Re: [patch] revert: [NET]: Fix races in net_rx_action vs netpoll

[Olaf Kirch]

On Monday 16 July 2007 23:40, Linus Torvalds wrote:
>  - The change seems to always set the LIST_FROZEN bit when calling 
>    ->poll(), and at least on e1000, the NAPI poll() routine ends up doing 
>    that netif_rx_complete(), so we're *guaranteed* to always take the 
>    early exit and not do anything.

That is only in the poll_napi case, where netpoll tries to push pending
frames. The default caller for dev->poll() is net_rx_action.

We only ever get into this particular code branch in poll_napi when
the RX_SCHED bit is already set, ie someone put the device on the NAPI
poll_list and raised the softirq. poll_napi is just doing manually what
net_rx_action would do anyway once the softirq is serviced.

>  - The early return from netif_rx_complete() ends up meaning that an 
>    edge-triggered interrupt isn't handled properly, and will this never 
>    happen again, since it never goes away.

Sorry, I may be sitting on my brain this morning, but I don't understand
how skipping netif_rx_complete would affect ACKing of interrupts.

Olaf
-- 
Olaf Kirch  |  --- o --- Nous sommes du soleil we love when we play
okir@xxxxxx |    / | \   sol.dhoop.naytheet.ah kin.ir.samse.qurax
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/