On Fri, Dec 26, 2008 at 7:59 PM, Rudolf Marek r.marek@assembler.cz wrote:
Hello,
Following patch fixes VIA SPI (VT8237S). It needs to have opcodes initialized same way as ICH7.
I am glad that the OPCODE patch is more useful than scratching my own itch :-)
Because its just a prefix opcode. And prefix opcodes are not checked. Question is now what to do. Silently complete command the spi_write_enable, when we find the opcode in prefix opcode? The prefix opcode should get executed with the prefix automatically.
For an ad hoc solution, we can ignore the request to execute a command in ich_spi_command if we find that command is a preop. Later when chip erase is executed, some predefined (in pops[] ) preop will be picked up.
However, I don't think it should be fixed in that way.
One of the problems behind is design inconsistency in different operations. For chip/block/sector erase operations, spi_write_enable is integrated into those operations; for byte program, spi_write_enable is not integrated.
yu ning