linux-kernel - Re: [PATCH v5 3/5] media: hantro: Add RK3588 VEPU121 support

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] [day] [month] [year] [list]
Message-ID: <14943967.O6BkTfRZtg@diego>
Date: Thu, 13 Jun 2024 10:01:09 +0200
From: Heiko Stübner <heiko@...ech.de>
To: Sebastian Reichel <sebastian.reichel@...labora.com>
Cc: Ezequiel Garcia <ezequiel@...guardiasur.com.ar>,
 Philipp Zabel <p.zabel@...gutronix.de>,
 Nicolas Frattaroli <frattaroli.nicolas@...il.com>,
 Rob Herring <robh@...nel.org>, Krzysztof Kozlowski <krzk+dt@...nel.org>,
 Conor Dooley <conor+dt@...nel.org>, Jianfeng Liu <liujianfeng1994@...il.com>,
 Emmanuel Gil Peyrot <linkmauve@...kmauve.fr>,
 Nicolas Dufresne <nicolas.dufresne@...labora.com>,
 linux-media@...r.kernel.org, linux-rockchip@...ts.infradead.org,
 devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
 kernel@...labora.com
Subject: Re: [PATCH v5 3/5] media: hantro: Add RK3588 VEPU121 support

Am Donnerstag, 13. Juni 2024, 00:44:38 CEST schrieb Sebastian Reichel:
> Hi,
> 
> On Wed, Jun 12, 2024 at 08:08:51PM GMT, Heiko Stübner wrote:
> > Am Mittwoch, 12. Juni 2024, 19:15:43 CEST schrieb Sebastian Reichel:
> > > Avoid exposing each of the 4 Hantro H1 cores separately to userspace.
> > > For now just expose the first one.
> > > 
> > > Signed-off-by: Sebastian Reichel <sebastian.reichel@...labora.com>
> > > ---
> > >  .../media/platform/verisilicon/hantro_drv.c   | 38 +++++++++++++++++++
> > >  1 file changed, 38 insertions(+)
> > > 
> > > diff --git a/drivers/media/platform/verisilicon/hantro_drv.c b/drivers/media/platform/verisilicon/hantro_drv.c
> > > index 34b123dafd89..b722a20c5fe3 100644
> > > --- a/drivers/media/platform/verisilicon/hantro_drv.c
> > > +++ b/drivers/media/platform/verisilicon/hantro_drv.c
> > > @@ -722,6 +722,7 @@ static const struct of_device_id of_hantro_match[] = {
> > >  	{ .compatible = "rockchip,rk3399-vpu", .data = &rk3399_vpu_variant, },
> > >  	{ .compatible = "rockchip,rk3568-vepu", .data = &rk3568_vepu_variant, },
> > >  	{ .compatible = "rockchip,rk3568-vpu", .data = &rk3568_vpu_variant, },
> > > +	{ .compatible = "rockchip,rk3588-vepu121", .data = &rk3568_vpu_variant, },
> > >  	{ .compatible = "rockchip,rk3588-av1-vpu", .data = &rk3588_vpu981_variant, },
> > >  #endif
> > >  #ifdef CONFIG_VIDEO_HANTRO_IMX8M
> > > @@ -992,6 +993,39 @@ static const struct media_device_ops hantro_m2m_media_ops = {
> > >  	.req_queue = v4l2_m2m_request_queue,
> > >  };
> > >  
> > > +/*
> > > + * Some SoCs, like RK3588 have multiple identical Hantro cores, but the
> > > + * kernel is currently missing support for multi-core handling. Exposing
> > > + * separate devices for each core to userspace is bad, since that does
> > > + * not allow scheduling tasks properly (and creates ABI). With this workaround
> > > + * the driver will only probe for the first core and early exit for the other
> > > + * cores. Once the driver gains multi-core support, the same technique
> > > + * for detecting the main core can be used to cluster all cores together.
> > > + */
> > > +static int hantro_disable_multicore(struct hantro_dev *vpu)
> > > +{
> > > +	const char *compatible;
> > > +	struct device_node *node;
> > > +	int ret;
> > > +
> > > +	/* Intentionally ignores the fallback strings */
> > > +	ret = of_property_read_string(vpu->dev->of_node, "compatible", &compatible);
> > > +	if (ret)
> > > +		return ret;
> > > +
> > > +	/* first compatible node found from the root node is considered the main core */
> > > +	node = of_find_compatible_node(NULL, NULL, compatible);
> > > +	if (!node)
> > > +		return -EINVAL; /* broken DT? */
> > > +
> > > +	if (vpu->dev->of_node != node) {
> > > +		dev_info(vpu->dev, "missing multi-core support, ignoring this instance\n");
> > > +		return -ENODEV;
> > > +	}
> > > +
> > > +	return 0;
> > > +}
> > > +
> > >  static int hantro_probe(struct platform_device *pdev)
> > >  {
> > >  	const struct of_device_id *match;
> > > @@ -1011,6 +1045,10 @@ static int hantro_probe(struct platform_device *pdev)
> > >  	match = of_match_node(of_hantro_match, pdev->dev.of_node);
> > >  	vpu->variant = match->data;
> > >  
> > > +	ret = hantro_disable_multicore(vpu);
> > > +	if (ret)
> > > +		return ret;
> > > +
> > 
> > I think this might be better as two patches?
> > 
> > As this patch stands, the disable-multicore handling is done for _all_
> > hantro variants, so part of me wants this to be labeled as such.
> > 
> > The whole reasoning is completely ok, but somehow having this under
> > the "add rk3588" umbrella feels strange ;-)
> 
> I can do that, but the 'rockchip,rk3588-vepu121' part is only needed
> because of the multicore handling. If the kernel already had this bit
> in the past, the RK3568 compatible could be used for RK3588 (as a
> fallback compatible), just like for VPU121.

I meant, you're doing hantro_disable_multicore() here also for everyone
else (i.MX etc), hence I'd like that to be a separate commit in this
series like:

----- 8< ------
media: hantro: Disable multi-core handling for the time being

The VSI doc for the Hantro codec describes the grouping of up to 4 instances.
The kernel currently doesn't handle multi-core processing .... foo bar ....
----- 8< ------

And then add rk3588 support on top of that.


Heiko