[v4] esp: fix asserts/segfaults discovered by fuzzer

[PATCH v4 for-6.0 10/12] esp: don't reset async_len directly in esp_select() if cancelling request

Posted by Mark Cave-Ayland 4 years, 10 months ago

Instead let the SCSI layer invoke the .cancel callback itself to cancel and
reset the request state.

Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Tested-by: Alexander Bulekov <alxndr@bu.edu>
---
 hw/scsi/esp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/hw/scsi/esp.c b/hw/scsi/esp.c
index 782c6ee357..3b9037e4f4 100644
--- a/hw/scsi/esp.c
+++ b/hw/scsi/esp.c
@@ -95,6 +95,7 @@ void esp_request_cancelled(SCSIRequest *req)
         scsi_req_unref(s->current_req);
         s->current_req = NULL;
         s->current_dev = NULL;
+        s->async_len = 0;
     }
 }
 
@@ -206,7 +207,6 @@ static int esp_select(ESPState *s)
     if (s->current_req) {
         /* Started a new command before the old one finished.  Cancel it.  */
         scsi_req_cancel(s->current_req);
-        s->async_len = 0;
     }
 
     s->current_dev = scsi_device_find(&s->bus, 0, target, 0);
-- 
2.20.1

Re: [PATCH v4 for-6.0 10/12] esp: don't reset async_len directly in esp_select() if cancelling request

Posted by Philippe Mathieu-Daudé 4 years, 10 months ago

On 4/7/21 9:57 PM, Mark Cave-Ayland wrote:
> Instead let the SCSI layer invoke the .cancel callback itself to cancel and
> reset the request state.
> 
> Signed-off-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
> Tested-by: Alexander Bulekov <alxndr@bu.edu>
> ---
>  hw/scsi/esp.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>