[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20250723113912.GA2880767@bhelgaas>
Date: Wed, 23 Jul 2025 06:39:12 -0500
From: Bjorn Helgaas <helgaas@...nel.org>
To: Madhavan Srinivasan <maddy@...ux.ibm.com>
Cc: Michael Ellerman <mpe@...erman.id.au>,
Mahesh J Salgaonkar <mahesh@...ux.ibm.com>,
linuxppc-dev <linuxppc-dev@...ts.ozlabs.org>,
linux-kernel <linux-kernel@...r.kernel.org>,
linux-pci <linux-pci@...r.kernel.org>,
christophe leroy <christophe.leroy@...roup.eu>,
Naveen N Rao <naveen@...nel.org>,
Bjorn Helgaas <bhelgaas@...gle.com>,
Shawn Anastasio <sanastasio@...torengineering.com>,
Timothy Pearson <tpearson@...torengineering.com>
Subject: Re: [PATCH v3 0/6] PowerNV PCIe Hotplug Driver Fixes
On Wed, Jul 23, 2025 at 04:30:18PM +0530, Madhavan Srinivasan wrote:
>
>
> On 7/23/25 2:17 AM, Bjorn Helgaas wrote:
> > On Thu, Jul 17, 2025 at 06:27:52PM -0500, Bjorn Helgaas wrote:
> >> On Tue, Jul 15, 2025 at 04:31:49PM -0500, Timothy Pearson wrote:
> >>> Hello all,
> >>>
> >>> This series includes several fixes for bugs in the PowerNV PCIe hotplug
> >>> driver that were discovered in testing with a Microsemi Switchtec PM8533
> >>> PFX 48xG3 PCIe switch on a PowerNV system, as well as one workaround for
> >>> PCIe switches that don't correctly implement slot presence detection
> >>> such as the aforementioned one. Without the workaround, the switch works
> >>> and downstream devices can be hot-unplugged, but the devices never come
> >>> back online after being plugged in again until the system is rebooted.
> >>> Other hotplug drivers (like pciehp_hpc) use a similar workaround.
> >>>
> >>> Also included are fixes for the EEH driver to make it hotplug safe,
> >>> and a small patch to enable all three attention indicator states per
> >>> the PCIe specification.
> >>>
> >>> Thanks,
> >>>
> >>> Shawn Anastasio (2):
> >>> PCI: pnv_php: Properly clean up allocated IRQs on unplug
> >>> PCI: pnv_php: Work around switches with broken presence detection
> >>>
> >>> Timothy Pearson (4):
> >>> powerpc/eeh: Export eeh_unfreeze_pe()
> >>> powerpc/eeh: Make EEH driver device hotplug safe
> >>> PCI: pnv_php: Fix surprise plug detection and recovery
> >>> PCI: pnv_php: Enable third attention indicator state
> >>>
> >>> arch/powerpc/kernel/eeh.c | 1 +
> >>> arch/powerpc/kernel/eeh_driver.c | 48 ++++--
> >>> arch/powerpc/kernel/eeh_pe.c | 10 +-
> >>> arch/powerpc/kernel/pci-hotplug.c | 3 +
> >>> drivers/pci/hotplug/pnv_php.c | 244 +++++++++++++++++++++++++++---
> >>> 5 files changed, 263 insertions(+), 43 deletions(-)
> >>
> >> I'm OK with this from a PCI perspective, and I optimistically put it
> >> on pci/hotplug.
> >>
> >> I'm happy to merge via the PCI tree, but would need acks from the
> >> powerpc folks for the arch/powerpc parts.
> >>
> >> Alternatively it could be merged via powerpc with my ack on the
> >> drivers/pci patches:
> >>
> >> Acked-by: Bjorn Helgaas <bhelgaas@...gle.com>
> >>
> >> If you do merge via powerpc, I made some comment formatting and commit
> >> log tweaks that I would like reflected in the drivers/pci part. These
> >> are on
> >> https://git.kernel.org/pub/scm/linux/kernel/git/pci/pci.git/log/?h=hotplug
> >
> > Powerpc folks: let me know how you want to handle this. I haven't
> > included it in pci/next yet because I don't have acks for the
> > arch/powerpc parts.
>
> Patchset looks fine to be.
>
> I am fine to take it via my tree since I already have your Acked-by.
OK, I'll assume this will be merged via your tree. Please cherry-pick
the drivers/pci patches from my tree to preserve my tweaks. I moved
them from pci/hotplug to pci/hotplug-pnv_php:
https://git.kernel.org/pub/scm/linux/kernel/git/pci/pci.git/log/?h=hotplug-pnv_php
Powered by blists - more mailing lists