On Mon, May 02, 2016 at 11:12:11AM +0200, Alexander Graf wrote:
[snip]
So Tom, how would you like to roll this? We can either
- Check raw after fs, default to fs and revert my patch or
- Leave fs boot broken (regression) or
- Leave raw boot broken (same as 2016.03)
Given that release is in 1 week, I'm wary on option 1. I also don't like regressions. So how about we revert my patch and fix it up with fs-before-raw boot for 2016.07?
So, we need to revert ef5ebe951bec72 which is what changed the raw offset. I think we also need to revert 22d90d560a2b which yes, will go back to breaking raw MLO + FS U-Boot. The problem here is that we have the fallback case of "load hard-coded size from MMC, assume u-boot.bin is raw". So raw will never fail. Marek posted a series late last week that would address this by letting us say that we want to call an invalid signature for U-Boot bad and continue on to other methods. I'm wondering if this doesn't go far enough actually and we should say that unless specifically set (like CONFIG_SPL_NAND_RAW_ONLY does), we don't assume it's good. It's been a long time since we didn't default to making a u-boot.img and SPL has "always" supported both. It was mainly legacy code for letting people mix-and-match SPL + X-Loader (and go back and forth). I'm leaning towards making this an opt-in thing and introduce CONFIG_SPL_MMC_BIN_ONLY (and rename the NAND one) as the logic is getting really really convoluted now.