Message ID | 20171202053723.32660-1-jakub.kicinski@netronome.com |
---|---|
State | Accepted, archived |
Delegated to: | David Miller |
Headers | show |
Series | [net] nfp: fix port stats for mac representors | expand |
On Fri, Dec 1, 2017 at 9:37 PM, Jakub Kicinski <jakub.kicinski@netronome.com> wrote: > From: Pieter Jansen van Vuuren <pieter.jansenvanvuuren@netronome.com> > > Previously we swapped the tx_packets, tx_bytes and tx_dropped counters > with rx_packets, rx_bytes and rx_dropped counters, respectively. This > behaviour is correct and expected for VF representors but it should not > be swapped for physical port mac representors. Ah, I forgot to point the finger. Should I repost? Fixes: eadfa4c3be99 ("nfp: add stats and xmit helpers for representors") > Signed-off-by: Pieter Jansen van Vuuren <pieter.jansenvanvuuren@netronome.com> > Reviewed-by: Simon Horman <simon.horman@netronome.com> > Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com>
From: Jakub Kicinski <jakub.kicinski@netronome.com> Date: Fri, 1 Dec 2017 21:37:23 -0800 > From: Pieter Jansen van Vuuren <pieter.jansenvanvuuren@netronome.com> > > Previously we swapped the tx_packets, tx_bytes and tx_dropped counters > with rx_packets, rx_bytes and rx_dropped counters, respectively. This > behaviour is correct and expected for VF representors but it should not > be swapped for physical port mac representors. > > Signed-off-by: Pieter Jansen van Vuuren <pieter.jansenvanvuuren@netronome.com> > Reviewed-by: Simon Horman <simon.horman@netronome.com> > Reviewed-by: Jakub Kicinski <jakub.kicinski@netronome.com> Applied.
From: Jakub Kicinski <jakub.kicinski@netronome.com> Date: Sat, 2 Dec 2017 02:03:31 -0800 > On Fri, Dec 1, 2017 at 9:37 PM, Jakub Kicinski > <jakub.kicinski@netronome.com> wrote: >> From: Pieter Jansen van Vuuren <pieter.jansenvanvuuren@netronome.com> >> >> Previously we swapped the tx_packets, tx_bytes and tx_dropped counters >> with rx_packets, rx_bytes and rx_dropped counters, respectively. This >> behaviour is correct and expected for VF representors but it should not >> be swapped for physical port mac representors. > > Ah, I forgot to point the finger. Should I repost? > > Fixes: eadfa4c3be99 ("nfp: add stats and xmit helpers for representors") I added this to the commit message, thanks.
diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_repr.c b/drivers/net/ethernet/netronome/nfp/nfp_net_repr.c index 924a05e05da0..78b36c67c232 100644 --- a/drivers/net/ethernet/netronome/nfp/nfp_net_repr.c +++ b/drivers/net/ethernet/netronome/nfp/nfp_net_repr.c @@ -84,16 +84,13 @@ nfp_repr_phy_port_get_stats64(struct nfp_port *port, { u8 __iomem *mem = port->eth_stats; - /* TX and RX stats are flipped as we are returning the stats as seen - * at the switch port corresponding to the phys port. - */ - stats->tx_packets = readq(mem + NFP_MAC_STATS_RX_FRAMES_RECEIVED_OK); - stats->tx_bytes = readq(mem + NFP_MAC_STATS_RX_IN_OCTETS); - stats->tx_dropped = readq(mem + NFP_MAC_STATS_RX_IN_ERRORS); + stats->tx_packets = readq(mem + NFP_MAC_STATS_TX_FRAMES_TRANSMITTED_OK); + stats->tx_bytes = readq(mem + NFP_MAC_STATS_TX_OUT_OCTETS); + stats->tx_dropped = readq(mem + NFP_MAC_STATS_TX_OUT_ERRORS); - stats->rx_packets = readq(mem + NFP_MAC_STATS_TX_FRAMES_TRANSMITTED_OK); - stats->rx_bytes = readq(mem + NFP_MAC_STATS_TX_OUT_OCTETS); - stats->rx_dropped = readq(mem + NFP_MAC_STATS_TX_OUT_ERRORS); + stats->rx_packets = readq(mem + NFP_MAC_STATS_RX_FRAMES_RECEIVED_OK); + stats->rx_bytes = readq(mem + NFP_MAC_STATS_RX_IN_OCTETS); + stats->rx_dropped = readq(mem + NFP_MAC_STATS_RX_IN_ERRORS); } static void