November 2023 - Linux-stable-mirror

[PATCH AUTOSEL 4.14] locking/ww_mutex/test: Fix potential workqueue corruption

by Sasha Levin

From: John Stultz <jstultz(a)google.com> [ Upstream commit bccdd808902f8c677317cec47c306e42b93b849e ] In some cases running with the test-ww_mutex code, I was seeing odd behavior where sometimes it seemed flush_workqueue was returning before all the work threads were finished. Often this would cause strange crashes as the mutexes would be freed while they were being used. Looking at the code, there is a lifetime problem as the controlling thread that spawns the work allocates the "struct stress" structures that are passed to the workqueue threads. Then when the workqueue threads are finished, they free the stress struct that was passed to them. Unfortunately the workqueue work_struct node is in the stress struct. Which means the work_struct is freed before the work thread returns and while flush_workqueue is waiting. It seems like a better idea to have the controlling thread both allocate and free the stress structures, so that we can be sure we don't corrupt the workqueue by freeing the structure prematurely. So this patch reworks the test to do so, and with this change I no longer see the early flush_workqueue returns. Signed-off-by: John Stultz <jstultz(a)google.com> Signed-off-by: Ingo Molnar <mingo(a)kernel.org> Link: https://lore.kernel.org/r/20230922043616.19282-3-jstultz@google.com Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- kernel/locking/test-ww_mutex.c | 20 ++++++++++++-------- 1 file changed, 12 insertions(+), 8 deletions(-) diff --git a/kernel/locking/test-ww_mutex.c b/kernel/locking/test-ww_mutex.c index 654977862b06b..8489a01f943e8 100644 --- a/kernel/locking/test-ww_mutex.c +++ b/kernel/locking/test-ww_mutex.c @@ -439,7 +439,6 @@ static void stress_inorder_work(struct work_struct *work) } while (!time_after(jiffies, stress->timeout)); kfree(order); - kfree(stress); } struct reorder_lock { @@ -504,7 +503,6 @@ static void stress_reorder_work(struct work_struct *work) list_for_each_entry_safe(ll, ln, &locks, link) kfree(ll); kfree(order); - kfree(stress); } static void stress_one_work(struct work_struct *work) @@ -525,8 +523,6 @@ static void stress_one_work(struct work_struct *work) break; } } while (!time_after(jiffies, stress->timeout)); - - kfree(stress); } #define STRESS_INORDER BIT(0) @@ -537,15 +533,24 @@ static void stress_one_work(struct work_struct *work) static int stress(int nlocks, int nthreads, unsigned int flags) { struct ww_mutex *locks; - int n; + struct stress *stress_array; + int n, count; locks = kmalloc_array(nlocks, sizeof(*locks), GFP_KERNEL); if (!locks) return -ENOMEM; + stress_array = kmalloc_array(nthreads, sizeof(*stress_array), + GFP_KERNEL); + if (!stress_array) { + kfree(locks); + return -ENOMEM; + } + for (n = 0; n < nlocks; n++) ww_mutex_init(&locks[n], &ww_class); + count = 0; for (n = 0; nthreads; n++) { struct stress *stress; void (*fn)(struct work_struct *work); @@ -569,9 +574,7 @@ static int stress(int nlocks, int nthreads, unsigned int flags) if (!fn) continue; - stress = kmalloc(sizeof(*stress), GFP_KERNEL); - if (!stress) - break; + stress = &stress_array[count++]; INIT_WORK(&stress->work, fn); stress->locks = locks; @@ -586,6 +589,7 @@ static int stress(int nlocks, int nthreads, unsigned int flags) for (n = 0; n < nlocks; n++) ww_mutex_destroy(&locks[n]); + kfree(stress_array); kfree(locks); return 0; -- 2.42.0

1 year, 3 months

1
0
0 0

[PATCH AUTOSEL 4.19] locking/ww_mutex/test: Fix potential workqueue corruption

by Sasha Levin

From: John Stultz <jstultz(a)google.com> [ Upstream commit bccdd808902f8c677317cec47c306e42b93b849e ] In some cases running with the test-ww_mutex code, I was seeing odd behavior where sometimes it seemed flush_workqueue was returning before all the work threads were finished. Often this would cause strange crashes as the mutexes would be freed while they were being used. Looking at the code, there is a lifetime problem as the controlling thread that spawns the work allocates the "struct stress" structures that are passed to the workqueue threads. Then when the workqueue threads are finished, they free the stress struct that was passed to them. Unfortunately the workqueue work_struct node is in the stress struct. Which means the work_struct is freed before the work thread returns and while flush_workqueue is waiting. It seems like a better idea to have the controlling thread both allocate and free the stress structures, so that we can be sure we don't corrupt the workqueue by freeing the structure prematurely. So this patch reworks the test to do so, and with this change I no longer see the early flush_workqueue returns. Signed-off-by: John Stultz <jstultz(a)google.com> Signed-off-by: Ingo Molnar <mingo(a)kernel.org> Link: https://lore.kernel.org/r/20230922043616.19282-3-jstultz@google.com Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- kernel/locking/test-ww_mutex.c | 20 ++++++++++++-------- 1 file changed, 12 insertions(+), 8 deletions(-) diff --git a/kernel/locking/test-ww_mutex.c b/kernel/locking/test-ww_mutex.c index 65a3b7e55b9fc..4fd05d9d5d6d1 100644 --- a/kernel/locking/test-ww_mutex.c +++ b/kernel/locking/test-ww_mutex.c @@ -439,7 +439,6 @@ static void stress_inorder_work(struct work_struct *work) } while (!time_after(jiffies, stress->timeout)); kfree(order); - kfree(stress); } struct reorder_lock { @@ -504,7 +503,6 @@ static void stress_reorder_work(struct work_struct *work) list_for_each_entry_safe(ll, ln, &locks, link) kfree(ll); kfree(order); - kfree(stress); } static void stress_one_work(struct work_struct *work) @@ -525,8 +523,6 @@ static void stress_one_work(struct work_struct *work) break; } } while (!time_after(jiffies, stress->timeout)); - - kfree(stress); } #define STRESS_INORDER BIT(0) @@ -537,15 +533,24 @@ static void stress_one_work(struct work_struct *work) static int stress(int nlocks, int nthreads, unsigned int flags) { struct ww_mutex *locks; - int n; + struct stress *stress_array; + int n, count; locks = kmalloc_array(nlocks, sizeof(*locks), GFP_KERNEL); if (!locks) return -ENOMEM; + stress_array = kmalloc_array(nthreads, sizeof(*stress_array), + GFP_KERNEL); + if (!stress_array) { + kfree(locks); + return -ENOMEM; + } + for (n = 0; n < nlocks; n++) ww_mutex_init(&locks[n], &ww_class); + count = 0; for (n = 0; nthreads; n++) { struct stress *stress; void (*fn)(struct work_struct *work); @@ -569,9 +574,7 @@ static int stress(int nlocks, int nthreads, unsigned int flags) if (!fn) continue; - stress = kmalloc(sizeof(*stress), GFP_KERNEL); - if (!stress) - break; + stress = &stress_array[count++]; INIT_WORK(&stress->work, fn); stress->locks = locks; @@ -586,6 +589,7 @@ static int stress(int nlocks, int nthreads, unsigned int flags) for (n = 0; n < nlocks; n++) ww_mutex_destroy(&locks[n]); + kfree(stress_array); kfree(locks); return 0; -- 2.42.0

1 year, 3 months

1
0
0 0

[PATCH AUTOSEL 5.4] locking/ww_mutex/test: Fix potential workqueue corruption

by Sasha Levin

From: John Stultz <jstultz(a)google.com> [ Upstream commit bccdd808902f8c677317cec47c306e42b93b849e ] In some cases running with the test-ww_mutex code, I was seeing odd behavior where sometimes it seemed flush_workqueue was returning before all the work threads were finished. Often this would cause strange crashes as the mutexes would be freed while they were being used. Looking at the code, there is a lifetime problem as the controlling thread that spawns the work allocates the "struct stress" structures that are passed to the workqueue threads. Then when the workqueue threads are finished, they free the stress struct that was passed to them. Unfortunately the workqueue work_struct node is in the stress struct. Which means the work_struct is freed before the work thread returns and while flush_workqueue is waiting. It seems like a better idea to have the controlling thread both allocate and free the stress structures, so that we can be sure we don't corrupt the workqueue by freeing the structure prematurely. So this patch reworks the test to do so, and with this change I no longer see the early flush_workqueue returns. Signed-off-by: John Stultz <jstultz(a)google.com> Signed-off-by: Ingo Molnar <mingo(a)kernel.org> Link: https://lore.kernel.org/r/20230922043616.19282-3-jstultz@google.com Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- kernel/locking/test-ww_mutex.c | 20 ++++++++++++-------- 1 file changed, 12 insertions(+), 8 deletions(-) diff --git a/kernel/locking/test-ww_mutex.c b/kernel/locking/test-ww_mutex.c index 3e82f449b4ff7..da36997d8742c 100644 --- a/kernel/locking/test-ww_mutex.c +++ b/kernel/locking/test-ww_mutex.c @@ -426,7 +426,6 @@ static void stress_inorder_work(struct work_struct *work) } while (!time_after(jiffies, stress->timeout)); kfree(order); - kfree(stress); } struct reorder_lock { @@ -491,7 +490,6 @@ static void stress_reorder_work(struct work_struct *work) list_for_each_entry_safe(ll, ln, &locks, link) kfree(ll); kfree(order); - kfree(stress); } static void stress_one_work(struct work_struct *work) @@ -512,8 +510,6 @@ static void stress_one_work(struct work_struct *work) break; } } while (!time_after(jiffies, stress->timeout)); - - kfree(stress); } #define STRESS_INORDER BIT(0) @@ -524,15 +520,24 @@ static void stress_one_work(struct work_struct *work) static int stress(int nlocks, int nthreads, unsigned int flags) { struct ww_mutex *locks; - int n; + struct stress *stress_array; + int n, count; locks = kmalloc_array(nlocks, sizeof(*locks), GFP_KERNEL); if (!locks) return -ENOMEM; + stress_array = kmalloc_array(nthreads, sizeof(*stress_array), + GFP_KERNEL); + if (!stress_array) { + kfree(locks); + return -ENOMEM; + } + for (n = 0; n < nlocks; n++) ww_mutex_init(&locks[n], &ww_class); + count = 0; for (n = 0; nthreads; n++) { struct stress *stress; void (*fn)(struct work_struct *work); @@ -556,9 +561,7 @@ static int stress(int nlocks, int nthreads, unsigned int flags) if (!fn) continue; - stress = kmalloc(sizeof(*stress), GFP_KERNEL); - if (!stress) - break; + stress = &stress_array[count++]; INIT_WORK(&stress->work, fn); stress->locks = locks; @@ -573,6 +576,7 @@ static int stress(int nlocks, int nthreads, unsigned int flags) for (n = 0; n < nlocks; n++) ww_mutex_destroy(&locks[n]); + kfree(stress_array); kfree(locks); return 0; -- 2.42.0

1 year, 3 months

1
0
0 0

[PATCH AUTOSEL 5.10] locking/ww_mutex/test: Fix potential workqueue corruption

by Sasha Levin

From: John Stultz <jstultz(a)google.com> [ Upstream commit bccdd808902f8c677317cec47c306e42b93b849e ] In some cases running with the test-ww_mutex code, I was seeing odd behavior where sometimes it seemed flush_workqueue was returning before all the work threads were finished. Often this would cause strange crashes as the mutexes would be freed while they were being used. Looking at the code, there is a lifetime problem as the controlling thread that spawns the work allocates the "struct stress" structures that are passed to the workqueue threads. Then when the workqueue threads are finished, they free the stress struct that was passed to them. Unfortunately the workqueue work_struct node is in the stress struct. Which means the work_struct is freed before the work thread returns and while flush_workqueue is waiting. It seems like a better idea to have the controlling thread both allocate and free the stress structures, so that we can be sure we don't corrupt the workqueue by freeing the structure prematurely. So this patch reworks the test to do so, and with this change I no longer see the early flush_workqueue returns. Signed-off-by: John Stultz <jstultz(a)google.com> Signed-off-by: Ingo Molnar <mingo(a)kernel.org> Link: https://lore.kernel.org/r/20230922043616.19282-3-jstultz@google.com Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- kernel/locking/test-ww_mutex.c | 20 ++++++++++++-------- 1 file changed, 12 insertions(+), 8 deletions(-) diff --git a/kernel/locking/test-ww_mutex.c b/kernel/locking/test-ww_mutex.c index 3e82f449b4ff7..da36997d8742c 100644 --- a/kernel/locking/test-ww_mutex.c +++ b/kernel/locking/test-ww_mutex.c @@ -426,7 +426,6 @@ static void stress_inorder_work(struct work_struct *work) } while (!time_after(jiffies, stress->timeout)); kfree(order); - kfree(stress); } struct reorder_lock { @@ -491,7 +490,6 @@ static void stress_reorder_work(struct work_struct *work) list_for_each_entry_safe(ll, ln, &locks, link) kfree(ll); kfree(order); - kfree(stress); } static void stress_one_work(struct work_struct *work) @@ -512,8 +510,6 @@ static void stress_one_work(struct work_struct *work) break; } } while (!time_after(jiffies, stress->timeout)); - - kfree(stress); } #define STRESS_INORDER BIT(0) @@ -524,15 +520,24 @@ static void stress_one_work(struct work_struct *work) static int stress(int nlocks, int nthreads, unsigned int flags) { struct ww_mutex *locks; - int n; + struct stress *stress_array; + int n, count; locks = kmalloc_array(nlocks, sizeof(*locks), GFP_KERNEL); if (!locks) return -ENOMEM; + stress_array = kmalloc_array(nthreads, sizeof(*stress_array), + GFP_KERNEL); + if (!stress_array) { + kfree(locks); + return -ENOMEM; + } + for (n = 0; n < nlocks; n++) ww_mutex_init(&locks[n], &ww_class); + count = 0; for (n = 0; nthreads; n++) { struct stress *stress; void (*fn)(struct work_struct *work); @@ -556,9 +561,7 @@ static int stress(int nlocks, int nthreads, unsigned int flags) if (!fn) continue; - stress = kmalloc(sizeof(*stress), GFP_KERNEL); - if (!stress) - break; + stress = &stress_array[count++]; INIT_WORK(&stress->work, fn); stress->locks = locks; @@ -573,6 +576,7 @@ static int stress(int nlocks, int nthreads, unsigned int flags) for (n = 0; n < nlocks; n++) ww_mutex_destroy(&locks[n]); + kfree(stress_array); kfree(locks); return 0; -- 2.42.0

1 year, 3 months

1
0
0 0

[PATCH AUTOSEL 5.15] locking/ww_mutex/test: Fix potential workqueue corruption

by Sasha Levin

From: John Stultz <jstultz(a)google.com> [ Upstream commit bccdd808902f8c677317cec47c306e42b93b849e ] In some cases running with the test-ww_mutex code, I was seeing odd behavior where sometimes it seemed flush_workqueue was returning before all the work threads were finished. Often this would cause strange crashes as the mutexes would be freed while they were being used. Looking at the code, there is a lifetime problem as the controlling thread that spawns the work allocates the "struct stress" structures that are passed to the workqueue threads. Then when the workqueue threads are finished, they free the stress struct that was passed to them. Unfortunately the workqueue work_struct node is in the stress struct. Which means the work_struct is freed before the work thread returns and while flush_workqueue is waiting. It seems like a better idea to have the controlling thread both allocate and free the stress structures, so that we can be sure we don't corrupt the workqueue by freeing the structure prematurely. So this patch reworks the test to do so, and with this change I no longer see the early flush_workqueue returns. Signed-off-by: John Stultz <jstultz(a)google.com> Signed-off-by: Ingo Molnar <mingo(a)kernel.org> Link: https://lore.kernel.org/r/20230922043616.19282-3-jstultz@google.com Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- kernel/locking/test-ww_mutex.c | 20 ++++++++++++-------- 1 file changed, 12 insertions(+), 8 deletions(-) diff --git a/kernel/locking/test-ww_mutex.c b/kernel/locking/test-ww_mutex.c index 3e82f449b4ff7..da36997d8742c 100644 --- a/kernel/locking/test-ww_mutex.c +++ b/kernel/locking/test-ww_mutex.c @@ -426,7 +426,6 @@ static void stress_inorder_work(struct work_struct *work) } while (!time_after(jiffies, stress->timeout)); kfree(order); - kfree(stress); } struct reorder_lock { @@ -491,7 +490,6 @@ static void stress_reorder_work(struct work_struct *work) list_for_each_entry_safe(ll, ln, &locks, link) kfree(ll); kfree(order); - kfree(stress); } static void stress_one_work(struct work_struct *work) @@ -512,8 +510,6 @@ static void stress_one_work(struct work_struct *work) break; } } while (!time_after(jiffies, stress->timeout)); - - kfree(stress); } #define STRESS_INORDER BIT(0) @@ -524,15 +520,24 @@ static void stress_one_work(struct work_struct *work) static int stress(int nlocks, int nthreads, unsigned int flags) { struct ww_mutex *locks; - int n; + struct stress *stress_array; + int n, count; locks = kmalloc_array(nlocks, sizeof(*locks), GFP_KERNEL); if (!locks) return -ENOMEM; + stress_array = kmalloc_array(nthreads, sizeof(*stress_array), + GFP_KERNEL); + if (!stress_array) { + kfree(locks); + return -ENOMEM; + } + for (n = 0; n < nlocks; n++) ww_mutex_init(&locks[n], &ww_class); + count = 0; for (n = 0; nthreads; n++) { struct stress *stress; void (*fn)(struct work_struct *work); @@ -556,9 +561,7 @@ static int stress(int nlocks, int nthreads, unsigned int flags) if (!fn) continue; - stress = kmalloc(sizeof(*stress), GFP_KERNEL); - if (!stress) - break; + stress = &stress_array[count++]; INIT_WORK(&stress->work, fn); stress->locks = locks; @@ -573,6 +576,7 @@ static int stress(int nlocks, int nthreads, unsigned int flags) for (n = 0; n < nlocks; n++) ww_mutex_destroy(&locks[n]); + kfree(stress_array); kfree(locks); return 0; -- 2.42.0

1 year, 3 months

1
0
0 0

[PATCH AUTOSEL 6.1 1/2] locking/ww_mutex/test: Fix potential workqueue corruption

by Sasha Levin

From: John Stultz <jstultz(a)google.com> [ Upstream commit bccdd808902f8c677317cec47c306e42b93b849e ] In some cases running with the test-ww_mutex code, I was seeing odd behavior where sometimes it seemed flush_workqueue was returning before all the work threads were finished. Often this would cause strange crashes as the mutexes would be freed while they were being used. Looking at the code, there is a lifetime problem as the controlling thread that spawns the work allocates the "struct stress" structures that are passed to the workqueue threads. Then when the workqueue threads are finished, they free the stress struct that was passed to them. Unfortunately the workqueue work_struct node is in the stress struct. Which means the work_struct is freed before the work thread returns and while flush_workqueue is waiting. It seems like a better idea to have the controlling thread both allocate and free the stress structures, so that we can be sure we don't corrupt the workqueue by freeing the structure prematurely. So this patch reworks the test to do so, and with this change I no longer see the early flush_workqueue returns. Signed-off-by: John Stultz <jstultz(a)google.com> Signed-off-by: Ingo Molnar <mingo(a)kernel.org> Link: https://lore.kernel.org/r/20230922043616.19282-3-jstultz@google.com Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- kernel/locking/test-ww_mutex.c | 20 ++++++++++++-------- 1 file changed, 12 insertions(+), 8 deletions(-) diff --git a/kernel/locking/test-ww_mutex.c b/kernel/locking/test-ww_mutex.c index 43efb2a041602..b1e25695185a4 100644 --- a/kernel/locking/test-ww_mutex.c +++ b/kernel/locking/test-ww_mutex.c @@ -466,7 +466,6 @@ static void stress_inorder_work(struct work_struct *work) } while (!time_after(jiffies, stress->timeout)); kfree(order); - kfree(stress); } struct reorder_lock { @@ -531,7 +530,6 @@ static void stress_reorder_work(struct work_struct *work) list_for_each_entry_safe(ll, ln, &locks, link) kfree(ll); kfree(order); - kfree(stress); } static void stress_one_work(struct work_struct *work) @@ -552,8 +550,6 @@ static void stress_one_work(struct work_struct *work) break; } } while (!time_after(jiffies, stress->timeout)); - - kfree(stress); } #define STRESS_INORDER BIT(0) @@ -564,15 +560,24 @@ static void stress_one_work(struct work_struct *work) static int stress(int nlocks, int nthreads, unsigned int flags) { struct ww_mutex *locks; - int n; + struct stress *stress_array; + int n, count; locks = kmalloc_array(nlocks, sizeof(*locks), GFP_KERNEL); if (!locks) return -ENOMEM; + stress_array = kmalloc_array(nthreads, sizeof(*stress_array), + GFP_KERNEL); + if (!stress_array) { + kfree(locks); + return -ENOMEM; + } + for (n = 0; n < nlocks; n++) ww_mutex_init(&locks[n], &ww_class); + count = 0; for (n = 0; nthreads; n++) { struct stress *stress; void (*fn)(struct work_struct *work); @@ -596,9 +601,7 @@ static int stress(int nlocks, int nthreads, unsigned int flags) if (!fn) continue; - stress = kmalloc(sizeof(*stress), GFP_KERNEL); - if (!stress) - break; + stress = &stress_array[count++]; INIT_WORK(&stress->work, fn); stress->locks = locks; @@ -613,6 +616,7 @@ static int stress(int nlocks, int nthreads, unsigned int flags) for (n = 0; n < nlocks; n++) ww_mutex_destroy(&locks[n]); + kfree(stress_array); kfree(locks); return 0; -- 2.42.0

1 year, 3 months

1
1
0 0

[PATCH AUTOSEL 6.5 1/5] locking/ww_mutex/test: Fix potential workqueue corruption

by Sasha Levin

From: John Stultz <jstultz(a)google.com> [ Upstream commit bccdd808902f8c677317cec47c306e42b93b849e ] In some cases running with the test-ww_mutex code, I was seeing odd behavior where sometimes it seemed flush_workqueue was returning before all the work threads were finished. Often this would cause strange crashes as the mutexes would be freed while they were being used. Looking at the code, there is a lifetime problem as the controlling thread that spawns the work allocates the "struct stress" structures that are passed to the workqueue threads. Then when the workqueue threads are finished, they free the stress struct that was passed to them. Unfortunately the workqueue work_struct node is in the stress struct. Which means the work_struct is freed before the work thread returns and while flush_workqueue is waiting. It seems like a better idea to have the controlling thread both allocate and free the stress structures, so that we can be sure we don't corrupt the workqueue by freeing the structure prematurely. So this patch reworks the test to do so, and with this change I no longer see the early flush_workqueue returns. Signed-off-by: John Stultz <jstultz(a)google.com> Signed-off-by: Ingo Molnar <mingo(a)kernel.org> Link: https://lore.kernel.org/r/20230922043616.19282-3-jstultz@google.com Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- kernel/locking/test-ww_mutex.c | 20 ++++++++++++-------- 1 file changed, 12 insertions(+), 8 deletions(-) diff --git a/kernel/locking/test-ww_mutex.c b/kernel/locking/test-ww_mutex.c index 93cca6e698600..7c5a8f05497f2 100644 --- a/kernel/locking/test-ww_mutex.c +++ b/kernel/locking/test-ww_mutex.c @@ -466,7 +466,6 @@ static void stress_inorder_work(struct work_struct *work) } while (!time_after(jiffies, stress->timeout)); kfree(order); - kfree(stress); } struct reorder_lock { @@ -531,7 +530,6 @@ static void stress_reorder_work(struct work_struct *work) list_for_each_entry_safe(ll, ln, &locks, link) kfree(ll); kfree(order); - kfree(stress); } static void stress_one_work(struct work_struct *work) @@ -552,8 +550,6 @@ static void stress_one_work(struct work_struct *work) break; } } while (!time_after(jiffies, stress->timeout)); - - kfree(stress); } #define STRESS_INORDER BIT(0) @@ -564,15 +560,24 @@ static void stress_one_work(struct work_struct *work) static int stress(int nlocks, int nthreads, unsigned int flags) { struct ww_mutex *locks; - int n; + struct stress *stress_array; + int n, count; locks = kmalloc_array(nlocks, sizeof(*locks), GFP_KERNEL); if (!locks) return -ENOMEM; + stress_array = kmalloc_array(nthreads, sizeof(*stress_array), + GFP_KERNEL); + if (!stress_array) { + kfree(locks); + return -ENOMEM; + } + for (n = 0; n < nlocks; n++) ww_mutex_init(&locks[n], &ww_class); + count = 0; for (n = 0; nthreads; n++) { struct stress *stress; void (*fn)(struct work_struct *work); @@ -596,9 +601,7 @@ static int stress(int nlocks, int nthreads, unsigned int flags) if (!fn) continue; - stress = kmalloc(sizeof(*stress), GFP_KERNEL); - if (!stress) - break; + stress = &stress_array[count++]; INIT_WORK(&stress->work, fn); stress->locks = locks; @@ -613,6 +616,7 @@ static int stress(int nlocks, int nthreads, unsigned int flags) for (n = 0; n < nlocks; n++) ww_mutex_destroy(&locks[n]); + kfree(stress_array); kfree(locks); return 0; -- 2.42.0

1 year, 3 months

1
4
0 0

[PATCH AUTOSEL 6.6 1/5] locking/ww_mutex/test: Fix potential workqueue corruption

by Sasha Levin

From: John Stultz <jstultz(a)google.com> [ Upstream commit bccdd808902f8c677317cec47c306e42b93b849e ] In some cases running with the test-ww_mutex code, I was seeing odd behavior where sometimes it seemed flush_workqueue was returning before all the work threads were finished. Often this would cause strange crashes as the mutexes would be freed while they were being used. Looking at the code, there is a lifetime problem as the controlling thread that spawns the work allocates the "struct stress" structures that are passed to the workqueue threads. Then when the workqueue threads are finished, they free the stress struct that was passed to them. Unfortunately the workqueue work_struct node is in the stress struct. Which means the work_struct is freed before the work thread returns and while flush_workqueue is waiting. It seems like a better idea to have the controlling thread both allocate and free the stress structures, so that we can be sure we don't corrupt the workqueue by freeing the structure prematurely. So this patch reworks the test to do so, and with this change I no longer see the early flush_workqueue returns. Signed-off-by: John Stultz <jstultz(a)google.com> Signed-off-by: Ingo Molnar <mingo(a)kernel.org> Link: https://lore.kernel.org/r/20230922043616.19282-3-jstultz@google.com Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- kernel/locking/test-ww_mutex.c | 20 ++++++++++++-------- 1 file changed, 12 insertions(+), 8 deletions(-) diff --git a/kernel/locking/test-ww_mutex.c b/kernel/locking/test-ww_mutex.c index 93cca6e698600..7c5a8f05497f2 100644 --- a/kernel/locking/test-ww_mutex.c +++ b/kernel/locking/test-ww_mutex.c @@ -466,7 +466,6 @@ static void stress_inorder_work(struct work_struct *work) } while (!time_after(jiffies, stress->timeout)); kfree(order); - kfree(stress); } struct reorder_lock { @@ -531,7 +530,6 @@ static void stress_reorder_work(struct work_struct *work) list_for_each_entry_safe(ll, ln, &locks, link) kfree(ll); kfree(order); - kfree(stress); } static void stress_one_work(struct work_struct *work) @@ -552,8 +550,6 @@ static void stress_one_work(struct work_struct *work) break; } } while (!time_after(jiffies, stress->timeout)); - - kfree(stress); } #define STRESS_INORDER BIT(0) @@ -564,15 +560,24 @@ static void stress_one_work(struct work_struct *work) static int stress(int nlocks, int nthreads, unsigned int flags) { struct ww_mutex *locks; - int n; + struct stress *stress_array; + int n, count; locks = kmalloc_array(nlocks, sizeof(*locks), GFP_KERNEL); if (!locks) return -ENOMEM; + stress_array = kmalloc_array(nthreads, sizeof(*stress_array), + GFP_KERNEL); + if (!stress_array) { + kfree(locks); + return -ENOMEM; + } + for (n = 0; n < nlocks; n++) ww_mutex_init(&locks[n], &ww_class); + count = 0; for (n = 0; nthreads; n++) { struct stress *stress; void (*fn)(struct work_struct *work); @@ -596,9 +601,7 @@ static int stress(int nlocks, int nthreads, unsigned int flags) if (!fn) continue; - stress = kmalloc(sizeof(*stress), GFP_KERNEL); - if (!stress) - break; + stress = &stress_array[count++]; INIT_WORK(&stress->work, fn); stress->locks = locks; @@ -613,6 +616,7 @@ static int stress(int nlocks, int nthreads, unsigned int flags) for (n = 0; n < nlocks; n++) ww_mutex_destroy(&locks[n]); + kfree(stress_array); kfree(locks); return 0; -- 2.42.0

1 year, 3 months

1
4
0 0

stable-rc/linux-5.10.y baseline: 152 runs, 2 regressions (v5.10.199-96-gfed6441dbe524)

by kernelci.org bot

stable-rc/linux-5.10.y baseline: 152 runs, 2 regressions (v5.10.199-96-gfed6441dbe524) Regressions Summary ------------------- platform | arch | lab | compiler | defconfig | regressions -------------------+-------+---------------+----------+-----------+------------ sun50i-h6-pine-h64 | arm64 | lab-clabbe | gcc-10 | defconfig | 1 sun50i-h6-pine-h64 | arm64 | lab-collabora | gcc-10 | defconfig | 1 Details: https://kernelci.org/test/job/stable-rc/branch/linux-5.10.y/kernel/v5.10.19… Test: baseline Tree: stable-rc Branch: linux-5.10.y Describe: v5.10.199-96-gfed6441dbe524 URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git SHA: fed6441dbe524de2cf3a6a40d5d65c369bf583a0 Test Regressions ---------------- platform | arch | lab | compiler | defconfig | regressions -------------------+-------+---------------+----------+-----------+------------ sun50i-h6-pine-h64 | arm64 | lab-clabbe | gcc-10 | defconfig | 1 Details: https://kernelci.org/test/plan/id/6549367b5716947be1efcf89 Results: 5 PASS, 1 FAIL, 1 SKIP Full config: defconfig Compiler: gcc-10 (aarch64-linux-gnu-gcc (Debian 10.2.1-6) 10.2.1 20210110) Plain log: https://storage.kernelci.org//stable-rc/linux-5.10.y/v5.10.199-96-gfed6441d… HTML log: https://storage.kernelci.org//stable-rc/linux-5.10.y/v5.10.199-96-gfed6441d… Rootfs: http://storage.kernelci.org/images/rootfs/buildroot/buildroot-baseline/2023… * baseline.bootrr.deferred-probe-empty: https://kernelci.org/test/case/id/6549367b5716947be1efcf92 failing since 26 days (last pass: v5.10.176-224-g10e9fd53dc59, first fail: v5.10.198) 2023-11-06T18:54:24.985357 <8>[ 16.934667] <LAVA_SIGNAL_ENDRUN 0_dmesg 443029_1.5.2.4.1> 2023-11-06T18:54:25.090403 / # # 2023-11-06T18:54:25.192134 export SHELL=/bin/sh 2023-11-06T18:54:25.192909 # 2023-11-06T18:54:25.294063 / # export SHELL=/bin/sh. /lava-443029/environment 2023-11-06T18:54:25.294757 2023-11-06T18:54:25.395947 / # . /lava-443029/environment/lava-443029/bin/lava-test-runner /lava-443029/1 2023-11-06T18:54:25.397064 2023-11-06T18:54:25.400425 / # /lava-443029/bin/lava-test-runner /lava-443029/1 2023-11-06T18:54:25.468861 + export 'TESTRUN_ID=1_bootrr' ... (11 line(s) more) platform | arch | lab | compiler | defconfig | regressions -------------------+-------+---------------+----------+-----------+------------ sun50i-h6-pine-h64 | arm64 | lab-collabora | gcc-10 | defconfig | 1 Details: https://kernelci.org/test/plan/id/65493683ec1bd0ef86efcefe Results: 5 PASS, 1 FAIL, 1 SKIP Full config: defconfig Compiler: gcc-10 (aarch64-linux-gnu-gcc (Debian 10.2.1-6) 10.2.1 20210110) Plain log: https://storage.kernelci.org//stable-rc/linux-5.10.y/v5.10.199-96-gfed6441d… HTML log: https://storage.kernelci.org//stable-rc/linux-5.10.y/v5.10.199-96-gfed6441d… Rootfs: http://storage.kernelci.org/images/rootfs/buildroot/buildroot-baseline/2023… * baseline.bootrr.deferred-probe-empty: https://kernelci.org/test/case/id/65493683ec1bd0ef86efcf07 failing since 26 days (last pass: v5.10.176-224-g10e9fd53dc59, first fail: v5.10.198) 2023-11-06T19:01:16.663939 / # # 2023-11-06T19:01:16.765985 export SHELL=/bin/sh 2023-11-06T19:01:16.766664 # 2023-11-06T19:01:16.867975 / # export SHELL=/bin/sh. /lava-11956273/environment 2023-11-06T19:01:16.868679 2023-11-06T19:01:16.970129 / # . /lava-11956273/environment/lava-11956273/bin/lava-test-runner /lava-11956273/1 2023-11-06T19:01:16.971206 2023-11-06T19:01:16.988129 / # /lava-11956273/bin/lava-test-runner /lava-11956273/1 2023-11-06T19:01:17.029706 + export 'TESTRUN_ID=1_bootrr' 2023-11-06T19:01:17.047268 + cd /lava-1195627<8>[ 18.282577] <LAVA_SIGNAL_STARTRUN 1_bootrr 11956273_1.5.2.4.5> ... (10 line(s) more)

1 year, 3 months

1
0
0 0

[PATCH] dlm: use kernel_connect() and kernel_bind()

by Jordan Rife

Recent changes to kernel_connect() and kernel_bind() ensure that callers are insulated from changes to the address parameter made by BPF SOCK_ADDR hooks. This patch wraps direct calls to ops->connect() and ops->bind() with kernel_connect() and kernel_bind() to protect callers in such cases. Link: https://lore.kernel.org/netdev/9944248dba1bce861375fcce9de663934d933ba9.cam… Fixes: d74bad4e74ee ("bpf: Hooks for sys_connect") Fixes: 4fbac77d2d09 ("bpf: Hooks for sys_bind") Cc: stable(a)vger.kernel.org Signed-off-by: Jordan Rife <jrife(a)google.com> --- fs/dlm/lowcomms.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/fs/dlm/lowcomms.c b/fs/dlm/lowcomms.c index 67f8dd8a05ef2..6296c62c10fa9 100644 --- a/fs/dlm/lowcomms.c +++ b/fs/dlm/lowcomms.c @@ -1817,8 +1817,8 @@ static int dlm_tcp_bind(struct socket *sock) memcpy(&src_addr, &dlm_local_addr[0], sizeof(src_addr)); make_sockaddr(&src_addr, 0, &addr_len); - result = sock->ops->bind(sock, (struct sockaddr *)&src_addr, - addr_len); + result = kernel_bind(sock, (struct sockaddr *)&src_addr, + addr_len); if (result < 0) { /* This *may* not indicate a critical error */ log_print("could not bind for connect: %d", result); @@ -1830,7 +1830,7 @@ static int dlm_tcp_bind(struct socket *sock) static int dlm_tcp_connect(struct connection *con, struct socket *sock, struct sockaddr *addr, int addr_len) { - return sock->ops->connect(sock, addr, addr_len, O_NONBLOCK); + return kernel_connect(sock, addr, addr_len, O_NONBLOCK); } static int dlm_tcp_listen_validate(void) @@ -1862,8 +1862,8 @@ static int dlm_tcp_listen_bind(struct socket *sock) /* Bind to our port */ make_sockaddr(&dlm_local_addr[0], dlm_config.ci_tcp_port, &addr_len); - return sock->ops->bind(sock, (struct sockaddr *)&dlm_local_addr[0], - addr_len); + return kernel_bind(sock, (struct sockaddr *)&dlm_local_addr[0], + addr_len); } static const struct dlm_proto_ops dlm_tcp_ops = { @@ -1888,12 +1888,12 @@ static int dlm_sctp_connect(struct connection *con, struct socket *sock, int ret; /* - * Make sock->ops->connect() function return in specified time, + * Make kernel_connect() function return in specified time, * since O_NONBLOCK argument in connect() function does not work here, * then, we should restore the default value of this attribute. */ sock_set_sndtimeo(sock->sk, 5); - ret = sock->ops->connect(sock, addr, addr_len, 0); + ret = kernel_connect(sock, addr, addr_len, 0); sock_set_sndtimeo(sock->sk, 0); return ret; } -- 2.42.0.869.gea05f2083d-goog

1 year, 3 months

1
0
0 0

2025

2024

2023

2022

2021

2020

2019

2018

2017

Linux-stable-mirror November 2023