The first three patches of the v11 Linux RISC-V AIA series can be
merged independently hence sending these patches as an independent
perparatory series.
(Refer, https://www.spinics.net/lists/devicetree/msg643764.html)
These patches can also be found in the riscv_aia_prep_v1 branch at:
https://github.com/avpatel/linux.git
Anup Patel (3):
RISC-V: Don't fail in riscv_of_parent_hartid() for disabled HARTs
of: property: Add fw_devlink support for msi-parent
irqchip/sifive-plic: Fix syscore registration for multi-socket systems
arch/riscv/kernel/cpu.c | 11 ++++++-----
drivers/irqchip/irq-sifive-plic.c | 7 ++++---
drivers/of/property.c | 2 ++
3 files changed, 12 insertions(+), 8 deletions(-)
--
2.34.1
The riscv_of_processor_hartid() used by riscv_of_parent_hartid() fails
for HARTs disabled in the DT. This results in the following warning
thrown by the RISC-V INTC driver for the E-core on SiFive boards:
[ 0.000000] riscv-intc: unable to find hart id for /cpus/cpu@0/interrupt-controller
The riscv_of_parent_hartid() is only expected to read the hartid from
the DT so we should directly call of_get_cpu_hwid() instead of calling
riscv_of_processor_hartid().
Fixes: ad635e723e17 ("riscv: cpu: Add 64bit hartid support on RV64")
Signed-off-by: Anup Patel <[email protected]>
Reviewed-by: Atish Patra <[email protected]>
---
arch/riscv/kernel/cpu.c | 11 ++++++-----
1 file changed, 6 insertions(+), 5 deletions(-)
diff --git a/arch/riscv/kernel/cpu.c b/arch/riscv/kernel/cpu.c
index c17dacb1141c..157ace8b262c 100644
--- a/arch/riscv/kernel/cpu.c
+++ b/arch/riscv/kernel/cpu.c
@@ -125,13 +125,14 @@ int __init riscv_early_of_processor_hartid(struct device_node *node, unsigned lo
*/
int riscv_of_parent_hartid(struct device_node *node, unsigned long *hartid)
{
- int rc;
-
for (; node; node = node->parent) {
if (of_device_is_compatible(node, "riscv")) {
- rc = riscv_of_processor_hartid(node, hartid);
- if (!rc)
- return 0;
+ *hartid = (unsigned long)of_get_cpu_hwid(node, 0);
+ if (*hartid == ~0UL) {
+ pr_warn("Found CPU without hart ID\n");
+ return -ENODEV;
+ }
+ return 0;
}
}
--
2.34.1
This allows fw_devlink to create device links between consumers of
a MSI and the supplier of the MSI.
Signed-off-by: Anup Patel <[email protected]>
Acked-by: Rob Herring <[email protected]>
Reviewed-by: Saravana Kannan <[email protected]>
---
drivers/of/property.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/of/property.c b/drivers/of/property.c
index cf8dacf3e3b8..afdaefbd03f6 100644
--- a/drivers/of/property.c
+++ b/drivers/of/property.c
@@ -1267,6 +1267,7 @@ DEFINE_SIMPLE_PROP(resets, "resets", "#reset-cells")
DEFINE_SIMPLE_PROP(leds, "leds", NULL)
DEFINE_SIMPLE_PROP(backlight, "backlight", NULL)
DEFINE_SIMPLE_PROP(panel, "panel", NULL)
+DEFINE_SIMPLE_PROP(msi_parent, "msi-parent", "#msi-cells")
DEFINE_SUFFIX_PROP(regulators, "-supply", NULL)
DEFINE_SUFFIX_PROP(gpio, "-gpio", "#gpio-cells")
@@ -1356,6 +1357,7 @@ static const struct supplier_bindings of_supplier_bindings[] = {
{ .parse_prop = parse_leds, },
{ .parse_prop = parse_backlight, },
{ .parse_prop = parse_panel, },
+ { .parse_prop = parse_msi_parent, },
{ .parse_prop = parse_gpio_compat, },
{ .parse_prop = parse_interrupts, },
{ .parse_prop = parse_regulators, },
--
2.34.1
Hi Palmer,
On Wed, Oct 25, 2023 at 7:58 PM Anup Patel <[email protected]> wrote:
>
> The first three patches of the v11 Linux RISC-V AIA series can be
> merged independently hence sending these patches as an independent
> perparatory series.
> (Refer, https://www.spinics.net/lists/devicetree/msg643764.html)
>
> These patches can also be found in the riscv_aia_prep_v1 branch at:
> https://github.com/avpatel/linux.git
>
> Anup Patel (3):
> RISC-V: Don't fail in riscv_of_parent_hartid() for disabled HARTs
> of: property: Add fw_devlink support for msi-parent
> irqchip/sifive-plic: Fix syscore registration for multi-socket systems
As mentioned on the patchwork call, these are the first three patches
of the v11 Linux AIA series.
Please consider this for the Linux-6.7 merge window.
>
> arch/riscv/kernel/cpu.c | 11 ++++++-----
> drivers/irqchip/irq-sifive-plic.c | 7 ++++---
> drivers/of/property.c | 2 ++
> 3 files changed, 12 insertions(+), 8 deletions(-)
>
> --
> 2.34.1
>
Regards,
Anup
On multi-socket systems, we will have a separate PLIC in each socket
so we should register syscore operation only once for multi-socket
systems.
Fixes: e80f0b6a2cf3 ("irqchip/irq-sifive-plic: Add syscore callbacks for hibernation")
Signed-off-by: Anup Patel <[email protected]>
---
drivers/irqchip/irq-sifive-plic.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/drivers/irqchip/irq-sifive-plic.c b/drivers/irqchip/irq-sifive-plic.c
index e1484905b7bd..5b7bc4fd9517 100644
--- a/drivers/irqchip/irq-sifive-plic.c
+++ b/drivers/irqchip/irq-sifive-plic.c
@@ -532,17 +532,18 @@ static int __init __plic_init(struct device_node *node,
}
/*
- * We can have multiple PLIC instances so setup cpuhp state only
- * when context handler for current/boot CPU is present.
+ * We can have multiple PLIC instances so setup cpuhp state
+ * and register syscore operations only when context handler
+ * for current/boot CPU is present.
*/
handler = this_cpu_ptr(&plic_handlers);
if (handler->present && !plic_cpuhp_setup_done) {
cpuhp_setup_state(CPUHP_AP_IRQ_SIFIVE_PLIC_STARTING,
"irqchip/sifive/plic:starting",
plic_starting_cpu, plic_dying_cpu);
+ register_syscore_ops(&plic_irq_syscore_ops);
plic_cpuhp_setup_done = true;
}
- register_syscore_ops(&plic_irq_syscore_ops);
pr_info("%pOFP: mapped %d interrupts with %d handlers for"
" %d contexts.\n", node, nr_irqs, nr_handlers, nr_contexts);
--
2.34.1
On Wed, Oct 25 2023 at 19:58, Anup Patel wrote:
> This allows fw_devlink to create device links between consumers of
> a MSI and the supplier of the MSI.
How is this related to the two fixes in this series?
On Wed, Oct 25 2023 at 19:58, Anup Patel wrote:
> The riscv_of_processor_hartid() used by riscv_of_parent_hartid() fails
> for HARTs disabled in the DT. This results in the following warning
> thrown by the RISC-V INTC driver for the E-core on SiFive boards:
>
> [ 0.000000] riscv-intc: unable to find hart id for /cpus/cpu@0/interrupt-controller
>
> The riscv_of_parent_hartid() is only expected to read the hartid from
> the DT so we should directly call of_get_cpu_hwid() instead of calling
We should? Or maybe not?
Please write precise changelogs and use imperative wording as documented
in Documentation/process.
The following commit has been merged into the irq/core branch of tip:
Commit-ID: f99b926f6543faeadba1b4524d8dc9c102489135
Gitweb: https://git.kernel.org/tip/f99b926f6543faeadba1b4524d8dc9c102489135
Author: Anup Patel <[email protected]>
AuthorDate: Wed, 25 Oct 2023 19:58:20 +05:30
Committer: Thomas Gleixner <[email protected]>
CommitterDate: Fri, 27 Oct 2023 10:09:15 +02:00
irqchip/sifive-plic: Fix syscore registration for multi-socket systems
Multi-socket systems have a separate PLIC in each socket, so __plic_init()
is invoked for each PLIC. __plic_init() registers syscore operations, which
obviously fails on the second invocation.
Move it into the already existing condition for installing the CPU hotplug
state so it is only invoked once when the first PLIC is initialized.
[ tglx: Massaged changelog ]
Fixes: e80f0b6a2cf3 ("irqchip/irq-sifive-plic: Add syscore callbacks for hibernation")
Signed-off-by: Anup Patel <[email protected]>
Signed-off-by: Thomas Gleixner <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
---
drivers/irqchip/irq-sifive-plic.c | 7 ++++---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/drivers/irqchip/irq-sifive-plic.c b/drivers/irqchip/irq-sifive-plic.c
index e148490..5b7bc4f 100644
--- a/drivers/irqchip/irq-sifive-plic.c
+++ b/drivers/irqchip/irq-sifive-plic.c
@@ -532,17 +532,18 @@ done:
}
/*
- * We can have multiple PLIC instances so setup cpuhp state only
- * when context handler for current/boot CPU is present.
+ * We can have multiple PLIC instances so setup cpuhp state
+ * and register syscore operations only when context handler
+ * for current/boot CPU is present.
*/
handler = this_cpu_ptr(&plic_handlers);
if (handler->present && !plic_cpuhp_setup_done) {
cpuhp_setup_state(CPUHP_AP_IRQ_SIFIVE_PLIC_STARTING,
"irqchip/sifive/plic:starting",
plic_starting_cpu, plic_dying_cpu);
+ register_syscore_ops(&plic_irq_syscore_ops);
plic_cpuhp_setup_done = true;
}
- register_syscore_ops(&plic_irq_syscore_ops);
pr_info("%pOFP: mapped %d interrupts with %d handlers for"
" %d contexts.\n", node, nr_irqs, nr_handlers, nr_contexts);
On Fri, Oct 27, 2023 at 1:29 PM Thomas Gleixner <[email protected]> wrote:
>
> On Wed, Oct 25 2023 at 19:58, Anup Patel wrote:
> > The riscv_of_processor_hartid() used by riscv_of_parent_hartid() fails
> > for HARTs disabled in the DT. This results in the following warning
> > thrown by the RISC-V INTC driver for the E-core on SiFive boards:
> >
> > [ 0.000000] riscv-intc: unable to find hart id for /cpus/cpu@0/interrupt-controller
> >
> > The riscv_of_parent_hartid() is only expected to read the hartid from
> > the DT so we should directly call of_get_cpu_hwid() instead of calling
>
> We should? Or maybe not?
>
> Please write precise changelogs and use imperative wording as documented
> in Documentation/process.
Sure, I will update the wording in the commit description.
Thanks,
Anup
Hi Thomas,
On Fri, Oct 27, 2023 at 1:27 PM Thomas Gleixner <[email protected]> wrote:
>
> On Wed, Oct 25 2023 at 19:58, Anup Patel wrote:
> > This allows fw_devlink to create device links between consumers of
> > a MSI and the supplier of the MSI.
>
> How is this related to the two fixes in this series?
The first three patches of the v11 RISC-V AIA series are all
fixes hence I sent them separately for the 6.7 merge window.
(https://lore.kernel.org/lkml/[email protected]/)
All three fixes are unrelated to each other and were discovered
during AIA driver development.
This patch fixes the probing order for platform devices having
inter-dependency based on "msi-parent" DT property.
For example, the AIA APLIC (wired-to-MSI bridge) platform
device depends on the AIA IMSIC (MSI controller) platform
device.
Are you okay with this patch going through the RISC-V tree ?
Regards,
Anup
On Fri, Oct 27 2023 at 21:01, Anup Patel wrote:
> On Fri, Oct 27, 2023 at 1:27 PM Thomas Gleixner <[email protected]> wrote:
>>
>> On Wed, Oct 25 2023 at 19:58, Anup Patel wrote:
>> > This allows fw_devlink to create device links between consumers of
>> > a MSI and the supplier of the MSI.
>>
>> How is this related to the two fixes in this series?
>
> The first three patches of the v11 RISC-V AIA series are all
> fixes hence I sent them separately for the 6.7 merge window.
> (https://lore.kernel.org/lkml/[email protected]/)
>
> All three fixes are unrelated to each other and were discovered
> during AIA driver development.
>
> This patch fixes the probing order for platform devices having
> inter-dependency based on "msi-parent" DT property.
> For example, the AIA APLIC (wired-to-MSI bridge) platform
> device depends on the AIA IMSIC (MSI controller) platform
> device.
Well, the changelog should tell that it is a fix and not make the
illusion that this is pure enablement....
> Are you okay with this patch going through the RISC-V tree ?
Fine with me.