[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <87bk12sadn.fsf@kurt.kurt.home>
Date: Thu, 05 Sep 2024 11:44:20 +0200
From: Kurt Kanzenbach <kurt@...utronix.de>
To: "Ruinskiy, Dima" <dima.ruinskiy@...el.com>, Tony Nguyen
<anthony.l.nguyen@...el.com>, davem@...emloft.net, kuba@...nel.org,
pabeni@...hat.com, edumazet@...gle.com, netdev@...r.kernel.org
Cc: sasha.neftin@...el.com, vitaly.lifshits@...el.com,
maciej.fijalkowski@...el.com, magnus.karlsson@...el.com, ast@...nel.org,
daniel@...earbox.net, hawk@...nel.org, john.fastabend@...il.com,
bpf@...r.kernel.org, bigeasy@...utronix.de, Vinicius Costa Gomes
<vinicius.gomes@...el.com>, Simon Horman <horms@...nel.org>, Mor Bar-Gabay
<morx.bar.gabay@...el.com>
Subject: Re: [PATCH net-next 2/6] igc: Get rid of spurious interrupts
On Thu Sep 05 2024, Dima Ruinskiy wrote:
> On 31/08/2024 0:04, Tony Nguyen wrote:
>> - wr32(IGC_ICS, IGC_ICS_RXDMT0);
>> + struct igc_ring *rx_ring = adapter->rx_ring[0];
>> +
>> + if (test_bit(IGC_RING_FLAG_RX_ALLOC_FAILED, &rx_ring->flags)) {
>> + clear_bit(IGC_RING_FLAG_RX_ALLOC_FAILED, &rx_ring->flags);
>> + wr32(IGC_ICS, IGC_ICS_RXDMT0);
>> + }
> I have some concerns specifically about this code (Legacy/MSI interrupt
> case). The code only checks the IGC_RING_FLAG_RX_ALLOC_FAILED flag of
> ring 0. What if the failure was on another ring? It seems proper to
> iterate over all Rx rings in the adapter (I believe igc can have up to 4).
In case of Legacy/MSI only one vector, one rx queue and one tx queue is
utilized. The MSI-X code has to check for all rings, which it does.
Thanks,
Kurt
Download attachment "signature.asc" of type "application/pgp-signature" (862 bytes)
Powered by blists - more mailing lists