[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [PATCH] Mini-OS: netfront: Fix rx ring starvation in network_rx

martin@xxxxxxxxxx said:
> Rather, shouldn't the final if() in the loop (where data is sent up) be:
>   if(rx->status > NETIF_RSP_NULL) { ... }
> *we can't use NETIF_RSP_OKAY as NETIF_RSP_NULL is defined to be 1 (yuck).
> I don't see any value in keeping the printk(), the netfront driver already
> disclaims supporting extras in comments.

I just cross-checked with include/xen/io/netif.h and the Linux kernel
xen-netfront.c. We will never see NETRXF_extra_info as we (correctly) don't
advertise any of the relevant feature-* toggles in Xenstore.

AFAICT we should also never see NETIF_RSP_NULL as according to the
documentation that is only used for responses to TX requests. If this is
indeed the case then the if can be 

  if (rx->status > NETIF_RSP_OKAY)

it can't be tested with == as rx->status is overloaded to contain the
rx packet size.


Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.