[Qemu-devel] [PATCH 1/2] serial: refine serial_thr_ipending_needed

[Paolo Bonzini <pbonzini@redhat.com>]

If the THR interrupt is disabled, there is no need to migrate thr_ipending
because LSR.THRE will be sampled again when the interrupt is enabled.
(This is the behavior that is not documented in the datasheet, but
relied on by Windows!)

Note that in this case IIR will never be 0x2 so, if thr_ipending were
to be one, QEMU would produce the subsection.

Reported-by: Igor Mammedov <imammedo@redhat.com>
Cc: qemu-stable@nongnu.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 hw/char/serial.c | 13 +++++++++++--
 1 file changed, 11 insertions(+), 2 deletions(-)

diff --git a/hw/char/serial.c b/hw/char/serial.c
index 6d522ff..3aca874 100644
--- a/hw/char/serial.c
+++ b/hw/char/serial.c
@@ -645,8 +645,17 @@ static int serial_post_load(void *opaque, int version_id)
 static bool serial_thr_ipending_needed(void *opaque)
 {
     SerialState *s = opaque;
-    bool expected_value = ((s->iir & UART_IIR_ID) == UART_IIR_THRI);
-    return s->thr_ipending != expected_value;
+
+    if (s->ier & UART_IER_THRI) {
+        bool expected_value = ((s->iir & UART_IIR_ID) == UART_IIR_THRI);
+        return s->thr_ipending != expected_value;
+    } else {
+        /* LSR.THRE will be sampled again when the interrupt is enabled.
+         * thr_ipending is not used in this case, do not migrate it.  It
+         * will be set to false on the destination machine.
+         */
+        return false;
+    }
 }
 
 const VMStateDescription vmstate_serial_thr_ipending = {
-- 
2.1.0