Add B extension

[ved-rivos]

Specification: https://github.com/riscv/riscv-b

[allenjbaum]

Yes, that's why Zmmul and M can both be set. Clearing misa.M is then just removing the DIV/REM instructions but not the MUL instructions. I don't know why anyone would ever do that, but mine is not to reason why....

…

On Sun, Jan 7, 2024 at 1:41 PM Ved Shanbhogue ***@***.***> wrote: ***@***.**** commented on this pull request. ------------------------------ In model/riscv_insts_zba.sail <#385 (comment)>: > @@ -71,8 +71,8 @@ /* ****************************************************************** */ union clause ast = RISCV_SLLIUW : (bits(6), regidx, regidx) -mapping clause encdec = RISCV_SLLIUW(shamt, rs1, rd) if haveZba() & sizeof(xlen) == 64 - <-> 0b000010 @ shamt @ rs1 @ 0b001 @ rd @ 0b0011011 if haveZba() & sizeof(xlen) == 64 +mapping clause encdec = RISCV_SLLIUW(shamt, rs1, rd) if (haveZba() | haveBExt()) & sizeof(xlen) == 64 But that not allow expressing behavior of clearing B clearly. For instance, if misa.B and haveZba was both 1 at reset. Then clearing B would leave Zba available. But if misa.B was 1 and haveZba was 0 at reset then clearing B would disable Zba. Forcing state of Zba to misa.B would not allow expressing this. — Reply to this email directly, view it on GitHub <#385 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AHPXVJXNODU5VNGOK5IACG3YNMJAXAVCNFSM6AAAAABBQPMYXSVHI2DSMVQWIX3LMV43YUDVNRWFEZLROVSXG5CSMV3GSZLXHMYTQMBXHA4DCNBQGI> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[rpsene]

@billmcspadden-riscv this spec was recently ratified, can this be merged?

[Alasdair]

Does the haveBExt function this defines not need to be used somewhere?

[ved-rivos]

Please see the conversation thread in this PR.

[github-actions]

Test Results

712 tests  ±0   712 ✅ ±0   0s ⏱️ ±0s
  6 suites ±0     0 💤 ±0 
  1 files   ±0     0 ❌ ±0 

Results for commit dc41919. ± Comparison against base commit 40ed0c5.

♻️ This comment has been updated with latest results.

[Timmmm]

LGTM, thanks! @jrtc27 does this look ok to you?

[Timmmm]

I'll merge this in a couple of days if nobody objects. @ved-rivos what would you say your highest priority PR is? I think it would be good to get some more merged if possible.

[ved-rivos]

@Timmmm - dont really have a priority/preference - maybe this one #393 and #395 and #335 ?