On Sun, Aug 13, 2017 at 05:25:19AM +0300, Siarhei Siamashka wrote:
This reverts commit 00bbe96ebabbc83777cd8d6c6fd2791c5c8cf619.
There were two major problems with this patch:
- It made OMAP3530 devices non-bootable, as reported and investigated by Derald D. Woods in https://lists.denx.de/pipermail/u-boot/2017-August/302192.html
- The SPL code size increased really a lot.
SPL size for omap3_beagle_defconfig build with GCC 6:
== before reverting == text data bss dec hex filename 49721 1505 203200 254426 3e1da spl/u-boot-spl
== after reverting == text data bss dec hex filename 49233 1501 203200 253934 3dfee spl/u-boot-spl
Signed-off-by: Siarhei Siamashka siarhei.siamashka@gmail.com Reported-by: Derald D. Woods woods.technical@gmail.com
But the problem is that every TI "OMAP" SoC has this functionality and needs it as well (off th top of my head I forget if this applies to keystone a well). The changes in 00bbe96ebabb came about because we need to call that function in other, iirc, fastboot-related code areas. I am open to further ways to re-work this so that get_device_type() is smaller, but not to going back to just the way it was before. Perhaps a larger comment block and #if/elif casing of the address, in the cases where it's always known at build time? Thanks!