On 11/12/2026 6:48 PM, Christian S. Lima wrote:
According to the Zicfilp specification the bit SPELP in the mstatus
register should be 0 (read-only) when s-mode is not present.

Resolves: https://gitlab.com/qemu-project/qemu/-/work_items/4456#note_4181630619

The code LGTM but this tag doesn't do much as is.  Gitlab uses the tag to 
automatically
close the issue when the patch lands upstream.  It can't do much with a comment.

Reading the bug I believe the reporter intention was to add this SPELP comment
as a follow-up of what you've done in 4456, i.e. this should be done in that 
same
scope.  If we upstream them separately there's a risk of Gitlab closing 4456 
and,
if this patch is left behind, the reporter can notice that the bug wasn't fully
resolved and reopen it.

My suggestion is to either squash this change into the 4456 patch or send it 
together
as a 2 patch series.  Both are fine, as long as we upstream both changes at the
same time.


Thanks,
Daniel


Signed-off-by: Christian S. Lima <[email protected]>
---
  target/riscv/tcg/csr.c | 5 ++++-
  1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/target/riscv/tcg/csr.c b/target/riscv/tcg/csr.c
index af444e3e6f..340c61a1c4 100644
--- a/target/riscv/tcg/csr.c
+++ b/target/riscv/tcg/csr.c
@@ -2124,7 +2124,10 @@ static RISCVException write_mstatus(CPURISCVState *env, 
int csrno,
/* If cfi lp extension is available, then apply cfi lp mask */
      if (env_archcpu(env)->cfg.ext_zicfilp) {
-        mask |= (MSTATUS_MPELP | MSTATUS_SPELP);
+        mask |= MSTATUS_MPELP;
+        if (riscv_has_ext(env, RVS)) {
+            mask |= MSTATUS_SPELP;
+        }
      }
mstatus = (mstatus & ~mask) | (val & mask);


Reply via email to