Skip to contentSkip to conversation

arm pl011 serial: support multi-irq request

PATCH In review Bing Fan 4 participants 9 messages Aug 13, 2021 – Aug 16, 2021 2 messages not in the archive

Subsystems (2)
lore
Share & tools
lore

Bing Fan

author
#0 lore raw

PATCH v6 1/1 arm pl011 serial: support multi-irq request

From: Bing Fan <[email protected]>

In order to make pl011 work better, multiple interrupts are required, such as TXIM, RXIM, RTIM, error interrupt(FE/PE/BE/OE); at the same time, pl011 to GIC does not merge the interrupt lines(each serial-interrupt corresponding to different GIC hardware interrupt), so need to enable and request multiple gic interrupt numbers in the driver.

---
 drivers/tty/serial/amba-pl011.c | 39 +++++++++++++++++++++++++++++++--
 1 file changed, 37 insertions(+), 2 deletions(-)
drivers/tty/serial/amba-pl011.c +37 -2
index e14f3378b8a0..eaac3431459c 100644
@@ -1701,6 +1701,41 @@ static void pl011_write_lcr_h(struct uart_amba_port *uap, unsigned int lcr_h)
 	}
 }
 
+static void pl011_release_multi_irqs(struct uart_amba_port *uap, unsigned int max_cnt)
+{
+	struct amba_device *amba_dev = container_of(uap->port.dev, struct amba_device, dev);
+	int i;
+
+	for (i = 0; i < max_cnt; i++)
+		if (amba_dev->irq[i])
+			free_irq(amba_dev->irq[i], uap);
+}
+
+static int pl011_allocate_multi_irqs(struct uart_amba_port *uap)
+{
+	int ret = 0;
+	int i;
+	unsigned int virq;
+	struct amba_device *amba_dev = container_of(uap->port.dev, struct amba_device, dev);
+
+	pl011_write(uap->im, uap, REG_IMSC);
+
+	for (i = 0; i < AMBA_NR_IRQS; i++) {
+		virq = amba_dev->irq[i];
+		if (virq == 0)
+			break;
+
+		ret = request_irq(virq, pl011_int, IRQF_SHARED, dev_name(&amba_dev->dev), uap);
+		if (ret) {
+			dev_err(uap->port.dev, "request %u interrupt failed\n", virq);
+			pl011_release_multi_irqs(uap, i - 1);
+			break;
+		}
+	}
+
+	return ret;
+}
+
 static int pl011_allocate_irq(struct uart_amba_port *uap)
 {
 	pl011_write(uap->im, uap, REG_IMSC);
@@ -1753,7 +1788,7 @@ static int pl011_startup(struct uart_port *port)
 	if (retval)
 		goto clk_dis;
 
-	retval = pl011_allocate_irq(uap);
+	retval = pl011_allocate_multi_irqs(uap);
 	if (retval)
 		goto clk_dis;
 
@@ -1864,7 +1899,7 @@ static void pl011_shutdown(struct uart_port *port)
 
 	pl011_dma_shutdown(uap);
 
-	free_irq(uap->port.irq, uap);
+	pl011_release_multi_irqs(uap, AMBA_NR_IRQS);
 
 	pl011_disable_uart(uap);
 
Signature · 1 line
2.17.1
maintainer
#1 lore raw
Quoted 46 lines · Bing Fan

From: Bing Fan <[email protected]>

In order to make pl011 work better, multiple interrupts are required, such as TXIM, RXIM, RTIM, error interrupt(FE/PE/BE/OE); at the same time, pl011 to GIC does not merge the interrupt lines(each serial-interrupt corresponding to different GIC hardware interrupt), so need to enable and request multiple gic interrupt numbers in the driver.

---
 drivers/tty/serial/amba-pl011.c | 39 +++++++++++++++++++++++++++++++--
 1 file changed, 37 insertions(+), 2 deletions(-)
drivers/tty/serial/amba-pl011.c +24 -0
index e14f3378b8a0..eaac3431459c 100644
@@ -1701,6 +1701,41 @@ static void pl011_write_lcr_h(struct uart_amba_port *uap, unsigned int lcr_h)
 	}
 }
 
+static void pl011_release_multi_irqs(struct uart_amba_port *uap, unsigned int max_cnt)
+{
+	struct amba_device *amba_dev = container_of(uap->port.dev, struct amba_device, dev);
+	int i;
+
+	for (i = 0; i < max_cnt; i++)
+		if (amba_dev->irq[i])
+			free_irq(amba_dev->irq[i], uap);
+}
+
+static int pl011_allocate_multi_irqs(struct uart_amba_port *uap)
+{
+	int ret = 0;
+	int i;
+	unsigned int virq;
+	struct amba_device *amba_dev = container_of(uap->port.dev, struct amba_device, dev);
+
+	pl011_write(uap->im, uap, REG_IMSC);
+
+	for (i = 0; i < AMBA_NR_IRQS; i++) {
+		virq = amba_dev->irq[i];
+		if (virq == 0)
+			break;
+

drivers/tty/serial/amba-pl011.c static void pl011_write_lcr_h(struct uart_amba_port *uap, unsigned int lcr_h) · line 1737

Quoted 10 lines
drivers/tty/serial/amba-pl011.c +10 -0
+		ret = request_irq(virq, pl011_int, IRQF_SHARED, dev_name(&amba_dev->dev), uap);
+		if (ret) {
+			dev_err(uap->port.dev, "request %u interrupt failed\n", virq);
+			pl011_release_multi_irqs(uap, i - 1);
+			break;
+		}
+	}
+
+	return ret;
+}

This function looks identical to pl011_allocate_irq(), so what is the difference here? Why is this still needed at all? What does it do that is different from pl011_allocate_irq()?

confused,

greg k-h

maintainer
#3 lore raw
In reply to a message not in the archive
Quoted 62 lines · Bing Fan
Quoted 51 lines · Greg KH
Quoted 50 lines · Bing Fan

From: Bing Fan <[email protected]>

In order to make pl011 work better, multiple interrupts are required, such as TXIM, RXIM, RTIM, error interrupt(FE/PE/BE/OE); at the same time, pl011 to GIC does not merge the interrupt lines(each serial-interrupt corresponding to different GIC hardware interrupt), so need to enable and request multiple gic interrupt numbers in the driver.

---
  drivers/tty/serial/amba-pl011.c | 39 +++++++++++++++++++++++++++++++--
  1 file changed, 37 insertions(+), 2 deletions(-)
drivers/tty/serial/amba-pl011.c +29 -0
index e14f3378b8a0..eaac3431459c 100644
@@ -1701,6 +1701,41 @@ static void pl011_write_lcr_h(struct uart_amba_port *uap, unsigned int lcr_h)
  	}
  }
+static void pl011_release_multi_irqs(struct uart_amba_port *uap, unsigned int max_cnt)
+{
+	struct amba_device *amba_dev = container_of(uap->port.dev, struct amba_device, dev);
+	int i;
+
+	for (i = 0; i < max_cnt; i++)
+		if (amba_dev->irq[i])
+			free_irq(amba_dev->irq[i], uap);
+}
+
+static int pl011_allocate_multi_irqs(struct uart_amba_port *uap)
+{
+	int ret = 0;
+	int i;
+	unsigned int virq;
+	struct amba_device *amba_dev = container_of(uap->port.dev, struct amba_device, dev);
+
+	pl011_write(uap->im, uap, REG_IMSC);
+
+	for (i = 0; i < AMBA_NR_IRQS; i++) {
+		virq = amba_dev->irq[i];
+		if (virq == 0)
+			break;
+
+		ret = request_irq(virq, pl011_int, IRQF_SHARED, dev_name(&amba_dev->dev), uap);
+		if (ret) {
+			dev_err(uap->port.dev, "request %u interrupt failed\n", virq);
+			pl011_release_multi_irqs(uap, i - 1);
+			break;
Quoted 8 lines

drivers/tty/serial/amba-pl011.c static void pl011_write_lcr_h(struct uart_amba_port *uap, unsigned int lcr_h) · line 1737

Quoted 5 lines
drivers/tty/serial/amba-pl011.c +5 -0
+		}
+	}
+
+	return ret;
+}

This function looks identical to pl011_allocate_irq(), so what is the difference here? Why is this still needed at all? What does it do that is different from pl011_allocate_irq()?

Quoted 2 lines

The v6-patch is based on master of git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty.git, not tty-next.

Always submit patches based on tty-next if you want them accepted into that tree.

Quoted 10 lines

As below, the pl011_allocate_irq function supports single irq request only, which different from pl011_allocate_multi_irqs.

static int pl011_allocate_irq(struct uart_amba_port *uap) {
    pl011_write(uap->im, uap, REG_IMSC);

    return request_irq(uap->port.irq, pl011_int, IRQF_SHARED, "uart-pl011", uap);
}

Ok, but that does not reflect what is in my tree to be merged for 5.15-rc1. What is wrong with the code in there that is incorrect and needs to be changed?

thanks,

greg k-h

#4 lore raw

Hi Greg,

Quoted 76 lines · Greg KH
Quoted 62 lines · Bing Fan
Quoted 51 lines · Greg KH
Quoted 50 lines · Bing Fan

From: Bing Fan <[email protected]>

In order to make pl011 work better, multiple interrupts are required, such as TXIM, RXIM, RTIM, error interrupt(FE/PE/BE/OE); at the same time, pl011 to GIC does not merge the interrupt lines(each serial-interrupt corresponding to different GIC hardware interrupt), so need to enable and request multiple gic interrupt numbers in the driver.

---
   drivers/tty/serial/amba-pl011.c | 39 +++++++++++++++++++++++++++++++--
   1 file changed, 37 insertions(+), 2 deletions(-)
drivers/tty/serial/amba-pl011.c +29 -0
index e14f3378b8a0..eaac3431459c 100644
@@ -1701,6 +1701,41 @@ static void pl011_write_lcr_h(struct uart_amba_port *uap, unsigned int lcr_h)
   	}
   }
+static void pl011_release_multi_irqs(struct uart_amba_port *uap, unsigned int max_cnt)
+{
+	struct amba_device *amba_dev = container_of(uap->port.dev, struct amba_device, dev);
+	int i;
+
+	for (i = 0; i < max_cnt; i++)
+		if (amba_dev->irq[i])
+			free_irq(amba_dev->irq[i], uap);
+}
+
+static int pl011_allocate_multi_irqs(struct uart_amba_port *uap)
+{
+	int ret = 0;
+	int i;
+	unsigned int virq;
+	struct amba_device *amba_dev = container_of(uap->port.dev, struct amba_device, dev);
+
+	pl011_write(uap->im, uap, REG_IMSC);
+
+	for (i = 0; i < AMBA_NR_IRQS; i++) {
+		virq = amba_dev->irq[i];
+		if (virq == 0)
+			break;
+
+		ret = request_irq(virq, pl011_int, IRQF_SHARED, dev_name(&amba_dev->dev), uap);
+		if (ret) {
+			dev_err(uap->port.dev, "request %u interrupt failed\n", virq);
+			pl011_release_multi_irqs(uap, i - 1);
+			break;
Quoted 8 lines

drivers/tty/serial/amba-pl011.c static void pl011_write_lcr_h(struct uart_amba_port *uap, unsigned int lcr_h) · line 1737

Quoted 5 lines
drivers/tty/serial/amba-pl011.c +5 -0
+		}
+	}
+
+	return ret;
+}

This function looks identical to pl011_allocate_irq(), so what is the difference here? Why is this still needed at all? What does it do that is different from pl011_allocate_irq()?

Quoted 2 lines

The v6-patch is based on master of git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty.git, not tty-next.

Always submit patches based on tty-next if you want them accepted into that tree.

Quoted 7 lines

As below, the pl011_allocate_irq function supports single irq request only, which different from pl011_allocate_multi_irqs.

static int pl011_allocate_irq(struct uart_amba_port *uap) {
    pl011_write(uap->im, uap, REG_IMSC);

Quoted 7 lines
Quoted 3 lines

    return request_irq(uap->port.irq, pl011_int, IRQF_SHARED, "uart-pl011",
uap);
}

Ok, but that does not reflect what is in my tree to be merged for 5.15-rc1. What is wrong with the code in there that is incorrect and needs to be changed?

As reported by Qian Cai, it blows up on ACPI-based systems by assuming port.dev is an amba_device when in fact in that situation it's a platform_device. If you're able to drop the current patch from your tree that would probably be the best thing to do for the moment.

Cheers,
Robin.

#7 lore raw
In reply to Bing Fan

[ +Russell as the listed PL011 maintainer ]

Quoted 20 lines · Bing Fan

From: Bing Fan <[email protected]>

In order to make pl011 work better, multiple interrupts are required, such as TXIM, RXIM, RTIM, error interrupt(FE/PE/BE/OE); at the same time, pl011 to GIC does not merge the interrupt lines(each serial-interrupt corresponding to different GIC hardware interrupt), so need to enable and request multiple gic interrupt numbers in the driver.

---
  drivers/tty/serial/amba-pl011.c | 39 +++++++++++++++++++++++++++++++--
  1 file changed, 37 insertions(+), 2 deletions(-)
drivers/tty/serial/amba-pl011.c +0 -0
index e14f3378b8a0..eaac3431459c 100644
@@ -1701,6 +1701,41 @@ static void pl011_write_lcr_h(struct uart_amba_port *uap, unsigned int lcr_h)
  	}

drivers/tty/serial/amba-pl011.c static void pl011_write_lcr_h(struct uart_amba_port *uap, unsigned int lcr_h) · line 1711

Quoted 10 lines
drivers/tty/serial/amba-pl011.c +8 -0
  }
 
+static void pl011_release_multi_irqs(struct uart_amba_port *uap, unsigned int max_cnt)
+{
+	struct amba_device *amba_dev = container_of(uap->port.dev, struct amba_device, dev);
+	int i;
+
+	for (i = 0; i < max_cnt; i++)
+		if (amba_dev->irq[i])
+			free_irq(amba_dev->irq[i], uap);

When you request the IRQs you break at the first zero, so this could potentially try to free IRQs that you haven't requested, if there happen to be any nonzero values beyond that. Maybe that can never happen, but there seems little need for deliberate inconsistency here.

drivers/tty/serial/amba-pl011.c static void pl011_write_lcr_h(struct uart_amba_port *uap, unsigned int lcr_h) · line 1723

Quoted 12 lines
drivers/tty/serial/amba-pl011.c +12 -0
+}
+
+static int pl011_allocate_multi_irqs(struct uart_amba_port *uap)
+{
+	int ret = 0;
+	int i;
+	unsigned int virq;
+	struct amba_device *amba_dev = container_of(uap->port.dev, struct amba_device, dev);
+
+	pl011_write(uap->im, uap, REG_IMSC);
+
+	for (i = 0; i < AMBA_NR_IRQS; i++) {

It's not clear where these extra IRQs are expected to come from given that the DT binding explicitly defines only one :/

drivers/tty/serial/amba-pl011.c static void pl011_write_lcr_h(struct uart_amba_port *uap, unsigned int lcr_h) · line 1728

Quoted 5 lines
drivers/tty/serial/amba-pl011.c +5 -0
+		virq = amba_dev->irq[i];
+		if (virq == 0)
+			break;
+
+		ret = request_irq(virq, pl011_int, IRQF_SHARED, dev_name(&amba_dev->dev), uap);

Note that using dev_name() here technically breaks user ABI - scripts looking in /proc for an irq named "uart-pl011" will no longer find it.

Furthermore, the "dev" cookie passed to request_irq is supposed to be globally unique, which "uap" isn't once you start registering it multiple times. If firmware did describe all the individual PL011 IRQ outputs on a system where they are muxed to the same physical IRQ anyway, you'd end up registering ambiguous IRQ actions here. Of course in practice you might still get away with that, but it is technically wrong.

Robin.

Quoted 32 lines
drivers/tty/serial/amba-pl011.c +12 -2
+		if (ret) {
+			dev_err(uap->port.dev, "request %u interrupt failed\n", virq);
+			pl011_release_multi_irqs(uap, i - 1);
+			break;
+		}
+	}
+
+	return ret;
+}
+
  static int pl011_allocate_irq(struct uart_amba_port *uap)
  {
  	pl011_write(uap->im, uap, REG_IMSC);
@@ -1753,7 +1788,7 @@ static int pl011_startup(struct uart_port *port)
  	if (retval)
  		goto clk_dis;
 
-	retval = pl011_allocate_irq(uap);
+	retval = pl011_allocate_multi_irqs(uap);
  	if (retval)
  		goto clk_dis;
 
@@ -1864,7 +1899,7 @@ static void pl011_shutdown(struct uart_port *port)
 
  	pl011_dma_shutdown(uap);
 
-	free_irq(uap->port.irq, uap);
+	pl011_release_multi_irqs(uap, AMBA_NR_IRQS);
 
  	pl011_disable_uart(uap);
 
maintainer
#5 lore raw
In reply to Robin Murphy
Quoted 83 lines · Robin Murphy

Hi Greg,

Quoted 80 lines · Greg KH
Quoted 62 lines · Bing Fan
Quoted 51 lines · Greg KH
Quoted 50 lines · Bing Fan

From: Bing Fan <[email protected]>

In order to make pl011 work better, multiple interrupts are required, such as TXIM, RXIM, RTIM, error interrupt(FE/PE/BE/OE); at the same time, pl011 to GIC does not merge the interrupt lines(each serial-interrupt corresponding to different GIC hardware interrupt), so need to enable and request multiple gic interrupt numbers in the driver.

---
   drivers/tty/serial/amba-pl011.c | 39 +++++++++++++++++++++++++++++++--
   1 file changed, 37 insertions(+), 2 deletions(-)
drivers/tty/serial/amba-pl011.c +29 -0
index e14f3378b8a0..eaac3431459c 100644
@@ -1701,6 +1701,41 @@ static void pl011_write_lcr_h(struct uart_amba_port *uap, unsigned int lcr_h)
   	}
   }
+static void pl011_release_multi_irqs(struct uart_amba_port *uap, unsigned int max_cnt)
+{
+	struct amba_device *amba_dev = container_of(uap->port.dev, struct amba_device, dev);
+	int i;
+
+	for (i = 0; i < max_cnt; i++)
+		if (amba_dev->irq[i])
+			free_irq(amba_dev->irq[i], uap);
+}
+
+static int pl011_allocate_multi_irqs(struct uart_amba_port *uap)
+{
+	int ret = 0;
+	int i;
+	unsigned int virq;
+	struct amba_device *amba_dev = container_of(uap->port.dev, struct amba_device, dev);
+
+	pl011_write(uap->im, uap, REG_IMSC);
+
+	for (i = 0; i < AMBA_NR_IRQS; i++) {
+		virq = amba_dev->irq[i];
+		if (virq == 0)
+			break;
+
+		ret = request_irq(virq, pl011_int, IRQF_SHARED, dev_name(&amba_dev->dev), uap);
+		if (ret) {
+			dev_err(uap->port.dev, "request %u interrupt failed\n", virq);
+			pl011_release_multi_irqs(uap, i - 1);
+			break;
Quoted 8 lines

drivers/tty/serial/amba-pl011.c static void pl011_write_lcr_h(struct uart_amba_port *uap, unsigned int lcr_h) · line 1737

Quoted 5 lines
drivers/tty/serial/amba-pl011.c +5 -0
+		}
+	}
+
+	return ret;
+}

This function looks identical to pl011_allocate_irq(), so what is the difference here? Why is this still needed at all? What does it do that is different from pl011_allocate_irq()?

Quoted 2 lines

The v6-patch is based on master of git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty.git, not tty-next.

Always submit patches based on tty-next if you want them accepted into that tree.

Quoted 10 lines

As below, the pl011_allocate_irq function supports single irq request only, which different from pl011_allocate_multi_irqs.

static int pl011_allocate_irq(struct uart_amba_port *uap) {
    pl011_write(uap->im, uap, REG_IMSC);

    return request_irq(uap->port.irq, pl011_int, IRQF_SHARED, "uart-pl011",
uap);
}

Quoted 8 lines
Quoted 3 lines

Ok, but that does not reflect what is in my tree to be merged for 5.15-rc1. What is wrong with the code in there that is incorrect and needs to be changed?

As reported by Qian Cai, it blows up on ACPI-based systems by assuming port.dev is an amba_device when in fact in that situation it's a platform_device. If you're able to drop the current patch from your tree that would probably be the best thing to do for the moment.

I have not seen any such bug report.

If something needs to be reverted in linux-next, (i.e. in my tty-next branch), please let me know. Ideally by sending a pathc to do so...

thanks,

greg k-h

#6 lore raw
Quoted 88 lines · Greg KH
Quoted 87 lines · Robin Murphy

Hi Greg,

Quoted 83 lines · Greg KH
Quoted 62 lines · Bing Fan
Quoted 51 lines · Greg KH
Quoted 50 lines · Bing Fan

From: Bing Fan <[email protected]>

In order to make pl011 work better, multiple interrupts are required, such as TXIM, RXIM, RTIM, error interrupt(FE/PE/BE/OE); at the same time, pl011 to GIC does not merge the interrupt lines(each serial-interrupt corresponding to different GIC hardware interrupt), so need to enable and request multiple gic interrupt numbers in the driver.

---
    drivers/tty/serial/amba-pl011.c | 39 +++++++++++++++++++++++++++++++--
    1 file changed, 37 insertions(+), 2 deletions(-)
drivers/tty/serial/amba-pl011.c +29 -0
index e14f3378b8a0..eaac3431459c 100644
@@ -1701,6 +1701,41 @@ static void pl011_write_lcr_h(struct uart_amba_port *uap, unsigned int lcr_h)
    	}
    }
+static void pl011_release_multi_irqs(struct uart_amba_port *uap, unsigned int max_cnt)
+{
+	struct amba_device *amba_dev = container_of(uap->port.dev, struct amba_device, dev);
+	int i;
+
+	for (i = 0; i < max_cnt; i++)
+		if (amba_dev->irq[i])
+			free_irq(amba_dev->irq[i], uap);
+}
+
+static int pl011_allocate_multi_irqs(struct uart_amba_port *uap)
+{
+	int ret = 0;
+	int i;
+	unsigned int virq;
+	struct amba_device *amba_dev = container_of(uap->port.dev, struct amba_device, dev);
+
+	pl011_write(uap->im, uap, REG_IMSC);
+
+	for (i = 0; i < AMBA_NR_IRQS; i++) {
+		virq = amba_dev->irq[i];
+		if (virq == 0)
+			break;
+
+		ret = request_irq(virq, pl011_int, IRQF_SHARED, dev_name(&amba_dev->dev), uap);
+		if (ret) {
+			dev_err(uap->port.dev, "request %u interrupt failed\n", virq);
+			pl011_release_multi_irqs(uap, i - 1);
+			break;
Quoted 8 lines

drivers/tty/serial/amba-pl011.c static void pl011_write_lcr_h(struct uart_amba_port *uap, unsigned int lcr_h) · line 1737

Quoted 5 lines
drivers/tty/serial/amba-pl011.c +5 -0
+		}
+	}
+
+	return ret;
+}

This function looks identical to pl011_allocate_irq(), so what is the difference here? Why is this still needed at all? What does it do that is different from pl011_allocate_irq()?

Quoted 2 lines

The v6-patch is based on master of git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty.git, not tty-next.

Always submit patches based on tty-next if you want them accepted into that tree.

Quoted 10 lines

As below, the pl011_allocate_irq function supports single irq request only, which different from pl011_allocate_multi_irqs.

static int pl011_allocate_irq(struct uart_amba_port *uap) {
    pl011_write(uap->im, uap, REG_IMSC);

    return request_irq(uap->port.irq, pl011_int, IRQF_SHARED, "uart-pl011",
uap);
}

Ok, but that does not reflect what is in my tree to be merged for 5.15-rc1. What is wrong with the code in there that is incorrect and needs to be changed?

Quoted 6 lines
Quoted 4 lines

As reported by Qian Cai, it blows up on ACPI-based systems by assuming port.dev is an amba_device when in fact in that situation it's a platform_device. If you're able to drop the current patch from your tree that would probably be the best thing to do for the moment.

I have not seen any such bug report.

It's the thread on the v5 patch that you've just replied to ;)

Quoted 2 lines

If something needs to be reverted in linux-next, (i.e. in my tty-next branch), please let me know. Ideally by sending a pathc to do so...

Since I'm only really involved here off the back of a drive-by review comment, I'll leave that up to Qian.

Cheers,
Robin.

#9 lore raw
In reply to a message not in the archive
Quoted 12 lines · Bing Fan

At present, i think a focus of our discussion is whether this patch is necessary.

As for the other points you mentioned, I think they can be used as code review comments.

Yes, as you described below, most dts files have only one interrupt, but not all platforms are like this.

The scene I'm encountering now is the latter: the interrupt lines of the uart is connected to the gic separately

so the dts should be define like this:

Quoted 10 lines

               duart1: serial@5E139000 {
                        compatible = "arm,pl011", "arm,primecell"; reg = <0x00 0x5E139000 0x0 0x1000>; interrupts = <GIC_SPI 178 IRQ_TYPE_LEVEL_HIGH>, <GIC_SPI 179 IRQ_TYPE_LEVEL_HIGH>, <GIC_SPI 180 IRQ_TYPE_LEVEL_HIGH>, <GIC_SPI 181 IRQ_TYPE_LEVEL_HIGH>; clocks = <&sysclk>;
                        clock-names = "apb_pclk";
                };

Apologies for being unclear - the point I was implying is that of course you can do that in practice, but if you run that DTS through make dtbs_check it will fail. The binding needs extending to make it valid to specify more than one interrupt, and that's a separate patch and discussion in itself (simply increasing "maxitems" for the "interrupts" property is not enough to be robust).

Robin.

Quoted 100 lines

The current tty-master code cannot meet this scenario, so I submitted this patch.

Quoted 93 lines · Robin Murphy

[ +Russell as the listed PL011 maintainer ]

Quoted 21 lines · Bing Fan

From: Bing Fan <[email protected]>

In order to make pl011 work better, multiple interrupts are required, such as TXIM, RXIM, RTIM, error interrupt(FE/PE/BE/OE); at the same time, pl011 to GIC does not merge the interrupt lines(each serial-interrupt corresponding to different GIC hardware interrupt), so need to enable and request multiple gic interrupt numbers in the driver.

---
  drivers/tty/serial/amba-pl011.c | 39 +++++++++++++++++++++++++++++++--

  1 file changed, 37 insertions(+), 2 deletions(-)
drivers/tty/serial/amba-pl011.c +0 -0
index e14f3378b8a0..eaac3431459c 100644
@@ -1701,6 +1701,41 @@ static void pl011_write_lcr_h(struct 
 uart_amba_port *uap, unsigned int lcr_h)
       }

drivers/tty/serial/amba-pl011.c static void pl011_write_lcr_h(struct · line 1712

Quoted 10 lines
drivers/tty/serial/amba-pl011.c +7 -0
   }
   +static void pl011_release_multi_irqs(struct uart_amba_port *uap, unsigned int max_cnt)
+{
+    struct amba_device *amba_dev = container_of(uap->port.dev, 
 struct amba_device, dev);
+    int i;
+
+    for (i = 0; i < max_cnt; i++)
+        if (amba_dev->irq[i])
+            free_irq(amba_dev->irq[i], uap);

When you request the IRQs you break at the first zero, so this could potentially try to free IRQs that you haven't requested, if there happen to be any nonzero values beyond that. Maybe that can never happen, but there seems little need for deliberate inconsistency here.

drivers/tty/serial/amba-pl011.c static void pl011_write_lcr_h(struct uart_amba_port *uap, unsigned int lcr_h) · line 1723

Quoted 13 lines
drivers/tty/serial/amba-pl011.c +12 -0
+}
+
+static int pl011_allocate_multi_irqs(struct uart_amba_port *uap)
+{
+    int ret = 0;
+    int i;
+    unsigned int virq;
+    struct amba_device *amba_dev = container_of(uap->port.dev, 
 struct amba_device, dev);
+
+    pl011_write(uap->im, uap, REG_IMSC);
+
+    for (i = 0; i < AMBA_NR_IRQS; i++) {

It's not clear where these extra IRQs are expected to come from given that the DT binding explicitly defines only one :/

Quoted 6 lines
drivers/tty/serial/amba-pl011.c +5 -0
+        virq = amba_dev->irq[i];
+        if (virq == 0)
+            break;
+
+        ret = request_irq(virq, pl011_int, IRQF_SHARED, 

dev_name(&amba_dev->dev), uap);

Note that using dev_name() here technically breaks user ABI - scripts looking in /proc for an irq named "uart-pl011" will no longer find it.

Furthermore, the "dev" cookie passed to request_irq is supposed to be globally unique, which "uap" isn't once you start registering it multiple times. If firmware did describe all the individual PL011 IRQ outputs on a system where they are muxed to the same physical IRQ anyway, you'd end up registering ambiguous IRQ actions here. Of course in practice you might still get away with that, but it is technically wrong.

Robin.

Quoted 27 lines

+        if (ret) {
+            dev_err(uap->port.dev, "request %u interrupt failed\n", virq);
+            pl011_release_multi_irqs(uap, i - 1);
+            break;
+        }
+    }

drivers/tty/serial/amba-pl011.c +4 -0
+
+    return ret;
+}
+

  static int pl011_allocate_irq(struct uart_amba_port *uap) {
      pl011_write(uap->im, uap, REG_IMSC);

drivers/tty/serial/amba-pl011.c +1 -0
@@ -1753,7 +1788,7 @@ static int pl011_startup(struct uart_port *port)
       if (retval)
           goto clk_dis;
   -    retval = pl011_allocate_irq(uap);
+    retval = pl011_allocate_multi_irqs(uap);
       if (retval)
           goto clk_dis;
   @@ -1864,7 +1899,7 @@ static void pl011_shutdown(struct uart_port *port)

        pl011_dma_shutdown(uap);
  -    free_irq(uap->port.irq, uap);
+    pl011_release_multi_irqs(uap, AMBA_NR_IRQS);
        pl011_disable_uart(uap);

#10 lore raw
In reply to Robin Murphy
Quoted 5 lines · Robin Murphy
Quoted 5 lines
drivers/tty/serial/amba-pl011.c +5 -0
+static int pl011_allocate_multi_irqs(struct uart_amba_port *uap)
+{
+	int ret = 0;
+	int i;
+	unsigned int virq;
Quoted 8 lines

drivers/tty/serial/amba-pl011.c static void pl011_write_lcr_h(struct uart_amba_port *uap, unsigned int lcr_h) · line 1723

Quoted 5 lines
drivers/tty/serial/amba-pl011.c +5 -0
+	struct amba_device *amba_dev = container_of(uap->port.dev, struct amba_device, dev);
+
+	pl011_write(uap->im, uap, REG_IMSC);
+
+	for (i = 0; i < AMBA_NR_IRQS; i++) {

It's not clear where these extra IRQs are expected to come from given that the DT binding explicitly defines only one :/

The DT binding (and driver) was written assuming that people wouldn't use the individual interrupts - but I guess someone decided it was a good idea to have a bazillion interrupt signals going to your interrupt controller from something as simple as a UART (which is permitted by the PL011 TRM.) It's only taken about 20 years for this to happen, so I think we should think we're lucky this hasn't come up before! :D

Quoted 9 lines

drivers/tty/serial/amba-pl011.c static void pl011_write_lcr_h(struct uart_amba_port *uap, unsigned int lcr_h) · line 1728

Quoted 5 lines
drivers/tty/serial/amba-pl011.c +5 -0
+		virq = amba_dev->irq[i];
+		if (virq == 0)
+			break;
+
+		ret = request_irq(virq, pl011_int, IRQF_SHARED, dev_name(&amba_dev->dev), uap);

Note that using dev_name() here technically breaks user ABI - scripts looking in /proc for an irq named "uart-pl011" will no longer find it.

Quoted 3 lines

Furthermore, the "dev" cookie passed to request_irq is supposed to be globally unique, which "uap" isn't once you start registering it multiple times.

There's no difference there.

First, the "private" used with request_irq() only has to be globally unique for the interrupt number being requested. Secondly, there is no way for two UARTs to share the same "uap" structure, and finally there is a 1:1 model between "uap" and "dev". So, I don't see a problem as far as whether we use "uap" or "dev" here.

Quoted 4 lines

If firmware did describe all the individual PL011 IRQ outputs on a system where they are muxed to the same physical IRQ anyway, you'd end up registering ambiguous IRQ actions here. Of course in practice you might still get away with that, but it is technically wrong.

Yes. This would also make a total nonsense of using multiple interrupt lines.

The whole point of using multiple interrupt lines from the UART is so the interrupt demultiplexing can be handled at the interrupt controller and their priorities can be decided there. If we adopt a software structure where we effectively register our "merged" interrupt handler for all these signals, then there is absolutely no benefit to using multiple interrupt signals, since that will override any priority, and we will still have the extra overhead of decoding which interrupt fired at the UART level.

Signature · 2 lines
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!