Hi Jakub On 26/07/23 9:20 am, Jakub Kicinski wrote: > On Mon, 24 Jul 2023 16:59:31 +0530 MD Danish Anwar wrote: >> + /* Rx */ >> + ICSSG_STATS(rx_packets, true), >> + ICSSG_STATS(rx_broadcast_frames, false), >> + ICSSG_STATS(rx_multicast_frames, false), > > There is a standard stat for mcast. > Sure. I will add multicast stats via .ndo_get_stats64 instead of ethtool. >> + ICSSG_STATS(rx_crc_errors, true), >> + ICSSG_STATS(rx_mii_error_frames, false), >> + ICSSG_STATS(rx_odd_nibble_frames, false), >> + ICSSG_STATS(rx_frame_max_size, false), >> + ICSSG_STATS(rx_max_size_error_frames, false), >> + ICSSG_STATS(rx_frame_min_size, false), >> + ICSSG_STATS(rx_min_size_error_frames, false), >> + ICSSG_STATS(rx_over_errors, true), >> + ICSSG_STATS(rx_class0_hits, false), >> + ICSSG_STATS(rx_class1_hits, false), >> + ICSSG_STATS(rx_class2_hits, false), >> + ICSSG_STATS(rx_class3_hits, false), >> + ICSSG_STATS(rx_class4_hits, false), >> + ICSSG_STATS(rx_class5_hits, false), >> + ICSSG_STATS(rx_class6_hits, false), >> + ICSSG_STATS(rx_class7_hits, false), >> + ICSSG_STATS(rx_class8_hits, false), >> + ICSSG_STATS(rx_class9_hits, false), >> + ICSSG_STATS(rx_class10_hits, false), >> + ICSSG_STATS(rx_class11_hits, false), >> + ICSSG_STATS(rx_class12_hits, false), >> + ICSSG_STATS(rx_class13_hits, false), >> + ICSSG_STATS(rx_class14_hits, false), >> + ICSSG_STATS(rx_class15_hits, false), >> + ICSSG_STATS(rx_smd_frags, false), >> + ICSSG_STATS(rx_bucket1_size, false), >> + ICSSG_STATS(rx_bucket2_size, false), >> + ICSSG_STATS(rx_bucket3_size, false), >> + ICSSG_STATS(rx_bucket4_size, false), > > Are the bucket sizes configurable? Can we set the bucket sizes > to standard RMON ones and use ethtool RMON stats? The bucket sizes are not configurable. Bucket size is read from hardware and is fixed. I don't think we can configure bucket size and use ethtool RMON stats. It's better to dump bucket sizes via ethtool -S. -- Thanks and Regards, Danish.
