[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [PATCH 2/3] xen-pcifront: Sanity check the MSI/MSI-X values



On Wed, 2011-02-16 at 22:17 +0000, Konrad Rzeszutek Wilk wrote:
> Check the returned vector values for any values that are
> odd or plain incorrect (say vector value zero), and if so
> print a warning. Also fixup the return values.
> 
> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@xxxxxxxxxx>
> ---
>  drivers/pci/xen-pcifront.c |   17 ++++++++++++++---
>  1 files changed, 14 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/pci/xen-pcifront.c b/drivers/pci/xen-pcifront.c
> index 3a5a6fc..6acf6ae 100644
> --- a/drivers/pci/xen-pcifront.c
> +++ b/drivers/pci/xen-pcifront.c
> @@ -277,13 +277,20 @@ static int pci_frontend_enable_msix(struct pci_dev *dev,
>       if (likely(!err)) {
>               if (likely(!op.value)) {
>                       /* we get the result */
> -                     for (i = 0; i < nvec; i++)
> +                     for (i = 0; i < nvec; i++) {
> +                             if (op.msix_entries[i].vector <= 0) {
> +                                     dev_warn(&dev->dev, "MSI-X entry %d" \
> +                                             " is zero!\n", i);

The test says "<= 0" but the text says "== 0". Perhaps
                                        dev_warn(&dev->dev, "MSI-X entry %d has 
invalid vector %d\n",
                                                 i, op.msix_entries[i].vector);

> 
> +                                     err = -EINVAL;
> +                                     continue;

Do we need / should we set *(*vector+i) to something to indicate its
invalidness rather than leave it potentially uninitialised?

> +                             }
>                               *(*vector+i) = op.msix_entries[i].vector;

BTW does the double indirection of the vector serve a purpose?
Everywhere I can see just updates *(*vector+i), I can't see any realloc
of the array itself etc.

Removing the extra level of indirection leads to vector[i] = foo instead
which is much easier on the eye.

> -                     return 0;
> +                     }
> +                     return err;
>               } else {
>                       printk(KERN_DEBUG "enable msix get value %x\n",
>                               op.value);
> -                     return op.value;
> +                     return err;

I think all of these "return err"s can now be pulled out to the end of
the function? makes it clearer what the return is.

>               }
>       } else {
>               dev_err(&dev->dev, "enable msix get err %x\n", err);
> @@ -325,6 +332,10 @@ static int pci_frontend_enable_msi(struct pci_dev *dev, 
> int **vector)
>       err = do_pci_op(pdev, &op);
>       if (likely(!err)) {
>               *(*vector) = op.value;
> +             if (op.value <= 0) {
> +                     dev_warn(&dev->dev, "MSI entry is zero!\n");

Same comment re <= vs == 0.

> +                     err = -EINVAL;
> +             }
>       } else {
>               dev_err(&dev->dev, "pci frontend enable msi failed for dev "
>                                   "%x:%x\n", op.bus, op.devfn);



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.