Re: [PATCHv2 2/4] net: ethernet: mediatek: do not force autonegiation at init

From: Greg Ungerer <gerg@kernel.org>
To: "René van Dorst" <opensource@vdorst.com>
Cc: sean.wang@mediatek.com, bjorn@mork.no, andrew@lunn.ch,
	vivien.didelot@savoirfairelinux.com, f.fainelli@gmail.com,
	netdev@vger.kernel.org, john@phrozen.org, neil@brown.name
Subject: Re: [PATCHv2 2/4] net: ethernet: mediatek: do not force autonegiation at init
Date: Tue, 15 Jan 2019 14:51:55 +1000	[thread overview]
Message-ID: <6fc5e1e2-c76a-b70f-4182-c86941110e10@kernel.org> (raw)
In-Reply-To: <20190114125557.Horde.ulAN5z7qLnDm0SUpZYJeQNb@www.vdorst.com>

Hi René,

On 14/1/19 10:55 pm, René van Dorst wrote:
> Quoting gerg@kernel.org:
> 
>> From: Greg Ungerer <gerg@kernel.org>
>>
>> Do not attempt to force a port phy auto-ngeotiation during the driver
>> init phase. It is not necessary and results in a warning at system
>> boot up:
>>
>> mtk_soc_eth 1e100000.ethernet: generated random MAC address be:e7:d4:9d:7d:b0
>> mtk_soc_eth 1e100000.ethernet: connected mac 0 to PHY at fixed-0:00 [uid=00000000, driver=Generic PHY]
>> ------------[ cut here ]------------
>> WARNING: CPU: 1 PID: 1 at drivers/net/phy/phy.c:548 phy_start_aneg+0xf0/0x1c0
>> called from state READY
>> Modules linked in:
>> CPU: 1 PID: 1 Comm: swapper/0 Not tainted 5.0.0-rc1-00005-ge3368ed2e336-dirty #11
>> Stack : 0000007a 00000000 00000000 807e0000 80710000 8006fb04 8072db28 0000000b
>>         00000000 00000000 8072d368 87c3db74 80770000 00000001 87c3db08 2892b90e
>>         00000000 00000000 80820000 00000000 00000000 2d363333 0000007b 00000000
>>         2d316372 00000000 00000000 00000000 80770000 00000000 80715e70 00000009
>>         00000224 80716e38 87d6a800 80660334 00000001 80340714 00000004 807e0004
>>         ...
>> Call Trace:
>> [<8000cfe0>] show_stack+0x94/0x12c
>> [<8060de6c>] dump_stack+0x8c/0xd0
>> [<80028430>] __warn+0x10c/0x120
>> [<80028488>] warn_slowpath_fmt+0x44/0x68
>> [<803adf78>] phy_start_aneg+0xf0/0x1c0
>> [<807aed8c>] mtk_init+0x434/0x4a4
>> [<8047310c>] register_netdevice+0x94/0x6d0
>> [<8047376c>] register_netdev+0x24/0x40
>> [<803bf4c0>] mtk_probe+0x2a0/0x930
>> [<80352d24>] platform_drv_probe+0x40/0x9c
>> [<80350e48>] really_probe+0x1f0/0x35c
>> [<803513e4>] __driver_attach+0x10c/0x114
>> [<8034ec88>] bus_for_each_dev+0x70/0xd0
>> [<8035023c>] bus_add_driver+0x1f0/0x25c
>> [<80351c70>] driver_register+0x80/0x144
>> [<800015f4>] do_one_initcall+0x50/0x1ac
>> [<80795e48>] kernel_init_freeable+0x184/0x26c
>> [<8062bbd4>] kernel_init+0x14/0x110
>> [<80007198>] ret_from_kernel_thread+0x14/0x1c
>> ---[ end trace 6e884e9f12ed14c0 ]---
>> mtk_soc_eth 1e100000.ethernet eth0: mediatek frame engine at 0xbe100000, irq 22
>>
>> Signed-off-by: Greg Ungerer <gerg@kernel.org>
>> ---
>>  drivers/net/ethernet/mediatek/mtk_eth_soc.c | 1 -
>>  1 file changed, 1 deletion(-)
>>
>> v2: first patch in this series for this change
>>
>> diff --git a/drivers/net/ethernet/mediatek/mtk_eth_soc.c b/drivers/net/ethernet/mediatek/mtk_eth_soc.c
>> index c7763a189619..e5a32d43a71c 100644
>> --- a/drivers/net/ethernet/mediatek/mtk_eth_soc.c
>> +++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.c
>> @@ -356,7 +356,6 @@ static int mtk_phy_connect(struct net_device *dev)
>>      linkmode_copy(dev->phydev->advertising, dev->phydev->supported);
>>      linkmode_set_bit(ETHTOOL_LINK_MODE_Autoneg_BIT,
>>               dev->phydev->advertising);
>> -    phy_start_aneg(dev->phydev);
>>
>>      of_node_put(np);
>>
>> -- 
>> 2.17.1
> 
> Hi Greg,
> 
> I think this is recently fix and applied, see https://patchwork.kernel.org/patch/10753537/

Yep, looks like that will do it. I will drop this one.

Regards
Greg