linux-kernel - Re: [PATCH v10 15/24] drm/mediatek: Remove ineffectual power management codes

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <2313a4b08f57ab24b48a13a31845eed7e0ab042d.camel@mediatek.com>
Date:   Thu, 19 Oct 2023 09:52:07 +0000
From:   Shawn Sung (宋孝謙) 
        <Shawn.Sung@...iatek.com>
To:     CK Hu (胡俊光) <ck.hu@...iatek.com>,
        "matthias.bgg@...il.com" <matthias.bgg@...il.com>,
        "angelogioacchino.delregno@...labora.com" 
        <angelogioacchino.delregno@...labora.com>,
        "robh+dt@...nel.org" <robh+dt@...nel.org>,
        "krzysztof.kozlowski+dt@...aro.org" 
        <krzysztof.kozlowski+dt@...aro.org>
CC:     "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "linux-mediatek@...ts.infradead.org" 
        <linux-mediatek@...ts.infradead.org>,
        "chunkuang.hu@...nel.org" <chunkuang.hu@...nel.org>,
        Jason-JH Lin (林睿祥) 
        <Jason-JH.Lin@...iatek.com>,
        "devicetree@...r.kernel.org" <devicetree@...r.kernel.org>,
        Nancy Lin (林欣螢) <Nancy.Lin@...iatek.com>,
        Moudy Ho (何宗原) <Moudy.Ho@...iatek.com>,
        "daniel@...ll.ch" <daniel@...ll.ch>,
        "p.zabel@...gutronix.de" <p.zabel@...gutronix.de>,
        "dri-devel@...ts.freedesktop.org" <dri-devel@...ts.freedesktop.org>,
        Nathan Lu (呂東霖) <Nathan.Lu@...iatek.com>,
        "airlied@...il.com" <airlied@...il.com>,
        "sean@...rly.run" <sean@...rly.run>,
        "linux-arm-kernel@...ts.infradead.org" 
        <linux-arm-kernel@...ts.infradead.org>,
        "fshao@...omium.org" <fshao@...omium.org>,
        "johnson.wang@...iatek.corp-partner.google.com" 
        <johnson.wang@...iatek.corp-partner.google.com>
Subject: Re: [PATCH v10 15/24] drm/mediatek: Remove ineffectual power
 management codes

Hi Angelo,

On Thu, 2023-10-19 at 11:07 +0200, AngeloGioacchino Del Regno wrote:
> Il 19/10/23 07:56, Hsiao Chien Sung ha scritto:
> > Display modules will be powered on when .atomic_enable(),
> > there is no need to do it again in mtk_crtc_ddp_hw_init().
> > Besides, the DRM devices are created manually when mtk-mmsys
> > is probed and there is no power domain linked to it.
> > 
> > Fixes: 119f5173628a ("drm/mediatek: Add DRM Driver for Mediatek SoC
> > MT8173.")
> > 
> > Signed-off-by: Hsiao Chien Sung <shawn.sung@...iatek.com>
> > ---
> >   drivers/gpu/drm/mediatek/mtk_drm_crtc.c | 18 +++---------------
> >   1 file changed, 3 insertions(+), 15 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/mediatek/mtk_drm_crtc.c
> > b/drivers/gpu/drm/mediatek/mtk_drm_crtc.c
> > index bc4cc75cca18..c7edd80be428 100644
> > --- a/drivers/gpu/drm/mediatek/mtk_drm_crtc.c
> > +++ b/drivers/gpu/drm/mediatek/mtk_drm_crtc.c
> > @@ -6,7 +6,6 @@
> >   #include <linux/clk.h>
> >   #include <linux/dma-mapping.h>
> >   #include <linux/mailbox_controller.h>
> > -#include <linux/pm_runtime.h>
> >   #include <linux/soc/mediatek/mtk-cmdq.h>
> >   #include <linux/soc/mediatek/mtk-mmsys.h>
> >   #include <linux/soc/mediatek/mtk-mutex.h>
> > @@ -362,22 +361,16 @@ static int mtk_crtc_ddp_hw_init(struct
> > mtk_drm_crtc *mtk_crtc, struct drm_atomic
> >   		drm_connector_list_iter_end(&conn_iter);
> >   	}
> >   
> > -	ret = pm_runtime_resume_and_get(crtc->dev->dev);
> > -	if (ret < 0) {
> > -		DRM_ERROR("Failed to enable power domain: %d\n", ret);
> > -		return ret;
> > -	}
> > -
> 
> Are you really sure that writes to DISP_REG_OVL_xxx and others in
> other modules,
> called by the .layer_config() callback, can be successfully done on
> an unpowered
> and/or unclocked module, on all MediaTek SoCs?
> This looks a bit odd.

Not sure if I get your point correctly. We removed this PM API because:

1. mtk_crtc_ddp_hw_init() is called by mtk_drm_crtc_atomic_enable(),
and the new inline function mtk_ddp_comp_power_on() is called before hw
init, we can make sure the power is on before configuring the hardware.

2. The device "crtc->dev->dev" here is assigned by the probe function
of mtk-mmsys, which will be look like "mediatek-drm.auto.13", and this
device is not linked to any power domain.

> 
> >   	ret = mtk_mutex_prepare(mtk_crtc->mutex);
> >   	if (ret < 0) {
> >   		DRM_ERROR("Failed to enable mutex clock: %d\n", ret);
> > -		goto err_pm_runtime_put;
> > +		goto error;
> >   	}
> >   
> >   	ret = mtk_crtc_ddp_clk_enable(mtk_crtc);
> >   	if (ret < 0) {
> >   		DRM_ERROR("Failed to enable component clocks: %d\n",
> > ret);
> > -		goto err_mutex_unprepare;
> > +		goto error;
> >   	}
> >   
> >   	for (i = 0; i < mtk_crtc->ddp_comp_nr - 1; i++) {
> > @@ -426,16 +419,13 @@ static int mtk_crtc_ddp_hw_init(struct
> > mtk_drm_crtc *mtk_crtc, struct drm_atomic
> >   
> 
> ...because you could otherwise just call pm_runtime_put() here,
> instead of removing
> the pm_runtime_resume_and_get() call, which is something I would
> advise to do.
> 
> Regards,
> Angelo
> 

Thanks,
Shawn