From: Bjørn Mork Date: Fri, 1 Nov 2013 10:16:49 +0000 (+0100) Subject: net: cdc_ncm: no point in filling up the NTBs if we send ZLPs X-Git-Tag: firefly_0821_release~176^2~4992^2~79^2~12 X-Git-Url: http://demsky.eecs.uci.edu/git/?a=commitdiff_plain;h=4d619f625a6011618de9fccc4b282d314d792897;p=firefly-linux-kernel-4.4.55.git net: cdc_ncm: no point in filling up the NTBs if we send ZLPs Padding NTBs to max size is part of the support for devices optimizing their DMA transfers. This optimization depends on max sized NTBs not being ZLP terminated. So we are much better off dropping the padding if we are going to send a ZLP anyway. Cc: Alexey Orishko Signed-off-by: Bjørn Mork Signed-off-by: David S. Miller --- diff --git a/drivers/net/usb/cdc_ncm.c b/drivers/net/usb/cdc_ncm.c index 5aa3e60851d0..42c86200f076 100644 --- a/drivers/net/usb/cdc_ncm.c +++ b/drivers/net/usb/cdc_ncm.c @@ -800,8 +800,12 @@ cdc_ncm_fill_tx_frame(struct usbnet *dev, struct sk_buff *skb, __le32 sign) * would be more efficient for USB HS mobile device with DMA * engine to receive a full size NTB, than canceling DMA * transfer and receiving a short packet. + * + * This optimization support is pointless if we end up sending + * a ZLP after full sized NTBs. */ - if (skb_out->len > CDC_NCM_MIN_TX_PKT) + if (!(dev->driver_info->flags & FLAG_SEND_ZLP) && + skb_out->len > CDC_NCM_MIN_TX_PKT) memset(skb_put(skb_out, ctx->tx_max - skb_out->len), 0, ctx->tx_max - skb_out->len); else if ((skb_out->len % dev->maxpacket) == 0)