[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <172908955984.387396.4029422650033805718.b4-ty@quicinc.com>
Date: Wed, 16 Oct 2024 07:39:19 -0700
From: Jeff Johnson <quic_jjohnson@...cinc.com>
To: <ath11k@...ts.infradead.org>, <linux-wireless@...r.kernel.org>,
<linux-kernel@...r.kernel.org>, Remi Pommarel <repk@...plefau.lt>
CC: Kalle Valo <kvalo@...nel.org>, Jeff Johnson <jjohnson@...nel.org>,
Praneesh P <quic_ppranees@...cinc.com>
Subject: Re: [PATCH v2] wifi: ath11k: Fix invalid ring usage in full
monitor mode
On Tue, 24 Sep 2024 21:41:19 +0200, Remi Pommarel wrote:
> On full monitor HW the monitor destination rxdma ring does not have the
> same descriptor format as in the "classical" mode. The full monitor
> destination entries are of hal_sw_monitor_ring type and fetched using
> ath11k_dp_full_mon_process_rx while the classical ones are of type
> hal_reo_entrance_ring and fetched with ath11k_dp_rx_mon_dest_process.
>
> Although both hal_sw_monitor_ring and hal_reo_entrance_ring are of same
> size, the offset to useful info (such as sw_cookie, paddr, etc) are
> different. Thus if ath11k_dp_rx_mon_dest_process gets called on full
> monitor destination ring, invalid skb buffer id will be fetched from DMA
> ring causing issues such as the following rcu_sched stall:
>
> [...]
Applied, thanks!
[1/1] wifi: ath11k: Fix invalid ring usage in full monitor mode
commit: befd716ed429b26eca7abde95da6195c548470de
Best regards,
--
Jeff Johnson <quic_jjohnson@...cinc.com>
Powered by blists - more mailing lists