mm, vmstat: Make sure mutex is a global static
Subsystems (2)
PATCH 1/1 mm, vmstat: Make sure mutex is a global static
The mutex in sysctl_vm_numa_stat_handler() needs to be a global static, not a stack variable, otherwise it doesn't serve any purpose. Also, reading the file with CONFIG_LOCKDEP=y will complain:
[ 63.258593] INFO: trying to register non-static key. [ 63.259113] the code is fine but needs lockdep annotation. [ 63.259596] turning off the locking correctness validator. [ 63.260073] CPU: 1 PID: 4102 Comm: perl Not tainted 4.14.0-rc8-next-20171107+ #419 [ 63.260769] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Ubuntu-1.8.2-1ubuntu1 04/01/2014 [ 63.261570] Call Trace: [ 63.261783] dump_stack+0x5f/0x86 [ 63.262062] register_lock_class+0xe4/0x550 [ 63.262408] ? __lock_acquire+0x308/0x1170 [ 63.262746] __lock_acquire+0x7e/0x1170 [ 63.263063] lock_acquire+0x9d/0x1d0 [ 63.263363] ? sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.263777] ? sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.264192] __mutex_lock+0xb8/0x9a0 [ 63.264488] ? sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.264942] ? sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.265398] ? sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.265840] sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.266270] proc_sys_call_handler+0xe3/0x100 [ 63.266655] __vfs_read+0x33/0x1b0 [ 63.266957] vfs_read+0xa6/0x150 [ 63.267244] SyS_read+0x55/0xc0 [ 63.267525] do_syscall_64+0x56/0x140 [ 63.267850] entry_SYSCALL64_slow_path+0x25/0x25
--- mm/vmstat.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
mm/vmstat.c +2 -1
@@ -72,11 +72,12 @@ static void invalid_numa_statistics(void)zero_global_numa_counters (); } +static DEFINE_MUTEX (vm_numa_stat_lock); +int sysctl_vm_numa_stat_handler (struct ctl_table *table,int write,void __user *buffer,size_t *length,loff_t *ppos) {int ret, oldval; -DEFINE_MUTEX (vm_numa_stat_lock);mutex_lock (&vm_numa_stat_lock);if (write)
Signature · 6 lines
2.7.4 -- Kees Cook Pixel Security
Quoted 3 lines · Kees Cook
The mutex in
sysctl_vm_numa_stat_handler()needs to be a global static, not a stack variable, otherwise it doesn't serve any purpose. Also, reading the file withCONFIG_LOCKDEP=ywill complain:
Oops, good catch.
Quoted 25 lines
[ 63.258593] INFO: trying to register non-static key. [ 63.259113] the code is fine but needs lockdep annotation. [ 63.259596] turning off the locking correctness validator. [ 63.260073] CPU: 1 PID: 4102 Comm: perl Not tainted 4.14.0-rc8-next-20171107+ #419 [ 63.260769] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Ubuntu-1.8.2-1ubuntu1 04/01/2014 [ 63.261570] Call Trace: [ 63.261783] dump_stack+0x5f/0x86 [ 63.262062] register_lock_class+0xe4/0x550 [ 63.262408] ? __lock_acquire+0x308/0x1170 [ 63.262746] __lock_acquire+0x7e/0x1170 [ 63.263063] lock_acquire+0x9d/0x1d0 [ 63.263363] ? sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.263777] ? sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.264192] __mutex_lock+0xb8/0x9a0 [ 63.264488] ? sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.264942] ? sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.265398] ? sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.265840] sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.266270] proc_sys_call_handler+0xe3/0x100 [ 63.266655] __vfs_read+0x33/0x1b0 [ 63.266957] vfs_read+0xa6/0x150 [ 63.267244] SyS_read+0x55/0xc0 [ 63.267525] do_syscall_64+0x56/0x140 [ 63.267850] entry_SYSCALL64_slow_path+0x25/0x25
Quoted 1 line
Fixes 920d5f77d1a25 ("mm, sysctl: make NUMA stats configurable")
Note that this hash is specific to particular next-$DATE as mmotm is reimported each day.
Quoted 15 lines
Cc Jesper Dangaard Brouer <[email protected]>Cc Dave Hansen <[email protected]>Cc Ying Huang <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Michal Hocko <[email protected]>Signed-off-by Kees Cook <[email protected]>--- mm/vmstat.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)mm/vmstat.c +0
-0
mm/vmstat.c static void invalid_numa_statistics(void) · line 79
Quoted 10 lines
mm/vmstat.c +2
-1@@ -72,11 +72,12 @@ static void invalid_numa_statistics(void)zero_global_numa_counters (); } +static DEFINE_MUTEX (vm_numa_stat_lock); +int sysctl_vm_numa_stat_handler (struct ctl_table *table,int write,void __user *buffer,size_t *length,loff_t *ppos) {int ret, oldval; -DEFINE_MUTEX (vm_numa_stat_lock);
Yeah it was Michal who suggested scoping the mutex here instead of global scope, but I think he didn't mean to remove the 'static' qualifier, and we both missed that in the review :(
So the scope under sysctl_vm_numa_stat_handler() should be okay, just with the 'static' added.
Quoted 4 lines
mm/vmstat.c +0
-0mutex_lock (&vm_numa_stat_lock);if (write)
Quoted 1 line · Vlastimil Babka
On 11/07/2017 10:38 PM, Kees Cook wrote:
[...]
Quoted 5 lines
Quoted 5 lines
mm/vmstat.c +2
-0+static DEFINE_MUTEX (vm_numa_stat_lock); +int sysctl_vm_numa_stat_handler (struct ctl_table *table,int write,void __user *buffer,size_t *length,loff_t *ppos) {
Quoted 8 lines
mm/vmstat.c static void invalid_numa_statistics(void) · line 79
Quoted 2 lines
mm/vmstat.c +0
-1int ret, oldval; -DEFINE_MUTEX (vm_numa_stat_lock);Yeah it was Michal who suggested scoping the mutex here instead of global scope, but I think he didn't mean to remove the 'static' qualifier, and we both missed that in the review :(
So the scope undersysctl_vm_numa_stat_handler()should be okay, just with the 'static' added.
exactly! Sorry that I've missed the missing static and thanks for catching that Kees.
Signature · 2 lines
Michal Hocko SUSE Labs
Quoted 32 lines · Vlastimil Babka
Quoted 3 lines · Kees Cook
The mutex in
sysctl_vm_numa_stat_handler()needs to be a global static, not a stack variable, otherwise it doesn't serve any purpose. Also, reading the file withCONFIG_LOCKDEP=ywill complain:Oops, good catch.
Quoted 25 lines
[ 63.258593] INFO: trying to register non-static key. [ 63.259113] the code is fine but needs lockdep annotation. [ 63.259596] turning off the locking correctness validator. [ 63.260073] CPU: 1 PID: 4102 Comm: perl Not tainted 4.14.0-rc8-next-20171107+ #419 [ 63.260769] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Ubuntu-1.8.2-1ubuntu1 04/01/2014 [ 63.261570] Call Trace: [ 63.261783] dump_stack+0x5f/0x86 [ 63.262062] register_lock_class+0xe4/0x550 [ 63.262408] ? __lock_acquire+0x308/0x1170 [ 63.262746] __lock_acquire+0x7e/0x1170 [ 63.263063] lock_acquire+0x9d/0x1d0 [ 63.263363] ? sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.263777] ? sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.264192] __mutex_lock+0xb8/0x9a0 [ 63.264488] ? sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.264942] ? sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.265398] ? sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.265840] sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.266270] proc_sys_call_handler+0xe3/0x100 [ 63.266655] __vfs_read+0x33/0x1b0 [ 63.266957] vfs_read+0xa6/0x150 [ 63.267244] SyS_read+0x55/0xc0 [ 63.267525] do_syscall_64+0x56/0x140 [ 63.267850] entry_SYSCALL64_slow_path+0x25/0x25
Quoted 4 lines
Quoted 1 line
Fixes 920d5f77d1a25 ("mm, sysctl: make NUMA stats configurable")Note that this hash is specific to particular next-$DATE as mmotm is reimported each day.
Ah yes, duh. :)
Quoted 19 lines
Quoted 19 lines
Cc Jesper Dangaard Brouer <[email protected]>Cc Dave Hansen <[email protected]>Cc Ying Huang <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Michal Hocko <[email protected]>Signed-off-by Kees Cook <[email protected]>--- mm/vmstat.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)mm/vmstat.c +0
-0@@ -72,11 +72,12 @@ static void invalid_numa_statistics(void)zero_global_numa_counters (); }
Quoted 13 lines
mm/vmstat.c static void invalid_numa_statistics(void) · line 79
Quoted 7 lines
mm/vmstat.c +2
-1+static DEFINE_MUTEX (vm_numa_stat_lock); +int sysctl_vm_numa_stat_handler (struct ctl_table *table,int write,void __user *buffer,size_t *length,loff_t *ppos) {int ret, oldval; -DEFINE_MUTEX (vm_numa_stat_lock);Yeah it was Michal who suggested scoping the mutex here instead of global scope, but I think he didn't mean to remove the 'static' qualifier, and we both missed that in the review :(
So the scope undersysctl_vm_numa_stat_handler()should be okay, just with the 'static' added.
That part is a matter of taste, I guess. :) But yes, static is important.
-Kees
Signature · 2 lines
Kees Cook Pixel Security
Quoted 4 lines · Kees Cook
The mutex in
sysctl_vm_numa_stat_handler()needs to be a global static, not a stack variable, otherwise it doesn't serve any purpose. Also, reading the file withCONFIG_LOCKDEP=ywill complain:
It's my mistake. Kees, thanks for catching it.
Quoted 55 lines
[ 63.258593] INFO: trying to register non-static key. [ 63.259113] the code is fine but needs lockdep annotation. [ 63.259596] turning off the locking correctness validator. [ 63.260073] CPU: 1 PID: 4102 Comm: perl Not tainted 4.14.0-rc8-next-20171107+ #419 [ 63.260769] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Ubuntu-1.8.2-1ubuntu1 04/01/2014 [ 63.261570] Call Trace: [ 63.261783] dump_stack+0x5f/0x86 [ 63.262062] register_lock_class+0xe4/0x550 [ 63.262408] ? __lock_acquire+0x308/0x1170 [ 63.262746] __lock_acquire+0x7e/0x1170 [ 63.263063] lock_acquire+0x9d/0x1d0 [ 63.263363] ? sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.263777] ? sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.264192] __mutex_lock+0xb8/0x9a0 [ 63.264488] ? sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.264942] ? sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.265398] ? sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.265840] sysctl_vm_numa_stat_handler+0x8f/0x2d0 [ 63.266270] proc_sys_call_handler+0xe3/0x100 [ 63.266655] __vfs_read+0x33/0x1b0 [ 63.266957] vfs_read+0xa6/0x150 [ 63.267244] SyS_read+0x55/0xc0 [ 63.267525] do_syscall_64+0x56/0x140 [ 63.267850] entry_SYSCALL64_slow_path+0x25/0x25Fixes 920d5f77d1a25 ("mm, sysctl: make NUMA stats configurable")Cc Jesper Dangaard Brouer <[email protected]>Cc Dave Hansen <[email protected]>Cc Ying Huang <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Michal Hocko <[email protected]>Signed-off-by Kees Cook <[email protected]>--- mm/vmstat.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)mm/vmstat.c +2
-1@@ -72,11 +72,12 @@ static void invalid_numa_statistics(void)zero_global_numa_counters (); } +static DEFINE_MUTEX (vm_numa_stat_lock); +int sysctl_vm_numa_stat_handler (struct ctl_table *table,int write,void __user *buffer,size_t *length,loff_t *ppos) {int ret, oldval; -DEFINE_MUTEX (vm_numa_stat_lock);mutex_lock (&vm_numa_stat_lock);if (write)
Quoted 2 lines · Kees Cook
Quoted 1 line · Vlastimil Babka
On 11/07/2017 10:38 PM, Kees Cook wrote:
[...]
Quoted 14 lines
Quoted 5 lines
Quoted 5 lines
mm/vmstat.c +2
-0+static DEFINE_MUTEX (vm_numa_stat_lock); +int sysctl_vm_numa_stat_handler (struct ctl_table *table,int write,void __user *buffer,size_t *length,loff_t *ppos) {Quoted 8 lines
Quoted 2 lines
int ret, oldval;-
DEFINE_MUTEX(vm_numa_stat_lock);Yeah it was Michal who suggested scoping the mutex here instead of global scope, but I think he didn't mean to remove the 'static' qualifier, and we both missed that in the review :(
So the scope undersysctl_vm_numa_stat_handler()should be okay, just with the 'static' added.
Quoted 1 line
That part is a matter of taste, I guess. :) But yes, static is important.
The primary reason I prefer function scope lock is that it is less tempting to reuse the lock for something else that way. But this is hardly something to insist on. So I am ok with the file scope as well.
Signature · 2 lines
Michal Hocko SUSE Labs