~65KB while mine can not). All bytes seems to be functional, I had zeroed out and erased the whole chip and I got 128KB 0x00 an 128KB 0xff. After a power cycle the random data written to the chip can be preserved.
That's very odd. Maybe they just relabelled the AT25F1024s as AT25F512? I am only aware of (non-functional) fakes the other way round (and you too obviously ;)
I'll change the ID of AT25F512 and commit that later. Thanks a lot for your efforts.