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: <737b07582ef2a4b2f134a1a931b6621ff96adb77.camel@pengutronix.de>
Subject: Re: [PATCH v10 07/11] drm/etnaviv: Add support for the dma coherent
 device
From:   Lucas Stach <l.stach@pengutronix.de>
To:     Sui Jingfeng <suijingfeng@loongson.cn>,
        Sui Jingfeng <18949883232@163.com>,
        Russell King <linux+etnaviv@armlinux.org.uk>,
        Christian Gmeiner <christian.gmeiner@gmail.com>,
        David Airlie <airlied@gmail.com>,
        Daniel Vetter <daniel@ffwll.ch>
Cc:     linux-kernel@vger.kernel.org, dri-devel@lists.freedesktop.org,
        etnaviv@lists.freedesktop.org,
        Philipp Zabel <p.zabel@pengutronix.de>,
        Bjorn Helgaas <bhelgaas@google.com>
Date:   Wed, 21 Jun 2023 18:12:16 +0200
In-Reply-To: <9c8afcb4-70c0-a920-2a78-78a9ac884c80@loongson.cn>
References: <20230620094716.2231414-1-18949883232@163.com>
         <20230620094716.2231414-8-18949883232@163.com>
         <8f74f0962c8bab6c832919a5340667c54e1a7ddc.camel@pengutronix.de>
         <aa73348d-5ec8-4ac0-2ec0-0cce24756c63@loongson.cn>
         <87c9576e6ca1b58fa94e0bc1a2f4be3847f0518c.camel@pengutronix.de>
         <9c8afcb4-70c0-a920-2a78-78a9ac884c80@loongson.cn>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 8BIT
User-Agent: Evolution 3.46.4 (3.46.4-1.fc37) 
MIME-Version: 1.0
Precedence: bulk

Am Mittwoch, dem 21.06.2023 um 23:41 +0800 schrieb Sui Jingfeng:
> On 2023/6/21 23:23, Lucas Stach wrote:
> > Am Mittwoch, dem 21.06.2023 um 22:44 +0800 schrieb Sui Jingfeng:
> > > Hi,
> > > 
> > > On 2023/6/21 18:00, Lucas Stach wrote:
> > > > > diff --git a/drivers/gpu/drm/etnaviv/etnaviv_drv.h b/drivers/gpu/drm/etnaviv/etnaviv_drv.h
> > > > > index 9cd72948cfad..644e5712c050 100644
> > > > > --- a/drivers/gpu/drm/etnaviv/etnaviv_drv.h
> > > > > +++ b/drivers/gpu/drm/etnaviv/etnaviv_drv.h
> > > > > @@ -46,6 +46,12 @@ struct etnaviv_drm_private {
> > > > >    	struct xarray active_contexts;
> > > > >    	u32 next_context_id;
> > > > >    
> > > > > +	/*
> > > > > +	 * If true, the GPU is capable of snooping cpu cache. Here, it
> > > > > +	 * also means that cache coherency is enforced by the hardware.
> > > > > +	 */
> > > > > +	bool dma_coherent;
> > > > > +
> > > > No need for this, I think. Just use dev_is_dma_coherent() where you
> > > > need to know this.
> > > > 
> > > No, we want this value cached by the driver.
> > > 
> > Why? dev_is_dma_coherent() is a header-only function with a single
> > pointer chasing operation. Your cache is also a single pointer chasing
> > access, just that we now need storage for this information in both
> > struct device and struct etnaviv_gpu.
> 
> 
> You don't need store it in struct etnaviv_gpu.
> 
> As this variable is shared across the device, so it is better to be put 
> in the struct etnaviv_drm_private.
> 
> I don't think another 4 bytes allocation is something what we can't pay for.
> 
> 
> My patch doesn't mentioned that it need to store it inside of struct 
> etnaviv_gpu, do I?

You are right, I was mistaken about the etnaviv struct this is added
to. However there is still the fundamental question: what's the gain of
this cache? The information is already available in struct device and
will be accessed with the same amount of loads if you care that much
about micro-optimization.

Regards,
Lucas