From mboxrd@z Thu Jan 1 00:00:00 1970 From: Heiner Kallweit Subject: Re: [PATCH net-next] net: phy: realtek: add suspend/resume callbacks for RTL8211B Date: Thu, 24 May 2018 23:42:07 +0200 Message-ID: References: <20180524205302.GB6762@lunn.ch> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit Cc: David Miller , Realtek linux nic maintainers , Hau , Florian Fainelli , "netdev@vger.kernel.org" , Kevin Hao To: Andrew Lunn Return-path: Received: from mail-wr0-f193.google.com ([209.85.128.193]:45493 "EHLO mail-wr0-f193.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1161476AbeEXVmQ (ORCPT ); Thu, 24 May 2018 17:42:16 -0400 Received: by mail-wr0-f193.google.com with SMTP id w3-v6so5673374wrl.12 for ; Thu, 24 May 2018 14:42:16 -0700 (PDT) In-Reply-To: <20180524205302.GB6762@lunn.ch> Sender: netdev-owner@vger.kernel.org List-ID: Am 24.05.2018 um 22:53 schrieb Andrew Lunn: > On Thu, May 24, 2018 at 10:40:12PM +0200, Heiner Kallweit wrote: >> Add RTL8211B suspend / resume callbacks. >> >> Signed-off-by: Heiner Kallweit >> --- >> This patch is based on my knowledge of the r8169 driver, and on some >> guessing. Therefore I'd appreciate a confirmation from Realtek. >> >> The integrated PHY in some chips supported by the r8169 driver uses >> a special sequence for power-down/-up. I have a board with a RTL8168D >> network chip (one of the chips using the special sequence) and there >> the PHY identifies as RTL8211B. So my guess is that this applies also >> to external RTL8211B PHY's. >> >> A hint for RTL8211B requiring a special sequence is that no suspend/ >> resume callbacks are defined yet in the Realtek PHY driver. >> Last but not least the non-standard usage of register MII_MMD_DATA >> is in line with the description of patch 0231b1a074c6. >> ("net: phy: realtek: Use the dummy stubs for MMD register access for rtl8211b") >> --- >> drivers/net/phy/realtek.c | 16 ++++++++++++++++ >> 1 file changed, 16 insertions(+) >> >> diff --git a/drivers/net/phy/realtek.c b/drivers/net/phy/realtek.c >> index 9f48ecf9c..082fb40c6 100644 >> --- a/drivers/net/phy/realtek.c >> +++ b/drivers/net/phy/realtek.c >> @@ -145,6 +145,20 @@ static int rtl8211f_config_init(struct phy_device *phydev) >> return phy_modify_paged(phydev, 0xd08, 0x11, RTL8211F_TX_DELAY, val); >> } >> >> +static int rtl8211b_suspend(struct phy_device *phydev) >> +{ >> + phy_write(phydev, MII_MMD_DATA, BIT(9)); > > Hi Heiner > > Using it like this suggests it is not actually MMD_DATA, it is > something else which just happens to use the same address as the > optional MMD_DATA. To make this clearer, it would be good to add > #defines for both the register address and this BIT(9). Is there any > vendor code you know of which might give you a clue for appropriate > names? > Vendor code (Realtek r8168 driver) just writes value 0x0200 to register 0x0E. I also would have preferred to assign proper names to register and bit. Maybe the Realtek people on cc can provide some information. > I guess this device also does not support EEE? Does phy_init_eee() > correctly figure this out? Is there a chance calling phy_init_eee() > might trigger a suspend? > Right, EEE isn't supported. phy_init_eee() figures this out correctly because read_mmd callback is set to new genphy_read_mmd_unsupported. Heiner > Andrew >