Received-SPF: pass (google.com: domain of linux-wireless+bounces-1637-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.199.223 as permitted sender) client-ip=147.75.199.223;
Message-ID: <c704b3cc-d6e0-4231-9602-47a3e89b2f1e@quicinc.com>
Date: Tue, 9 Jan 2024 09:04:40 -0800
Precedence: bulk
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: [PATCH v2] wifi: ath11k: fix IOMMU errors on buffer rings
Content-Language: en-US
To: Kalle Valo <kvalo@kernel.org>, Zhenghao Gu <imguzh@gmail.com>
CC: <ath11k@lists.infradead.org>, <linux-wireless@vger.kernel.org>
References: <20231212031914.47339-1-imguzh@gmail.com>
 <8734v6ecp1.fsf@kernel.org>
From: Jeff Johnson <quic_jjohnson@quicinc.com>
In-Reply-To: <8734v6ecp1.fsf@kernel.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit

On 1/9/2024 7:41 AM, Kalle Valo wrote:
> Zhenghao Gu <imguzh@gmail.com> writes:
> 
>> virt_to_phys doesn't work on systems with IOMMU enabled,
>> which have non-identity physical-to-IOVA mappings.
> 
> Can you give an example of such system? Just curious where you are
> seeing this.
> 
>> It leads to IO_PAGE_FAULTs like this:
>> [IO_PAGE_FAULT domain=0x0023 address=0x1cce00000 flags=0x0020]
>> and no link can be established.
> 
> What do you mean with link in this context? Are you talking about 802.11
> association?
> 
>> This patch changes that to dma_map_single(), which works correctly.
> 
> Good catch. And virt_to_phys() documentation even says this:
> 
>  *	This function does not give bus mappings for DMA transfers. In
>  *	almost all conceivable cases a device driver should not be using
>  *	this function
> 
>> Tested-on: QCN9074 hw1.0 PCI WLAN.HK.2.7.0.1-01744-QCAHKSWPL_SILICONZ-1
>> Signed-off-by: Zhenghao Gu <imguzh@gmail.com>
> 
> Jeff, are you ok with this?
> 
> I did some cosmetics changes in the pending branch (removed unnecessary
> parenthesis, reverse xmas tree etc), please check:
> 
> https://git.kernel.org/pub/scm/linux/kernel/git/kvalo/ath.git/commit/?h=pending&id=fefa43d63e1928fce6e8c2bb626900e9ce98ca69
> 
LGTM, incorporates the v1 feedback from the engineering team

Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com>