xen/pcifront: Fix mysterious crashes when NUMA locality information was extracted.

Occasionaly PV guests would crash with:

pciback 0000:00:00.1: Xen PCI mapped GSI0 to IRQ16
BUG: unable to handle kernel paging request at 0000000d1a8c0be0
.. snip..
  <ffffffff8139ce1b>] find_next_bit+0xb/0x10
  [<ffffffff81387f22>] cpumask_next_and+0x22/0x40
  [<ffffffff813c1ef8>] pci_device_probe+0xb8/0x120
  [<ffffffff81529097>] ? driver_sysfs_add+0x77/0xa0
  [<ffffffff815293e4>] driver_probe_device+0x1a4/0x2d0
  [<ffffffff813c1ddd>] ? pci_match_device+0xdd/0x110
  [<ffffffff81529657>] __device_attach_driver+0xa7/0xb0
  [<ffffffff815295b0>] ? __driver_attach+0xa0/0xa0
  [<ffffffff81527622>] bus_for_each_drv+0x62/0x90
  [<ffffffff8152978d>] __device_attach+0xbd/0x110
  [<ffffffff815297fb>] device_attach+0xb/0x10
  [<ffffffff813b75ac>] pci_bus_add_device+0x3c/0x70
  [<ffffffff813b7618>] pci_bus_add_devices+0x38/0x80
  [<ffffffff813dc34e>] pcifront_scan_root+0x13e/0x1a0
  [<ffffffff817a0692>] pcifront_backend_changed+0x262/0x60b
  [<ffffffff814644c6>] ? xenbus_gather+0xd6/0x160
  [<ffffffff8120900f>] ? put_object+0x2f/0x50
  [<ffffffff81465c1d>] xenbus_otherend_changed+0x9d/0xa0
  [<ffffffff814678ee>] backend_changed+0xe/0x10
  [<ffffffff81463a28>] xenwatch_thread+0xc8/0x190
  [<ffffffff810f22f0>] ? woken_wake_function+0x10/0x10

which was the result of two things:

When we call pci_scan_root_bus we would pass in 'sd' (sysdata)
pointer which was an 'pcifront_sd' structure. However in the
pci_device_add it expects that the 'sd' is 'struct sysdata' and
sets the dev->node to what is in sd->node (offset 4):

set_dev_node(&dev->dev, pcibus_to_node(bus));

 __pcibus_to_node(const struct pci_bus *bus)
{
        const struct pci_sysdata *sd = bus->sysdata;

        return sd->node;
}

However our structure was pcifront_sd which had nothing at that
offset:

struct pcifront_sd {
        int                        domain;    /*     0     4 */
        /* XXX 4 bytes hole, try to pack */
        struct pcifront_device *   pdev;      /*     8     8 */
}

That is an hole - filled with garbage as we used kmalloc instead of
kzalloc (the second problem).

This patch fixes the issue by:
 1) Use kzalloc to initialize to a well known state.
 2) Put 'struct pci_sysdata' at the start of 'pcifront_sd'. That
    way access to the 'node' will access the right offset.

Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Cc: <stable@vger.kernel.org>
Reviewed-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Signed-off-by: David Vrabel <david.vrabel@citrix.com>
This commit is contained in:
Konrad Rzeszutek Wilk 2016-02-11 16:10:26 -05:00 committed by David Vrabel
parent 2cfec6a2f9
commit 4d8c8bd6f2

View file

@ -53,7 +53,7 @@ struct pcifront_device {
};
struct pcifront_sd {
int domain;
struct pci_sysdata sd;
struct pcifront_device *pdev;
};
@ -67,7 +67,9 @@ static inline void pcifront_init_sd(struct pcifront_sd *sd,
unsigned int domain, unsigned int bus,
struct pcifront_device *pdev)
{
sd->domain = domain;
/* Because we do not expose that information via XenBus. */
sd->sd.node = first_online_node;
sd->sd.domain = domain;
sd->pdev = pdev;
}
@ -468,8 +470,8 @@ static int pcifront_scan_root(struct pcifront_device *pdev,
dev_info(&pdev->xdev->dev, "Creating PCI Frontend Bus %04x:%02x\n",
domain, bus);
bus_entry = kmalloc(sizeof(*bus_entry), GFP_KERNEL);
sd = kmalloc(sizeof(*sd), GFP_KERNEL);
bus_entry = kzalloc(sizeof(*bus_entry), GFP_KERNEL);
sd = kzalloc(sizeof(*sd), GFP_KERNEL);
if (!bus_entry || !sd) {
err = -ENOMEM;
goto err_out;