On Tue, Jul 3, 2018 at 5:26 PM Pingfan Liu <kernelfans@xxxxxxxxx> wrote:
>
> On Tue, Jul 3, 2018 at 3:51 PM Lukas Wunner <lukas@xxxxxxxxx> wrote:
> >
> > On Tue, Jul 03, 2018 at 02:50:40PM +0800, Pingfan Liu wrote:
> > > commit 52cdbdd49853 ("driver core: correct device's shutdown order")
> > > places an assumption of supplier<-consumer order on the process of probe.
> > > But it turns out to break down the parent <- child order in some scene.
> > > E.g in pci, a bridge is enabled by pci core, and behind it, the devices
> > > have been probed. Then comes the bridge's module, which enables extra
> > > feature(such as hotplug) on this bridge. This will break the
> > > parent<-children order and cause failure when "kexec -e" in some scenario.
> > >
> > > The detailed description of the scenario:
> > > An IBM Power9 machine on which, two drivers portdrv_pci and shpchp(a mod)
> > > match the PCI_CLASS_BRIDGE_PCI, but neither of them success to probe due
> > > to some issue. For this case, the bridge is moved after its children in
> > > devices_kset. Then, when "kexec -e", a ata-disk behind the bridge can not
> > > write back buffer in flight due to the former shutdown of the bridge which
> > > clears the BusMaster bit.
> >
> > If you revert commit cc27b735ad3a ("PCI/portdrv: Turn off PCIe services
> > during shutdown"), does the issue go away?
>
> Yes, it is gone.
Have not figured out why the issue was gone. But I think it just cover
some fault.
re-fetch the boot log of mainline kernel without any patch, and filter
out the pci domain 0004
grep "devices_kset: Moving 0004:" newlog.txt
[ 2.114986] devices_kset: Moving 0004:00:00.0 to end of list <---
pcie port drive's probe, but it failed
[ 2.115192] devices_kset: Moving 0004:01:00.0 to end of list
[ 2.115591] devices_kset: Moving 0004:02:02.0 to end of list
[ 2.115923] devices_kset: Moving 0004:02:0a.0 to end of list
[ 2.116141] devices_kset: Moving 0004:02:0b.0 to end of list
[ 2.116358] devices_kset: Moving 0004:02:0c.0 to end of list
[ 3.181860] devices_kset: Moving 0004:03:00.0 to end of list <---
the ata disk controller which sits behind the bridge
[ 10.267081] devices_kset: Moving 0004:00:00.0 to end of list <---
shpc_probe() on this bridge, failed too.
Hence we have the bridge (parent) after the child in devices_kset.
Thanks,
Pingfan
