iv/linux
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

firewire: ohci: avoid separate DMA mapping for small AT payloads

Browse Source 
For AT packet payloads of up to eight bytes, we have enough unused space
in the DMA descriptors list so that we can put a copy of the payload
there and thus avoid having to create a separate streaming DMA mapping
for the payload buffer.

In a CPU-bound microbenchmark that just sends 8-byte packets, bandwidth
was measured to increase by 5.7 %, from 1009 KB/s to 1067 KB/s.  In
practice, the only performance-sensitive usage of small asynchronous
packets is the SBP-2 driver's write to the ORB_POINTER register during
SCSI command submission.

Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de>
This commit is contained in:
Clemens Ladisch 2011-04-11 09:57:54 +02:00 committed by Stefan Richter
parent ecf8328e59
commit da28947e7e
1 changed files with 17 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
drivers/firewire/ohci.c
View File

@ -1218,6 +1218,7 @@ static void context_stop(struct context *ctx)
}
struct driver_data {
u8 inline_data[8];
struct fw_packet *packet;
};
@ -1301,20 +1302,28 @@ static int at_context_queue_packet(struct context *ctx,
return -1;
}
BUILD_BUG_ON(sizeof(struct driver_data) > sizeof(struct descriptor));
driver_data = (struct driver_data *) &d[3];
driver_data->packet = packet;
packet->driver_data = driver_data;
if (packet->payload_length > 0) {
payload_bus =
dma_map_single(ohci->card.device, packet->payload,
packet->payload_length, DMA_TO_DEVICE);
if (packet->payload_length > sizeof(driver_data->inline_data)) {
payload_bus = dma_map_single(ohci->card.device,
packet->payload,
packet->payload_length,
DMA_TO_DEVICE);
if (dma_mapping_error(ohci->card.device, payload_bus)) {
packet->ack = RCODE_SEND_ERROR;
return -1;
}
packet->payload_bus = payload_bus;
packet->payload_mapped = true;
} else {
memcpy(driver_data->inline_data, packet->payload,
packet->payload_length);
payload_bus = d_bus + 3 * sizeof(*d);
}
d[2].req_count = cpu_to_le16(packet->payload_length);
d[2].data_address = cpu_to_le32(payload_bus);