Received-SPF: pass (google.com: domain of linux-wireless-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Kalle Valo <kvalo@kernel.org>
To:     Jeff Johnson <quic_jjohnson@quicinc.com>
Cc:     <linux-wireless@vger.kernel.org>, <ath12k@lists.infradead.org>
Subject: Re: [PATCH 40/50] wifi: ath12k: add qmi.h
References: <20220812161003.27279-1-kvalo@kernel.org>
        <20220812161003.27279-41-kvalo@kernel.org>
        <f6e2c694-4803-be69-ec05-08541b718b8f@quicinc.com>
Date:   Tue, 08 Nov 2022 17:06:48 +0200
In-Reply-To: <f6e2c694-4803-be69-ec05-08541b718b8f@quicinc.com> (Jeff
        Johnson's message of "Thu, 18 Aug 2022 17:25:42 -0700")
Message-ID: <87r0ydcwon.fsf@kernel.org>
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/26.1 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain
Precedence: bulk

Jeff Johnson <quic_jjohnson@quicinc.com> writes:

> On 8/12/2022 9:09 AM, Kalle Valo wrote:
>> From: Kalle Valo <quic_kvalo@quicinc.com>
>>
>> (Patches split into one patch per file for easier review, but the final
>> commit will be one big patch. See the cover letter for more info.)
>>
>> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
>> ---
>>   drivers/net/wireless/ath/ath12k/qmi.h | 610 ++++++++++++++++++++++++++++++++++
> [...]
>> +struct qmi_wlanfw_ce_tgt_pipe_cfg_s_v01 {
>> +	__le32 pipe_num;
>> +	__le32 pipe_dir;
>> +	__le32 nentries;
>> +	__le32 nbytes_max;
>> +	__le32 flags;
>> +};
>> +
>> +struct qmi_wlanfw_ce_svc_pipe_cfg_s_v01 {
>> +	__le32 service_id;
>> +	__le32 pipe_dir;
>> +	__le32 pipe_num;
>> +};
>
> curious why the above two structs use __le32 definitions but none of
> the others do. I thought the QMI interface would use the qmi_elem_info
> to properly encode/decode in host order on both sides of the pipe and
> hence we should expect host order on both sides

That is a good question, I was also expecting QMI to be in host order.
In ce.h I see this:

/* Establish a mapping between a service/direction and a pipe.
 * Configuration information for a Copy Engine pipe and services.
 * Passed from Host to Target through QMI message and must be in
 * little endian format.
 */
struct service_to_pipe {
	__le32 service_id;
	__le32 pipedir;
	__le32 pipenum;
};

But I'm not convinced about that comments. Thoughts?

-- 
https://patchwork.kernel.org/project/linux-wireless/list/

https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches