On Fri, Jan 17, 2014 at 22:42 +0530, Jagan Teki wrote:
On Fri, Jan 17, 2014 at 10:01 PM, Marek Vasut marex@denx.de wrote:
Anyway, I feel we're sinking deeper and deeper into this sh*t, we should instead take a step back and re-think the whole approach until we break it even more.
Yes - will shrink once we plan for new approach. But I'm unclear with new SPI-NOR.
Regarding this specific patch: I assume what Marek suggested was to restrict the "SPI slave" information to what's specific to an SPI slave. It's just not true that every SPI slave is a flash chip (an assumption which QSPI developers appear to fall for rather easily).
I was about to make a similar comment, that trimming the identifiers so rigorously leads to code that only "initiated" people can read. Even those who want to learn have no chance, there would not be a legend of any kind (except for the commit message, which soon is buried and not obvious to look up). And even with the legend, it's tedious to train the casual co-developer to those specific abbreviations, which may not even be in wide spread use outside of the U-Boot code base.
So I agree with Marek that we should take a deep breath, and be aware of the consequences before taking a specific direction (and having a clear direction would also be beneficial).
A more involved answer I will send to the quad SPI thread.
virtually yours Gerhard Sittig