linux-arm-msm.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
To: Yassine Oudjana <y.oudjana@protonmail.com>,
	Bjorn Andersson <bjorn.andersson@linaro.org>
Cc: Andy Gross <agross@kernel.org>, Ohad Ben-Cohen <ohad@wizery.com>,
	Mathieu Poirier <mathieu.poirier@linaro.org>,
	Rob Herring <robh+dt@kernel.org>,
	AngeloGioacchino Del Regno 
	<angelogioacchino.delregno@somainline.org>,
	Konrad Dybcio <konrad.dybcio@somainline.org>,
	Raffaele Tranquillini <raffaele.tranquillini@gmail.com>,
	linux-arm-msm@vger.kernel.org, linux-remoteproc@vger.kernel.org,
	devicetree@vger.kernel.org, phone-devel@vger.kernel.org,
	linux-kernel@vger.kernel.org
Subject: Re: [PATCH v2 1/5] arm64: dts: qcom: msm8996: Revamp reserved memory
Date: Wed, 3 Nov 2021 16:55:08 +0300	[thread overview]
Message-ID: <05b1f668-4dbf-36f7-76ef-6559e277d3c6@linaro.org> (raw)
In-Reply-To: <91b184266e545efcc5969fe6661b50da82351119.camel@protonmail.com>

On 03/11/2021 05:42, Yassine Oudjana wrote:
> On Wed, 2021-11-03 at 03:50 +0400, Dmitry Baryshkov wrote:
>> On 24/10/2021 06:04, Bjorn Andersson wrote:
>>> On Sun 26 Sep 14:06 CDT 2021, Yassine Oudjana wrote:
>>>
>>>> Fix a total overlap between zap_shader_region and slpi_region, and rename
>>>> all regions to match the naming convention in other Qualcomm SoC device trees.
>>>>
>>>> Signed-off-by: Yassine Oudjana <y.oudjana@protonmail.com>
>>>
>>> FYI, I like this series, but I held off applying it because I wanted to
>>> verify that the shuffling of the memory regions works on the existing
>>> 8996 boards.
>>>
>>> Unfortunately it didn't work, either with or without the shuffling on
>>> the db820c - and I've not found the time to figure out why that is. I
>>> hope to get back to this shortly (or that someone else will figure it
>>> out and provide a tested-by)
>>
>> I gave this a test too on my db820c. Usually the board MSS will crash
>> after ~0.1 - 0.15 seconds after booting up, then during recovery the
>> board will crash/reboot somewhere at the end of q6v5_mss_load() (and
>> typically after successful q6v5_rmb_mba_wait() call.
> 
>> Occasionally (approximately 1 of 20) the MSS will not crash, presenting
>> PDS service to the userspace. Even in this state it doesn't seem to be
>> able to lock the gps location (but this might be related to the big UART
>> mezzanine sitting on top of the board).
> 
> I've had MSS crash on xiaomi-scorpio too, but far less often. It seemed
> like some sort of race condition, as it only happened when ADSP and MSS
> were booted at the same time. To workaround this, I delayed loading of
> the rmtfs service to leave some time between booting ADSP and MSS.

For the tests to get 1:20 I've disabled ADSP (and WiFi/BT) completely. 
Without that I think I never got MSS to boot successfully. I think this 
is some kind of power/regulators/clock with ADSP and MSS draining too 
much power.

Note to myself: check rmtfs/partitions usage on db820c.

> 
>>
>> Unfortunately there seem to be no SLPI firmware for the db820c, so I can
>> not test slpi.
>>
>> A notice regarding the patchset itself. It looks like pil_q6v5_mss.c
>> driver misses mx and cx proxy power domains for the MSS_MSM8996 case.
> 
> I didn't notice that. I guess they stay on and MSS is able to boot
> anyway. I'll add them similar to PATCH 2/5.

Thanks!

> 
>>
>>>
>>> Regards,
>>> Bjorn
>>>
>>>> ---
>>>>    .../dts/qcom/msm8996-sony-xperia-tone.dtsi    | 18 ++++--
>>>>    .../boot/dts/qcom/msm8996-xiaomi-common.dtsi  | 18 +++---
>>>>    arch/arm64/boot/dts/qcom/msm8996.dtsi         | 63 ++++++++++---------
>>>>    3 files changed, 55 insertions(+), 44 deletions(-)
>>>>
>>>> diff --git a/arch/arm64/boot/dts/qcom/msm8996-sony-xperia-tone.dtsi b/arch/arm64/boot/dts/qcom/msm8996-sony-xperia-tone.dtsi
>>>> index 507396c4d23b..4c26e66f0610 100644
>>>> --- a/arch/arm64/boot/dts/qcom/msm8996-sony-xperia-tone.dtsi
>>>> +++ b/arch/arm64/boot/dts/qcom/msm8996-sony-xperia-tone.dtsi
>>>> @@ -13,9 +13,10 @@
>>>>    #include <dt-bindings/pinctrl/qcom,pmic-gpio.h>
>>>>    #include <dt-bindings/pinctrl/qcom,pmic-mpp.h>
>>>>
>>>> -/delete-node/ &slpi_region;
>>>> -/delete-node/ &venus_region;
>>>> -/delete-node/ &zap_shader_region;
>>>> +/delete-node/ &adsp_mem;
>>>> +/delete-node/ &slpi_mem;
>>>> +/delete-node/ &venus_mem;
>>>> +/delete-node/ &gpu_mem;
>>>>
>>>>    / {
>>>>    	qcom,msm-id = <246 0x30001>; /* MSM8996 V3.1 (Final) */
>>>> @@ -46,18 +47,23 @@ cont_splash_mem: memory@83401000 {
>>>>    			no-map;
>>>>    		};
>>>>
>>>> -		zap_shader_region: gpu@90400000 {
>>>> +		adsp_mem: adsp@8ea00000 {
>>>> +			reg = <0x0 0x8ea00000 0x0 0x1a00000>;
>>>> +			no-map;
>>>> +		};
>>>> +
>>>> +		gpu_mem: gpu@90400000 {
>>>>    			compatible = "shared-dma-pool";
>>>>    			reg = <0x0 0x90400000 0x0 0x2000>;
>>>>    			no-map;
>>>>    		};
>>>>
>>>> -		slpi_region: memory@90500000 {
>>>> +		slpi_mem: memory@90500000 {
>>>>    			reg = <0 0x90500000 0 0xa00000>;
>>>>    			no-map;
>>>>    		};
>>>>
>>>> -		venus_region: memory@90f00000 {
>>>> +		venus_mem: memory@90f00000 {
>>>>    			reg = <0 0x90f00000 0 0x500000>;
>>>>    			no-map;
>>>>    		};
>>>> diff --git a/arch/arm64/boot/dts/qcom/msm8996-xiaomi-common.dtsi b/arch/arm64/boot/dts/qcom/msm8996-xiaomi-common.dtsi
>>>> index d239b01b8505..a5e7bccadba2 100644
>>>> --- a/arch/arm64/boot/dts/qcom/msm8996-xiaomi-common.dtsi
>>>> +++ b/arch/arm64/boot/dts/qcom/msm8996-xiaomi-common.dtsi
>>>> @@ -66,32 +66,32 @@ memory@88800000 {
>>>>
>>>>    		/* This platform has all PIL regions offset by 0x1400000 */
>>>>    		/delete-node/ mpss@88800000;
>>>> -		mpss_region: mpss@89c00000 {
>>>> +		mpss_mem: mpss@89c00000 {
>>>>    			reg = <0x0 0x89c00000 0x0 0x6200000>;
>>>>    			no-map;
>>>>    		};
>>>>
>>>>    		/delete-node/ adsp@8ea00000;
>>>> -		adsp_region: adsp@8ea00000 {
>>>> +		adsp_mem: adsp@8fe00000 {
>>>>    			reg = <0x0 0x8fe00000 0x0 0x1b00000>;
>>>>    			no-map;
>>>>    		};
>>>>
>>>> -		/delete-node/ slpi@90b00000;
>>>> -		slpi_region: slpi@91900000 {
>>>> +		/delete-node/ slpi@90500000;
>>>> +		slpi_mem: slpi@91900000 {
>>>>    			reg = <0x0 0x91900000 0x0 0xa00000>;
>>>>    			no-map;
>>>>    		};
>>>>
>>>> -		/delete-node/ gpu@8f200000;
>>>> -		zap_shader_region: gpu@92300000 {
>>>> +		/delete-node/ gpu@90f00000;
>>>> +		gpu_mem: gpu@92300000 {
>>>>    			compatible = "shared-dma-pool";
>>>>    			reg = <0x0 0x92300000 0x0 0x2000>;
>>>>    			no-map;
>>>>    		};
>>>>
>>>>    		/delete-node/ venus@91000000;
>>>> -		venus_region: venus@90400000 {
>>>> +		venus_mem: venus@92400000 {
>>>>    			reg = <0x0 0x92400000 0x0 0x500000>;
>>>>    			no-map;
>>>>    		};
>>>> @@ -107,7 +107,7 @@ ramoops@92900000 {
>>>>    			pmsg-size = <0x40000>;
>>>>    		};
>>>>
>>>> -		/delete-node/ rmtfs@86700000;
>>>> +		/delete-node/ rmtfs;
>>>>    		rmtfs@f6c00000 {
>>>>    			compatible = "qcom,rmtfs-mem";
>>>>    			reg = <0 0xf6c00000 0 0x200000>;
>>>> @@ -118,7 +118,7 @@ rmtfs@f6c00000 {
>>>>    		};
>>>>
>>>>    		/delete-node/ mba@91500000;
>>>> -		mba_region: mba@f6f00000 {
>>>> +		mba_mem: mba@f6f00000 {
>>>>    			reg = <0x0 0xf6f00000 0x0 0x100000>;
>>>>    			no-map;
>>>>    		};
>>>> diff --git a/arch/arm64/boot/dts/qcom/msm8996.dtsi b/arch/arm64/boot/dts/qcom/msm8996.dtsi
>>>> index eb3ec5ff46eb..1495fff6ffc9 100644
>>>> --- a/arch/arm64/boot/dts/qcom/msm8996.dtsi
>>>> +++ b/arch/arm64/boot/dts/qcom/msm8996.dtsi
>>>> @@ -384,60 +384,65 @@ reserved-memory {
>>>>    		#size-cells = <2>;
>>>>    		ranges;
>>>>
>>>> -		mba_region: mba@91500000 {
>>>> -			reg = <0x0 0x91500000 0x0 0x200000>;
>>>> +		hyp_mem: memory@85800000 {
>>>> +			reg = <0x0 0x85800000 0x0 0x600000>;
>>>>    			no-map;
>>>>    		};
>>>>
>>>> -		slpi_region: slpi@90b00000 {
>>>> -			reg = <0x0 0x90b00000 0x0 0xa00000>;
>>>> +		xbl_mem: memory@85e00000 {
>>>> +			reg = <0x0 0x85e00000 0x0 0x200000>;
>>>>    			no-map;
>>>>    		};
>>>>
>>>> -		venus_region: venus@90400000 {
>>>> -			reg = <0x0 0x90400000 0x0 0x700000>;
>>>> +		smem_mem: smem-mem@86000000 {
>>>> +			reg = <0x0 0x86000000 0x0 0x200000>;
>>>>    			no-map;
>>>>    		};
>>>>
>>>> -		adsp_region: adsp@8ea00000 {
>>>> -			reg = <0x0 0x8ea00000 0x0 0x1a00000>;
>>>> +		tz_mem: memory@86200000 {
>>>> +			reg = <0x0 0x86200000 0x0 0x2600000>;
>>>>    			no-map;
>>>>    		};
>>>>
>>>> -		mpss_region: mpss@88800000 {
>>>> -			reg = <0x0 0x88800000 0x0 0x6200000>;
>>>> +		rmtfs_mem: rmtfs {
>>>> +			compatible = "qcom,rmtfs-mem";
>>>> +
>>>> +			size = <0x0 0x200000>;
>>>> +			alloc-ranges = <0x0 0xa0000000 0x0 0x2000000>;
>>>>    			no-map;
>>>> +
>>>> +			qcom,client-id = <1>;
>>>> +			qcom,vmid = <15>;
>>>>    		};
>>>>
>>>> -		smem_mem: smem-mem@86000000 {
>>>> -			reg = <0x0 0x86000000 0x0 0x200000>;
>>>> +		mpss_mem: mpss@88800000 {
>>>> +			reg = <0x0 0x88800000 0x0 0x6200000>;
>>>>    			no-map;
>>>>    		};
>>>>
>>>> -		memory@85800000 {
>>>> -			reg = <0x0 0x85800000 0x0 0x800000>;
>>>> +		adsp_mem: adsp@8ea00000 {
>>>> +			reg = <0x0 0x8ea00000 0x0 0x1b00000>;
>>>>    			no-map;
>>>>    		};
>>>>
>>>> -		memory@86200000 {
>>>> -			reg = <0x0 0x86200000 0x0 0x2600000>;
>>>> +		slpi_mem: slpi@90500000 {
>>>> +			reg = <0x0 0x90500000 0x0 0xa00000>;
>>>>    			no-map;
>>>>    		};
>>>>
>>>> -		rmtfs@86700000 {
>>>> -			compatible = "qcom,rmtfs-mem";
>>>> -
>>>> -			size = <0x0 0x200000>;
>>>> -			alloc-ranges = <0x0 0xa0000000 0x0 0x2000000>;
>>>> +		gpu_mem: gpu@90f00000 {
>>>> +			compatible = "shared-dma-pool";
>>>> +			reg = <0x0 0x90f00000 0x0 0x100000>;
>>>>    			no-map;
>>>> +		};
>>>>
>>>> -			qcom,client-id = <1>;
>>>> -			qcom,vmid = <15>;
>>>> +		venus_mem: venus@91000000 {
>>>> +			reg = <0x0 0x91000000 0x0 0x500000>;
>>>> +			no-map;
>>>>    		};
>>>>
>>>> -		zap_shader_region: gpu@8f200000 {
>>>> -			compatible = "shared-dma-pool";
>>>> -			reg = <0x0 0x90b00000 0x0 0xa00000>;
>>>> +		mba_mem: mba@91500000 {
>>>> +			reg = <0x0 0x91500000 0x0 0x200000>;
>>>>    			no-map;
>>>>    		};
>>>>    	};
>>>> @@ -1013,7 +1018,7 @@ opp-133000000 {
>>>>    			};
>>>>
>>>>    			zap-shader {
>>>> -				memory-region = <&zap_shader_region>;
>>>> +				memory-region = <&gpu_mem>;
>>>>    			};
>>>>    		};
>>>>
>>>> @@ -2001,7 +2006,7 @@ venus: video-codec@c00000 {
>>>>    				 <&venus_smmu 0x2c>,
>>>>    				 <&venus_smmu 0x2d>,
>>>>    				 <&venus_smmu 0x31>;
>>>> -			memory-region = <&venus_region>;
>>>> +			memory-region = <&venus_mem>;
>>>>    			status = "disabled";
>>>>
>>>>    			video-decoder {
>>>> @@ -3008,7 +3013,7 @@ adsp_pil: remoteproc@9300000 {
>>>>    			clocks = <&xo_board>;
>>>>    			clock-names = "xo";
>>>>
>>>> -			memory-region = <&adsp_region>;
>>>> +			memory-region = <&adsp_mem>;
>>>>
>>>>    			qcom,smem-states = <&smp2p_adsp_out 0>;
>>>>    			qcom,smem-state-names = "stop";
>>>> --
>>>> 2.33.0
>>>>
>>>>
>>
>>
>> --
>> With best wishes
>> Dmitry
> 
> 
> 


-- 
With best wishes
Dmitry

  reply	other threads:[~2021-11-03 13:55 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-09-26 19:06 [PATCH v2 0/5] msm8996: Enable support for MSS and SLPI Yassine Oudjana
2021-09-26 19:06 ` [PATCH v2 1/5] arm64: dts: qcom: msm8996: Revamp reserved memory Yassine Oudjana
2021-10-24  3:04   ` Bjorn Andersson
2021-11-02 23:50     ` Dmitry Baryshkov
2021-11-03  2:42       ` Yassine Oudjana
2021-11-03 13:55         ` Dmitry Baryshkov [this message]
2022-04-03 19:40   ` Dmitry Baryshkov
2021-09-26 19:06 ` [PATCH v2 2/5] remoteproc: qcom: pas: Use the same init resources for MSM8996 and MSM8998 Yassine Oudjana
2021-09-27 23:13   ` (subset) " Bjorn Andersson
2021-09-26 19:06 ` [PATCH v2 3/5] arm64: dts: qcom: msm8996: Unify smp2p naming Yassine Oudjana
2022-04-03 19:41   ` Dmitry Baryshkov
2021-09-26 19:06 ` [PATCH v2 4/5] arm64: dts: qcom: msm8996: Add MSS and SLPI Yassine Oudjana
2022-04-03 19:41   ` Dmitry Baryshkov
2021-09-26 19:06 ` [PATCH v2 5/5] arm64: dts: qcom: msm8996-xiaomi-*: Enable " Yassine Oudjana
2022-04-03 19:41   ` Dmitry Baryshkov

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=05b1f668-4dbf-36f7-76ef-6559e277d3c6@linaro.org \
    --to=dmitry.baryshkov@linaro.org \
    --cc=agross@kernel.org \
    --cc=angelogioacchino.delregno@somainline.org \
    --cc=bjorn.andersson@linaro.org \
    --cc=devicetree@vger.kernel.org \
    --cc=konrad.dybcio@somainline.org \
    --cc=linux-arm-msm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-remoteproc@vger.kernel.org \
    --cc=mathieu.poirier@linaro.org \
    --cc=ohad@wizery.com \
    --cc=phone-devel@vger.kernel.org \
    --cc=raffaele.tranquillini@gmail.com \
    --cc=robh+dt@kernel.org \
    --cc=y.oudjana@protonmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).