Hi Titus,
On 21/1/25 11:36, Titus Rwantare wrote:
This has been useful when debugging when unsure if the guest is
generating i2c traffic.
Signed-off-by: Titus Rwantare <tit...@google.com>
---
hw/misc/i2c-echo.c | 13 +++++++++++++
hw/misc/trace-events | 5 +++++
2 files changed, 18 insertions(+)
diff --git a/hw/misc/i2c-echo.c b/hw/misc/i2c-echo.c
index 5ae3d0817e..06110e0f8e 100644
--- a/hw/misc/i2c-echo.c
+++ b/hw/misc/i2c-echo.c
@@ -13,6 +13,7 @@
#include "qemu/main-loop.h"
#include "block/aio.h"
#include "hw/i2c/i2c.h"
+#include "trace.h"
#define TYPE_I2C_ECHO "i2c-echo"
OBJECT_DECLARE_SIMPLE_TYPE(I2CEchoState, I2C_ECHO)
@@ -80,11 +81,19 @@ static int i2c_echo_event(I2CSlave *s, enum i2c_event event)
case I2C_START_RECV:
state->pos = 0;
+ trace_i2c_echo_event(DEVICE(s)->canonical_path, "I2C_START_RECV");
break;
case I2C_START_SEND:
state->pos = 0;
+ trace_i2c_echo_event(DEVICE(s)->canonical_path, "I2C_START_SEND");
+ break;
+
+ case I2C_START_SEND_ASYNC:
+ state->pos = 0;
+
+ trace_i2c_echo_event(DEVICE(s)->canonical_path,
"I2C_START_SEND_ASYNC");
This patch is doing a bit more that what it describes, since now
START_SEND_ASYNC returns 0x00 instead of 0xff (besides updating pos).
Please split as one patch per logical change.
Thanks!
Phil.
break;
case I2C_FINISH:
