On 13.11.2017 15:35, Mike Looijmans wrote: > On 10-11-17 11:58, Michal Simek wrote: >> Hi, >> >> this series is trying to cleanup ps7_init* file that we don't need to >> have the same copy of the same functions in different locations. >> This work is done based on solution from Topic.nl for miami boards >> where format was changed a little bit to save one word in config data >> segment. > > Nice to see one's work appreciated. A bit of comment below. > >> At the same time older method of simply copying files to particular >> folder is still working. >> Please test this in your board to make sure I didn't break anything. >> I have tested it on zybo/zc702/zc706. >> >> Thanks, >> Michal >> >> >> Michal Simek (9): >> arm: zynq: Add missing ps7_post_config declaration >> arm: zynq: Enable debug uart on zc706 >> arm: zynq: Remove ps7_debug code >> arm: zynq: Move ps7_* to separate file >> arm: zynq: Get rid of ps7_reset_apu() for syzygy board >> arm: zynq: Move common ps7_init* initialization to arch code >> arm: zynq: Add ps7GetSiliconVersion() to ps7_spl_init >> arm: zynq: Convert EMIT_WRITE to EMIT_MASKWRITE > > There's actually room for one more opcode, so I would recommend > implementing MASK_WRITE instead of forcing everything to be done in > read-modify-write mode. > > #define OPCODE_MASKWRITE 3U > #define EMIT_MASKWRITE(addr, val) OPCODE_MASKWRITE | addr, val
Ah right - it can be done like that. Not an issue. > > This "constant table" approach could also be implemented for the zynqmp > platforms, cutting down the size of the SPL considerably. Probably. > > (Only thing remaining is to have Vivado output these files properly...) Which is the biggest problem. I don't think this will happen. Thanks, Michal _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de https://lists.denx.de/listinfo/u-boot
