On 7/30/26 4:30 AM, Abhigyan Kumar wrote:
tdata_csr_write is approachable by guest programs. In case of unassigned
values written to a tdata CSR in debug mode, the switch-case falls
through straight to default causing an immediate abort. The risc-v spec
says:

     All tdata registers follow write-any-read-legal semantics. If a debugger
     writes an unsupported configuration, the register will read back a value
     that is supported (which may simply be a disabled trigger). This means
     that a debugger must always read back values it writes to tdata
     registers, unless it already knows what is supported.

This fix correctly ensures that the WARL behaviour of tdata CSRs.

Fixes: a42bd001 ("target/riscv: debug: Determine the trigger type from
tdata1.type")
Signed-off-by: Abhigyan Kumar <[email protected]>
---
  target/riscv/tcg/debug.c | 3 ++-
  1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/target/riscv/tcg/debug.c b/target/riscv/tcg/debug.c
index 3c0fe7010..ae493235d 100644
--- a/target/riscv/tcg/debug.c
+++ b/target/riscv/tcg/debug.c
@@ -913,7 +913,8 @@ void tdata_csr_write(CPURISCVState *env, int tdata_index, 
target_ulong val)
                        trigger_type);
          break;
      default:
-        g_assert_not_reached();
+        qemu_log_mask(LOG_GUEST_ERROR, "trigger type: %d is unassigned\n",
+                trigger_type);
      }
  }

Greetings. It has been over two weeks since I submitted this patch. It's my humble request to know the status of the review/acceptance for my patch. Please inform me about any required amends.

Thank you for your patience and attention.

Reply via email to