Re: [PATCH v4 02/23] hw/char/sh_serial: Use hw_error instead of fprintf and abort

[Thomas Huth]

On 29/10/2021 07.38, Philippe Mathieu-Daudé wrote:
On 10/28/21 21:27, BALATON Zoltan wrote:
It does the same with dumping some more state but avoids calling abort
directly and printing to stderr from the device model.

hw_error() is unfortunately misnamed, it is meant for CPU code,
and we want to get ride of it. What you probably want here is
error_report() which also reports to the monitor.

Looking at the text of the messages, maybe it would be even better to use 
qemu_log_mask(LOG_UNIMP, ...) or qemu_log_mask(LOG_GUEST_ERROR, ...) ?

 Thomas


Signed-off-by: BALATON Zoltan <bala...@eik.bme.hu>
---
  hw/char/sh_serial.c | 9 +++------
  1 file changed, 3 insertions(+), 6 deletions(-)

diff --git a/hw/char/sh_serial.c b/hw/char/sh_serial.c
index 1b1e6a6a04..dbefb51d71 100644
--- a/hw/char/sh_serial.c
+++ b/hw/char/sh_serial.c
@@ -26,6 +26,7 @@
   */
  
  #include "qemu/osdep.h"
+#include "hw/hw.h"
  #include "hw/irq.h"
  #include "hw/sh4/sh.h"
  #include "chardev/char-fe.h"
@@ -200,9 +201,7 @@ static void sh_serial_write(void *opaque, hwaddr offs,
          }
      }
  
-    fprintf(stderr, "sh_serial: unsupported write to 0x%02"
-            HWADDR_PRIx "\n", offs);
-    abort();
+    hw_error("sh_serial: unsupported write to 0x%02"HWADDR_PRIx"\n", offs);
  }
  
  static uint64_t sh_serial_read(void *opaque, hwaddr offs,
@@ -307,9 +306,7 @@ static uint64_t sh_serial_read(void *opaque, hwaddr offs,
  #endif
  
      if (ret & ~((1 << 16) - 1)) {
-        fprintf(stderr, "sh_serial: unsupported read from 0x%02"
-                HWADDR_PRIx "\n", offs);
-        abort();
+        hw_error("sh_serial: unsupported read from 0x%02"HWADDR_PRIx"\n", 
offs);
      }
  
      return ret;