[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20160928211632.GA32406@lunn.ch>
Date: Wed, 28 Sep 2016 23:16:32 +0200
From: Andrew Lunn <andrew@...n.ch>
To: Sergei Shtylyov <sergei.shtylyov@...entembedded.com>
Cc: Florian Fainelli <f.fainelli@...il.com>,
Vivien Didelot <vivien.didelot@...oirfairelinux.com>,
netdev <netdev@...r.kernel.org>
Subject: Re: [PATCH RFC 2/6] net: phy: Use threaded IRQ, to allow IRQ from
sleeping devices
On Wed, Sep 28, 2016 at 02:38:03PM +0300, Sergei Shtylyov wrote:
> Hello.
>
> On 9/28/2016 11:32 AM, Andrew Lunn wrote:
>
> >The interrupt lines from PHYs maybe connected to I2C bus expanders, or
> >from switches on MDIO busses. Such interrupts are sourced from devices
> >which sleep, so use threaded interrupts. Threaded interrupts require
> >that the interrupt requester also uses the threaded API. Change the
> >phylib to use the threaded API, which is backwards compatible with
> >none-threaded IRQs.
> >
> >Signed-off-by: Andrew Lunn <andrew@...n.ch>
> >---
> > drivers/net/phy/phy.c | 7 +++----
> > 1 file changed, 3 insertions(+), 4 deletions(-)
> >
> >diff --git a/drivers/net/phy/phy.c b/drivers/net/phy/phy.c
> >index c6f66832a1a6..5c29ed72f721 100644
> >--- a/drivers/net/phy/phy.c
> >+++ b/drivers/net/phy/phy.c
> >@@ -722,10 +722,9 @@ phy_err:
> > int phy_start_interrupts(struct phy_device *phydev)
> > {
> > atomic_set(&phydev->irq_disable, 0);
> >- if (request_irq(phydev->irq, phy_interrupt,
> >- IRQF_SHARED,
> >- "phy_interrupt",
> >- phydev) < 0) {
> >+ if (request_threaded_irq(phydev->irq, NULL, phy_interrupt,
> >+ IRQF_ONESHOT, "phy_interrupt",
>
> What about IRQF_SHARED?
IRQF_ONESHOT | IRQF_SHARED does work.
However, it requires all uses of the interrupt use the same flags.
The commit messages suggests IRQF_SHARED is there because of boards
which have multiple PHYs using one IRQ. That case will work, since all
PHYs will be using this code.
What will be an issue is if the sharing is between a PHY and something
else. That something else will also need to use IRQF_ONESHOT. Anybody
know if this situation actually exists?
Thanks
Andrew
Powered by blists - more mailing lists