+-- On Tue, 31 May 2016, Peter Maydell wrote --+
| This is a FIFO *read*, and the trace message is for the case when the FIFO
| is too full, which can't happen for a read.
Okay.
| OK, so we need to fix these code paths so that they keep all of the read,
| write and size values in sync (or jus
On 31 May 2016 at 06:20, P J P wrote:
> Hello Peter,
>
> +-- On Mon, 30 May 2016, Peter Maydell wrote --+
> | > +} else if (s->ti_rptr < TI_BUFSZ) {
> | > s->rregs[ESP_FIFO] = s->ti_buf[s->ti_rptr++];
> | > +} else {
> | > +trace_esp_error
Hello Peter,
+-- On Mon, 30 May 2016, Peter Maydell wrote --+
| > +} else if (s->ti_rptr < TI_BUFSZ) {
| > s->rregs[ESP_FIFO] = s->ti_buf[s->ti_rptr++];
| > +} else {
| > +trace_esp_error_fifo_overrun();
|
| Isn't this an underrun, not an
On 30 May 2016 at 19:58, P J P wrote:
> From: Prasad J Pandit
>
> The 53C9X Fast SCSI Controller(FSC) comes with internal 16-byte
> FIFO buffers. One is used to handle commands and other is for
> information transfer. While reading/writing to these buffers
> an index into 's->ti_buf[TI_BUFSZ=16]'
From: Prasad J Pandit
The 53C9X Fast SCSI Controller(FSC) comes with internal 16-byte
FIFO buffers. One is used to handle commands and other is for
information transfer. While reading/writing to these buffers
an index into 's->ti_buf[TI_BUFSZ=16]' could exceed its size,
as a check was missing to