Re: [RFC 03/11] net: phy: refactor c45 phy identification sequence

[Russell King - ARM Linux admin]

On Sat, May 23, 2020 at 12:32:59PM -0500, Jeremy Linton wrote:
> Hi,
> 
> On 5/23/20 10:28 AM, Andrew Lunn wrote:
> > On Fri, May 22, 2020 at 04:30:51PM -0500, Jeremy Linton wrote:
> > > Lets factor out the phy id logic, and make it generic
> > > so that it can be used for c22 and c45.
> > > 
> > > Signed-off-by: Jeremy Linton <jeremy.linton@xxxxxxx>
> > > ---
> > >   drivers/net/phy/phy_device.c | 65 +++++++++++++++++++-----------------
> > >   1 file changed, 35 insertions(+), 30 deletions(-)
> > > 
> > > diff --git a/drivers/net/phy/phy_device.c b/drivers/net/phy/phy_device.c
> > > index 7746c07b97fe..f0761fa5e40b 100644
> > > --- a/drivers/net/phy/phy_device.c
> > > +++ b/drivers/net/phy/phy_device.c
> > > @@ -695,6 +695,29 @@ static int get_phy_c45_devs_in_pkg(struct mii_bus *bus, int addr, int dev_addr,
> > >   	return 0;
> > >   }
> > > +static int _get_phy_id(struct mii_bus *bus, int addr, int dev_addr,
> > > +		       u32 *phy_id, bool c45)
> > 
> > Hi Jeremy
> > 
> > How about read_phy_id() so you can avoid the _ prefix.
> > 
> > >   static bool valid_phy_id(int val)
> > >   {
> > >   	return (val > 0 && ((val & 0x1fffffff) != 0x1fffffff));
> > > @@ -715,17 +738,17 @@ static bool valid_phy_id(int val)
> > >    */
> > >   static int get_phy_c45_ids(struct mii_bus *bus, int addr, u32 *phy_id,
> > >   			   struct phy_c45_device_ids *c45_ids) {
> > > -	int phy_reg;
> > > -	int i, reg_addr;
> > > +	int ret;
> > > +	int i;
> > >   	const int num_ids = ARRAY_SIZE(c45_ids->device_ids);
> > >   	u32 *devs = &c45_ids->devices_in_package;
> > >   	/* Find first non-zero Devices In package. Device zero is reserved
> > >   	 * for 802.3 c45 complied PHYs, so don't probe it at first.
> > >   	 */
> > > -	for (i = 1; i < num_ids && *devs == 0; i++) {
> > > -		phy_reg = get_phy_c45_devs_in_pkg(bus, addr, i, devs);
> > > -		if (phy_reg < 0)
> > > +	for (i = 0; i < num_ids && *devs == 0; i++) {
> > > +		ret = get_phy_c45_devs_in_pkg(bus, addr, i, devs);
> > > +		if (ret < 0)
> > >   			return -EIO;
> > 
> > Renaming reg_addr to ret does not belong in this patch.
> > 
> 
> Looks like I changed the loop index in this patch while shuffling things
> around yesterday too. The "for (i = 1/0.." change belongs in 5/11 as well.

Indeed; that's a change of behaviour - see the CS4315/CS4340 workaround.

Note that MMD 0 is explicitly marked "Reserved" in 802.3, so we
shouldn't be probing it in this way.

Also note that bit 0 of the "devices in package" does not mean that
MMD 0 is accessible; it means that the clause 22 registers are
present:

  "Bit 5.0 is used to indicate that Clause 22 functionality has been
  implemented within a Clause 45 electrical interface device."

which basically means Clause 22 protocol over Clause 45 electrical
interface.

So, we should be careful about poking in MMD 0 if 5.0 is set.

Some places that will break are:

- phylink_phy_read() / phylink_phy_write()

- phy_restart_aneg() / phy_config_aneg() if we have an clause 45
  MDIO interface that can't issue clause 22 MDIO cycles with a PHY
  that sets 5.0.

These comments apply more to your patch 4, but you brought up the
MMD 0 accesses in this patch, so I thought I'd provide it here.

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTC for 0.8m (est. 1762m) line in suburbia: sync at 13.1Mbps down 424kbps up