* Re: [PATCHv3 09/10] PCI: Unify device inaccessible
[not found] ` <e0606dfcf8780bf994432dc373581fdf0af18f8e.camel@kernel.crashing.org>
@ 2025-04-18 3:55 ` Lukas Wunner
0 siblings, 0 replies; only message in thread
From: Lukas Wunner @ 2025-04-18 3:55 UTC (permalink / raw)
To: Benjamin Herrenschmidt
Cc: Keith Busch, Linux PCI, Bjorn Helgaas, Sinan Kaya, Thomas Tai,
poza, Christoph Hellwig, Mika Westerberg, Mahesh J Salgaonkar,
Oliver O'Halloran, Madhavan Srinivasan, Michael Ellerman,
linuxppc-dev
[cc += PowerPC / EEH maintainers]
Hi Ben,
On Tue, Sep 25, 2018 at 11:10:01AM +1000, Benjamin Herrenschmidt wrote:
> On Tue, 2018-09-18 at 17:57 -0600, Keith Busch wrote:
> > + * pci_dev_set_io_state - Set the new error state if possible.
> > + *
> > + * @dev - pci device to set new error_state
> > + * @new - the state we want dev to be in
> > + *
> > + * Must be called with device_lock held.
>
> Any reason why you don't do cmpxchg as I originally suggested (sorry
> I've been away and may have missed some previous emails)
>
> This won't work for PowerPC EEH. We will change the state from a
> readl() so at interrupt time or any other context.
>
> We really need the cmpxchg variant.
Independently from your request, pci_dev_set_io_state() was
converted to cmpxchg() in 2023 with commit 74ff8864cc84
("PCI: hotplug: Allow marking devices as disconnected during
bind/unbind").
So you may now amend EEH to use pcie_do_recovery() or whatever
you needed this for.
I had kept your e-mail in my inbox as a reminder that there's a
remaining issue here and just came across it while clearing out
other messages.
Thanks,
Lukas
^ permalink raw reply [flat|nested] only message in thread