* [PATCH v9 4/4] PCI: Allow extend_bridge_window() to shrink resource if necessary
@ 2019-10-29 15:29 Nicholas Johnson
2019-10-30 12:07 ` mika.westerberg
0 siblings, 1 reply; 4+ messages in thread
From: Nicholas Johnson @ 2019-10-29 15:29 UTC (permalink / raw)
To: linux-kernel; +Cc: linux-pci, bhelgaas, mika.westerberg, corbet, benh, logang
Remove checks for resource size in extend_bridge_window(). This is
necessary to allow the pci_bus_distribute_available_resources() to
function when the kernel parameter pci=hpmemsize=nn[KMG] is used to
allocate resources. Because the kernel parameter sets the size of all
hotplug bridges to be the same, there are problems when nested hotplug
bridges are encountered. Fitting a downstream hotplug bridge with size X
and normal bridges with non-zero size Y into parent hotplug bridge with
size X is impossible, and hence the downstream hotplug bridge needs to
shrink to fit into its parent.
Add check for if bridge is extended or shrunken and adjust pci_dbg to
reflect this.
Reset the resource if its new size is zero (if we have run out of a
bridge window resource) to prevent the PCI resource assignment code from
attempting to assign a zero-sized resource.
Rename extend_bridge_window() to adjust_bridge_window() to reflect the
fact that the window can now shrink.
Signed-off-by: Nicholas Johnson <nicholas.johnson-opensource@outlook.com.au>
---
drivers/pci/setup-bus.c | 23 +++++++++++++++--------
1 file changed, 15 insertions(+), 8 deletions(-)
diff --git a/drivers/pci/setup-bus.c b/drivers/pci/setup-bus.c
index fe8b2c715..f8cd54584 100644
--- a/drivers/pci/setup-bus.c
+++ b/drivers/pci/setup-bus.c
@@ -1814,7 +1814,7 @@ void __init pci_assign_unassigned_resources(void)
}
}
-static void extend_bridge_window(struct pci_dev *bridge, struct resource *res,
+static void adjust_bridge_window(struct pci_dev *bridge, struct resource *res,
struct list_head *add_list,
resource_size_t new_size)
{
@@ -1823,13 +1823,20 @@ static void extend_bridge_window(struct pci_dev *bridge, struct resource *res,
if (res->parent)
return;
- if (resource_size(res) >= new_size)
- return;
+ if (new_size > resource_size(res)) {
+ add_size = new_size - resource_size(res);
+ pci_dbg(bridge, "bridge window %pR extended by %pa\n", res,
+ &add_size);
+ } else if (new_size < resource_size(res)) {
+ add_size = resource_size(res) - new_size;
+ pci_dbg(bridge, "bridge window %pR shrunken by %pa\n", res,
+ &add_size);
+ }
- add_size = new_size - resource_size(res);
- pci_dbg(bridge, "bridge window %pR extended by %pa\n", res, &add_size);
res->end = res->start + new_size - 1;
remove_from_list(add_list, res);
+ if (!new_size)
+ reset_resource(res);
}
static void pci_bus_distribute_available_resources(struct pci_bus *bus,
@@ -1865,9 +1872,9 @@ static void pci_bus_distribute_available_resources(struct pci_bus *bus,
* Update the resources to fill as much remaining resource space in the
* parent bridge as possible, while considering alignment.
*/
- extend_bridge_window(bridge, io_res, add_list, resource_size(&io));
- extend_bridge_window(bridge, mmio_res, add_list, resource_size(&mmio));
- extend_bridge_window(bridge, mmio_pref_res, add_list,
+ adjust_bridge_window(bridge, io_res, add_list, resource_size(&io));
+ adjust_bridge_window(bridge, mmio_res, add_list, resource_size(&mmio));
+ adjust_bridge_window(bridge, mmio_pref_res, add_list,
resource_size(&mmio_pref));
/*
--
2.23.0
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH v9 4/4] PCI: Allow extend_bridge_window() to shrink resource if necessary
2019-10-29 15:29 [PATCH v9 4/4] PCI: Allow extend_bridge_window() to shrink resource if necessary Nicholas Johnson
@ 2019-10-30 12:07 ` mika.westerberg
2019-10-30 13:22 ` Nicholas Johnson
0 siblings, 1 reply; 4+ messages in thread
From: mika.westerberg @ 2019-10-30 12:07 UTC (permalink / raw)
To: Nicholas Johnson; +Cc: linux-kernel, linux-pci, bhelgaas, corbet, benh, logang
On Tue, Oct 29, 2019 at 03:29:21PM +0000, Nicholas Johnson wrote:
> Remove checks for resource size in extend_bridge_window(). This is
> necessary to allow the pci_bus_distribute_available_resources() to
> function when the kernel parameter pci=hpmemsize=nn[KMG] is used to
> allocate resources. Because the kernel parameter sets the size of all
> hotplug bridges to be the same, there are problems when nested hotplug
> bridges are encountered. Fitting a downstream hotplug bridge with size X
> and normal bridges with non-zero size Y into parent hotplug bridge with
> size X is impossible, and hence the downstream hotplug bridge needs to
> shrink to fit into its parent.
>
> Add check for if bridge is extended or shrunken and adjust pci_dbg to
> reflect this.
>
> Reset the resource if its new size is zero (if we have run out of a
> bridge window resource) to prevent the PCI resource assignment code from
> attempting to assign a zero-sized resource.
>
> Rename extend_bridge_window() to adjust_bridge_window() to reflect the
> fact that the window can now shrink.
>
> Signed-off-by: Nicholas Johnson <nicholas.johnson-opensource@outlook.com.au>
> ---
> drivers/pci/setup-bus.c | 23 +++++++++++++++--------
> 1 file changed, 15 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/pci/setup-bus.c b/drivers/pci/setup-bus.c
> index fe8b2c715..f8cd54584 100644
> --- a/drivers/pci/setup-bus.c
> +++ b/drivers/pci/setup-bus.c
> @@ -1814,7 +1814,7 @@ void __init pci_assign_unassigned_resources(void)
> }
> }
>
> -static void extend_bridge_window(struct pci_dev *bridge, struct resource *res,
> +static void adjust_bridge_window(struct pci_dev *bridge, struct resource *res,
> struct list_head *add_list,
> resource_size_t new_size)
> {
> @@ -1823,13 +1823,20 @@ static void extend_bridge_window(struct pci_dev *bridge, struct resource *res,
> if (res->parent)
> return;
>
> - if (resource_size(res) >= new_size)
> - return;
> + if (new_size > resource_size(res)) {
> + add_size = new_size - resource_size(res);
> + pci_dbg(bridge, "bridge window %pR extended by %pa\n", res,
> + &add_size);
> + } else if (new_size < resource_size(res)) {
> + add_size = resource_size(res) - new_size;
> + pci_dbg(bridge, "bridge window %pR shrunken by %pa\n", res,
> + &add_size);
> + }
Do we need to care about new_size == resource_size(res)?
Also there are several calls of resource_size(res) above so probably
worth storing it into a helper variable.
> - add_size = new_size - resource_size(res);
> - pci_dbg(bridge, "bridge window %pR extended by %pa\n", res, &add_size);
> res->end = res->start + new_size - 1;
> remove_from_list(add_list, res);
> + if (!new_size)
> + reset_resource(res);
> }
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH v9 4/4] PCI: Allow extend_bridge_window() to shrink resource if necessary
2019-10-30 12:07 ` mika.westerberg
@ 2019-10-30 13:22 ` Nicholas Johnson
2019-10-30 13:27 ` mika.westerberg
0 siblings, 1 reply; 4+ messages in thread
From: Nicholas Johnson @ 2019-10-30 13:22 UTC (permalink / raw)
To: mika.westerberg; +Cc: linux-kernel, linux-pci, bhelgaas, corbet, benh, logang
On Wed, Oct 30, 2019 at 02:07:05PM +0200, mika.westerberg@linux.intel.com wrote:
> On Tue, Oct 29, 2019 at 03:29:21PM +0000, Nicholas Johnson wrote:
> > Remove checks for resource size in extend_bridge_window(). This is
> > necessary to allow the pci_bus_distribute_available_resources() to
> > function when the kernel parameter pci=hpmemsize=nn[KMG] is used to
> > allocate resources. Because the kernel parameter sets the size of all
> > hotplug bridges to be the same, there are problems when nested hotplug
> > bridges are encountered. Fitting a downstream hotplug bridge with size X
> > and normal bridges with non-zero size Y into parent hotplug bridge with
> > size X is impossible, and hence the downstream hotplug bridge needs to
> > shrink to fit into its parent.
> >
> > Add check for if bridge is extended or shrunken and adjust pci_dbg to
> > reflect this.
> >
> > Reset the resource if its new size is zero (if we have run out of a
> > bridge window resource) to prevent the PCI resource assignment code from
> > attempting to assign a zero-sized resource.
> >
> > Rename extend_bridge_window() to adjust_bridge_window() to reflect the
> > fact that the window can now shrink.
> >
> > Signed-off-by: Nicholas Johnson <nicholas.johnson-opensource@outlook.com.au>
> > ---
> > drivers/pci/setup-bus.c | 23 +++++++++++++++--------
> > 1 file changed, 15 insertions(+), 8 deletions(-)
> >
> > diff --git a/drivers/pci/setup-bus.c b/drivers/pci/setup-bus.c
> > index fe8b2c715..f8cd54584 100644
> > --- a/drivers/pci/setup-bus.c
> > +++ b/drivers/pci/setup-bus.c
> > @@ -1814,7 +1814,7 @@ void __init pci_assign_unassigned_resources(void)
> > }
> > }
> >
> > -static void extend_bridge_window(struct pci_dev *bridge, struct resource *res,
> > +static void adjust_bridge_window(struct pci_dev *bridge, struct resource *res,
> > struct list_head *add_list,
> > resource_size_t new_size)
> > {
> > @@ -1823,13 +1823,20 @@ static void extend_bridge_window(struct pci_dev *bridge, struct resource *res,
> > if (res->parent)
> > return;
> >
> > - if (resource_size(res) >= new_size)
> > - return;
> > + if (new_size > resource_size(res)) {
> > + add_size = new_size - resource_size(res);
> > + pci_dbg(bridge, "bridge window %pR extended by %pa\n", res,
> > + &add_size);
> > + } else if (new_size < resource_size(res)) {
> > + add_size = resource_size(res) - new_size;
> > + pci_dbg(bridge, "bridge window %pR shrunken by %pa\n", res,
> > + &add_size);
> > + }
>
> Do we need to care about new_size == resource_size(res)?
Sorry, I skipped over this before re-posting. If we are not changing the
size, does it need to be logged in pci_dbg? I could change it to ">=",
meaning if the size does not change then it will be "extended" by 0 in
pci_dbg.
Let me know if you wish for this to happen. But I am easy either way.
If you are fine with how it is, PATCH v10 should have addressed
everything so far.
>
> Also there are several calls of resource_size(res) above so probably
> worth storing it into a helper variable.
>
> > - add_size = new_size - resource_size(res);
> > - pci_dbg(bridge, "bridge window %pR extended by %pa\n", res, &add_size);
> > res->end = res->start + new_size - 1;
> > remove_from_list(add_list, res);
> > + if (!new_size)
> > + reset_resource(res);
> > }
Cheers
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH v9 4/4] PCI: Allow extend_bridge_window() to shrink resource if necessary
2019-10-30 13:22 ` Nicholas Johnson
@ 2019-10-30 13:27 ` mika.westerberg
0 siblings, 0 replies; 4+ messages in thread
From: mika.westerberg @ 2019-10-30 13:27 UTC (permalink / raw)
To: Nicholas Johnson; +Cc: linux-kernel, linux-pci, bhelgaas, corbet, benh, logang
On Wed, Oct 30, 2019 at 01:22:44PM +0000, Nicholas Johnson wrote:
> On Wed, Oct 30, 2019 at 02:07:05PM +0200, mika.westerberg@linux.intel.com wrote:
> > On Tue, Oct 29, 2019 at 03:29:21PM +0000, Nicholas Johnson wrote:
> > > Remove checks for resource size in extend_bridge_window(). This is
> > > necessary to allow the pci_bus_distribute_available_resources() to
> > > function when the kernel parameter pci=hpmemsize=nn[KMG] is used to
> > > allocate resources. Because the kernel parameter sets the size of all
> > > hotplug bridges to be the same, there are problems when nested hotplug
> > > bridges are encountered. Fitting a downstream hotplug bridge with size X
> > > and normal bridges with non-zero size Y into parent hotplug bridge with
> > > size X is impossible, and hence the downstream hotplug bridge needs to
> > > shrink to fit into its parent.
> > >
> > > Add check for if bridge is extended or shrunken and adjust pci_dbg to
> > > reflect this.
> > >
> > > Reset the resource if its new size is zero (if we have run out of a
> > > bridge window resource) to prevent the PCI resource assignment code from
> > > attempting to assign a zero-sized resource.
> > >
> > > Rename extend_bridge_window() to adjust_bridge_window() to reflect the
> > > fact that the window can now shrink.
> > >
> > > Signed-off-by: Nicholas Johnson <nicholas.johnson-opensource@outlook.com.au>
> > > ---
> > > drivers/pci/setup-bus.c | 23 +++++++++++++++--------
> > > 1 file changed, 15 insertions(+), 8 deletions(-)
> > >
> > > diff --git a/drivers/pci/setup-bus.c b/drivers/pci/setup-bus.c
> > > index fe8b2c715..f8cd54584 100644
> > > --- a/drivers/pci/setup-bus.c
> > > +++ b/drivers/pci/setup-bus.c
> > > @@ -1814,7 +1814,7 @@ void __init pci_assign_unassigned_resources(void)
> > > }
> > > }
> > >
> > > -static void extend_bridge_window(struct pci_dev *bridge, struct resource *res,
> > > +static void adjust_bridge_window(struct pci_dev *bridge, struct resource *res,
> > > struct list_head *add_list,
> > > resource_size_t new_size)
> > > {
> > > @@ -1823,13 +1823,20 @@ static void extend_bridge_window(struct pci_dev *bridge, struct resource *res,
> > > if (res->parent)
> > > return;
> > >
> > > - if (resource_size(res) >= new_size)
> > > - return;
> > > + if (new_size > resource_size(res)) {
> > > + add_size = new_size - resource_size(res);
> > > + pci_dbg(bridge, "bridge window %pR extended by %pa\n", res,
> > > + &add_size);
> > > + } else if (new_size < resource_size(res)) {
> > > + add_size = resource_size(res) - new_size;
> > > + pci_dbg(bridge, "bridge window %pR shrunken by %pa\n", res,
> > > + &add_size);
> > > + }
> >
> > Do we need to care about new_size == resource_size(res)?
> Sorry, I skipped over this before re-posting. If we are not changing the
> size, does it need to be logged in pci_dbg? I could change it to ">=",
> meaning if the size does not change then it will be "extended" by 0 in
> pci_dbg.
>
> Let me know if you wish for this to happen. But I am easy either way.
No need.
> If you are fine with how it is, PATCH v10 should have addressed
> everything so far.
Yes, v10 looks fine by me.
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2019-10-30 13:27 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-10-29 15:29 [PATCH v9 4/4] PCI: Allow extend_bridge_window() to shrink resource if necessary Nicholas Johnson
2019-10-30 12:07 ` mika.westerberg
2019-10-30 13:22 ` Nicholas Johnson
2019-10-30 13:27 ` mika.westerberg
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).