Received-SPF: pass (google.com: domain of linux-wireless+bounces-3858-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) client-ip=147.75.80.249;
Message-ID: <d67fb4f4-aea2-4668-aac4-6e7eca8db4fa@gmail.com>
Date: Wed, 21 Feb 2024 04:38:10 -0800
Precedence: bulk
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: [PATCH] wifi: ath10k: poll service ready message before failing
Content-Language: en-US
To: Baochen Qiang <quic_bqiang@quicinc.com>, ath10k@lists.infradead.org
Cc: linux-wireless@vger.kernel.org
References: <20240221031729.2707-1-quic_bqiang@quicinc.com>
From: James Prestwood <prestwoj@gmail.com>
In-Reply-To: <20240221031729.2707-1-quic_bqiang@quicinc.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit

Hi Baochen,

On 2/20/24 7:17 PM, Baochen Qiang wrote:
> Currently host relies on CE interrupts to get notified that
> the service ready message is ready. This results in timeout
> issue if the interrupt is not fired, due to some unknown
> reasons. See below logs:
>
> [76321.937866] ath10k_pci 0000:02:00.0: wmi service ready event not received
> ...
> [76322.016738] ath10k_pci 0000:02:00.0: Could not init core: -110
>
> And finally it causes WLAN interface bring up failure.
>
> Change to give it one more chance here by polling CE rings,
> before failing directly.
>
> Tested-on: QCA6174 hw3.2 PCI WLAN.RM.4.4.1-00157-QCARMSWPZ-1
>
> Fixes: 5e3dd157d7e7 ("ath10k: mac80211 driver for Qualcomm Atheros 802.11ac CQA98xx devices")
> Reported-by: James Prestwood <prestwoj@gmail.com>
> Link: https://lore.kernel.org/linux-wireless/304ce305-fbe6-420e-ac2a-d61ae5e6ca1a@gmail.com/
> Signed-off-by: Baochen Qiang <quic_bqiang@quicinc.com>
> ---
>   drivers/net/wireless/ath/ath10k/wmi.c | 22 +++++++++++++++++++---
>   1 file changed, 19 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/net/wireless/ath/ath10k/wmi.c b/drivers/net/wireless/ath/ath10k/wmi.c
> index ddf15717d504..bf6cb2c73128 100644
> --- a/drivers/net/wireless/ath/ath10k/wmi.c
> +++ b/drivers/net/wireless/ath/ath10k/wmi.c
> @@ -1763,12 +1763,28 @@ void ath10k_wmi_put_wmi_channel(struct ath10k *ar, struct wmi_channel *ch,
>   
>   int ath10k_wmi_wait_for_service_ready(struct ath10k *ar)
>   {
> -	unsigned long time_left;
> +	unsigned long time_left, i;
>   
>   	time_left = wait_for_completion_timeout(&ar->wmi.service_ready,
>   						WMI_SERVICE_READY_TIMEOUT_HZ);
> -	if (!time_left)
> -		return -ETIMEDOUT;
> +	if (!time_left) {
> +		/* Sometimes the PCI HIF doesn't receive interrupt
> +		 * for the service ready message even if the buffer
> +		 * was completed. PCIe sniffer shows that it's
> +		 * because the corresponding CE ring doesn't fires
> +		 * it. Workaround here by polling CE rings once.
> +		 */
> +		ath10k_warn(ar, "failed to receive service ready completion, polling..\n");
> +
> +		for (i = 0; i < CE_COUNT; i++)
> +			ath10k_hif_send_complete_check(ar, i, 1);
> +
> +		time_left = wait_for_completion_timeout(&ar->wmi.service_ready,
> +							WMI_SERVICE_READY_TIMEOUT_HZ);
> +		if (!time_left)
> +			return -ETIMEDOUT;
> +	}
> +
>   	return 0;
>   }
>   
>
> base-commit: 707e306f3573fa321ae197d77366578e4566cff5

Thank you for looking at this I will test this and see if it resolves 
the problem we're seeing but since its somewhat rare it may take me a 
bit to validate.

Is this any different than just trying to bring up the interface again 
from userspace? I could be wrong, but my concern with this is that when 
I retried in userspace things got into a very odd state:

  - IWD starts

  - ifdown interface

  - ifup interface, timeout -110

  - Retry ifup, success

  - Authenticate/associate succeed

  - 4-way handshake fails because the device never received the 1/4 frame.

IWD would then retry indefinitely with auth/assoc succeeding but never 
receiving any 4-way handshake frames. The only way to get things working 
again was reloading the ath10k driver/reboot. Maybe this patch is 
different because its waiting for the initial request and no issuing a 
second one? Just wanted to point that out in case it sheds any light.

Thanks,

James