On 30/10/2018 06:47, Hervé Poussineau wrote:
> Le 29/10/2018 à 14:39, Mark Cave-Ayland a écrit :
>> From: Laurent Vivier <laur...@vivier.eu>
>>
>> Co-developed-by: Mark Cave-Ayland <mark.cave-ayl...@ilande.co.uk>
>> Signed-off-by: Mark Cave-Ayland <mark.cave-ayl...@ilande.co.uk>
>> Signed-off-by: Laurent Vivier <laur...@vivier.eu>
>> ---
>> hw/scsi/esp.c | 292
>> +++++++++++++++++++++++++++++++++++++++++++++-----
>> include/hw/scsi/esp.h | 7 ++
>> 2 files changed, 270 insertions(+), 29 deletions(-)
>>
>
> [...]
>
>> @@ -409,6 +564,7 @@ uint64_t esp_reg_read(ESPState *s, uint32_t saddr)
>> s->ti_size--;
>> s->rregs[ESP_FIFO] = s->ti_buf[s->ti_rptr++];
>> }
>> + esp_raise_irq(s);
>> if (s->ti_rptr == s->ti_wptr) {
>> s->ti_rptr = 0;
>> s->ti_wptr = 0;
>
> You're now raising irq (not drq) when reading registers. In my opinion, this
> deserves
> its own patch with an explanation.
Oooh good spot. Laurent, can you explain a bit more as to how the pseudo-DMA is
wired
up for ESP?
ATB,
Mark.
