On Wed, Jan 24, 2018 at 4:20 PM, Marek Vasut marek.vasut@gmail.com wrote:
Split sh_eth_recv into two functions, one which checks whether a packet was received and one which handles the received packet. This is done in preparation for DM support, which handles these two parts separately.
Signed-off-by: Marek Vasut marek.vasut+renesas@gmail.com Cc: Nobuhiro Iwamatsu iwamatsu@nigauri.org Cc: Joe Hershberger joe.hershberger@ni.com
drivers/net/sh_eth.c | 73 +++++++++++++++++++++++++++++++++------------------- 1 file changed, 46 insertions(+), 27 deletions(-)
diff --git a/drivers/net/sh_eth.c b/drivers/net/sh_eth.c index 7b11a5a0d3..99eab4c688 100644 --- a/drivers/net/sh_eth.c +++ b/drivers/net/sh_eth.c @@ -118,39 +118,58 @@ static int sh_eth_send_legacy(struct eth_device *dev, void *packet, int len) return sh_eth_send_common(eth, packet, len); }
-static int sh_eth_recv_common(struct sh_eth_dev *eth) +static int sh_eth_recv_start(struct sh_eth_dev *eth) { int port = eth->port, len = 0;
This "port" local variable seems pretty useless.
struct sh_eth_info *port_info = ð->port_info[port];
uchar *packet; /* Check if the rx descriptor is ready */ invalidate_cache(port_info->rx_desc_cur, sizeof(struct rx_desc_s));if (!(port_info->rx_desc_cur->rd0 & RD_RACT)) {/* Check for errors */if (!(port_info->rx_desc_cur->rd0 & RD_RFE)) {len = port_info->rx_desc_cur->rd1 & 0xffff;packet = (uchar *)ADDR_TO_P2(port_info->rx_desc_cur->rd2);invalidate_cache(packet, len);net_process_received_packet(packet, len);}/* Make current descriptor available again */if (port_info->rx_desc_cur->rd0 & RD_RDLE)port_info->rx_desc_cur->rd0 = RD_RACT | RD_RDLE;elseport_info->rx_desc_cur->rd0 = RD_RACT;flush_cache_wback(port_info->rx_desc_cur,sizeof(struct rx_desc_s));/* Point to the next descriptor */port_info->rx_desc_cur++;if (port_info->rx_desc_cur >=port_info->rx_desc_base + NUM_RX_DESC)port_info->rx_desc_cur = port_info->rx_desc_base;}
if (port_info->rx_desc_cur->rd0 & RD_RACT)return -EINVAL;/* Check for errors */if (port_info->rx_desc_cur->rd0 & RD_RFE)return -EINVAL;len = port_info->rx_desc_cur->rd1 & 0xffff;
Just return here directly and delete the "len" variable.
return len;+}
+static void sh_eth_recv_finish(struct sh_eth_dev *eth) +{
struct sh_eth_info *port_info = ð->port_info[eth->port];/* Make current descriptor available again */if (port_info->rx_desc_cur->rd0 & RD_RDLE)port_info->rx_desc_cur->rd0 = RD_RACT | RD_RDLE;elseport_info->rx_desc_cur->rd0 = RD_RACT;flush_cache_wback(port_info->rx_desc_cur,sizeof(struct rx_desc_s));/* Point to the next descriptor */port_info->rx_desc_cur++;if (port_info->rx_desc_cur >=port_info->rx_desc_base + NUM_RX_DESC)port_info->rx_desc_cur = port_info->rx_desc_base;+}
+static int sh_eth_recv_common(struct sh_eth_dev *eth) +{
int port = eth->port, len = 0;
I would also remove "port" here.
struct sh_eth_info *port_info = ð->port_info[port];uchar *packet = (uchar *)ADDR_TO_P2(port_info->rx_desc_cur->rd2);len = sh_eth_recv_start(eth);if (len > 0) {invalidate_cache(packet, len);net_process_received_packet(packet, len);sh_eth_recv_finish(eth);} else
I thought checkpatch would want you to have matching braces here. I think it looks cleaner at least. Probably a moot point though since...
len = 0;
Are you expecting "len" to be negative? If not, it will already be 0.
From the function above, it looks like it will always be >= 0.
/* Restart the receiver if disabled */ if (!(sh_eth_read(port_info, EDRRR) & EDRRR_R))-- 2.15.1
U-Boot mailing list U-Boot@lists.denx.de https://lists.denx.de/listinfo/u-boot