Skip to contentSkip to conversation

RISC-V: Add CONFIG_PREEMPT=y to defconfig

PATCH Guenter Roeck 3 participants 10 messages Dec 20, 2018 – Jan 8, 2019 RISC-V

lore
Share & tools
lore
#0 lore raw
Quoted 6 lines · Palmer Dabbelt

I don't remember who suggested this, but I was talking to someone at Plumbers about all the horrible bugs we've had lurking around in our port despite it still getting to userspace and they suggested that I enable CONFIG_PREEMPT in our defconfig as it'll help shake out more issues.

Quoted 6 lines

This patch enables CONFIG_PREEMPT in the single RISC-V defconfig. It still passes my tests (ie, booting Fedora on QEMU). Maybe I'm too paranoid, but I still don't trust our port enough to just pull the trigger on this without at least accumulating some "Tested-by"s on it.

I can only give you a

Fails-miserably-for: Guenter Roeck <[email protected]>

Not the new defconfig itself, but the new defconfig plus a variety of selftest options. I don't know what exactly is failing, but here is my list of enabled selftests on top of defconfig:

CONFIG_CRYPTO_MANAGER_DISABLE_TESTS=y
CONFIG_CRC32_SELFTEST=y
CONFIG_DEBUG_LOCKING_API_SELFTESTS=y
CONFIG_DEBUG_NMI_SELFTEST=y
CONFIG_DEBUG_RODATA_TEST=y
CONFIG_DEBUG_TLBFLUSH=y
CONFIG_DMATEST=y
CONFIG_GLOB_SELFTEST=y
CONFIG_OF_UNITTEST=y
CONFIG_PCI_EPF_TEST=y
CONFIG_PCI_ENDPOINT_TEST=y
CONFIG_PROVE_LOCKING=y
CONFIG_PROVE_RCU=y
CONFIG_RCU_EQS_DEBUG=y
CONFIG_STATIC_KEYS_SELFTEST=y
CONFIG_STRING_SELFTEST=y
CONFIG_TEST_BITMAP=y
CONFIG_TEST_FIRMWARE=y
CONFIG_TEST_SORT=y
CONFIG_TEST_SYSCTL=y
CONFIG_TEST_UUID=y
CONFIG_USB_TEST=y
CONFIG_USB_EHSET_TEST_FIXTURE=y
CONFIG_USB_LINK_LAYER_TEST=y
CONFIG_WW_MUTEX_SELFTEST=y
CONFIG_TORTURE_TEST=y
CONFIG_LOCK_TORTURE_TEST=y
CONFIG_RCU_TORTURE_TEST=y
CONFIG_RBTREE_TEST=y
CONFIG_INTERVAL_TREE_TEST=y

Tests pass after disabling those tests (or after reverting your patch).

Guenter

Quoted 22 lines
---
 arch/riscv/configs/defconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/riscv/configs/defconfig +1 -1
index ef4f15df9adf..8869471af0fc 100644
@@ -1,5 +1,6 @@
 CONFIG_SYSVIPC=y
 CONFIG_POSIX_MQUEUE=y
+CONFIG_PREEMPT=y
 CONFIG_IKCONFIG=y
 CONFIG_IKCONFIG_PROC=y
 CONFIG_CGROUPS=y
@@ -77,4 +78,3 @@ CONFIG_NFS_V4_2=y
 CONFIG_ROOT_NFS=y
 CONFIG_CRYPTO_USER_API_HASH=y
 CONFIG_PRINTK_TIME=y
-# CONFIG_RCU_TRACE is not set
Signature · 1 line
2.7.4

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

#1 lore raw
Quoted 44 lines · Guenter Roeck
Quoted 6 lines · Palmer Dabbelt

I don't remember who suggested this, but I was talking to someone at Plumbers about all the horrible bugs we've had lurking around in our port despite it still getting to userspace and they suggested that I enable CONFIG_PREEMPT in our defconfig as it'll help shake out more issues.

Quoted 6 lines

This patch enables CONFIG_PREEMPT in the single RISC-V defconfig. It still passes my tests (ie, booting Fedora on QEMU). Maybe I'm too paranoid, but I still don't trust our port enough to just pull the trigger on this without at least accumulating some "Tested-by"s on it.

I can only give you a

Fails-miserably-for: Guenter Roeck <[email protected]>

Not the new defconfig itself, but the new defconfig plus a variety of selftest options. I don't know what exactly is failing, but here is my list of enabled selftests on top of defconfig:

CONFIG_CRYPTO_MANAGER_DISABLE_TESTS=y
CONFIG_CRC32_SELFTEST=y
CONFIG_DEBUG_LOCKING_API_SELFTESTS=y
CONFIG_DEBUG_NMI_SELFTEST=y
CONFIG_DEBUG_RODATA_TEST=y
CONFIG_DEBUG_TLBFLUSH=y
CONFIG_DMATEST=y
CONFIG_GLOB_SELFTEST=y
CONFIG_OF_UNITTEST=y
CONFIG_PCI_EPF_TEST=y
CONFIG_PCI_ENDPOINT_TEST=y
CONFIG_PROVE_LOCKING=y
CONFIG_PROVE_RCU=y
CONFIG_RCU_EQS_DEBUG=y
CONFIG_STATIC_KEYS_SELFTEST=y
CONFIG_STRING_SELFTEST=y
CONFIG_TEST_BITMAP=y
CONFIG_TEST_FIRMWARE=y
CONFIG_TEST_SORT=y
CONFIG_TEST_SYSCTL=y
CONFIG_TEST_UUID=y
CONFIG_USB_TEST=y

Quoted 9 lines

CONFIG_USB_EHSET_TEST_FIXTURE=y
CONFIG_USB_LINK_LAYER_TEST=y
CONFIG_WW_MUTEX_SELFTEST=y
CONFIG_TORTURE_TEST=y
CONFIG_LOCK_TORTURE_TEST=y
CONFIG_RCU_TORTURE_TEST=y
CONFIG_RBTREE_TEST=y
CONFIG_INTERVAL_TREE_TEST=y

Update: defconfig+CONFIG_LOCK_TORTURE_TEST=y causes the hang if CONFIG_PREEMPT=y.

Guenter

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

#2 lore raw
Quoted 55 lines · Guenter Roeck
Quoted 44 lines · Guenter Roeck
Quoted 6 lines · Palmer Dabbelt

I don't remember who suggested this, but I was talking to someone at Plumbers about all the horrible bugs we've had lurking around in our port despite it still getting to userspace and they suggested that I enable CONFIG_PREEMPT in our defconfig as it'll help shake out more issues.

Quoted 6 lines

This patch enables CONFIG_PREEMPT in the single RISC-V defconfig. It still passes my tests (ie, booting Fedora on QEMU). Maybe I'm too paranoid, but I still don't trust our port enough to just pull the trigger on this without at least accumulating some "Tested-by"s on it.

I can only give you a

Fails-miserably-for: Guenter Roeck <[email protected]>

Not the new defconfig itself, but the new defconfig plus a variety of selftest options. I don't know what exactly is failing, but here is my list of enabled selftests on top of defconfig:

CONFIG_CRYPTO_MANAGER_DISABLE_TESTS=y
CONFIG_CRC32_SELFTEST=y
CONFIG_DEBUG_LOCKING_API_SELFTESTS=y
CONFIG_DEBUG_NMI_SELFTEST=y
CONFIG_DEBUG_RODATA_TEST=y
CONFIG_DEBUG_TLBFLUSH=y
CONFIG_DMATEST=y
CONFIG_GLOB_SELFTEST=y
CONFIG_OF_UNITTEST=y
CONFIG_PCI_EPF_TEST=y
CONFIG_PCI_ENDPOINT_TEST=y
CONFIG_PROVE_LOCKING=y
CONFIG_PROVE_RCU=y
CONFIG_RCU_EQS_DEBUG=y
CONFIG_STATIC_KEYS_SELFTEST=y
CONFIG_STRING_SELFTEST=y
CONFIG_TEST_BITMAP=y
CONFIG_TEST_FIRMWARE=y
CONFIG_TEST_SORT=y
CONFIG_TEST_SYSCTL=y
CONFIG_TEST_UUID=y
CONFIG_USB_TEST=y

Quoted 9 lines

CONFIG_USB_EHSET_TEST_FIXTURE=y
CONFIG_USB_LINK_LAYER_TEST=y
CONFIG_WW_MUTEX_SELFTEST=y
CONFIG_TORTURE_TEST=y
CONFIG_LOCK_TORTURE_TEST=y
CONFIG_RCU_TORTURE_TEST=y
CONFIG_RBTREE_TEST=y
CONFIG_INTERVAL_TREE_TEST=y

Quoted 3 lines

Update: defconfig+CONFIG_LOCK_TORTURE_TEST=y causes the hang if CONFIG_PREEMPT=y.

Another update. CONFIG_RCU_TORTURE_TEST=y or CONFIG_WW_MUTEX_SELFTEST=y also result in hangs, only a bit later. I have now disabled all three test options in my riscv boot tests if CONFIG_PREEMPT=y. The lack of error reports in the future does _not_ mean that the problem has magically disappeared. I just find it pointless to keep testing a known broken configuration, even more so since doing so would hide other potential problems.

Guenter

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

#3 lore raw
Quoted 60 lines · On Tue, 25 Dec 2018 19:23:50 PST (-0800), [email protected] wrote:
Quoted 55 lines · Guenter Roeck
Quoted 44 lines · Guenter Roeck
Quoted 6 lines · Palmer Dabbelt

I don't remember who suggested this, but I was talking to someone at Plumbers about all the horrible bugs we've had lurking around in our port despite it still getting to userspace and they suggested that I enable CONFIG_PREEMPT in our defconfig as it'll help shake out more issues.

Quoted 6 lines

This patch enables CONFIG_PREEMPT in the single RISC-V defconfig. It still passes my tests (ie, booting Fedora on QEMU). Maybe I'm too paranoid, but I still don't trust our port enough to just pull the trigger on this without at least accumulating some "Tested-by"s on it.

I can only give you a

Fails-miserably-for: Guenter Roeck <[email protected]>

Not the new defconfig itself, but the new defconfig plus a variety of selftest options. I don't know what exactly is failing, but here is my list of enabled selftests on top of defconfig:

CONFIG_CRYPTO_MANAGER_DISABLE_TESTS=y
CONFIG_CRC32_SELFTEST=y
CONFIG_DEBUG_LOCKING_API_SELFTESTS=y
CONFIG_DEBUG_NMI_SELFTEST=y
CONFIG_DEBUG_RODATA_TEST=y
CONFIG_DEBUG_TLBFLUSH=y
CONFIG_DMATEST=y
CONFIG_GLOB_SELFTEST=y
CONFIG_OF_UNITTEST=y
CONFIG_PCI_EPF_TEST=y
CONFIG_PCI_ENDPOINT_TEST=y
CONFIG_PROVE_LOCKING=y
CONFIG_PROVE_RCU=y
CONFIG_RCU_EQS_DEBUG=y
CONFIG_STATIC_KEYS_SELFTEST=y
CONFIG_STRING_SELFTEST=y
CONFIG_TEST_BITMAP=y
CONFIG_TEST_FIRMWARE=y
CONFIG_TEST_SORT=y
CONFIG_TEST_SYSCTL=y
CONFIG_TEST_UUID=y
CONFIG_USB_TEST=y

Quoted 9 lines

CONFIG_USB_EHSET_TEST_FIXTURE=y
CONFIG_USB_LINK_LAYER_TEST=y
CONFIG_WW_MUTEX_SELFTEST=y
CONFIG_TORTURE_TEST=y
CONFIG_LOCK_TORTURE_TEST=y
CONFIG_RCU_TORTURE_TEST=y
CONFIG_RBTREE_TEST=y
CONFIG_INTERVAL_TREE_TEST=y

Quoted 3 lines

Update: defconfig+CONFIG_LOCK_TORTURE_TEST=y causes the hang if CONFIG_PREEMPT=y.

Quoted 7 lines

Another update. CONFIG_RCU_TORTURE_TEST=y or CONFIG_WW_MUTEX_SELFTEST=y also result in hangs, only a bit later. I have now disabled all three test options in my riscv boot tests if CONFIG_PREEMPT=y. The lack of error reports in the future does _not_ mean that the problem has magically disappeared. I just find it pointless to keep testing a known broken configuration, even more so since doing so would hide other potential problems.

Thanks. I agree it's pointless to test more, as there appear to be lots of bugs to fix already and this will just mask other bugs.

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

#4 lore raw

Palmer Dabbelt <[email protected]> 於 2018年12月27日 週四 上午1:19寫道:

Quoted 70 lines · On Tue, 25 Dec 2018 19:23:50 PST (-0800), [email protected] wrote:
Quoted 60 lines · Guenter Roeck
Quoted 55 lines · Guenter Roeck
Quoted 44 lines · Palmer Dabbelt
Quoted 6 lines

I don't remember who suggested this, but I was talking to someone at Plumbers about all the horrible bugs we've had lurking around in our port despite it still getting to userspace and they suggested that I enable CONFIG_PREEMPT in our defconfig as it'll help shake out more issues.

Quoted 6 lines

This patch enables CONFIG_PREEMPT in the single RISC-V defconfig. It still passes my tests (ie, booting Fedora on QEMU). Maybe I'm too paranoid, but I still don't trust our port enough to just pull the trigger on this without at least accumulating some "Tested-by"s on it.

I can only give you a

Fails-miserably-for: Guenter Roeck <[email protected]>

Not the new defconfig itself, but the new defconfig plus a variety of selftest options. I don't know what exactly is failing, but here is my list of enabled selftests on top of defconfig:

CONFIG_CRYPTO_MANAGER_DISABLE_TESTS=y
CONFIG_CRC32_SELFTEST=y
CONFIG_DEBUG_LOCKING_API_SELFTESTS=y
CONFIG_DEBUG_NMI_SELFTEST=y
CONFIG_DEBUG_RODATA_TEST=y
CONFIG_DEBUG_TLBFLUSH=y
CONFIG_DMATEST=y
CONFIG_GLOB_SELFTEST=y
CONFIG_OF_UNITTEST=y
CONFIG_PCI_EPF_TEST=y
CONFIG_PCI_ENDPOINT_TEST=y
CONFIG_PROVE_LOCKING=y
CONFIG_PROVE_RCU=y
CONFIG_RCU_EQS_DEBUG=y
CONFIG_STATIC_KEYS_SELFTEST=y
CONFIG_STRING_SELFTEST=y
CONFIG_TEST_BITMAP=y
CONFIG_TEST_FIRMWARE=y
CONFIG_TEST_SORT=y
CONFIG_TEST_SYSCTL=y
CONFIG_TEST_UUID=y
CONFIG_USB_TEST=y

Quoted 9 lines

CONFIG_USB_EHSET_TEST_FIXTURE=y
CONFIG_USB_LINK_LAYER_TEST=y
CONFIG_WW_MUTEX_SELFTEST=y
CONFIG_TORTURE_TEST=y
CONFIG_LOCK_TORTURE_TEST=y
CONFIG_RCU_TORTURE_TEST=y
CONFIG_RBTREE_TEST=y
CONFIG_INTERVAL_TREE_TEST=y

Quoted 3 lines

Update: defconfig+CONFIG_LOCK_TORTURE_TEST=y causes the hang if CONFIG_PREEMPT=y.

Quoted 7 lines

Another update. CONFIG_RCU_TORTURE_TEST=y or CONFIG_WW_MUTEX_SELFTEST=y also result in hangs, only a bit later. I have now disabled all three test options in my riscv boot tests if CONFIG_PREEMPT=y. The lack of error reports in the future does _not_ mean that the problem has magically disappeared. I just find it pointless to keep testing a known broken configuration, even more so since doing so would hide other potential problems.

Quoted 3 lines

Thanks. I agree it's pointless to test more, as there appear to be lots of bugs to fix already and this will just mask other bugs.

Last week, I encountered some locking issues in the kernel, so I tried to test the
kernel with some relative selftests. Since I had already set CONFIG_PREEMPT=y, I encountered the same results in RCU_TORTURE_TEST and LOCK_TORTURE_TEST. I tried to debug it and found the hang was due to the lack of _TIF_NEED_RESCHED check for kernel threads in entry.S. I will send a patch to solve this problem and I hope you can give me some comments if you find some
mistakes.

Vincent Chen

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

#5 lore raw
Quoted 75 lines · Vincent Chen

Palmer Dabbelt <[email protected]> 於 2018年12月27日 週四 上午1:19寫道:

Quoted 70 lines · On Tue, 25 Dec 2018 19:23:50 PST (-0800), [email protected] wrote:
Quoted 60 lines · Guenter Roeck
Quoted 55 lines · Guenter Roeck
Quoted 44 lines · Palmer Dabbelt
Quoted 6 lines

I don't remember who suggested this, but I was talking to someone at Plumbers about all the horrible bugs we've had lurking around in our port despite it still getting to userspace and they suggested that I enable CONFIG_PREEMPT in our defconfig as it'll help shake out more issues.

Quoted 6 lines

This patch enables CONFIG_PREEMPT in the single RISC-V defconfig. It still passes my tests (ie, booting Fedora on QEMU). Maybe I'm too paranoid, but I still don't trust our port enough to just pull the trigger on this without at least accumulating some "Tested-by"s on it.

I can only give you a

Fails-miserably-for: Guenter Roeck <[email protected]>

Not the new defconfig itself, but the new defconfig plus a variety of selftest options. I don't know what exactly is failing, but here is my list of enabled selftests on top of defconfig:

CONFIG_CRYPTO_MANAGER_DISABLE_TESTS=y
CONFIG_CRC32_SELFTEST=y
CONFIG_DEBUG_LOCKING_API_SELFTESTS=y
CONFIG_DEBUG_NMI_SELFTEST=y
CONFIG_DEBUG_RODATA_TEST=y
CONFIG_DEBUG_TLBFLUSH=y
CONFIG_DMATEST=y
CONFIG_GLOB_SELFTEST=y
CONFIG_OF_UNITTEST=y
CONFIG_PCI_EPF_TEST=y
CONFIG_PCI_ENDPOINT_TEST=y
CONFIG_PROVE_LOCKING=y
CONFIG_PROVE_RCU=y
CONFIG_RCU_EQS_DEBUG=y
CONFIG_STATIC_KEYS_SELFTEST=y
CONFIG_STRING_SELFTEST=y
CONFIG_TEST_BITMAP=y
CONFIG_TEST_FIRMWARE=y
CONFIG_TEST_SORT=y
CONFIG_TEST_SYSCTL=y
CONFIG_TEST_UUID=y
CONFIG_USB_TEST=y

Quoted 9 lines

CONFIG_USB_EHSET_TEST_FIXTURE=y
CONFIG_USB_LINK_LAYER_TEST=y
CONFIG_WW_MUTEX_SELFTEST=y
CONFIG_TORTURE_TEST=y
CONFIG_LOCK_TORTURE_TEST=y
CONFIG_RCU_TORTURE_TEST=y
CONFIG_RBTREE_TEST=y
CONFIG_INTERVAL_TREE_TEST=y

Quoted 3 lines

Update: defconfig+CONFIG_LOCK_TORTURE_TEST=y causes the hang if CONFIG_PREEMPT=y.

Quoted 7 lines

Another update. CONFIG_RCU_TORTURE_TEST=y or CONFIG_WW_MUTEX_SELFTEST=y also result in hangs, only a bit later. I have now disabled all three test options in my riscv boot tests if CONFIG_PREEMPT=y. The lack of error reports in the future does _not_ mean that the problem has magically disappeared. I just find it pointless to keep testing a known broken configuration, even more so since doing so would hide other potential problems.

Quoted 3 lines

Thanks. I agree it's pointless to test more, as there appear to be lots of bugs to fix already and this will just mask other bugs.

Quoted 10 lines

Last week, I encountered some locking issues in the kernel, so I tried to test the
kernel with some relative selftests. Since I had already set CONFIG_PREEMPT=y, I encountered the same results in RCU_TORTURE_TEST and LOCK_TORTURE_TEST. I tried to debug it and found the hang was due to the lack of _TIF_NEED_RESCHED check for kernel threads in entry.S. I will send a patch to solve this problem and I hope you can give me some comments if you find some
mistakes.

I'll be happy to run a set of tests if you Cc: me on the patch.

Guenter

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

#6 lore raw

Guenter Roeck <[email protected]> 於 2019年1月3日 週四 上午11:51寫道:

Quoted 88 lines · Vincent Chen
Quoted 75 lines

Palmer Dabbelt <[email protected]> 於 2018年12月27日 週四 上午1:19寫道:

Quoted 70 lines · On Tue, 25 Dec 2018 19:23:50 PST (-0800), [email protected] wrote:
Quoted 60 lines · Guenter Roeck
Quoted 55 lines · Guenter Roeck
Quoted 44 lines · Palmer Dabbelt
Quoted 6 lines

I don't remember who suggested this, but I was talking to someone at Plumbers about all the horrible bugs we've had lurking around in our port despite it still getting to userspace and they suggested that I enable CONFIG_PREEMPT in our defconfig as it'll help shake out more issues.

Quoted 6 lines

This patch enables CONFIG_PREEMPT in the single RISC-V defconfig. It still passes my tests (ie, booting Fedora on QEMU). Maybe I'm too paranoid, but I still don't trust our port enough to just pull the trigger on this without at least accumulating some "Tested-by"s on it.

I can only give you a

Fails-miserably-for: Guenter Roeck <[email protected]>

Not the new defconfig itself, but the new defconfig plus a variety of selftest options. I don't know what exactly is failing, but here is my list of enabled selftests on top of defconfig:

CONFIG_CRYPTO_MANAGER_DISABLE_TESTS=y
CONFIG_CRC32_SELFTEST=y
CONFIG_DEBUG_LOCKING_API_SELFTESTS=y
CONFIG_DEBUG_NMI_SELFTEST=y
CONFIG_DEBUG_RODATA_TEST=y
CONFIG_DEBUG_TLBFLUSH=y
CONFIG_DMATEST=y
CONFIG_GLOB_SELFTEST=y
CONFIG_OF_UNITTEST=y
CONFIG_PCI_EPF_TEST=y
CONFIG_PCI_ENDPOINT_TEST=y
CONFIG_PROVE_LOCKING=y
CONFIG_PROVE_RCU=y
CONFIG_RCU_EQS_DEBUG=y
CONFIG_STATIC_KEYS_SELFTEST=y
CONFIG_STRING_SELFTEST=y
CONFIG_TEST_BITMAP=y
CONFIG_TEST_FIRMWARE=y
CONFIG_TEST_SORT=y
CONFIG_TEST_SYSCTL=y
CONFIG_TEST_UUID=y
CONFIG_USB_TEST=y

Quoted 9 lines

CONFIG_USB_EHSET_TEST_FIXTURE=y
CONFIG_USB_LINK_LAYER_TEST=y
CONFIG_WW_MUTEX_SELFTEST=y
CONFIG_TORTURE_TEST=y
CONFIG_LOCK_TORTURE_TEST=y
CONFIG_RCU_TORTURE_TEST=y
CONFIG_RBTREE_TEST=y
CONFIG_INTERVAL_TREE_TEST=y

Quoted 3 lines

Update: defconfig+CONFIG_LOCK_TORTURE_TEST=y causes the hang if CONFIG_PREEMPT=y.

Quoted 7 lines

Another update. CONFIG_RCU_TORTURE_TEST=y or CONFIG_WW_MUTEX_SELFTEST=y also result in hangs, only a bit later. I have now disabled all three test options in my riscv boot tests if CONFIG_PREEMPT=y. The lack of error reports in the future does _not_ mean that the problem has magically disappeared. I just find it pointless to keep testing a known broken configuration, even more so since doing so would hide other potential problems.

Quoted 3 lines

Thanks. I agree it's pointless to test more, as there appear to be lots of bugs to fix already and this will just mask other bugs.

Quoted 10 lines

Last week, I encountered some locking issues in the kernel, so I tried to test the
kernel with some relative selftests. Since I had already set CONFIG_PREEMPT=y, I encountered the same results in RCU_TORTURE_TEST and LOCK_TORTURE_TEST. I tried to debug it and found the hang was due to the lack of _TIF_NEED_RESCHED check for kernel threads in entry.S. I will send a patch to solve this problem and I hope you can give me some comments if you find some
mistakes.

Quoted 3 lines

I'll be happy to run a set of tests if you Cc: me on the patch.

Guenter

Sorry, I have sent this patch but I forget to add you to the Cc list. I will send you this patch again and please help me to run a set of tests. Thanks

Vincent Chen

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

#7 lore raw
Quoted 89 lines · Vincent Chen

Guenter Roeck <[email protected]> 於 2019年1月3日 週四 上午11:51寫道:

Quoted 88 lines · Vincent Chen
Quoted 85 lines

Palmer Dabbelt <[email protected]> 於 2018年12月27日 週四 上午1:19寫道:

Quoted 70 lines · On Tue, 25 Dec 2018 19:23:50 PST (-0800), [email protected] wrote:
Quoted 60 lines · Guenter Roeck
Quoted 55 lines · Guenter Roeck
Quoted 44 lines · Palmer Dabbelt
Quoted 6 lines

I don't remember who suggested this, but I was talking to someone at Plumbers about all the horrible bugs we've had lurking around in our port despite it still getting to userspace and they suggested that I enable CONFIG_PREEMPT in our defconfig as it'll help shake out more issues.

Quoted 6 lines

This patch enables CONFIG_PREEMPT in the single RISC-V defconfig. It still passes my tests (ie, booting Fedora on QEMU). Maybe I'm too paranoid, but I still don't trust our port enough to just pull the trigger on this without at least accumulating some "Tested-by"s on it.

I can only give you a

Fails-miserably-for: Guenter Roeck <[email protected]>

Not the new defconfig itself, but the new defconfig plus a variety of selftest options. I don't know what exactly is failing, but here is my list of enabled selftests on top of defconfig:

CONFIG_CRYPTO_MANAGER_DISABLE_TESTS=y
CONFIG_CRC32_SELFTEST=y
CONFIG_DEBUG_LOCKING_API_SELFTESTS=y
CONFIG_DEBUG_NMI_SELFTEST=y
CONFIG_DEBUG_RODATA_TEST=y
CONFIG_DEBUG_TLBFLUSH=y
CONFIG_DMATEST=y
CONFIG_GLOB_SELFTEST=y
CONFIG_OF_UNITTEST=y
CONFIG_PCI_EPF_TEST=y
CONFIG_PCI_ENDPOINT_TEST=y
CONFIG_PROVE_LOCKING=y
CONFIG_PROVE_RCU=y
CONFIG_RCU_EQS_DEBUG=y
CONFIG_STATIC_KEYS_SELFTEST=y
CONFIG_STRING_SELFTEST=y
CONFIG_TEST_BITMAP=y
CONFIG_TEST_FIRMWARE=y
CONFIG_TEST_SORT=y
CONFIG_TEST_SYSCTL=y
CONFIG_TEST_UUID=y
CONFIG_USB_TEST=y

Quoted 9 lines

CONFIG_USB_EHSET_TEST_FIXTURE=y
CONFIG_USB_LINK_LAYER_TEST=y
CONFIG_WW_MUTEX_SELFTEST=y
CONFIG_TORTURE_TEST=y
CONFIG_LOCK_TORTURE_TEST=y
CONFIG_RCU_TORTURE_TEST=y
CONFIG_RBTREE_TEST=y
CONFIG_INTERVAL_TREE_TEST=y

Quoted 3 lines

Update: defconfig+CONFIG_LOCK_TORTURE_TEST=y causes the hang if CONFIG_PREEMPT=y.

Quoted 7 lines

Another update. CONFIG_RCU_TORTURE_TEST=y or CONFIG_WW_MUTEX_SELFTEST=y also result in hangs, only a bit later. I have now disabled all three test options in my riscv boot tests if CONFIG_PREEMPT=y. The lack of error reports in the future does _not_ mean that the problem has magically disappeared. I just find it pointless to keep testing a known broken configuration, even more so since doing so would hide other potential problems.

Quoted 3 lines

Thanks. I agree it's pointless to test more, as there appear to be lots of bugs to fix already and this will just mask other bugs.

Last week, I encountered some locking issues in the kernel, so I tried to test the
kernel with some relative selftests. Since I had already set CONFIG_PREEMPT=y, I encountered the same results in RCU_TORTURE_TEST and LOCK_TORTURE_TEST. I tried to debug it and found the hang was due to the lack of _TIF_NEED_RESCHED check for kernel threads in entry.S. I will send a patch to solve this problem and I hope you can give me some comments if you find some
mistakes.

Quoted 6 lines
Quoted 3 lines

I'll be happy to run a set of tests if you Cc: me on the patch.

Guenter

Sorry, I have sent this patch but I forget to add you to the Cc list. I will send you this patch again and please help me to run a set of tests.

No need to resend; I picked it up from patchwork. Running a test now.

Guenter

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

#8 lore raw
In reply to Vincent Chen
Quoted 75 lines · On Wed, 02 Jan 2019 18:59:11 PST (-0800), [email protected] wrote:

Palmer Dabbelt <[email protected]> 於 2018年12月27日 週四 上午1:19寫道:

Quoted 70 lines · On Tue, 25 Dec 2018 19:23:50 PST (-0800), [email protected] wrote:
Quoted 60 lines · Guenter Roeck
Quoted 55 lines · Guenter Roeck
Quoted 44 lines · Palmer Dabbelt
Quoted 6 lines

I don't remember who suggested this, but I was talking to someone at Plumbers about all the horrible bugs we've had lurking around in our port despite it still getting to userspace and they suggested that I enable CONFIG_PREEMPT in our defconfig as it'll help shake out more issues.

Quoted 6 lines

This patch enables CONFIG_PREEMPT in the single RISC-V defconfig. It still passes my tests (ie, booting Fedora on QEMU). Maybe I'm too paranoid, but I still don't trust our port enough to just pull the trigger on this without at least accumulating some "Tested-by"s on it.

I can only give you a

Fails-miserably-for: Guenter Roeck <[email protected]>

Not the new defconfig itself, but the new defconfig plus a variety of selftest options. I don't know what exactly is failing, but here is my list of enabled selftests on top of defconfig:

CONFIG_CRYPTO_MANAGER_DISABLE_TESTS=y
CONFIG_CRC32_SELFTEST=y
CONFIG_DEBUG_LOCKING_API_SELFTESTS=y
CONFIG_DEBUG_NMI_SELFTEST=y
CONFIG_DEBUG_RODATA_TEST=y
CONFIG_DEBUG_TLBFLUSH=y
CONFIG_DMATEST=y
CONFIG_GLOB_SELFTEST=y
CONFIG_OF_UNITTEST=y
CONFIG_PCI_EPF_TEST=y
CONFIG_PCI_ENDPOINT_TEST=y
CONFIG_PROVE_LOCKING=y
CONFIG_PROVE_RCU=y
CONFIG_RCU_EQS_DEBUG=y
CONFIG_STATIC_KEYS_SELFTEST=y
CONFIG_STRING_SELFTEST=y
CONFIG_TEST_BITMAP=y
CONFIG_TEST_FIRMWARE=y
CONFIG_TEST_SORT=y
CONFIG_TEST_SYSCTL=y
CONFIG_TEST_UUID=y
CONFIG_USB_TEST=y

Quoted 9 lines

CONFIG_USB_EHSET_TEST_FIXTURE=y
CONFIG_USB_LINK_LAYER_TEST=y
CONFIG_WW_MUTEX_SELFTEST=y
CONFIG_TORTURE_TEST=y
CONFIG_LOCK_TORTURE_TEST=y
CONFIG_RCU_TORTURE_TEST=y
CONFIG_RBTREE_TEST=y
CONFIG_INTERVAL_TREE_TEST=y

Quoted 3 lines

Update: defconfig+CONFIG_LOCK_TORTURE_TEST=y causes the hang if CONFIG_PREEMPT=y.

Quoted 7 lines

Another update. CONFIG_RCU_TORTURE_TEST=y or CONFIG_WW_MUTEX_SELFTEST=y also result in hangs, only a bit later. I have now disabled all three test options in my riscv boot tests if CONFIG_PREEMPT=y. The lack of error reports in the future does _not_ mean that the problem has magically disappeared. I just find it pointless to keep testing a known broken configuration, even more so since doing so would hide other potential problems.

Quoted 3 lines

Thanks. I agree it's pointless to test more, as there appear to be lots of bugs to fix already and this will just mask other bugs.

Quoted 9 lines

Last week, I encountered some locking issues in the kernel, so I tried to test the
kernel with some relative selftests. Since I had already set CONFIG_PREEMPT=y, I encountered the same results in RCU_TORTURE_TEST and LOCK_TORTURE_TEST. I tried to debug it and found the hang was due to the lack of _TIF_NEED_RESCHED check for kernel threads in entry.S. I will send a patch to solve this problem and I hope you can give me some comments if you find some
mistakes.

Thanks!

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

#9 lore raw
In reply to Guenter Roeck
Quoted 12 lines · On Wed, 19 Dec 2018 17:45:11 PST (-0800), [email protected] wrote:
Quoted 11 lines · Palmer Dabbelt

I don't remember who suggested this, but I was talking to someone at Plumbers about all the horrible bugs we've had lurking around in our port despite it still getting to userspace and they suggested that I enable CONFIG_PREEMPT in our defconfig as it'll help shake out more issues.

This patch enables CONFIG_PREEMPT in the single RISC-V defconfig. It still passes my tests (ie, booting Fedora on QEMU). Maybe I'm too paranoid, but I still don't trust our port enough to just pull the trigger on this without at least accumulating some "Tested-by"s on it.

Quoted 5 lines
Quoted 1 line

I can only give you a

Fails-miserably-for: Guenter Roeck <[email protected]>

Ah, perfect -- I get a lot more worried when things like this pass then when they fail :)

Quoted 35 lines

Not the new defconfig itself, but the new defconfig plus a variety of selftest options. I don't know what exactly is failing, but here is my list of enabled selftests on top of defconfig:

CONFIG_CRYPTO_MANAGER_DISABLE_TESTS=y
CONFIG_CRC32_SELFTEST=y
CONFIG_DEBUG_LOCKING_API_SELFTESTS=y
CONFIG_DEBUG_NMI_SELFTEST=y
CONFIG_DEBUG_RODATA_TEST=y
CONFIG_DEBUG_TLBFLUSH=y
CONFIG_DMATEST=y
CONFIG_GLOB_SELFTEST=y
CONFIG_OF_UNITTEST=y
CONFIG_PCI_EPF_TEST=y
CONFIG_PCI_ENDPOINT_TEST=y
CONFIG_PROVE_LOCKING=y
CONFIG_PROVE_RCU=y
CONFIG_RCU_EQS_DEBUG=y
CONFIG_STATIC_KEYS_SELFTEST=y
CONFIG_STRING_SELFTEST=y
CONFIG_TEST_BITMAP=y
CONFIG_TEST_FIRMWARE=y
CONFIG_TEST_SORT=y
CONFIG_TEST_SYSCTL=y
CONFIG_TEST_UUID=y
CONFIG_USB_TEST=y
CONFIG_USB_EHSET_TEST_FIXTURE=y
CONFIG_USB_LINK_LAYER_TEST=y
CONFIG_WW_MUTEX_SELFTEST=y
CONFIG_TORTURE_TEST=y
CONFIG_LOCK_TORTURE_TEST=y
CONFIG_RCU_TORTURE_TEST=y
CONFIG_RBTREE_TEST=y
CONFIG_INTERVAL_TREE_TEST=y

Quoted 1 line

Tests pass after disabling those tests (or after reverting your patch).

Great. I'll take a look when I have time, and I definitely won't merge the patch.

Thanks for testing!

Quoted 17 lines

Guenter

Quoted 14 lines
---
 arch/riscv/configs/defconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/riscv/configs/defconfig +1 -0
index ef4f15df9adf..8869471af0fc 100644
@@ -1,5 +1,6 @@
 CONFIG_SYSVIPC=y
 CONFIG_POSIX_MQUEUE=y
+CONFIG_PREEMPT=y
 CONFIG_IKCONFIG=y
 CONFIG_IKCONFIG_PROC=y
Quoted 8 lines
Quoted 8 lines
arch/riscv/configs/defconfig +0 -2
 CONFIG_CGROUPS=y
@@ -77,4 +78,3 @@ CONFIG_NFS_V4_2=y
 CONFIG_ROOT_NFS=y
 CONFIG_CRYPTO_USER_API_HASH=y
 CONFIG_PRINTK_TIME=y
-# CONFIG_RCU_TRACE is not set
--

2.7.4

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