[PATCH v5 05/12] powerpc/eeh: EEH for pSeries hot plug

[Oliver O'Halloran]

On Fri, Sep 20, 2019 at 6:28 AM Nathan Lynch <nathanl at linux.ibm.com> wrote:
>
> Hello Sam,
>
> Sam Bobroff <sbobroff at linux.ibm.com> writes:
>
> With this change, I get a crash (use after free by the looks of it) when
> I remove and then add a pci device in qemu:
>
> $ qemu-system-ppc64 -M pseries -append 'debug console=hvc0' \
>   -nographic -vga none -m 1G,slots=32,maxmem=1024G -smp 2 \
>   -kernel vmlinux -initrd ~/b/br/ppc64le-initramfs/images/rootfs.cpio \
>   -nic model=e1000

is there anything special in your kernel config? I tested this with
pseries_le_defconfig and couldn't hit the crash.

>
> ...
>
> # echo 1 > /sys/devices/pci0000:00/0000:00:00.0/remove ; \
>   echo 1 > /sys/devices/pci0000:00/pci_bus/0000:00/rescan
>
> pci 0000:00:00.0: Removing from iommu group 0
> pci 0000:00:00.0: [8086:100e] type 00 class 0x020000
> pci 0000:00:00.0: reg 0x10: [mem 0x200080000000-0x20008001ffff]
> pci 0000:00:00.0: reg 0x14: [io  0x10040-0x1007f]
> pci 0000:00:00.0: reg 0x30: [mem 0x200080040000-0x20008007ffff pref]
> pci 0000:00:00.0: Adding to iommu group 0
> pci 0000:00:00.0: BAR 6: assigned [mem 0x200080000000-0x20008003ffff pref]
> pci 0000:00:00.0: BAR 0: assigned [mem 0x200080040000-0x20008005ffff]
> pci 0000:00:00.0: BAR 1: assigned [io  0x10000-0x1003f]
> e1000 0000:00:00.0 eth0: (PCI:33MHz:32-bit) 52:54:00:12:34:56
> e1000 0000:00:00.0 eth0: Intel(R) PRO/1000 Network Connection
> pci 0000:00:00.0: Removing from iommu group 0
> pci 0000:00:00.0: [8086:100e] type 00 class 0x020000
> pci 0000:00:00.0: reg 0x10: [mem 0x200080040000-0x20008005ffff]
> pci 0000:00:00.0: reg 0x14: [io  0x10000-0x1003f]
> pci 0000:00:00.0: reg 0x30: [mem 0x200080040000-0x20008007ffff pref]
> pci 0000:00:00.0: BAR 6: assigned [mem 0x200080000000-0x20008003ffff pref]
> pci 0000:00:00.0: BAR 0: assigned [mem 0x200080040000-0x20008005ffff]
> pci 0000:00:00.0: BAR 1: assigned [io  0x10000-0x1003f]
> BUG: Unable to handle kernel data access at 0x6b6b6b6b6b6b6bfb
> Faulting instruction address: 0xc000000000597270
> Oops: Kernel access of bad area, sig: 11 [#1]
> LE PAGE_SIZE=64K MMU=Hash SMP NR_CPUS=2048 NUMA pSeries
> Modules linked in:
> CPU: 0 PID: 2464 Comm: pci-probe-vs-cp Not tainted 5.3.0-rc2-00092-gf381d5711f09 #76
> NIP:  c000000000597270 LR: c000000000599470 CTR: c0000000002030b0
> REGS: c00000003ee4f650 TRAP: 0380   Not tainted  (5.3.0-rc2-00092-gf381d5711f09)
> MSR:  8000000000009033 <SF,EE,ME,IR,DR,RI,LE>  CR: 24002442  XER: 00000000
> CFAR: c00000000059946c IRQMASK: 0
> GPR00: c000000000599470 c00000003ee4f8e0 c000000003317a00 6b6b6b6b6b6b6b6b
> GPR04: c000000001d0fa38 0000000000000000 0000000000000000 221a64979a66f870
> GPR08: c00000000347b398 0000000000000000 c00000000336e070 ffffffffffffffff
> GPR12: 0000000000002000 c000000004060000 0000000000000000 0000000000000000
> GPR16: 00000000100a78d8 00007fffe9fdff96 00000000100a7898 0000000000000000
> GPR20: 0000000000000000 00000000100e0ff0 0000000000000000 00000000100e0fe8
> GPR24: 0000000000000000 000001002ae50260 c000000001d0fa38 6b6b6b6b6b6b6b6b
> GPR28: fffffffffffffff2 c000000001d0fa38 0000000000000000 c000000003118c18
> NIP [c000000000597270] kernfs_find_ns+0x50/0x3d0
> LR [c000000000599470] kernfs_remove_by_name_ns+0x60/0xe0
> Call Trace:
> [c00000003ee4f8e0] [c00000000020950c] lockdep_hardirqs_on+0x10c/0x210 (unreliable)
> [c00000003ee4f970] [c000000000599470] kernfs_remove_by_name_ns+0x60/0xe0
> [c00000003ee4fa00] [c00000000059ca08] sysfs_remove_file_ns+0x28/0x40
> [c00000003ee4fa20] [c000000000cbd70c] device_remove_file+0x2c/0x40
> [c00000003ee4fa40] [c000000000051480] eeh_sysfs_remove_device+0x50/0xf0
> [c00000003ee4fa80] [c00000000004a594] eeh_add_device_late.part.7+0x84/0x220
> [c00000003ee4fb00] [c0000000000e94f0] pseries_pcibios_bus_add_device+0x60/0xb0
> [c00000003ee4fb70] [c00000000006fc40] pcibios_bus_add_device+0x40/0x60
> [c00000003ee4fb90] [c000000000bc5220] pci_bus_add_device+0x30/0x100
> [c00000003ee4fc00] [c000000000bc5344] pci_bus_add_devices+0x54/0xb0
> [c00000003ee4fc40] [c000000000bca058] pci_rescan_bus+0x48/0x70
> [c00000003ee4fc70] [c000000000bd9adc] dev_bus_rescan_store+0xcc/0x100
> [c00000003ee4fcb0] [c000000000cbc9d8] dev_attr_store+0x38/0x60
> [c00000003ee4fcd0] [c00000000059c460] sysfs_kf_write+0x70/0xb0
> [c00000003ee4fd10] [c00000000059aa98] kernfs_fop_write+0xf8/0x280
> [c00000003ee4fd60] [c0000000004b3e5c] __vfs_write+0x3c/0x70
> [c00000003ee4fd80] [c0000000004b81f0] vfs_write+0xd0/0x220
> [c00000003ee4fdd0] [c0000000004b85ac] ksys_write+0x7c/0x140
> [c00000003ee4fe20] [c00000000000bc6c] system_call+0x5c/0x70
>
> FWIW during boot the EEH core reports:
>
>   EEH: No capable adapters found: recovery disabled.
>
> > diff --git a/arch/powerpc/kernel/eeh.c b/arch/powerpc/kernel/eeh.c
> > index ca8b0c58a6a7..87edac6f2fd9 100644
> > --- a/arch/powerpc/kernel/eeh.c
> > +++ b/arch/powerpc/kernel/eeh.c
> > @@ -1272,7 +1272,7 @@ void eeh_add_device_late(struct pci_dev *dev)
> >       struct pci_dn *pdn;
> >       struct eeh_dev *edev;
> >
> > -     if (!dev || !eeh_enabled())
> > +     if (!dev)
> >               return;
> >
> >       pr_debug("EEH: Adding device %s\n", pci_name(dev));
>
> Reverting this hunk works around (fixes?) it.