Re: [PATCH ath-next v2 03/17] wifi: ath12k: Add hardware parameter for QCN6432

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] wifi: ath12k: missing qmi_max_chunk_size for QCN6432 breaks memory allocation fallback
--

commit 417c2f74682574a9e982b5753b7747e20b63cda7
Author: Sowmiya Sree Elavalagan <quic_ssreeela@quicinc.com>

wifi: ath12k: Add hardware parameter for QCN6432

This commit adds hardware parameters for the QCN6432 WiFi device in the ath12k
driver. It introduces a new hardware revision definition and configures the
ath12k_hw_params structure for QCN6432 by reusing parameters from IPQ5332 and
QCN9274.

> diff --git a/drivers/net/wireless/ath/ath12k/wifi7/hw.c b/drivers/net/wireless/ath/ath12k/wifi7/hw.c
> index b1c17cc197dfd..494c7b3ae1acf 100644
> --- a/drivers/net/wireless/ath/ath12k/wifi7/hw.c
> +++ b/drivers/net/wireless/ath/ath12k/wifi7/hw.c
> @@ -906,6 +906,91 @@ static const struct ath12k_hw_params ath12k_wifi7_hw_params[] = {

[ ... ]

>  	},
> +	{
> +		.name = "qcn6432 hw1.0",
> +		.hw_rev = ATH12K_HW_QCN6432_HW10,

[Severity: Medium]
Does this struct initialization need to include .qmi_max_chunk_size?

If .qmi_max_chunk_size is not initialized, it will default to 0. During
firmware memory allocation in ath12k_qmi_alloc_chunk(), if
dma_alloc_coherent() fails, the driver falls back based on this threshold:

	if (!vaddr) {
		if (chunk->size > ab->hw_params->qmi_max_chunk_size) {
			ab->qmi.target_mem_delayed = true;
			...
			ath12k_qmi_free_target_mem_chunk(ab);
			return -EAGAIN;
		}
		...
		return -ENOMEM;
	}

Because qmi_max_chunk_size is 0, any failed chunk size will be strictly
greater than 0. Will this cause the driver to indefinitely return -EAGAIN,
asking the firmware to supply progressively smaller chunk requests, rather
than correctly terminating the failing QMI negotiation with -ENOMEM?

> +		.fw = {
> +			.dir = "QCN6432/hw1.0",
> +			.board_size = 256 * 1024,
> +			.cal_offset = 128 * 1024,

[ ... ]

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261016-mpp-bringup-v2-0-f510c33bfcf5@oss.qualcomm.com?part=3