From: Marco Felsch <m.felsch@pengutronix.de>
To: Andrzej Hajda <a.hajda@samsung.com>
Cc: Andy Shevchenko <andriy.shevchenko@linux.intel.com>,
Grant Likely <grant.likely@arm.com>,
Saravana Kannan <saravanak@google.com>,
artem.bityutskiy@linux.intel.com, balbi@kernel.org,
broonie@kernel.org, fntoth@gmail.com, gregkh@linuxfoundation.org,
linux-kernel@vger.kernel.org, linux-pm@vger.kernel.org,
peter.ujfalusi@ti.com, rafael@kernel.org,
kernel-team@android.com, nd <nd@arm.com>,
kernel@pengutronix.de
Subject: Re: [PATCH v3] driver core: Break infinite loop when deferred probe can't be satisfied
Date: Tue, 9 Jun 2020 14:10:29 +0200 [thread overview]
Message-ID: <20200609121029.nfhgilpu5meoygoa@pengutronix.de> (raw)
In-Reply-To: <b413d39f-71c4-d291-276d-1087baf07080@samsung.com>
On 20-06-09 11:27, Andrzej Hajda wrote:
>
> On 09.06.2020 08:45, Marco Felsch wrote:
> > On 20-06-08 13:11, Andrzej Hajda wrote:
> >> On 08.06.2020 11:17, Marco Felsch wrote:
> >>> On 20-03-26 18:31, Andy Shevchenko wrote:
> >>>> On Thu, Mar 26, 2020 at 03:01:22PM +0000, Grant Likely wrote:
> >>>>> On 25/03/2020 12:51, Andy Shevchenko wrote:
> >>>>>> On Tue, Mar 24, 2020 at 08:29:01PM -0700, Saravana Kannan wrote:
> >>>>>>> On Tue, Mar 24, 2020 at 5:38 AM Andy Shevchenko <andriy.shevchenko@linux.intel.com> wrote:
> >>>>>>>> Consider the following scenario.
> >>>>>>>>
> >>>>>>>> The main driver of USB OTG controller (dwc3-pci), which has the following
> >>>>>>>> functional dependencies on certain platform:
> >>>>>>>> - ULPI (tusb1210)
> >>>>>>>> - extcon (tested with extcon-intel-mrfld)
> >>>>>>>>
> >>>>>>>> Note, that first driver, tusb1210, is available at the moment of
> >>>>>>>> dwc3-pci probing, while extcon-intel-mrfld is built as a module and
> >>>>>>>> won't appear till user space does something about it.
> >>>>>>>>
> >>>>>>>> This is depicted by kernel configuration excerpt:
> >>>>>>>>
> >>>>>>>> CONFIG_PHY_TUSB1210=y
> >>>>>>>> CONFIG_USB_DWC3=y
> >>>>>>>> CONFIG_USB_DWC3_ULPI=y
> >>>>>>>> CONFIG_USB_DWC3_DUAL_ROLE=y
> >>>>>>>> CONFIG_USB_DWC3_PCI=y
> >>>>>>>> CONFIG_EXTCON_INTEL_MRFLD=m
> >>>>>>>>
> >>>>>>>> In the Buildroot environment the modules are probed by alphabetical ordering
> >>>>>>>> of their modaliases. The latter comes to the case when USB OTG driver will be
> >>>>>>>> probed first followed by extcon one.
> >>>>>>>>
> >>>>>>>> So, if the platform anticipates extcon device to be appeared, in the above case
> >>>>>>>> we will get deferred probe of USB OTG, because of ordering.
> >>>>>>>>
> >>>>>>>> Since current implementation, done by the commit 58b116bce136 ("drivercore:
> >>>>>>>> deferral race condition fix") counts the amount of triggered deferred probe,
> >>>>>>>> we never advance the situation -- the change makes it to be an infinite loop.
> >>>>>>> Hi Andy,
> >>>>>>>
> >>>>>>> I'm trying to understand this sequence of steps. Sorry if the questions
> >>>>>>> are stupid -- I'm not very familiar with USB/PCI stuff.
> >>>>>> Thank you for looking into this. My answer below.
> >>>>>>
> >>>>>> As a first thing I would like to tell that there is another example of bad
> >>>>>> behaviour of deferred probe with no relation to USB. The proposed change also
> >>>>>> fixes that one (however, less possible to find in real life).
> >>>>>>
> >>>>>>>> ---8<---8<---
> >>>>>>>>
> >>>>>>>> [ 22.187127] driver_deferred_probe_trigger <<< 1
> >>>>>>>>
> >>>>>>>> ...here is the late initcall triggers deferred probe...
> >>>>>>>>
> >>>>>>>> [ 22.191725] platform dwc3.0.auto: deferred_probe_work_func in deferred list
> >>>>>>>>
> >>>>>>>> ...dwc3.0.auto is the only device in the deferred list...
> >>>>>>> Ok, dwc3.0.auto is the only unprobed device at this point?
> >>>>>> Correct.
> >>>>>>
> >>>>>>>> [ 22.198727] platform dwc3.0.auto: deferred_probe_work_func 1 <<< counter 1
> >>>>>>>>
> >>>>>>>> ...the counter before mutex is unlocked is kept the same...
> >>>>>>>>
> >>>>>>>> [ 22.205663] platform dwc3.0.auto: Retrying from deferred list
> >>>>>>>>
> >>>>>>>> ...mutes has been unlocked, we try to re-probe the driver...
> >>>>>>>>
> >>>>>>>> [ 22.211487] bus: 'platform': driver_probe_device: matched device dwc3.0.auto with driver dwc3
> >>>>>>>> [ 22.220060] bus: 'platform': really_probe: probing driver dwc3 with device dwc3.0.auto
> >>>>>>>> [ 22.238735] bus: 'ulpi': driver_probe_device: matched device dwc3.0.auto.ulpi with driver tusb1210
> >>>>>>>> [ 22.247743] bus: 'ulpi': really_probe: probing driver tusb1210 with device dwc3.0.auto.ulpi
> >>>>>>>> [ 22.256292] driver: 'tusb1210': driver_bound: bound to device 'dwc3.0.auto.ulpi'
> >>>>>>>> [ 22.263723] driver_deferred_probe_trigger <<< 2
> >>>>>>>>
> >>>>>>>> ...the dwc3.0.auto probes ULPI, we got successful bound and bumped counter...
> >>>>>>>>
> >>>>>>>> [ 22.268304] bus: 'ulpi': really_probe: bound device dwc3.0.auto.ulpi to driver tusb1210
> >>>>>>> So where did this dwc3.0.auto.ulpi come from?
> >>>>>>> Looks like the device is created by dwc3_probe() through this call flow:
> >>>>>>> dwc3_probe() -> dwc3_core_init() -> dwc3_core_ulpi_init() ->
> >>>>>>> dwc3_ulpi_init() -> ulpi_register_interface() -> ulpi_register()
> >>>>>> Correct.
> >>>>>>
> >>>>>>>> [ 22.276697] platform dwc3.0.auto: Driver dwc3 requests probe deferral
> >>>>>>> Can you please point me to which code patch actually caused the probe
> >>>>>>> deferral?
> >>>>>> Sure, it's in drd.c.
> >>>>>>
> >>>>>> if (device_property_read_string(dev, "linux,extcon-name", &name) == 0) {
> >>>>>> edev = extcon_get_extcon_dev(name);
> >>>>>> if (!edev)
> >>>>>> return ERR_PTR(-EPROBE_DEFER);
> >>>>>> return edev;
> >>>>>> }
> >>>>>>
> >>>>>>>> ...but extcon driver is still missing...
> >>>>>>>>
> >>>>>>>> [ 22.283174] platform dwc3.0.auto: Added to deferred list
> >>>>>>>> [ 22.288513] platform dwc3.0.auto: driver_deferred_probe_add_trigger local counter: 1 new counter 2
> >>>>>>> I'm not fully aware of all the USB implications, but if extcon is
> >>>>>>> needed, why can't that check be done before we add and probe the ulpi
> >>>>>>> device? That'll avoid this whole "fake" probing and avoid the counter
> >>>>>>> increase. And avoid the need for this patch that's touching the code
> >>>>>>> code that's already a bit delicate.
> >>>>>>> Also, with my limited experience with all the possible drivers in the
> >>>>>>> kernel, it's weird that the ulpi device is added and probed before we
> >>>>>>> make sure the parent device (dwc3.0.auto) can actually probe
> >>>>>>> successfully.
> >>>>>> As I said above the deferred probe trigger has flaw on its own.
> >>>>>> Even if we fix for USB case, there is (and probably will be) others.
> >>>>> Right here is the driver design bug. A driver's probe() hook should *not*
> >>>>> return -EPROBE_DEFER after already creating child devices which may have
> >>>>> already been probed.
> >>>> Any documentation statement for this requirement?
> >>>>
> >>>> By the way, I may imagine other mechanisms that probe the driver on other CPU
> >>>> at the same time (let's consider parallel modprobes). The current code has a
> >>>> flaw with that.
> >>> Hi,
> >>>
> >>> sorry for picking this up again but I stumbled above the same issue
> >>> within the driver imx/drm driver which is using the component framework.
> >>> I end up in a infinity boot loop if I enabled the HDMI (which is the
> >>> DesignWare bridge device) and the LVDS support and the LVDS bind return
> >>> with EPROBE_DEFER. There are no words within the component framework docs
> >>> which says that this is forbidden. Of course we can work-around the
> >>> driver-core framework but IMHO this shouldn't be the way to go. I do not
> >>> say that we should revert the commit introducing the regression but we
> >>> should address this not only by extending the docs since the most
> >>> drm-drivers are using the component framework and can end up in the same
> >>> situation.
> >> I am not sure why do you think this is similar issue.
> > Because I see trying to bind the device over and over..
> >
> >> Please describe the issue in more detail. Which drivers defers probe and
> >> why, and why do you have infinite loop.
> > As said I'm currently on the imx-drm driver. The iMX6 devices are
> > using the synopsis HDMI IP core and so they are using this bridge device
> > driver (drivers/gpu/drm/bridge/synopsys/). The imx-drm driver can be
> > build module wise. As example I enabled the LDB and the HDMI support.
> > The HDMI driver is composed as platform driver with different
> > (sub-)drivers and devices. Those devices are populated by the HDMI core
> > driver _probe() function and triggers a driver_deferred_probe_trigger()
> > after the driver successfully probed. The LDB driver bind() returns
> > -EPROBE_DEFER because the panel we are looking for depends on a defered
> > regulator device. Now the defered probe code tries to probe the defered
> > devices again because the local-trigger count was changed by the HDMI
> > driver and we are in the never ending loop.
> >
> >> In general deferring probe from bind is not forbidden, but it should be
> >> used carefully (as everything in kernel :) ). Fixing deferring probe
> >> issues in many cases it is a matter of figuring out 'dependency loops'
> >> and breaking them by splitting device initialization into more than one
> >> phase.
> > We are on the way of splitting the imx-drm driver but there are many
> > other DRM drivers using the component framework. As far as I can see the
> > sunxi8 driver is component based and uses the same HDMI driver. I'm with
> > Andy that we should fix that on the common/core place.
>
>
> I have looked at the drivers and I see the main issue I see is that imx
> drivers performs resource acquisition in bind phase.
As I said we are working on this.
> I think rule of
> thumb should be "do not expose yourself, until you are ready", which in
> this case means "do not call component_add, until resources are
> acquired" - ie resource acquisition should be performed in probe.
Hm.. there are is no documentation which forbid this use-case. I thought
that the component framework bind() equals the driver probe() function..
> I use
> this approach mainly to avoid multiple deferred re-probes, but it should
> solve also this issue, so even if there will be solution to "deferred
> probe issues" in core it would be good to fix imx drivers.
Pls, see my above comments. It is not only the imx driver. Also we
shouldn't expect that driver-developers will follow a rule which is
not written somewhere.
Regards,
Marco
> Regards
>
> Andrzej
>
>
> >
> > Regards,
> > Marco
> >
> >> Regards
> >>
> >> Andrzej
> >>
> >>
> >>>>> It can be solved by refactoring the driver probe routine. If a resource is
> >>>>> required to be present, then check that it is available early; before
> >>>>> registering child devices.
> >>>> We fix one and leave others.
> >>> E.g. the imx-drm and the sunxi driver...
> >>>
> >>> Regards,
> >>> Marco
> >>>
> >>>>> The proposed solution to modify driver core is fragile and susceptible to
> >>>>> side effects from other probe paths. I don't think it is the right approach.
> >>>> Have you tested it on your case? Does it fix the issue?
> >>>>
next prev parent reply other threads:[~2020-06-09 12:10 UTC|newest]
Thread overview: 34+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-03-24 17:57 [PATCH v3] driver core: Break infinite loop when deferred probe can't be satisfied Andy Shevchenko
2020-03-25 3:29 ` Saravana Kannan
2020-03-25 12:51 ` Andy Shevchenko
2020-03-25 22:08 ` Saravana Kannan
2020-03-26 8:39 ` Rafael J. Wysocki
2020-03-26 9:45 ` Peter Ujfalusi
2020-03-26 12:03 ` Andy Shevchenko
2020-03-26 13:45 ` Grant Likely
2020-03-26 14:23 ` Andy Shevchenko
2020-03-26 11:57 ` Andy Shevchenko
2020-03-26 13:48 ` Grant Likely
2020-03-26 18:45 ` Saravana Kannan
2020-03-26 11:54 ` Andy Shevchenko
2020-03-26 14:46 ` Grant Likely
2020-03-26 19:55 ` Saravana Kannan
2020-03-26 15:01 ` Grant Likely
2020-03-26 15:20 ` Grant Likely
2020-03-26 16:31 ` Andy Shevchenko
2020-03-26 16:39 ` Greg KH
2020-03-26 18:06 ` Grant Likely
2020-03-27 8:03 ` Greg KH
2020-03-27 12:37 ` Grant Likely
2020-03-27 12:51 ` Greg KH
2020-06-08 9:17 ` Marco Felsch
2020-06-08 11:11 ` Andrzej Hajda
2020-06-09 6:45 ` Marco Felsch
2020-06-09 7:30 ` Saravana Kannan
2020-06-09 9:27 ` Andrzej Hajda
2020-06-09 12:10 ` Marco Felsch [this message]
2020-06-09 13:02 ` Andrzej Hajda
2020-06-09 13:16 ` Mark Brown
2020-06-08 11:13 ` Andy Shevchenko
2020-06-08 11:59 ` Marco Felsch
2020-06-08 12:11 ` Andy Shevchenko
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20200609121029.nfhgilpu5meoygoa@pengutronix.de \
--to=m.felsch@pengutronix.de \
--cc=a.hajda@samsung.com \
--cc=andriy.shevchenko@linux.intel.com \
--cc=artem.bityutskiy@linux.intel.com \
--cc=balbi@kernel.org \
--cc=broonie@kernel.org \
--cc=fntoth@gmail.com \
--cc=grant.likely@arm.com \
--cc=gregkh@linuxfoundation.org \
--cc=kernel-team@android.com \
--cc=kernel@pengutronix.de \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pm@vger.kernel.org \
--cc=nd@arm.com \
--cc=peter.ujfalusi@ti.com \
--cc=rafael@kernel.org \
--cc=saravanak@google.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).