netdev - Re: [PATCH net 0/1] net: phy: Fix interrupt mask loss on resume from hibernation

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]

Date:   Fri, 22 Jan 2021 16:19:51 +0100
From:   Heiner Kallweit <hkallweit1@...il.com>
To:     Laurent Badel <laurentbadel@...on.com>, netdev@...r.kernel.org,
        Andrew Lunn <andrew@...n.ch>,
        Russell King <linux@...linux.org.uk>,
        "David S . Miller" <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>, linux-kernel@...r.kernel.org,
        "Rafael J . Wysocki" <rjw@...ysocki.net>,
        Pavel Machek <pavel@....cz>, linux-pm@...r.kernel.org
Subject: Re: [PATCH net 0/1] net: phy: Fix interrupt mask loss on resume from
 hibernation

On 22.01.2021 15:35, Laurent Badel wrote:
> Some PHYs such as SMSC LAN87xx clear the interrupt mask register on
> software reset. Since mdio_bus_phy_restore() calls phy_init_hw() which
> does a software reset of the PHY, these PHYs will lose their interrupt 
> mask configuration on resuming from hibernation.

The (optional) software reset is done via soft_reset callback.
So if the PHY in question needs special treatment after a soft reset,
why not add it to the soft_reset callback?

> 
> I initially reconfigured only the PHY interrupt mask using 
> phydev->config_intr(), which worked fine with PM_DEBUG/test_resume, but
> there seems to be an issue when resuming from a real hibernation, by which
> the interrupt type is not set appropriately (in this case 
> IRQ_TYPE_LEVEL_LOW). Calling irq_set_irq_type() directly from sysfs 

This sounds to me like a lower level driver (e.g. for GPIO / interrupt
controller) not resuming properly from hibernation. Supposedly things
like edge/level high/low/both are stored per interrupt line in a register
of the interrupt controller, and the controller would have to restore
the register value on resume from hibernation. You may want to have
a look at that driver.

> restored the PHY functionality immediately suggesting that everything is
> otherwise well configured. Therefore this patch suggests freeing and
> re-requesting the interrupt, to guarantee proper interrupt configuration.
> 
> Laurent Badel (1):
>   net: phy: Reconfigure PHY interrupt in mdio_bus_phy_restore()
> 
>  drivers/net/phy/phy_device.c | 9 +++++++++
>  1 file changed, 9 insertions(+)
>