[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <dfcc8baa-c0a3-c554-a8cf-75702a1c4cad@linaro.org>
Date: Thu, 9 Feb 2023 15:54:53 +0200
From: Dmitry Baryshkov <dmitry.baryshkov@...aro.org>
To: Bryan O'Donoghue <bryan.odonoghue@...aro.org>,
Konrad Dybcio <konrad.dybcio@...aro.org>,
Amit Pundir <amit.pundir@...aro.org>
Cc: Bjorn Andersson <andersson@...nel.org>,
Andy Gross <agross@...nel.org>,
Rob Herring <robh+dt@...nel.org>,
Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
Caleb Connolly <caleb.connolly@...aro.org>,
linux-arm-msm <linux-arm-msm@...r.kernel.org>,
dt <devicetree@...r.kernel.org>,
lkml <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] arm64: dts: qcom: sdm845-db845c: Mark cont splash memory
region as reserved
On 09/02/2023 14:22, Bryan O'Donoghue wrote:
> On 09/02/2023 12:11, Bryan O'Donoghue wrote:
>>> If the bootloader splash is enabled then this memory is used until the
>>> DPU driver instructs MDP5 pipes to suck data from a newly assigned
>>> address,
>>> so there's a short window where it is.
>>
>> It seems a shame to reserve 30 something megabytes of memory for
>> continuous splash unless we are actually using it is my point.
>>
>> If I'm running headless its just wasted memory.
>
> Couldn't we
>
> 1. Find reserved continuous splash memory
> 2. Fee it in the MDP when we make the transition
Qualcomm has investigated freeing the MDP/DPU cont_splash memory, but I
fear that the end result was that it is not _that_ easy to free it. It
is marked as reserved/no-map, so the kernel doesn't think about it as a
memory region. Adding it back required hacking around that.
--
With best wishes
Dmitry
Powered by blists - more mailing lists