Re: [ISSUE: sky2 - rx error] Link stops working under heavy traffic load connected to a mv88e6176

[Rafa Corvillo]

On 08/05/17 14:38, Andrew Lunn wrote:
static unsigned sky2_get_rx_threshold(struct sky2_port *sky2)
{
         unsigned size;
         /* Space needed for frame data + headers rounded up */
         size = roundup(sky2->netdev->mtu + ETH_HLEN + VLAN_HLEN, 8);

         /* Stopping point for hardware truncation */
         return (size - 8) / sizeof(u32);
}

This is not going to be big enough for a frame with a DSA header.


Then, would be a good fix add 8 bytes to the size variable in this function?

Yes. Also look at the transmit code, is there again a limit based on
the MTU.

Hi Andrew,

Adding 8 bytes (sky2->netdev->mtu + ETH_HLEN + VLAN_HLEN + 8 
(EDSA_HLEN)) does not fix the error, because the interface keep having a 
maximum length of 1518 bytes (sky2->netdev->mtu + ETH_HLEN + VLAN_HLEN).

The polling function of sky2 driver (sky2_poll) calls to the function 
sky2_status_intr (with the parameter struct sky2_hw *hw). The 
sky2_status_intr function gets the status of the list elements (struct 
sky2_status_le) from the sky2_hw parameter and, from the sky2_status_le, 
gets the maximum length (1518) and the status code (0x5f20010). When the 
latter function (sky2_status_intr) calls to the sky2_receive function 
with the parameters length and status, it reports about the error (rx 
error, status 0x5f20010
length 1518).

I don't know who sets the maximum length (1518) and the status code 
(0x5f20010) of the packets. Is it possible that these values to be set 
outside the sky2 code?

Thanks,

Rafa


Settings for marvell:
         Supported ports: [ TP ]
         Supported link modes:   10baseT/Half 10baseT/Full
                                 100baseT/Half 100baseT/Full
                                 1000baseT/Half 1000baseT/Full
         Supported pause frame use: No
         Supports auto-negotiation: Yes
         Advertised link modes:  10baseT/Half 10baseT/Full
                                 100baseT/Half 100baseT/Full
                                 1000baseT/Half 1000baseT/Full
         Advertised pause frame use: No
         Advertised auto-negotiation: No
         Speed: 1000Mb/s
         Duplex: Full
         Port: Twisted Pair
         PHYAD: 0
         Transceiver: internal
         Auto-negotiation: on
         MDI-X: Unknown
         Supports Wake-on: pg
         Wake-on: d
         Current message level: 0x000000ff (255)
                                drv probe link timer ifdown ifup
rx_err tx_err
         Link detected: yes


So this suggests there is a real PHY there, and it is
auto-negotiating.

What we cannot see is the status for the PHY it connects to. But since
this PHY has established a link, the other PHY is probably O.K. It is
just a bit unsafe, since you are relying on reset behaviour. There is
nothing in software configuring the second PHY to make it
auto-negotiate.

        Andrew