Re: [PATCH] PCI: fix sriov enabling with virtual bus

From: Bjorn Helgaas
Date: Thu Oct 30 2014 - 13:09:23 EST


[+cc Rafael, linux-acpi]

On Wed, Oct 29, 2014 at 03:26:10PM -0700, Yinghai Lu wrote:
> While enabling sriov for intel igb device got:
>
> sca05-0a81fda5:~ # echo 7 > /sys/bus/pci/devices/0000\:09\:00.0/sriov_numvfs
> [ 729.612191] pci 0000:0a:10.0: [8086:1520] type 00 class 0x020000
> [ 729.619002] BUG: unable to handle kernel NULL pointer dereference at 00000000000000e8
> [ 729.627767] IP: [<ffffffff815901cb>] pci_get_hp_params+0x5b/0x640
> [ 729.634593] PGD 0
> [ 729.636844] Oops: 0000 [#1] SMP
> [ 729.640466] Modules linked in:
> ...
> [ 729.795268] Call Trace:
> [ 729.798007] [<ffffffff81ea0843>] ? pci_mmcfg_read+0x123/0x140
> [ 729.804519] [<ffffffff81ea0770>] ? pci_mmcfg_read+0x50/0x140
> [ 729.810942] [<ffffffff815692a3>] pci_configure_device+0x33/0x350
> [ 729.817744] [<ffffffff8156aba4>] pci_device_add+0x24/0x160
> [ 729.823965] [<ffffffff8158f3bb>] pci_enable_sriov+0x4db/0x7d0
> [ 729.830486] [<ffffffff81b63f54>] ? igb_pci_enable_sriov+0xe4/0x200
> [ 729.837481] [<ffffffff81b63f7f>] igb_pci_enable_sriov+0x10f/0x200
> [ 729.844386] [<ffffffff8155129c>] ? _kstrtoull+0x2c/0x80
> [ 729.850315] [<ffffffff81b640a5>] igb_pci_sriov_configure+0x35/0x40
> [ 729.857318] [<ffffffff815752b5>] sriov_numvfs_store+0xe5/0x140
> [ 729.863934] [<ffffffff817e1f88>] dev_attr_store+0x18/0x30
> [ 729.870063] [<ffffffff812624d8>] sysfs_kf_write+0x48/0x60
> [ 729.876186] [<ffffffff812617ef>] ? kernfs_fop_write+0xaf/0x170
> [ 729.882797] [<ffffffff81261827>] kernfs_fop_write+0xe7/0x170
> [ 729.889222] [<ffffffff811ef66b>] vfs_write+0xcb/0x1c0
> [ 729.894958] [<ffffffff811f0019>] SyS_write+0x49/0xb0
> ...
> [ 729.943531] ---[ end trace 7cf0cdb66637665a ]---
>
> and pci_get_hp_params+0x5b point to
> 0xffffffff815901cb is in pci_get_hp_params (include/linux/device.h:815).
> 810 #include <linux/pm_wakeup.h>
> 811
> 812 static inline const char *dev_name(const struct device *dev)
> 813 {
> 814 /* Use the init name until the kobject becomes available */
> 815 if (dev->init_name)
> 816 return dev->init_name;
> 817
> 818 return kobject_name(&dev->kobj);
>
> The root cause:
> Now pci_configure_device/pci_get_hp_params will be called for every pci_dev,
> including VF that is under virtual bus. But virtual bus does not have bridge
> set. So we can not refer pbus->self->dev directly.

This raises the question of what the correct behavior should be. Your
patch certainly avoids the NULL pointer dereference. It does so by making
acpi_pci_get_bridge_handle() fail gracefully, which means we will not look
for _HPP/_HPX for VF devices. Is that what we want?

Most of the fields included in _HPX are read-only or not applicable for
VFs, but we need to at least ask the question of whether we want to
completely ignore _HPX for VFs. If we do, I think maybe we should make
that more explicit in the code, e.g., by adding an explicit test of
dev->is_virtfn, instead of relying on this special case behavior of
acpi_pci_get_bridge_handle() that in turn depends on the obscure property
of a VF not having a bridge device.

Personally, I think that since the _HPX spec doesn't mention VFs at all, we
might want to assume that _HPX should apply to VFs, just like it applies to
PFs. I can imagine future _HPX record formats, or non-ACPI firmware
configuration hints, that *would* apply to VFs, so it seems like it would
be pretty arbitrary to say "we won't configure VF devices at all."

> Add checking with pbus->slef and bail out early.
>
> Fixing: commit 6cd33649fa83 ("PCI: Add pci_configure_device() during enumeration")

Thanks for including this, but why not use the same format everybody else
does:

Fixes: 6cd33649fa83 ("PCI: Add pci_configure_device() during enumeration")

> Signed-off-by: Yinghai Lu <yinghai@xxxxxxxxxx>
>
> ---
> include/linux/pci-acpi.h | 6 +++++-
> 1 file changed, 5 insertions(+), 1 deletion(-)
>
> Index: linux-2.6/include/linux/pci-acpi.h
> ===================================================================
> --- linux-2.6.orig/include/linux/pci-acpi.h
> +++ linux-2.6/include/linux/pci-acpi.h
> @@ -44,8 +44,12 @@ static inline acpi_handle acpi_pci_get_b
>
> if (pci_is_root_bus(pbus))
> dev = pbus->bridge;
> - else
> + else {
> + if (!pbus->self)
> + return NULL;
> +
> dev = &pbus->self->dev;
> + }
>
> dev_printk(KERN_DEBUG, &pbus->dev, " bridge ACPI_HANDLE %p : %s\n",
> ACPI_HANDLE(dev), dev_name(dev));
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/