[PATCH v4] cpu-topology: Don't error on more than CONFIG_NR_CPUS CPUs in device tree

[PATCH v4] cpu-topology: Don't error on more than CONFIG_NR_CPUS CPUs in device tree

[Zeng Tao]

When the kernel is configured with CONFIG_NR_CPUS smaller than the
number of CPU nodes in the device tree(DT), all the CPU nodes parsing
done to fetch topology information will fail. This is not reasonable
as it is legal to have all the physical CPUs in the system in the DT.

Let us just skip such CPU DT nodes that are not used in the kernel
rather than returning an error.

Reviewed-by: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Zeng Tao <prime.zeng@hisilicon.com>
---
Changelog:
v3->v4:
 -Fix the Review comments by sudeep, including subject, commit log, code
 comments, and also apply the stick the Review-by tag.
v2->v3:
 -Fix the Review comments by sudeep, including fix typo, remove redundant check
 logic, change the warning print level etc.
v1->v2:
 -Remove redundant -ENODEV assignment in get_cpu_for_node
 -Add comment to describe the get_cpu_for_node return values
 -Add skip process for cpu threads
 -Update the commit log with more detail
---
 drivers/base/arch_topology.c | 20 +++++++++++++++-----
 1 file changed, 15 insertions(+), 5 deletions(-)

diff --git a/drivers/base/arch_topology.c b/drivers/base/arch_topology.c
index 5fe44b3..c175fe3 100644
--- a/drivers/base/arch_topology.c
+++ b/drivers/base/arch_topology.c
@@ -248,6 +248,16 @@ core_initcall(free_raw_capacity);
 #endif
 
 #if defined(CONFIG_ARM64) || defined(CONFIG_RISCV)
+/*
+ * This function returns the logic cpu number of the node.
+ * There are basically three kinds of return values:
+ * (1) logic cpu number which is > 0.
+ * (2) -ENODEV when the device tree(DT) node is valid and found in the DT but
+ * there is no possible logical CPU in the kernel to match. This happens
+ * when CONFIG_NR_CPUS is configure to be smaller than the number of
+ * CPU nodes in DT. We need to just ignore this case.
+ * (3) -1 if the node does not exist in the device tree
+ */
 static int __init get_cpu_for_node(struct device_node *node)
 {
 	struct device_node *cpu_node;
@@ -261,7 +271,8 @@ static int __init get_cpu_for_node(struct device_node *node)
 	if (cpu >= 0)
 		topology_parse_cpu_capacity(cpu_node, cpu);
 	else
-		pr_crit("Unable to find CPU node for %pOF\n", cpu_node);
+		pr_info("CPU node for %pOF exist but the possible cpu range is :%*pbl\n",
+			cpu_node, cpumask_pr_args(cpu_possible_mask));
 
 	of_node_put(cpu_node);
 	return cpu;
@@ -286,9 +297,8 @@ static int __init parse_core(struct device_node *core, int package_id,
 				cpu_topology[cpu].package_id = package_id;
 				cpu_topology[cpu].core_id = core_id;
 				cpu_topology[cpu].thread_id = i;
-			} else {
-				pr_err("%pOF: Can't get CPU for thread\n",
-				       t);
+			} else if (cpu != -ENODEV) {
+				pr_err("%pOF: Can't get CPU for thread\n", t);
 				of_node_put(t);
 				return -EINVAL;
 			}
@@ -307,7 +317,7 @@ static int __init parse_core(struct device_node *core, int package_id,
 
 		cpu_topology[cpu].package_id = package_id;
 		cpu_topology[cpu].core_id = core_id;
-	} else if (leaf) {
+	} else if (leaf && cpu != -ENODEV) {
 		pr_err("%pOF: Can't get CPU for leaf core\n", core);
 		return -EINVAL;
 	}
-- 
2.8.1

Re: [PATCH v4] cpu-topology: Don't error on more than CONFIG_NR_CPUS CPUs in device tree

[Sudeep Holla]

On Fri, Jan 17, 2020 at 09:52:52AM +0800, Zeng Tao wrote:
> When the kernel is configured with CONFIG_NR_CPUS smaller than the
> number of CPU nodes in the device tree(DT), all the CPU nodes parsing
> done to fetch topology information will fail. This is not reasonable
> as it is legal to have all the physical CPUs in the system in the DT.
>
> Let us just skip such CPU DT nodes that are not used in the kernel
> rather than returning an error.
>
> Reviewed-by: Sudeep Holla <sudeep.holla@arm.com>
> Signed-off-by: Zeng Tao <prime.zeng@hisilicon.com>

Hi Greg,

Can you pick this patch for v5.6 ?

--
Regards,
Sudeep

Re: [PATCH v4] cpu-topology: Don't error on more than CONFIG_NR_CPUS CPUs in device tree

[Greg Kroah-Hartman]

On Fri, Jan 17, 2020 at 10:19:57AM +0000, Sudeep Holla wrote:
> On Fri, Jan 17, 2020 at 09:52:52AM +0800, Zeng Tao wrote:
> > When the kernel is configured with CONFIG_NR_CPUS smaller than the
> > number of CPU nodes in the device tree(DT), all the CPU nodes parsing
> > done to fetch topology information will fail. This is not reasonable
> > as it is legal to have all the physical CPUs in the system in the DT.
> >
> > Let us just skip such CPU DT nodes that are not used in the kernel
> > rather than returning an error.
> >
> > Reviewed-by: Sudeep Holla <sudeep.holla@arm.com>
> > Signed-off-by: Zeng Tao <prime.zeng@hisilicon.com>
> 
> Hi Greg,
> 
> Can you pick this patch for v5.6 ?

oops, didn't realize this was for me, sorry, will go do so now.

greg k-h

Re: [PATCH v4] cpu-topology: Don't error on more than CONFIG_NR_CPUS CPUs in device tree

[Sudeep Holla]

On Fri, Jan 17, 2020 at 03:21:35PM +0100, Greg Kroah-Hartman wrote:
> On Fri, Jan 17, 2020 at 10:19:57AM +0000, Sudeep Holla wrote:
> > On Fri, Jan 17, 2020 at 09:52:52AM +0800, Zeng Tao wrote:
> > > When the kernel is configured with CONFIG_NR_CPUS smaller than the
> > > number of CPU nodes in the device tree(DT), all the CPU nodes parsing
> > > done to fetch topology information will fail. This is not reasonable
> > > as it is legal to have all the physical CPUs in the system in the DT.
> > >
> > > Let us just skip such CPU DT nodes that are not used in the kernel
> > > rather than returning an error.
> > >
> > > Reviewed-by: Sudeep Holla <sudeep.holla@arm.com>
> > > Signed-off-by: Zeng Tao <prime.zeng@hisilicon.com>
> >
> > Hi Greg,
> >
> > Can you pick this patch for v5.6 ?
>
> oops, didn't realize this was for me, sorry, will go do so now.
>

No worries, it just got baked this morning :). I just dropped you a note
before I could forget.

--
Regards,
Sudeep