Re: [PATCH 5.17 127/298] driver core: Fix wait_for_device_probe() & deferred_probe_timeout interaction

From: Geert Uytterhoeven
Date: Wed Aug 16 2023 - 06:15:56 EST


On Wed, Aug 16, 2023 at 12:10 PM Geert Uytterhoeven
<geert@xxxxxxxxxxxxxx> wrote:
> On Wed, Aug 16, 2023 at 11:39 AM Shreeya Patel
> <shreeya.patel@xxxxxxxxxxxxx> wrote:
> > On 13/06/22 15:40, Greg Kroah-Hartman wrote:
> > > From: Saravana Kannan<saravanak@xxxxxxxxxx>
> > >
> > > [ Upstream commit 5ee76c256e928455212ab759c51d198fedbe7523 ]
> > >
> > > Mounting NFS rootfs was timing out when deferred_probe_timeout was
> > > non-zero [1]. This was because ip_auto_config() initcall times out
> > > waiting for the network interfaces to show up when
> > > deferred_probe_timeout was non-zero. While ip_auto_config() calls
> > > wait_for_device_probe() to make sure any currently running deferred
> > > probe work or asynchronous probe finishes, that wasn't sufficient to
> > > account for devices being deferred until deferred_probe_timeout.
> > >
> > > Commit 35a672363ab3 ("driver core: Ensure wait_for_device_probe() waits
> > > until the deferred_probe_timeout fires") tried to fix that by making
> > > sure wait_for_device_probe() waits for deferred_probe_timeout to expire
> > > before returning.
> > >
> > > However, if wait_for_device_probe() is called from the kernel_init()
> > > context:
> > >
> > > - Before deferred_probe_initcall() [2], it causes the boot process to
> > > hang due to a deadlock.
> > >
> > > - After deferred_probe_initcall() [3], it blocks kernel_init() from
> > > continuing till deferred_probe_timeout expires and beats the point of
> > > deferred_probe_timeout that's trying to wait for userspace to load
> > > modules.
> > >
> > > Neither of this is good. So revert the changes to
> > > wait_for_device_probe().
> > >
> > > [1] -https://lore.kernel.org/lkml/TYAPR01MB45443DF63B9EF29054F7C41FD8C60@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx/
> > > [2] -https://lore.kernel.org/lkml/YowHNo4sBjr9ijZr@dev-arch.thelio-3990X/
> > > [3] -https://lore.kernel.org/lkml/Yo3WvGnNk3LvLb7R@xxxxxxxxxxxxx/
> >
> > Hi Saravana, Greg,
> >
> >
> > KernelCI found this patch causes the baseline.bootrr.deferred-probe-empty test to fail on r8a77960-ulcb,
> > see the following details for more information.
>
> Commit 9be4cbd09da820a2 ("driver core: Set default deferred_probe_timeout
> back to 0.") in v5.19 contains a reference to the same commit as
> mentioned in the Fixes tag. Does backporting that help?

Anyway, remembering the days (weeks?) spent in investigating
subtle issues with fw_devlinks and deferred probe, collecting all the
fixes for backporting to stable may be a very hard job...

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds