netdev - Re: [PATCH] net: dsa: mv88e6xxx: Make *_c45 callbacks agree with phy_*

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] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAO-L_46kqBrDdYP7p3He0cBF1OP7TJKnhYK1NR_gMZf2n_928A@mail.gmail.com>
Date: Tue, 16 Jan 2024 17:51:13 -0800
From: Tim Menninger <tmenninger@...estorage.com>
To: Andrew Lunn <andrew@...n.ch>
Cc: f.fainelli@...il.com, olteanv@...il.com, davem@...emloft.net, 
	edumazet@...gle.com, kuba@...nel.org, pabeni@...hat.com, 
	netdev@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] net: dsa: mv88e6xxx: Make *_c45 callbacks agree with
 phy_*_c45 callbacks

On Tue, Jan 16, 2024 at 3:21 PM Andrew Lunn <andrew@...n.ch> wrote:
>
> > Hi Andrew,
> >
> > It bubbles up as EIO (the translation happens in get_phy_c45_ids when
> > get_phy_c45_devs_in_pkg fails) and ultimately causes the probe to fail.
> >
> > The EIO causes the scan to stop and fail immediately - the way I read
> > mdiobus_scan_bus_c45, only ENODEV is permissible.
>
> O.K. At minimum, this should be added to the commit message.
>
> However, i'm wondering if this is the correct fix. I would prefer that
> the scan code just acts on the -EOPNOTSUPP the same was as
> -ENODEV. Maybe the error code from phy_c45_probe_present() should be
> returned as is. And mdiobus_scan_bus_c45() is extended to handle
> -EOPNOTSUPP ?
>
>             Andrew

Noted about the commit message.

To return -EOPNOTSUPP high enough up that the mdiobus_scan function(s) can
directly handle it would mean at minimum, these functions have -EOPNOTSUPP
added to their respective list of possible return values:
    - get_phy_c45_ids (static)
    - phy_get_c45_ids
    - get_phy_device
    - mdiobus_scan (static)
    - mdiobus_scan_c22 (static)
    - mdiobus_scan_c45 (static)

I didn't look beyond to see whether any callers of phy_get_c45_ids or
get_phy_device also return errors as-are, but it feels a little broad in
scope to me.

My impression is still that the read_c45 function should agree with the
phy_read_c45 function, but that isn't a hill I care to die on if you still
think otherwise. Thoughts?