Feedback Comments for C spec

[riscv]

Feedback Comments for C spec

[kasanovic]

rd=0 is not allowed for either C.ADDI or C.ADDIW. C.NOP is treated as a separate encoding in table, though implementations can implement as C.ADDI with destination of 0.

Agree on CA format - will be added.

[hirooih]

Hi,

I've been completely forgetting about this ticket.

I referred v.2.1 when I wrote this. I've checked v2.2 but it remains same.

v2.2 P. 76 Integer Register-Immediate Operations

The rd/rs1 field C.ADDI is described as "dest", not "dest !=0".

P.79 NOP Instruction

C.NOP is encoded as c.addi x0, 0 and so expands to addi x0, x0, 0.

P.82 Table12.5

Both C.NOP and C.ADDI are descrbied. The rd/rs1 field C.ADDI is described as "dest != 0", not "dest".

P. 76 and P. 79 are consistent. But they are not consistent with P.82 a little bit.

[aswaterman]

If this is a new issue, please open a new ticket

[hirooih]

If this is a new issue, please open a new ticket

No. It is an old issue opened on April, 2017.
Please take your time to read and reopen it.

[hirooih]

Just a note.
I've confirmed all issues I reported are fixed on riscv-spec-imfdqc-ratification-20190305.pdf.
Thank you.