On 10/10/21 8:24 PM, Sergey Shtylyov wrote: [...] >>>>>> Update EMAC configuration mode comment from "PAUSE prohibition" >>>>>> to "EMAC Mode: PAUSE prohibition; Duplex; TX; RX; CRC Pass Through; >>>>>> Promiscuous". >>>>>> >>>>>> Signed-off-by: Biju Das <biju.das.jz@xxxxxxxxxxxxxx> >>>>>> Suggested-by: Sergey Shtylyov <s.shtylyov@xxxxxx> >>>>>> --- >>>>>> v1->v2: >>>>>> * No change >>>>>> V1: >>>>>> * New patch. >>>>>> --- >>>>>> drivers/net/ethernet/renesas/ravb_main.c | 2 +- >>>>>> 1 file changed, 1 insertion(+), 1 deletion(-) >>>>>> >>>>>> diff --git a/drivers/net/ethernet/renesas/ravb_main.c >>>>>> b/drivers/net/ethernet/renesas/ravb_main.c >>>>>> index 9a770a05c017..b78aca235c37 100644 >>>>>> --- a/drivers/net/ethernet/renesas/ravb_main.c >>>>>> +++ b/drivers/net/ethernet/renesas/ravb_main.c >>>>>> @@ -519,7 +519,7 @@ static void ravb_emac_init_gbeth(struct >>>>>> net_device >>>>> *ndev) >>>>>> /* Receive frame limit set register */ >>>>>> ravb_write(ndev, GBETH_RX_BUFF_MAX + ETH_FCS_LEN, RFLR); >>>>>> >>>>>> - /* PAUSE prohibition */ >>>>>> + /* EMAC Mode: PAUSE prohibition; Duplex; TX; RX; CRC Pass Through; >>>>>> +Promiscuous */ >>>>> >>>>> Promiscuous mode, really? Why?! >>>> >>>> This is TOE related, >> >> I meant the context here is TOE register related. That is what I meant. >> >>> >>> The promiscuous mode is supported by _all_ Ethernet controllers, I >>> think. >>> >>>> and is recommendation from BSP team. >>> >>> On what grounds? >> >> The reference implementation has this on. Any way it is good catch. >> I will turn it off and check. >> >> by looking at the RJ LED's there is not much activity and packet statistics also show not much activity by default. >> >> How can we check, it is overloading the controller? So that I can compare with and without this setting > > Maybe it doesn't get overloaded that simply, but definitely the promiscuous mode is not the thing > for the normal driver use... > >>>> If you think it is wrong. >>>> I can take this out. Please let me know. Currently the board is booting >>> and everything works without issues. >>> >>> Please do take it out. It'll needlessly overload the controller when >>> there's much traffic on the local network. >> >> >> I can see much activity only on RJ45 LED's when I call tcpdump or by setting IP link set eth0 promisc on. >> Otherwise there is no traffic at all. > > Sounds like the kernel initially sets the RX mode with IFF_PROMISC = 0 and thus clear ECMR.PRM but I don't > see where it does this? Could you instrument ravb_set_tx_mode() plz? Sorry, ravb_set_rx_mode(), of/c. :-) >> Regards, >> Biju > > [...] MBR, Sergey
