On 8/14/2026 12:33 PM, Philippe Mathieu-Daudé wrote:
Hi,
(merged as 9b63a54dd2c0ea9a667c9e9dc662755b9feffa7e)
On 11/6/26 07:28, Brian Cain wrote:
From: Brian Cain <[email protected]>
Co-authored-by: Sid Manning <[email protected]>
Reviewed-by: Taylor Simpson <[email protected]>
Signed-off-by: Brian Cain <[email protected]>
---
target/hexagon/cpu.c | 1 -
target/hexagon/cpu_helper.c | 354 ++++++++++++++++++++++++++++++++++++
target/hexagon/op_helper.c | 33 +++-
3 files changed, 384 insertions(+), 4 deletions(-)
create mode 100644 target/hexagon/cpu_helper.c
+static void hexagon_resume_thread(CPUHexagonState *env)
+{
+ CPUState *cs = env_cpu(env);
+ clear_wait_mode(env);
+ /*
+ * The wait instruction keeps the PC pointing to itself
+ * so that it has an opportunity to check for interrupts.
+ *
+ * When we come out of wait mode, adjust the PC to the
+ * next executable instruction.
+ */
+ env->gpr[HEX_REG_PC] = env->wait_next_pc;
+ cs = env_cpu(env);
+ ASSERT_DIRECT_TO_GUEST_UNSET(env, cs->exception_index);
+ cs->halted = false;
+ cs->exception_index = HEX_EVENT_NONE;
+ qemu_cpu_kick(cs);
Why not call cpu_resume()?
If you don't mind I'll send a patch to clean that up.
Sure - that seems reasonable.
Thanks!
+}
