Skip to contentSkip to conversation

irqchip: xilinx: Switch to generic domain handler

PATCH In review Michal Simek 2 participants 6 messages Feb 24, 2020 – Mar 9, 2020

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

v2 cover irqchip: xilinx: Switch to generic domain handler

Hi,

this series is based on cascade mode patch sent by Mubin
(https://lkml.org/lkml/2020/2/11/888 - v3 series).

The first patch is just fixing error patch. The second is converting microblaze do_IRQ() to generic IRQ handler with appropriate changes in xilinx intc driver.

Also removes concurrent_irq global variable which wasn't wired anywhere but it stores number of concurrent IRQs handled by one call. There is option to get it back if needed but I haven't seen it in other archs that's why I have removed it too.

Thanks,
Michal

Changes in v2:
- Merge generic irq multi handler(v1 2/3) and domain irq patch (v1 3/3) from together
- Add likely() suggested by Marc

Michal Simek (2):
irqchip: xilinx: Fill error code when irq domain registration fails
irqchip: xilinx: Enable generic irq multi handler

 arch/microblaze/Kconfig           |  2 ++
 arch/microblaze/include/asm/irq.h |  3 ---
 arch/microblaze/kernel/irq.c      | 21 +------------------
 drivers/irqchip/irq-xilinx-intc.c | 35 ++++++++++++++++++-------------
 4 files changed, 24 insertions(+), 37 deletions(-)
Signature · 1 line
2.25.1
#1 lore raw

PATCH v2 1/2 irqchip: xilinx: Fill error code when irq domain registration fails

There is no ret filled in case of irq_domain_add_linear() failure.

---

Changes in v2: None

 drivers/irqchip/irq-xilinx-intc.c | 1 +
 1 file changed, 1 insertion(+)
drivers/irqchip/irq-xilinx-intc.c +1 -0
index 51f461d2934f..cf1bb470d7b5 100644
@@ -230,6 +230,7 @@ static int __init xilinx_intc_of_init(struct device_node *intc,
 						  &xintc_irq_domain_ops, irqc);
 	if (!irqc->root_domain) {
 		pr_err("irq-xilinx: Unable to create IRQ domain\n");
+		ret = -EINVAL;
 		goto error;
 	}
 
Signature · 1 line
2.25.1
#2 lore raw

PATCH v2 2/2 irqchip: xilinx: Enable generic irq multi handler

In reply to Michal Simek

Register default arch handler via driver instead of directly pointing to xilinx intc controller. This patch makes architecture code more generic.

Driver calls generic domain specific irq handler which does the most of things self. Also get rid of concurrent_irq counting which hasn't been exported anywhere.
Based on this loop was also optimized by using do/while loop instead of goto loop.

---

Changes in v2:
- Merge generic irq multi handler(v1 2/3) and domain irq patch (v1 3/3) from together
- Add likely() suggested by Marc

 arch/microblaze/Kconfig           |  2 ++
 arch/microblaze/include/asm/irq.h |  3 ---
 arch/microblaze/kernel/irq.c      | 21 +------------------
 drivers/irqchip/irq-xilinx-intc.c | 34 ++++++++++++++++++-------------
 4 files changed, 23 insertions(+), 37 deletions(-)
arch/microblaze/Kconfig +2 -0
index 6a331bd57ea8..242f58ec086b 100644
@@ -47,6 +47,8 @@ config MICROBLAZE
 	select CPU_NO_EFFICIENT_FFS
 	select MMU_GATHER_NO_RANGE if MMU
 	select SPARSE_IRQ
+	select GENERIC_IRQ_MULTI_HANDLER
+	select HANDLE_DOMAIN_IRQ
 
 # Endianness selection
 choice
arch/microblaze/include/asm/irq.h +0 -3
index eac2fb4b3fb9..5166f0893e2b 100644
@@ -14,7 +14,4 @@
 struct pt_regs;
 extern void do_IRQ(struct pt_regs *regs);
 
-/* should be defined in each interrupt controller driver */
-extern unsigned int xintc_get_irq(void);
-
 #endif /* _ASM_MICROBLAZE_IRQ_H */
arch/microblaze/kernel/irq.c +1 -20
index 903dad822fad..0b37dde60a1e 100644
@@ -20,29 +20,10 @@
 #include <linux/irqchip.h>
 #include <linux/of_irq.h>
 
-static u32 concurrent_irq;
-
 void __irq_entry do_IRQ(struct pt_regs *regs)
 {
-	unsigned int irq;
-	struct pt_regs *old_regs = set_irq_regs(regs);
 	trace_hardirqs_off();
-
-	irq_enter();
-	irq = xintc_get_irq();
-next_irq:
-	BUG_ON(!irq);
-	generic_handle_irq(irq);
-
-	irq = xintc_get_irq();
-	if (irq != -1U) {
-		pr_debug("next irq: %d\n", irq);
-		++concurrent_irq;
-		goto next_irq;
-	}
-
-	irq_exit();
-	set_irq_regs(old_regs);
+	handle_arch_irq(regs);
 	trace_hardirqs_on();
 }
 
drivers/irqchip/irq-xilinx-intc.c +20 -14
index cf1bb470d7b5..2de573ee9764 100644
@@ -125,20 +125,6 @@ static unsigned int xintc_get_irq_local(struct xintc_irq_chip *irqc)
 	return irq;
 }
 
-unsigned int xintc_get_irq(void)
-{
-	u32 hwirq;
-	unsigned int irq = -1;
-
-	hwirq = xintc_read(primary_intc, IVR);
-	if (hwirq != -1U)
-		irq = irq_find_mapping(primary_intc->root_domain, hwirq);
-
-	pr_debug("irq-xilinx: hwirq=%d, irq=%d\n", hwirq, irq);
-
-	return irq;
-}
-
 static int xintc_map(struct irq_domain *d, unsigned int irq, irq_hw_number_t hw)
 {
 	struct xintc_irq_chip *irqc = d->host_data;
@@ -178,6 +164,25 @@ static void xil_intc_irq_handler(struct irq_desc *desc)
 	chained_irq_exit(chip, desc);
 }
 
+static void xil_intc_handle_irq(struct pt_regs *regs)
+{
+	u32 hwirq;
+	struct xintc_irq_chip *irqc = primary_intc;
+
+	do {
+		hwirq = xintc_read(irqc, IVR);
+		if (likely(hwirq != -1U)) {
+			int ret;
+
+			ret = handle_domain_irq(irqc->root_domain, hwirq, regs);
+			WARN_ONCE(ret, "Unhandled HWIRQ %d\n", hwirq);
+			continue;
+		}
+
+		break;
+	} while (1);
+}
+
 static int __init xilinx_intc_of_init(struct device_node *intc,
 					     struct device_node *parent)
 {
@@ -248,6 +253,7 @@ static int __init xilinx_intc_of_init(struct device_node *intc,
 	} else {
 		primary_intc = irqc;
 		irq_set_default_host(primary_intc->root_domain);
+		set_handle_irq(xil_intc_handle_irq);
 	}
 
 	return 0;
Signature · 1 line
2.25.1
#3 lore raw
In reply to Michal Simek
Quoted 5 lines · On Mon, 24 Feb 2020 13:05:12 +0100 Michal Simek <[email protected]> wrote:

Hi,

this series is based on cascade mode patch sent by Mubin
(https://lkml.org/lkml/2020/2/11/888 - v3 series).

Quoted 7 lines

The first patch is just fixing error patch. The second is converting microblaze do_IRQ() to generic IRQ handler with appropriate changes in xilinx intc driver.

Also removes concurrent_irq global variable which wasn't wired anywhere but it stores number of concurrent IRQs handled by one call. There is option to get it back if needed but I haven't seen it in other archs that's why I have removed it too.

Queued for 5.7.

Thanks,

	M.
Signature · 1 line
Jazz is not dead. It just smells funny...
#4 lore raw
Quoted 11 lines · On Sun, 8 Mar 2020 14:01:26 +0000 Marc Zyngier <[email protected]> wrote:
Quoted 8 lines · On Mon, 24 Feb 2020 13:05:12 +0100 Michal Simek <[email protected]> wrote:

Hi,

this series is based on cascade mode patch sent by Mubin
(https://lkml.org/lkml/2020/2/11/888 - v3 series).

The first patch is just fixing error patch. The second is converting microblaze do_IRQ() to generic IRQ handler with appropriate changes in xilinx intc driver.

Quoted 6 lines
Quoted 4 lines

Also removes concurrent_irq global variable which wasn't wired anywhere but it stores number of concurrent IRQs handled by one call. There is option to get it back if needed but I haven't seen it in other archs that's why I have removed it too.

Queued for 5.7.

Scratch that, this doesn't apply to mainline because of the above dependency (and said dependency hasn't been reposted after Thomas' review). I've now dropped it. Please resubmit a full series that contains all the pre-requisite to be applied on mainline.

	M.
Signature · 1 line
Jazz is not dead. It just smells funny...
#5 lore raw
Quoted 19 lines · Marc Zyngier
Quoted 16 lines · On Sun, 8 Mar 2020 14:01:26 +0000 Marc Zyngier <[email protected]> wrote:
Quoted 12 lines · On Mon, 24 Feb 2020 13:05:12 +0100 Michal Simek <[email protected]> wrote:

Hi,

this series is based on cascade mode patch sent by Mubin
(https://lkml.org/lkml/2020/2/11/888 - v3 series).

The first patch is just fixing error patch. The second is converting microblaze do_IRQ() to generic IRQ handler with appropriate changes in xilinx intc driver.

Also removes concurrent_irq global variable which wasn't wired anywhere but it stores number of concurrent IRQs handled by one call. There is option to get it back if needed but I haven't seen it in other archs that's why I have removed it too.

Quoted 6 lines
Quoted 1 line

Queued for 5.7.

Scratch that, this doesn't apply to mainline because of the above dependency (and said dependency hasn't been reposted after Thomas' review). I've now dropped it. Please resubmit a full series that contains all the pre-requisite to be applied on mainline.

I have asked Mubin to send these two patches as series with cascade mode support.

Thanks,
Michal

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel