swapfile: initialize spinlock for swap_cluster_info
Subsystems (2)
PATCH 1/1 swapfile: initialize spinlock for swap_cluster_info
We changed swap_cluster_info lock from bit_spin_lock to spinlock so we need to initialize the spinlock before the using. Otherwise, lockdep is broken.
---
Andrew,
I think it's no worth to add this patch to separate commit.
If you don't mind, it's okay to fold this patch to mm-swap-add-cluster-lock-v5. Thanks.
mm/swapfile.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-)
mm/swapfile.c +7 -2
@@ -2762,6 +2762,7 @@ SYSCALL_DEFINE2(swapon, const char __user *, specialfile, int, swap_flags)if (p->bdev &&blk_queue_nonrot (bdev_get_queue (p->bdev))) {int cpu; +unsigned long ci, nr_cluster; p->flags |=SWP_SOLIDSTATE ;/*
@@ -2769,13 +2770,17 @@ SYSCALL_DEFINE2(swapon, const char __user *, specialfile, int, swap_flags)* SSD */ p->cluster_next =1 + (prandom_u32 () % p->highest_bit); + nr_cluster =DIV_ROUND_UP (maxpages,SWAPFILE_CLUSTER ); - cluster_info =vzalloc (DIV_ROUND_UP (maxpages, -SWAPFILE_CLUSTER ) *sizeof (*cluster_info)); + cluster_info =vzalloc (nr_cluster *sizeof (*cluster_info));if (!cluster_info) { error = -ENOMEM ; goto bad_swap; } + +for (ci =0 ; ci < nr_cluster; ci++) +spin_lock_init (&((cluster_info + ci)->lock)); + p->percpu_cluster = alloc_percpu(struct percpu_cluster);if (!p->percpu_cluster) { error = -ENOMEM ;
Signature · 7 lines
2.7.4 -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
Hi, Minchan,
Quoted 8 lines · Minchan Kim
We changed
swap_cluster_infolock frombit_spin_lockto spinlock so we need to initialize the spinlock before the using. Otherwise, lockdep is broken.Cc Tim Chen <[email protected]>Cc Huang Ying <[email protected]>Cc Hugh Dickins <[email protected]>Signed-off-by Minchan Kim <[email protected]>
Good catch! Thanks a lot for your fixing!
Best Regards,
Huang, Ying
Quoted 31 lines
---
Andrew,
I think it's no worth to add this patch to separate commit.
If you don't mind, it's okay to fold this patch to mm-swap-add-cluster-lock-v5. Thanks.mm/swapfile.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-)mm/swapfile.c +3
-2@@ -2762,6 +2762,7 @@ SYSCALL_DEFINE2(swapon, const char __user *, specialfile, int, swap_flags)if (p->bdev &&blk_queue_nonrot (bdev_get_queue (p->bdev))) {int cpu; +unsigned long ci, nr_cluster; p->flags |=SWP_SOLIDSTATE ;/* @@ -2769,13 +2770,17 @@ SYSCALL_DEFINE2(swapon, const char __user *, specialfile, int, swap_flags)* SSD */ p->cluster_next =1 + (prandom_u32 () % p->highest_bit); + nr_cluster =DIV_ROUND_UP (maxpages,SWAPFILE_CLUSTER ); - cluster_info =vzalloc (DIV_ROUND_UP (maxpages, -SWAPFILE_CLUSTER ) *sizeof (*cluster_info)); + cluster_info =vzalloc (nr_cluster *sizeof (*cluster_info));if (!cluster_info) {
mm/swapfile.c SYSCALL_DEFINE2(swapon, const char __user *, specialfile, int, swap_flags) · line 2786
Quoted 10 lines
mm/swapfile.c +4
-0error = -ENOMEM ; goto bad_swap; } + +for (ci =0 ; ci < nr_cluster; ci++) +spin_lock_init (&((cluster_info + ci)->lock)); + p->percpu_cluster = alloc_percpu(struct percpu_cluster);if (!p->percpu_cluster) { error = -ENOMEM ;
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to [email protected]. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>