On 4 September 2012 19:16, Stefan Weil <s...@weilnetz.de> wrote: > Am 04.09.2012 20:12, schrieb Stefan Weil: >> Am 04.09.2012 19:57, schrieb Peter Maydell: >>> Checked against the data sheet -- last documented register is at >>> offset $1F0, so correcting the offset check rather than the array >>> length is the correct fix.
>> Then m5206_mbar_width should be shortened to 124 elements >> (0x1f0 / 4) _and_ the offset check needs a correction. Why bother? The relevant offsets will hit the hw_error() cases in m5206_mbar_read() and m5206_mbar_write() anyway, the same as for the other cases where there are "holes" in the register space. The only reason we're doing these checks here is to avoid overrunning the width array... > Sorry, 125 elements, of course. Or are there undocumented > registers at 0x1f4, 0x1f8 and 0x1fc? If there were, I wouldn't know, because they aren't documented :-) -- PMM
