netdev - Re: [RFC] net: dsa: qca8k: CPU port broken with commit 5502b218e001 ("net: phy: use phy_resolve_aneg_linkmode in genphy_read

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] [day] [month] [year] [list]

Date:   Thu, 28 Feb 2019 17:32:03 +0100
From:   Michal Vokáč <michal.vokac@...ft.com>
To:     Heiner Kallweit <hkallweit1@...il.com>
Cc:     "David S. Miller" <davem@...emloft.net>,
        Florian Fainelli <f.fainelli@...il.com>,
        netdev@...r.kernel.org, Christian Lamparter <chunkeey@...il.com>,
        Andrew Lunn <andrew@...n.ch>, Vinod Koul <vkoul@...nel.org>
Subject: Re: [RFC] net: dsa: qca8k: CPU port broken with commit 5502b218e001
 ("net: phy: use phy_resolve_aneg_linkmode in genphy_read_status")

Added Christian to the CC list.
I see you also work with the QCA833x switch. Your comments are welcome!

On 20. 02. 19 19:37, Heiner Kallweit wrote:
> On 20.02.2019 16:02, Michal Vokáč wrote:
>> Hi,
>>
>> Another issue in a row with networking on imx6dl-yapp4 platform [1]
>> that uses QCA8334 Ethernet switch.
>>
>> Very recently, with Vinod and Andrew, we solved an issue with
>> RGMII_ID mode by patch[2][3]. I tested those with next-20190215
>> and it worked just fine.
>>
>> The patch[2] was merged into next-20190220 so I tested the latest version.
>> Now the cpu port does not work again. I tracked it down to this commit
>> 5502b218e001 ("net: phy: use phy_resolve_aneg_linkmode in
>> genphy_read_status") [4]
>>
>> If I revert the offending commit, cpu port works fine. I suspect the
>> problem is on the qca8k driver side but I am not really sure.
>> AFAICT autonegotiation is not available on the QCA833x cpu port (MAC0).
>>
> Patch[4] only affects a code path with AUTONEG_ENABLE == phydev->autoneg.
> IMO this shouldn't be the case for the cpu port (like you state).
> Seems like somewhere a phydev->autoneg = AUTONEG_DISABLE is missing.
>
> Not clear to me is where the difference with this patch comes from.
> It would be helpful to know (w/o patch[4]):
> - which values the register reads in genphy_read_status() return
> - which bits are set in phydev->advertising

Hello Heiner,

just for the record, this is what I noted when debugging the issue
and planned to send before the weekend..

next-20190228 does not work because of different issue. I will create
another thread for that.

next-20190222 works fine even without 0f3b1cf23f0e ("net: phy: fix reading
fixed phy status")!

next-20190220 does not work. It seems there is some problem with fixed-link
handling and/or with how the switch indicates its autonegotiation capability.

To make the CPU port work with next-20190220, I can either do:
  - force genphy_read_status() to go the AUTONEG_ENABLE != phydev->autoneg
    path, or
  - prevent the qca8k_adjust_link() from writing into the switch registers.
    More details below.

1) Linux version next-20190220, with patch[4]

- genphy_config_init() reads CPU port MII_BMSR register, value: 0x12c
- The bit BMSR_ANEGCAPABLE is set (!?) and so the genphy_read_status()
   takes the AUTONEG_ENABLE == phydev->autoneg path.
- genphy_read_status() returns with phydev->advertising = 0x0
- qca8k_adjust_link() function is called.
- I can bring up the cpu port but it DOES NOT work.
- genphy_read_status() returns with phydev->advertising = 0x22e0
   (bit 13,9,7,6,5 set)
- If I remove the qca8k_write() in qca8k_adjust_link(), or remove
   the qca8k_adjust_link() totaly, the CPU port DOES work.
   
2) Linux version next-20190220, without patch[4]

- genphy_config_init() reads CPU port MII_BMSR register, value: 0x12c
- The bit BMSR_ANEGCAPABLE is set (!?) and so the genphy_read_status()
   takes the AUTONEG_ENABLE path.
- genphy_read_status() returns with phydev->advertising = 0x0
- The qca8k_adjust_link() function is called.
- I can bring up the cpu port and it DOES work.
- genphy_read_status() returns with phydev->advertising = 0x22e0
   (bit 13,9,7,6,5 set)

Not sure if this will be helpful in any way as it is related to the
now outdated versions next-20190220 and next-20190222.

Best regards,
Michal

>> Any ideas what may be the root cause of the problem?
>>
>> Thank you,
>> Michal
>>
>> [1] https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=87489ec3a77f3e01bcf0d46e353ae7112ec8c4f0
>> [2] https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=a968b5e9d5879f9535d6099505f9e14abcafb623
>> [3] https://lore.kernel.org/patchwork/patch/1043817/
>> [4] https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/?id=5502b218e001
>>