Skip to contentSkip to conversation

mmc: sdhci-of-aspeed: test: Fix dependencies when KUNIT=m

PATCH In review David Gow 4 participants 6 messages Jul 15, 2022 – Aug 3, 2022

Subsystems (2)
lore
Share & tools
lore

David Gow

author
#0 lore raw

PATCH 1/1 mmc: sdhci-of-aspeed: test: Fix dependencies when KUNIT=m

While the sdhci-of-aspeed KUnit tests do work when builtin, and do work when KUnit itself is being built as a module, the two together break.

This is because the KUnit tests (understandably) depend on KUnit, so a built-in test cannot build if KUnit is a module.

Fix this by adding a dependency on (MMC_SDHCI_OF_ASPEED=m || KUNIT=y), which only excludes this one problematic configuration.

This was reported on a nasty openrisc-randconfig run by the kernel test robot, though for some reason (compiler optimisations removing the test code?) I wasn't able to reproduce it locally on x86:
https://lore.kernel.org/linux-mm/202207140122.fzhlf60k-lkp at intel.com/T/

Fixes 291cd54e5b05 ("mmc: sdhci-of-aspeed: test: Use kunit_test_suite() macro")
---
 drivers/mmc/host/Kconfig | 1 +
 1 file changed, 1 insertion(+)
drivers/mmc/host/Kconfig +1 -0
index 10c563999d3d..e63608834411 100644
@@ -171,6 +171,7 @@ config MMC_SDHCI_OF_ASPEED
 config MMC_SDHCI_OF_ASPEED_TEST
 	bool "Tests for the ASPEED SDHCI driver" if !KUNIT_ALL_TESTS
 	depends on MMC_SDHCI_OF_ASPEED && KUNIT
+	depends on (MMC_SDHCI_OF_ASPEED=m || KUNIT=y)
 	default KUNIT_ALL_TESTS
 	help
 	  Enable KUnit tests for the ASPEED SDHCI driver. Select this
Signature · 1 line
2.37.0.170.g444d1eabd0-goog
#1 lore raw
Quoted 26 lines · David Gow

While the sdhci-of-aspeed KUnit tests do work when builtin, and do work when KUnit itself is being built as a module, the two together break.

This is because the KUnit tests (understandably) depend on KUnit, so a built-in test cannot build if KUnit is a module.

Fix this by adding a dependency on (MMC_SDHCI_OF_ASPEED=m || KUNIT=y), which only excludes this one problematic configuration.

This was reported on a nasty openrisc-randconfig run by the kernel test robot, though for some reason (compiler optimisations removing the test code?) I wasn't able to reproduce it locally on x86:
https://lore.kernel.org/linux-mm/202207140122.fzhlf60k-lkp at intel.com/T/

Fixes 291cd54e5b05 ("mmc: sdhci-of-aspeed: test: Use kunit_test_suite() macro")
---
 drivers/mmc/host/Kconfig | 1 +
 1 file changed, 1 insertion(+)
drivers/mmc/host/Kconfig +0 -0
index 10c563999d3d..e63608834411 100644

drivers/mmc/host/Kconfig config MMC_SDHCI_OF_ASPEED · line 174

Quoted 4 lines
drivers/mmc/host/Kconfig +1 -0
@@ -171,6 +171,7 @@ config MMC_SDHCI_OF_ASPEED
 config MMC_SDHCI_OF_ASPEED_TEST
 	bool "Tests for the ASPEED SDHCI driver" if !KUNIT_ALL_TESTS
 	depends on MMC_SDHCI_OF_ASPEED && KUNIT
+	depends on (MMC_SDHCI_OF_ASPEED=m || KUNIT=y)

Should this replace the line above? Isn't it just more constrained?

Regardless, thanks for your work here, the kunit integration with the ASPEED SDHCI driver bothered me a lot when I wrote it.

Andrew

author
#2 lore raw
Quoted 25 lines · Andrew Jeffery
Quoted 21 lines · David Gow

While the sdhci-of-aspeed KUnit tests do work when builtin, and do work when KUnit itself is being built as a module, the two together break.

This is because the KUnit tests (understandably) depend on KUnit, so a built-in test cannot build if KUnit is a module.

Fix this by adding a dependency on (MMC_SDHCI_OF_ASPEED=m || KUNIT=y), which only excludes this one problematic configuration.

This was reported on a nasty openrisc-randconfig run by the kernel test robot, though for some reason (compiler optimisations removing the test code?) I wasn't able to reproduce it locally on x86:
https://lore.kernel.org/linux-mm/202207140122.fzhlf60k-lkp at intel.com/T/

Fixes 291cd54e5b05 ("mmc: sdhci-of-aspeed: test: Use kunit_test_suite() macro")
---
 drivers/mmc/host/Kconfig | 1 +
 1 file changed, 1 insertion(+)
Quoted 12 lines

drivers/mmc/host/Kconfig config MMC_SDHCI_OF_ASPEED · line 174

Quoted 9 lines
drivers/mmc/host/Kconfig +1 -0
index 10c563999d3d..e63608834411 100644
@@ -171,6 +171,7 @@ config MMC_SDHCI_OF_ASPEED
 config MMC_SDHCI_OF_ASPEED_TEST
      bool "Tests for the ASPEED SDHCI driver" if !KUNIT_ALL_TESTS
      depends on MMC_SDHCI_OF_ASPEED && KUNIT
+     depends on (MMC_SDHCI_OF_ASPEED=m || KUNIT=y)

Should this replace the line above? Isn't it just more constrained?

We need both lines. The first ensures that both KUNIT and MMC_SDHCI_OF_ASPEED are available, and the second just targets the case where KUNIT=m and MMC_SDHCI_OF_ASPEED=y.
If we got rid of the first line, we could end up compiling this without KUnit at all (if MMC_SDHCI_OF_ASPEED=m).

Quoted 2 lines

Regardless, thanks for your work here, the kunit integration with the ASPEED SDHCI driver bothered me a lot when I wrote it.

No worries: we're all still figuring out exactly how these sorts of tests should interact with modules, so it's been a great real-world example for us to experiment with.

Cheers,
-- David
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 4003 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://lists.ozlabs.org/pipermail/linux-aspeed/attachments/20220719/652b7e72/attachment.p7s>

#3 lore raw
Quoted 39 lines · David Gow
Quoted 25 lines · Andrew Jeffery
Quoted 21 lines · David Gow

While the sdhci-of-aspeed KUnit tests do work when builtin, and do work when KUnit itself is being built as a module, the two together break.

This is because the KUnit tests (understandably) depend on KUnit, so a built-in test cannot build if KUnit is a module.

Fix this by adding a dependency on (MMC_SDHCI_OF_ASPEED=m || KUNIT=y), which only excludes this one problematic configuration.

This was reported on a nasty openrisc-randconfig run by the kernel test robot, though for some reason (compiler optimisations removing the test code?) I wasn't able to reproduce it locally on x86:
https://lore.kernel.org/linux-mm/202207140122.fzhlf60k-lkp at intel.com/T/

Fixes 291cd54e5b05 ("mmc: sdhci-of-aspeed: test: Use kunit_test_suite() macro")
---
 drivers/mmc/host/Kconfig | 1 +
 1 file changed, 1 insertion(+)
Quoted 12 lines

drivers/mmc/host/Kconfig config MMC_SDHCI_OF_ASPEED · line 174

Quoted 9 lines
drivers/mmc/host/Kconfig +1 -0
index 10c563999d3d..e63608834411 100644
@@ -171,6 +171,7 @@ config MMC_SDHCI_OF_ASPEED
 config MMC_SDHCI_OF_ASPEED_TEST
      bool "Tests for the ASPEED SDHCI driver" if !KUNIT_ALL_TESTS
      depends on MMC_SDHCI_OF_ASPEED && KUNIT
+     depends on (MMC_SDHCI_OF_ASPEED=m || KUNIT=y)

Should this replace the line above? Isn't it just more constrained?

Quoted 5 lines

We need both lines. The first ensures that both KUNIT and MMC_SDHCI_OF_ASPEED are available, and the second just targets the case where KUNIT=m and MMC_SDHCI_OF_ASPEED=y.
If we got rid of the first line, we could end up compiling this without KUnit at all (if MMC_SDHCI_OF_ASPEED=m).

Ah, yes. Fair enough!

Acked-by on v1 1/1
#4 lore raw
In reply to David Gow
Quoted 10 lines · David Gow

While the sdhci-of-aspeed KUnit tests do work when builtin, and do work when KUnit itself is being built as a module, the two together break.

This is because the KUnit tests (understandably) depend on KUnit, so a built-in test cannot build if KUnit is a module.

Fix this by adding a dependency on (MMC_SDHCI_OF_ASPEED=m || KUNIT=y), which only excludes this one problematic configuration.

Quoted 8 lines

This was reported on a nasty openrisc-randconfig run by the kernel test robot, though for some reason (compiler optimisations removing the test code?) I wasn't able to reproduce it locally on x86:
https://lore.kernel.org/linux-mm/202207140122.fzhlf60k-lkp at intel.com/T/

Fixes 291cd54e5b05 ("mmc: sdhci-of-aspeed: test: Use kunit_test_suite() macro")

I assume this should go together with the other recent kunit patches, so please add:

Kind regards
Uffe

Quoted 19 lines
---
 drivers/mmc/host/Kconfig | 1 +
 1 file changed, 1 insertion(+)
drivers/mmc/host/Kconfig +1 -1
index 10c563999d3d..e63608834411 100644
@@ -171,6 +171,7 @@ config MMC_SDHCI_OF_ASPEED
 config MMC_SDHCI_OF_ASPEED_TEST
        bool "Tests for the ASPEED SDHCI driver" if !KUNIT_ALL_TESTS
        depends on MMC_SDHCI_OF_ASPEED && KUNIT
+       depends on (MMC_SDHCI_OF_ASPEED=m || KUNIT=y)
        default KUNIT_ALL_TESTS
        help
          Enable KUnit tests for the ASPEED SDHCI driver. Select this
--

2.37.0.170.g444d1eabd0-goog

Acked-by on v1 1/1
#5 lore raw
In reply to David Gow
Quoted 10 lines · David Gow

While the sdhci-of-aspeed KUnit tests do work when builtin, and do work when KUnit itself is being built as a module, the two together break.

This is because the KUnit tests (understandably) depend on KUnit, so a built-in test cannot build if KUnit is a module.

Fix this by adding a dependency on (MMC_SDHCI_OF_ASPEED=m || KUNIT=y), which only excludes this one problematic configuration.

Quoted 8 lines

This was reported on a nasty openrisc-randconfig run by the kernel test robot, though for some reason (compiler optimisations removing the test code?) I wasn't able to reproduce it locally on x86:
https://lore.kernel.org/linux-mm/202207140122.fzhlf60k-lkp at intel.com/T/

Fixes 291cd54e5b05 ("mmc: sdhci-of-aspeed: test: Use kunit_test_suite() macro")