On Tue, May 31, 2011 at 11:17:45PM -0700, Yinghai Lu wrote:
> On 05/31/2011 06:10 PM, Ben Hutchings wrote:
> >
> > In this configuration it can be ignored completely, but I don't think
> > there's any generic way to determine that.
I am tempted to say -- the platform is buggy, since it is asking for more
resources than are available. Maybe we should provide a kernel command line
flag to ignore sriov resources?
>
....
>
> [PATCH] PCI: Make assign resource to unassigned SRIOV BAR to be optional
>
> make them into add list, so it will make other resource could get resource
> allocation at first.
>
> still try to make several try with them by adding them into fail list.
Under what scenario will a retry ever succeed?
>
> Signed-off-by: Yinghai Lu <yinghai@xxxxxxxxxx>
>
> ---
> drivers/pci/setup-bus.c | 28 ++++++++++++++++++++++++----
> 1 file changed, 24 insertions(+), 4 deletions(-)
>
> Index: linux-2.6/drivers/pci/setup-bus.c
> ===================================================================
> --- linux-2.6.orig/drivers/pci/setup-bus.c
> +++ linux-2.6/drivers/pci/setup-bus.c
> @@ -120,13 +120,16 @@ static inline void reset_resource(struct
> * resources
> * @head : head of the list tracking requests with allocated
> * resources
> + * @fail_head : head of the list tracking requests failed to allocate
> + * resources
> *
> * Walk through each element of the add_head and try to procure
> * additional resources for the element, provided the element
> * is in the head list.
> */
> static void adjust_resources_sorted(struct resource_list_x *add_head,
> - struct resource_list *head)
> + struct resource_list *head,
> + struct resource_list_x *fail_head)
> {
> struct resource *res;
> struct resource_list_x *list, *tmp, *prev;
> @@ -153,9 +156,17 @@ static void adjust_resources_sorted(stru
> idx = res - &list->dev->resource[0];
> add_size=list->add_size;
> if (!resource_size(res) && add_size) {
> - res->end = res->start + add_size - 1;
> - if(pci_assign_resource(list->dev, idx))
> + res->end = res->start + add_size - 1;
> + if(pci_assign_resource(list->dev, idx)) {
> +#ifdef CONFIG_PCI_IOV
> + /* add to failed list for next loop */
> + if (fail_head && idx >= PCI_IOV_RESOURCES &&
> + idx <= PCI_IOV_RESOURCE_END)
> + add_to_failed_list(fail_head, list->dev,
> + res);
> +#endif
do we need this? if we cannot allocate them, just ignore them. Anyway its a nice-to-have resource.
> reset_resource(res);
> + }
> } else if (add_size) {
> adjust_resource(res, res->start,
> resource_size(res) + add_size);
> @@ -212,7 +223,7 @@ static void __assign_resources_sorted(st
> /* Try to satisfy any additional nice-to-have resource
> requests */
> if (add_head)
> - adjust_resources_sorted(add_head, head);
> + adjust_resources_sorted(add_head, head, fail_head);
> free_list(resource_list, head);
> }
>
> @@ -662,6 +673,15 @@ static int pbus_size_mem(struct pci_bus
> if (r->parent || (r->flags & mask) != type)
> continue;
> r_size = resource_size(r);
> +#ifdef CONFIG_PCI_IOV
> + /* put SRIOV requested res to good to have list */
---> nice-to-have
> + if (add_head && i >= PCI_IOV_RESOURCES &&
> + i <= PCI_IOV_RESOURCE_END) {
> + r->end = r->start - 1;
> + add_to_list(add_head, dev, r, r_size);
> + r_size = 0;
> + }
> +#endif
> /* For bridges size != alignment */
> align = pci_resource_alignment(dev, r);
> order = __ffs(align) - 20;
I will try this patch on my setup and report back.
Thanks,
RP
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
