Received-SPF: pass (google.com: domain of linux-kernel+bounces-68738-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.199.223 as permitted sender) client-ip=147.75.199.223;
From: =?UTF-8?q?Ilpo=20J=C3=A4rvinen?= <ilpo.jarvinen@linux.intel.com>
Date: Fri, 16 Feb 2024 16:23:29 +0200 (EET)
To: "Maciej W. Rozycki" <macro@orcam.me.uk>
cc: Bjorn Helgaas <bhelgaas@google.com>, linux-pci@vger.kernel.org, 
    LKML <linux-kernel@vger.kernel.org>, 
    Mika Westerberg <mika.westerberg@linux.intel.com>
Subject: Re: [PATCH 1/1] PCI: Cleanup link activation wait logic
In-Reply-To: <alpine.DEB.2.21.2402161349350.3971@angie.orcam.me.uk>
Message-ID: <853a63bd-74cb-e19b-24b7-426a0fdd9003@linux.intel.com>
References: <20240202134108.4096-1-ilpo.jarvinen@linux.intel.com> <alpine.DEB.2.21.2402021359450.15781@angie.orcam.me.uk> <ce73f41a-b529-726f-ee4e-9d0e0cee3320@linux.intel.com> <alpine.DEB.2.21.2402161349350.3971@angie.orcam.me.uk>
Precedence: bulk
MIME-Version: 1.0
Content-Type: multipart/mixed; boundary="8323328-291145476-1708093409=:1097"

  This message is in MIME format.  The first part should be readable text,
  while the remaining parts are likely unreadable without MIME-aware tools.

--8323328-291145476-1708093409=:1097
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: QUOTED-PRINTABLE

On Fri, 16 Feb 2024, Maciej W. Rozycki wrote:
> On Fri, 16 Feb 2024, Ilpo J=C3=A4rvinen wrote:
>=20
> > >  You change the logic here in that the second conditional isn't run i=
f the=20
> > > first has not.  This is wrong, unclamping is not supposed to rely on =
LBMS.=20
> > > It is supposed to be always run and any failure has to be reported to=
o, as=20
> > > a retraining error.
> >=20
> > Now that (I think) I fully understand the intent of the second=20
> > condition/block one additional question occurred to me.
> >=20
> > How is the 2nd condition even supposed to work in the current place whe=
n=20
> > firmware has pre-arranged the 2.5GT/s resctriction? Wouldn't the link c=
ome=20
> > up fine in that case and the quirk code is not called at all since the=
=20
> > link came up successfully?
>=20
>  The quirk is called unconditionally from `pci_device_add', so an attempt=
=20
> to unclamp will always happen with a working link for qualifying devices.

Ah, thanks. I'd stared the other two calls enough of times I'd forgotten=20
the 3rd one even existed.

> > Yet another thing in this quirk code I don't like is how it can leaves =
the=20
> > target speed to 2.5GT/s when the quirk fails to get the link working=20
> > (which actually does happen in the disconnection cases because DLLLA wo=
n't=20
> > be set so the target speed will not be restored).
>=20
>  I chose to leave the target speed at the most recent setting, because th=
e=20
> link doesn't work in that case anyway, so I concluded it doesn't matter,=
=20
> but reduces messing with the device; technically you should retrain again=
=20
> afterwards.  I'm not opposed to changing this if you have a use case.

It remains suboptimally set in a case where something is plugged again=20
into that port, for Thunderbolt it doesn't matter as the PCIe speed picked=
=20
is quite bogus anyway, but disconnect then plug something again is not=20
limited to Thunderbolt.

I've no immediate plans on changing it now but it may come relevant when=20
attempting to make the bandwidth controller to trigger the quirk. To me=20
there are two quirks, not just one so I might have to split them to make=20
it better suited for triggering them from bwctrl.

--=20
 i.

--8323328-291145476-1708093409=:1097--