rv64: bug in encoding of bge?

[Paulo Matos]

On 09/01/2019 21:48, Jim Wilson wrote:
> On 1/9/19 6:10 AM, Paulo Matos wrote:
>> 0000000000000000 <foo>:
>>     0:   00c5d063                bge     a1,a2,0 <foo>
>>     4:   feb65ee3                bge     a2,a1,0 <foo>
>>     8:   fec5fce3                bgeu    a1,a2,0 <foo>
>>     c:   feb67ae3                bgeu    a2,a1,0 <foo>
>>
>>
>> The first encoding looks fine but the first byte of the others (0xfe)
>> just don't look correct. For example, for the second instruction I would
>> have expected:
>> Â Â Â  4:Â Â  00b65063Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â  bge a2,a1,0
> 
> The upper bits hold part of the offset.  The first instruction branches
> to itself so the offset is 0.  The following branches go backwards to
> address 0, so the offset is negative.  That causes the upper 7 bits to
> all be 1.
> 
Of course, I was totally confused because the dump shows
 bge     a2,a1,0
and my eyes were glossing over the <foo> at the end of each line.
Therefore I expected that all instructions should be looping to themselves.

But the original code has them all looping back to foo.

Got it, thanks.
-- 
Paulo Matos