On Mon, 18 Nov 2019, Finn Thain wrote:
> On Sun, 17 Nov 2019, Kars de Jong wrote:
>
> > The only [time when] the driver reads these registers is after a data
> > transfer. These are done using DMA on all Zorro boards, so I don't
> > think there's a risk of stale values from a PIO transfer there.
> >
>
> I'm not entirely sure that the chip is unaffected by stale counter
> values.
>
> (Stale transfer counter values are distinct from stale transfer count
> register values. Both are addressed by the patch.)
>
Sorry -- I should have said, "both were _intended_ to be addressed by the
patch". But, as Michael pointed out, the DMA NOP command was missing from
the v1 patch. Please see revised patch below.
diff --git a/drivers/scsi/esp_scsi.c b/drivers/scsi/esp_scsi.c
index bb88995a12c7..82d49f0f09df 100644
--- a/drivers/scsi/esp_scsi.c
+++ b/drivers/scsi/esp_scsi.c
@@ -2835,6 +2835,10 @@ void esp_send_pio_cmd(struct esp *esp, u32 addr, u32 esp_count,
cmd &= ~ESP_CMD_DMA;
esp->send_cmd_error = 0;
+ esp_write8(0, ESP_TCLOW);
+ esp_write8(0, ESP_TCMED);
+ scsi_esp_cmd(esp, ESP_CMD_NULL | ESP_CMD_DMA);
+
if (write) {
u8 *dst = (u8 *)addr;
u8 mask = ~(phase == ESP_MIP ? ESP_INTR_FDONE : ESP_INTR_BSERV);
diff --git a/drivers/scsi/mac_esp.c b/drivers/scsi/mac_esp.c
index 1c78bc10c790..797579247e47 100644
--- a/drivers/scsi/mac_esp.c
+++ b/drivers/scsi/mac_esp.c
@@ -361,8 +361,6 @@ static int esp_mac_probe(struct platform_device *dev)
esp->flags = ESP_FLAG_NO_DMA_MAP;
if (mep->pdma_io == NULL) {
printk(KERN_INFO PFX "using PIO for controller %d\n", dev->id);
- esp_write8(0, ESP_TCLOW);
- esp_write8(0, ESP_TCMED);
esp->flags |= ESP_FLAG_DISABLE_SYNC;
mac_esp_ops.send_dma_cmd = esp_send_pio_cmd;
} else {
