Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <4ec997ab-8a6f-76e7-20e7-b7f778124c87@linaro.org>
Date:   Tue, 22 Feb 2022 10:05:17 -0600
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.5.0
Subject: Re: [PATCH v3 23/25] bus: mhi: ep: Add support for queueing SKBs to
 the host
Content-Language: en-US
From:   Alex Elder <elder@linaro.org>
To:     Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Cc:     mhi@lists.linux.dev, quic_hemantk@quicinc.com,
        quic_bbhatt@quicinc.com, quic_jhugo@quicinc.com,
        vinod.koul@linaro.org, bjorn.andersson@linaro.org,
        dmitry.baryshkov@linaro.org, quic_vbadigan@quicinc.com,
        quic_cang@quicinc.com, quic_skananth@quicinc.com,
        linux-arm-msm@vger.kernel.org, linux-kernel@vger.kernel.org
References: <20220212182117.49438-1-manivannan.sadhasivam@linaro.org>
 <20220212182117.49438-24-manivannan.sadhasivam@linaro.org>
 <766e6568-0b80-c745-dd8f-7f401fb0422d@linaro.org>
 <20220222143825.GH5029@thinkpad>
 <f3248103-9450-dcf0-719d-77c6dcd85bfe@linaro.org>
In-Reply-To: <f3248103-9450-dcf0-719d-77c6dcd85bfe@linaro.org>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 2/22/22 9:18 AM, Alex Elder wrote:
> 
>> But anyway, there is no need to check for CHAIN flag while writing to 
>> host.
>> CHAIN flag is only used or even make sense when host writes data to 
>> device, so
> 
> I'm not sure that's correct, but I don't want to get into that issue here.
> We can talk about that separately.

I just wanted to send a short followup here.  My comments
were based on a misunderstanding, and Mani cleared it up
for me.  For host receives, the MHI specification states
that a packet that requires more than the size of the
buffer in a single TRE leads to an overflow event being
generated from the device to the host.  The buffer on
the TRE is filled, and subsequent packet data is written
to the next TRE's buffer (assuming it's present).

This differs from one feature of IPA and its GSI transfer
rings.  I won't explain that here, to avoid any confusion.

Mani explained things to me, and he's going to send an
updated series, which I'll review.

					-Alex