Skip to contentSkip to conversation

misc: pci_endpoint_test: Defer IRQ allocation until ioctl(PCITEST_SET_IRQTYPE)

PATCH RESEND Applied Niklas Cassel 3 participants 6 messages Apr 16, 2025 – Apr 23, 2025

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

PATCH RESEND 1/1 misc: pci_endpoint_test: Defer IRQ allocation until ioctl(PCITEST_SET_IRQTYPE)

Commit a402006d48a9 ("misc: pci_endpoint_test: Remove global 'irq_type' and 'no_msi'") changed so that the default IRQ vector requested by pci_endpoint_test_probe() was no longer the module param 'irq_type', but instead test->irq_type. test->irq_type is by default IRQ_TYPE_UNDEFINED (until someone calls ioctl(PCITEST_SET_IRQTYPE)).

However, the commit also changed so that after initializing test->irq_type to IRQ_TYPE_UNDEFINED, it also overrides it with driver_data->irq_type, if the PCI device and vendor ID provides driver_data.

This causes a regression for PCI device and vendor IDs that do not provide driver_data, and the driver now fails to probe on such platforms.

Considering that the pci endpoint selftests and the old pcitest.sh always call ioctl(PCITEST_SET_IRQTYPE) before performing any test that requires IRQs, simply remove the allocation of IRQs in pci_endpoint_test_probe(), and defer it until ioctl(PCITEST_SET_IRQTYPE) has been called.

A positive side effect of this is that even if the endpoint controller has issues with IRQs, the user can do still do all the tests/ioctls() that do not require working IRQs, e.g. PCITEST_BAR and PCITEST_BARS.

This also means that we can remove the now unused irq_type from driver_data. The irq_type will always be the one configured by the user using ioctl(PCITEST_SET_IRQTYPE). (A user that does not know, or care which irq_type that is used, can use PCITEST_IRQ_TYPE_AUTO. This has superseded the need for a default irq_type in driver_data.)

Fixes a402006d48a9c ("misc: pci_endpoint_test: Remove global 'irq_type' and 'no_msi'")
Reviewed-and-tested-by Frank Li <[email protected]>
---
 drivers/misc/pci_endpoint_test.c | 21 +--------------------
 1 file changed, 1 insertion(+), 20 deletions(-)
drivers/misc/pci_endpoint_test.c +1 -20
index d294850a35a1..c4e5e2c977be 100644
@@ -122,7 +122,6 @@ struct pci_endpoint_test {
 struct pci_endpoint_test_data {
 	enum pci_barno test_reg_bar;
 	size_t alignment;
-	int irq_type;
 };
 
 static inline u32 pci_endpoint_test_readl(struct pci_endpoint_test *test,
@@ -948,7 +947,6 @@ static int pci_endpoint_test_probe(struct pci_dev *pdev,
 		test_reg_bar = data->test_reg_bar;
 		test->test_reg_bar = test_reg_bar;
 		test->alignment = data->alignment;
-		test->irq_type = data->irq_type;
 	}
 
 	init_completion(&test->irq_raised);
@@ -970,10 +968,6 @@ static int pci_endpoint_test_probe(struct pci_dev *pdev,
 
 	pci_set_master(pdev);
 
-	ret = pci_endpoint_test_alloc_irq_vectors(test, test->irq_type);
-	if (ret)
-		goto err_disable_irq;
-
 	for (bar = 0; bar < PCI_STD_NUM_BARS; bar++) {
 		if (pci_resource_flags(pdev, bar) & IORESOURCE_MEM) {
 			base = pci_ioremap_bar(pdev, bar);
@@ -1009,10 +1003,6 @@ static int pci_endpoint_test_probe(struct pci_dev *pdev,
 		goto err_ida_remove;
 	}
 
-	ret = pci_endpoint_test_request_irq(test);
-	if (ret)
-		goto err_kfree_test_name;
-
 	pci_endpoint_test_get_capabilities(test);
 
 	misc_device = &test->miscdev;
@@ -1020,7 +1010,7 @@ static int pci_endpoint_test_probe(struct pci_dev *pdev,
 	misc_device->name = kstrdup(name, GFP_KERNEL);
 	if (!misc_device->name) {
 		ret = -ENOMEM;
-		goto err_release_irq;
+		goto err_kfree_test_name;
 	}
 	misc_device->parent = &pdev->dev;
 	misc_device->fops = &pci_endpoint_test_fops;
@@ -1036,9 +1026,6 @@ static int pci_endpoint_test_probe(struct pci_dev *pdev,
 err_kfree_name:
 	kfree(misc_device->name);
 
-err_release_irq:
-	pci_endpoint_test_release_irq(test);
-
 err_kfree_test_name:
 	kfree(test->name);
 
@@ -1051,8 +1038,6 @@ static int pci_endpoint_test_probe(struct pci_dev *pdev,
 			pci_iounmap(pdev, test->bar[bar]);
 	}
 
-err_disable_irq:
-	pci_endpoint_test_free_irq_vectors(test);
 	pci_release_regions(pdev);
 
 err_disable_pdev:
@@ -1092,23 +1077,19 @@ static void pci_endpoint_test_remove(struct pci_dev *pdev)
 static const struct pci_endpoint_test_data default_data = {
 	.test_reg_bar = BAR_0,
 	.alignment = SZ_4K,
-	.irq_type = PCITEST_IRQ_TYPE_MSI,
 };
 
 static const struct pci_endpoint_test_data am654_data = {
 	.test_reg_bar = BAR_2,
 	.alignment = SZ_64K,
-	.irq_type = PCITEST_IRQ_TYPE_MSI,
 };
 
 static const struct pci_endpoint_test_data j721e_data = {
 	.alignment = 256,
-	.irq_type = PCITEST_IRQ_TYPE_MSI,
 };
 
 static const struct pci_endpoint_test_data rk3588_data = {
 	.alignment = SZ_64K,
-	.irq_type = PCITEST_IRQ_TYPE_MSI,
 };
 
 /*
Signature · 1 line
2.49.0
#1 lore raw
Quoted 11 lines · Niklas Cassel

Commit a402006d48a9 ("misc: pci_endpoint_test: Remove global 'irq_type' and 'no_msi'") changed so that the default IRQ vector requested by pci_endpoint_test_probe() was no longer the module param 'irq_type', but instead test->irq_type. test->irq_type is by default IRQ_TYPE_UNDEFINED (until someone calls ioctl(PCITEST_SET_IRQTYPE)).

However, the commit also changed so that after initializing test->irq_type to IRQ_TYPE_UNDEFINED, it also overrides it with driver_data->irq_type, if the PCI device and vendor ID provides driver_data.

[...]

Applied, thanks!

[1/1] misc: pci_endpoint_test: Defer IRQ allocation until ioctl(PCITEST_SET_IRQTYPE)

      commit: 9d564bf7ab67ec326ec047d2d95087d8d888f9b1

Best regards,

Signature · 1 line
Manivannan Sadhasivam <[email protected]>
#2 lore raw
Quoted 11 lines · Manivannan Sadhasivam
Quoted 10 lines · Niklas Cassel

Commit a402006d48a9 ("misc: pci_endpoint_test: Remove global 'irq_type' and 'no_msi'") changed so that the default IRQ vector requested by pci_endpoint_test_probe() was no longer the module param 'irq_type', but instead test->irq_type. test->irq_type is by default IRQ_TYPE_UNDEFINED (until someone calls ioctl(PCITEST_SET_IRQTYPE)).

However, the commit also changed so that after initializing test->irq_type to IRQ_TYPE_UNDEFINED, it also overrides it with driver_data->irq_type, if the PCI device and vendor ID provides driver_data.

Quoted 6 lines
Quoted 1 line

[...]

Applied, thanks!

[1/1] misc: pci_endpoint_test: Defer IRQ allocation until ioctl(PCITEST_SET_IRQTYPE)

      commit: 9d564bf7ab67ec326ec047d2d95087d8d888f9b1

a402006d48a9c ("misc: pci_endpoint_test: Remove global 'irq_type' and 'no_msi'") appeared in v6.15-rc1, and apparently caused a regression that some driver fails to probe on such platforms.

Since a402006d48a9c caused a regression and hasn't appeared in a release yet, this sounds like a candidate for v6.15 via pci/for-linus?

I can move it if you agree. I *would* like to have a little more detail about the regression, e.g., the affected driver, so I can justify merging this after the merge window.

Bjorn

#3 lore raw
Quoted 19 lines · Bjorn Helgaas
Quoted 11 lines · Manivannan Sadhasivam
Quoted 10 lines · Niklas Cassel

Commit a402006d48a9 ("misc: pci_endpoint_test: Remove global 'irq_type' and 'no_msi'") changed so that the default IRQ vector requested by pci_endpoint_test_probe() was no longer the module param 'irq_type', but instead test->irq_type. test->irq_type is by default IRQ_TYPE_UNDEFINED (until someone calls ioctl(PCITEST_SET_IRQTYPE)).

However, the commit also changed so that after initializing test->irq_type to IRQ_TYPE_UNDEFINED, it also overrides it with driver_data->irq_type, if the PCI device and vendor ID provides driver_data.

Quoted 6 lines
Quoted 1 line

[...]

Applied, thanks!

[1/1] misc: pci_endpoint_test: Defer IRQ allocation until ioctl(PCITEST_SET_IRQTYPE)

      commit: 9d564bf7ab67ec326ec047d2d95087d8d888f9b1
Quoted 6 lines

a402006d48a9c ("misc: pci_endpoint_test: Remove global 'irq_type' and 'no_msi'") appeared in v6.15-rc1, and apparently caused a regression that some driver fails to probe on such platforms.

Since a402006d48a9c caused a regression and hasn't appeared in a release yet, this sounds like a candidate for v6.15 via pci/for-linus?

I agree:
https://lore.kernel.org/linux-pci/Z_0mUhHzGqNrMBGg@ryzen/

Quoted 4 lines

I can move it if you agree. I *would* like to have a little more detail about the regression, e.g., the affected driver, so I can justify merging this after the merge window.

All drivers without any driver_data defined, i.e.:

        { PCI_DEVICE(PCI_VENDOR_ID_FREESCALE, PCI_DEVICE_ID_IMX8),},
        { PCI_DEVICE(PCI_VENDOR_ID_RENESAS, PCI_DEVICE_ID_RENESAS_R8A774A1),},
        { PCI_DEVICE(PCI_VENDOR_ID_RENESAS, PCI_DEVICE_ID_RENESAS_R8A774B1),},
        { PCI_DEVICE(PCI_VENDOR_ID_RENESAS, PCI_DEVICE_ID_RENESAS_R8A774C0),},
        { PCI_DEVICE(PCI_VENDOR_ID_RENESAS, PCI_DEVICE_ID_RENESAS_R8A774E1),},

For those platforms, without this patch, you will get:
pci-endpoint-test 0001:01:00.0: Invalid IRQ type selected pci-endpoint-test 0001:01:00.0: probe with driver pci-endpoint-test failed with error -22

Kind regards,
Niklas

#4 lore raw
Quoted 42 lines · Niklas Cassel
Quoted 19 lines · Bjorn Helgaas
Quoted 11 lines · Manivannan Sadhasivam
Quoted 10 lines · Niklas Cassel

Commit a402006d48a9 ("misc: pci_endpoint_test: Remove global 'irq_type' and 'no_msi'") changed so that the default IRQ vector requested by pci_endpoint_test_probe() was no longer the module param 'irq_type', but instead test->irq_type. test->irq_type is by default IRQ_TYPE_UNDEFINED (until someone calls ioctl(PCITEST_SET_IRQTYPE)).

However, the commit also changed so that after initializing test->irq_type to IRQ_TYPE_UNDEFINED, it also overrides it with driver_data->irq_type, if the PCI device and vendor ID provides driver_data.

Quoted 6 lines
Quoted 1 line

[...]

Applied, thanks!

[1/1] misc: pci_endpoint_test: Defer IRQ allocation until ioctl(PCITEST_SET_IRQTYPE)

      commit: 9d564bf7ab67ec326ec047d2d95087d8d888f9b1
Quoted 6 lines

a402006d48a9c ("misc: pci_endpoint_test: Remove global 'irq_type' and 'no_msi'") appeared in v6.15-rc1, and apparently caused a regression that some driver fails to probe on such platforms.

Since a402006d48a9c caused a regression and hasn't appeared in a release yet, this sounds like a candidate for v6.15 via pci/for-linus?

I agree:
https://lore.kernel.org/linux-pci/Z_0mUhHzGqNrMBGg@ryzen/

Quoted 3 lines

I can move it if you agree. I *would* like to have a little more detail about the regression, e.g., the affected driver, so I can justify merging this after the merge window.

All drivers without any driver_data defined, i.e.:

        { PCI_DEVICE(PCI_VENDOR_ID_FREESCALE, PCI_DEVICE_ID_IMX8),},
        { PCI_DEVICE(PCI_VENDOR_ID_RENESAS, PCI_DEVICE_ID_RENESAS_R8A774A1),},
        { PCI_DEVICE(PCI_VENDOR_ID_RENESAS, PCI_DEVICE_ID_RENESAS_R8A774B1),},
        { PCI_DEVICE(PCI_VENDOR_ID_RENESAS, PCI_DEVICE_ID_RENESAS_R8A774C0),},
        { PCI_DEVICE(PCI_VENDOR_ID_RENESAS, PCI_DEVICE_ID_RENESAS_R8A774E1),},
Quoted 3 lines

For those platforms, without this patch, you will get:
pci-endpoint-test 0001:01:00.0: Invalid IRQ type selected pci-endpoint-test 0001:01:00.0: probe with driver pci-endpoint-test failed with error -22

So I guess this means it's "only" the host side pci_endpoint_test_driver that won't probe, and the problem only happens on the host controllers listed above or controllers with programmable Vendor/Device ID that are set to match one of these?

I cherry picked this patch to pci/for-linus for v6.15, thanks!

Mani, this was the only patch on pci/misc-endpoint, so I deleted the branch.

Bjorn

#5 lore raw
Quoted 9 lines · Bjorn Helgaas
Quoted 4 lines

For those platforms, without this patch, you will get:
pci-endpoint-test 0001:01:00.0: Invalid IRQ type selected pci-endpoint-test 0001:01:00.0: probe with driver pci-endpoint-test failed with error -22

So I guess this means it's "only" the host side pci_endpoint_test_driver that won't probe, and the problem only happens on the host controllers listed above or controllers with programmable Vendor/Device ID that are set to match one of these?

Correct!

Kind regards,
Niklas