netdev - RE: [EXTERNAL] Re: [PATCH 2/2] Reset PHY in phy_init

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: <CH2PR17MB35427EF2FAE4E31FCA144F89DFAA0@CH2PR17MB3542.namprd17.prod.outlook.com>
Date:   Thu, 30 Apr 2020 09:29:48 +0000
From:   "Badel, Laurent" <LaurentBadel@...on.com>
To:     Heiner Kallweit <hkallweit1@...il.com>,
        "fugang.duan@....com" <fugang.duan@....com>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        "andrew@...n.ch" <andrew@...n.ch>,
        "f.fainelli@...il.com" <f.fainelli@...il.com>,
        "linux@...linux.org.uk" <linux@...linux.org.uk>,
        "richard.leitner@...data.com" <richard.leitner@...data.com>,
        "davem@...emloft.net" <davem@...emloft.net>,
        "alexander.levin@...rosoft.com" <alexander.levin@...rosoft.com>,
        "gregkh@...uxfoundation.org" <gregkh@...uxfoundation.org>
CC:     "Quette, Arnaud" <ArnaudQuette@...on.com>
Subject: RE: [EXTERNAL]  Re: [PATCH 2/2] Reset PHY in phy_init_hw() before
 interrupt configuration

> 

-----------------------------
Eaton Industries Manufacturing GmbH ~ Registered place of business: Route de la Longeraie 7, 1110, Morges, Switzerland 

-----------------------------

-----Original Message-----
> From: Heiner Kallweit <hkallweit1@...il.com>
> Sent: Wednesday, April 29, 2020 7:06 PM
> To: Badel, Laurent <LaurentBadel@...on.com>; fugang.duan@....com;
> netdev@...r.kernel.org; andrew@...n.ch; f.fainelli@...il.com;
> linux@...linux.org.uk; richard.leitner@...data.com;
> davem@...emloft.net; alexander.levin@...rosoft.com;
> gregkh@...uxfoundation.org
> Cc: Quette, Arnaud <ArnaudQuette@...on.com>
> Subject: [EXTERNAL] Re: [PATCH 2/2] Reset PHY in phy_init_hw() before
> interrupt configuration
> 
> On 29.04.2020 11:03, Badel, Laurent wrote:
> > Description: this patch adds a reset of the PHY in phy_init_hw()
> > for PHY drivers bearing the PHY_RST_AFTER_CLK_EN flag.
> >
> > Rationale: due to the PHY reset reverting the interrupt mask to default,
> > it is necessary to either perform the reset before PHY configuration,
> > or re-configure the PHY after reset. This patch implements the former
> > as it is simpler and more generic.
> >
> > Fixes: 1b0a83ac04e383e3bed21332962b90710fcf2828 ("net: fec: add
> phy_reset_after_clk_enable() support")
> > Signed-off-by: Laurent Badel <laurentbadel@...on.com>
> >
> > ---
> >  drivers/net/phy/phy_device.c | 7 +++++--
> >  1 file changed, 5 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/net/phy/phy_device.c b/drivers/net/phy/phy_device.c
> > index 28e3c5c0e..2cc511364 100644
> > --- a/drivers/net/phy/phy_device.c
> > +++ b/drivers/net/phy/phy_device.c
> > @@ -1082,8 +1082,11 @@ int phy_init_hw(struct phy_device *phydev)
> >  {
> >  	int ret = 0;
> >
> > -	/* Deassert the reset signal */
> > -	phy_device_reset(phydev, 0);
> > +	/* Deassert the reset signal
> > +	 * If the PHY needs a reset, do it now
> > +	 */
> > +	if (!phy_reset_after_clk_enable(phydev))
> 
> If reset is asserted when entering phy_init_hw(), then
> phy_reset_after_clk_enable() basically becomes a no-op.
> Still it should work as expected due to the reset signal being
> deasserted. It would be worth describing in the comment
> why the code still works in this case.
> 

Thank you for the comment, this is a very good point. 
I will make sure to include some description when resubmitting. 
I had previously tested this and what I saw was that the first 
time you bring up the interface, the reset is not asserted so 
that phy_reset_after_clk_enable() is effective. 
The subsequent times the interface is brought up, the reset 
is already asserted when entering phy_init_hw(), so that 
it becomes a no-op as you correctly pointed out. However, 
that didn't cause any problem on my board, presumably because 
in that case the clock is already running when the PHY comes 
out of reset. 
I will re-test this carefully against the 'net' tree, though,
before coming to conclusions.  

> > +		phy_device_reset(phydev, 0);
> >
> >  	if (!phydev->drv)
> >  		return 0;
> >