On Sun, May 11, 2008 at 06:17:48PM +0200, Carl-Daniel Hailfinger wrote:
On 11.05.2008 18:01, Fredrik Tolf wrote:
Yes, the flashrom code with that patch seems to detect my chip and read its contents quite well. However, the same code does not detect the original chip on the board??? (an MX25L4005) satisfactory, and it seems to be a conflict with the SPI code.
Peter? This is a side effect of r3291, the "multiple flash chip" patch.
Yes, 3291 exposes this.
All of our SPI chips where a generic per-vendor fallback detection exists (that's all of them right now) will exhibit this problem.
I had reservations about "unknown SPI chip" and this is why.
This is also why I don't like fake flash chip entries for forcing reads from undetectable chips.
Both simply do not accurately describe or model what is actually going on and just end up confusing users and code.
It is ridiculous to claim that an "unknown chip" has been detected. If a chip had been positively detected it would not be unknown. This is behavior/language I could expect from commercial OS vendors.
I think we have to stop saying this and instead flashrom needs infrastructure and commands that can correctly model and handle SPI situation where the memory bus master is no longer transparent.
Fredrik, you should be able to work around this by calling flashrom -c MX25L4005 The patch I sent does not have anything to do with the MX25L4005 problem you're seeing because that problem was introduced in r3291. Can you signoff this patch since it works for you?
Please do sign off the patch if it works for you Fredrik. There are issues in flashrom, but as Carl-Daniel writes they are not closely related and I think the proposed RES change can go in for now.
//Peter