Please try this patch:
--- drivers/scsi/aic7xxx.c.~1~ Fri Feb 18 11:59:12 2000
+++ drivers/scsi/aic7xxx.c Sat Feb 19 04:00:04 2000
@@ -719,6 +719,11 @@
#define aic7xxx_position(cmd) ((cmd)->SCp.have_data_in)
/*
+ * The stored DMA mapping for single-buffer data transfers.
+ */
+#define aic7xxx_mapping(cmd) ((cmd)->SCp.phase)
+
+/*
* So we can keep track of our host structs
*/
static struct aic7xxx_host *first_aic7xxx = NULL;
@@ -2902,9 +2907,16 @@
pci_unmap_sg(p->pdev, sg, cmd->use_sg, scsi_to_pci_dma_dir(cmd->sc_data_direction));
}
else if (cmd->request_bufflen)
- pci_unmap_single(p->pdev, le32_to_cpu(scb->sg_list[0].address),
- le32_to_cpu(scb->sg_list[0].length),
+ pci_unmap_single(p->pdev, aic7xxx_mapping(cmd),
+ cmd->request_bufflen,
scsi_to_pci_dma_dir(cmd->sc_data_direction));
+ if (scb->flags & SCB_SENSE)
+ {
+ pci_unmap_single(p->pdev,
+ le32_to_cpu(scb->sg_list[0].address),
+ sizeof(cmd->sense_buffer),
+ PCI_DMA_FROMDEVICE);
+ }
if (scb->flags & SCB_RECOVERY_SCB)
{
p->flags &= ~AHC_ABORT_PENDING;
@@ -4889,7 +4901,7 @@
scb->sg_list[0].address =
cpu_to_le32(pci_map_single(p->pdev, sense_buffer,
sizeof(cmd->sense_buffer),
- scsi_to_pci_dma_dir(cmd->sc_data_direction)));
+ PCI_DMA_FROMDEVICE));
hscb->data_pointer = scb->sg_list[0].address;
scb->flags |= SCB_SENSE;
@@ -10986,6 +10998,7 @@
unsigned int address = pci_map_single(p->pdev, cmd->request_buffer,
cmd->request_bufflen,
scsi_to_pci_dma_dir(cmd->sc_data_direction));
+ aic7xxx_mapping(cmd) = address;
scb->sg_list[0].address = cpu_to_le32(address);
scb->sg_list[0].length = cpu_to_le32(cmd->request_bufflen);
scb->sg_count = 1;
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sat Feb 19 2000 - 07:17:22 EST