linux-kernel - Re: [PATCH 2/2] phy: phy-rockchip-samsung-hdptx: Don't use dt aliases to determine phy-id

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: <11259672.BaYr0rKQ5T@diego>
Date: Fri, 06 Dec 2024 14:53:12 +0100
From: Heiko Stübner <heiko@...ech.de>
To: vkoul@...nel.org, kishon@...nel.org,
 Cristian Ciocaltea <cristian.ciocaltea@...labora.com>
Cc: linux-phy@...ts.infradead.org, linux-arm-kernel@...ts.infradead.org,
 linux-rockchip@...ts.infradead.org, linux-kernel@...r.kernel.org,
 quentin.schulz@...rry.de, Heiko Stuebner <heiko.stuebner@...rry.de>
Subject:
 Re: [PATCH 2/2] phy: phy-rockchip-samsung-hdptx: Don't use dt aliases to
 determine phy-id

Hi Cristian,

Am Freitag, 6. Dezember 2024, 12:26:35 CET schrieb Cristian Ciocaltea:
> On 12/6/24 12:34 PM, Heiko Stuebner wrote:
> > From: Heiko Stuebner <heiko.stuebner@...rry.de>
> > 
> > The phy needs to know its identity in the system (phy0 or phy1 on rk3588)
> > for some actions and the driver currently contains code abusing of_alias
> > for that.
> > 
> > Devicetree aliases are always optional and should not be used for core
> > device functionality, so instead keep a list of phys on a soc in the
> > of_device_data and find the phy-id by comparing against the mapped
> > register-base.
> > 
> > Fixes: c4b09c562086 ("phy: phy-rockchip-samsung-hdptx: Add clock provider support")
> > Signed-off-by: Heiko Stuebner <heiko.stuebner@...rry.de>
> > ---
> >  .../phy/rockchip/phy-rockchip-samsung-hdptx.c | 50 ++++++++++++++++---
> >  1 file changed, 44 insertions(+), 6 deletions(-)
> > 
> > diff --git a/drivers/phy/rockchip/phy-rockchip-samsung-hdptx.c b/drivers/phy/rockchip/phy-rockchip-samsung-hdptx.c
> > index c5c64c209e96..b137f8c4d157 100644
> > --- a/drivers/phy/rockchip/phy-rockchip-samsung-hdptx.c
> > +++ b/drivers/phy/rockchip/phy-rockchip-samsung-hdptx.c
> > @@ -385,11 +385,22 @@ enum rk_hdptx_reset {
> >  	RST_MAX
> >  };
> 
> [...]
> 
> > +
> > +	/* find the phy-id from the io address */
> > +	hdptx->phy_id = -ENODEV;
> > +	for (id = 0; id < hdptx->cfgs->num_phys; id++) {
> > +		if (res->start == hdptx->cfgs->phy_ids[id]) {
> > +			hdptx->phy_id = id;
> > +			break;
> > +		}
> > +	}
> > +
> > +	if (hdptx->phy_id < 0)
> > +		return dev_err_probe(dev, -ENODEV, "no matching device found\n");
> 
> Maybe we could simply fallback to assume phy1 doesn't exist in this
> case, which avoids the need to provide a match data with a single entry.

Personally I'm a fan of consistent behaviour, not things working
accidentially ;-) . See the usbdp phy for example, also declaring just
one phy using the same mechanism.

Also I really don't trust the hdptxphy-grf being stable over time.
Rockchip engineers always move bts around in those, so there will be a
need for platform-data at some point anyway.


> Regardless,
> 
> Reviewed-by: Cristian Ciocaltea <cristian.ciocaltea@...labora.com>

thanks :-)

Heiko