diff options
| author | Russell King <rmk@dyn-67.arm.linux.org.uk> | 2007-03-04 20:21:38 +0000 |
|---|---|---|
| committer | Russell King <rmk+kernel@arm.linux.org.uk> | 2007-03-04 20:33:31 +0000 |
| commit | 23d046f43a05155e050a68f3ad1f19b672713374 (patch) | |
| tree | 5c32d8fde63c81a76083d2a550d5ba02ba44a873 /drivers/scsi/arm/powertec.c | |
| parent | 6b4df7ee1f636f0dbf3896235582559c86cb18aa (diff) | |
[ARM] ARM SCSI: Don't try to dma_map_sg too many scatterlist entries
An off-by-one bug meant we were always trying to map one too many
scatterlist entries. This was mostly harmless prior to the checks
going in to consistent_sync(), but now causes the kernel to BUG.
Also, powertec.c was missing an assignment to info->ec.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Diffstat (limited to 'drivers/scsi/arm/powertec.c')
| -rw-r--r-- | drivers/scsi/arm/powertec.c | 5 |
1 files changed, 3 insertions, 2 deletions
diff --git a/drivers/scsi/arm/powertec.c b/drivers/scsi/arm/powertec.c index f9cd20bfb95..159047a3499 100644 --- a/drivers/scsi/arm/powertec.c +++ b/drivers/scsi/arm/powertec.c @@ -148,10 +148,10 @@ powertecscsi_dma_setup(struct Scsi_Host *host, struct scsi_pointer *SCp, map_dir = DMA_FROM_DEVICE, dma_dir = DMA_MODE_READ; - dma_map_sg(dev, info->sg, bufs + 1, map_dir); + dma_map_sg(dev, info->sg, bufs, map_dir); disable_dma(dmach); - set_dma_sg(dmach, info->sg, bufs + 1); + set_dma_sg(dmach, info->sg, bufs); set_dma_mode(dmach, dma_dir); enable_dma(dmach); return fasdma_real_all; @@ -342,6 +342,7 @@ powertecscsi_probe(struct expansion_card *ec, const struct ecard_id *id) info->base = base; powertecscsi_terminator_ctl(host, term[ec->slot_no]); + info->ec = ec; info->info.scsi.io_base = base + POWERTEC_FAS216_OFFSET; info->info.scsi.io_shift = POWERTEC_FAS216_SHIFT; info->info.scsi.irq = ec->irq; |