[Devel] [PATCH RH8 1/2] ext4: Care about reserves in case of IO thread
Kirill Tkhai
ktkhai at virtuozzo.com
Tue Jun 15 13:19:17 MSK 2021
Kthread has root permittions, and it may exceed
ext4 reserves (see ext4_has_free_clusters(),
also see tune2fs -r).
Introduce a new flag PF_IO_THREAD, which indicates
a thread of *loop blocks devices, and care about
it in ext4_has_free_clusters().
Loop is the first user.
Note, that ms kernel (06.2021) still has free PF_xxx
bits, so I choosed to use one in 4.18.
Signed-off-by: Kirill Tkhai <ktkhai at virtuozzo.com>
---
drivers/block/loop.c | 2 +-
fs/ext4/balloc.c | 4 ++--
include/linux/sched.h | 1 +
3 files changed, 4 insertions(+), 3 deletions(-)
diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index 37ec5aba9ed2..86814c9d30af 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -925,7 +925,7 @@ static void loop_unprepare_queue(struct loop_device *lo)
static int loop_kthread_worker_fn(void *worker_ptr)
{
- current->flags |= PF_LESS_THROTTLE | PF_MEMALLOC_NOIO;
+ current->flags |= PF_LESS_THROTTLE | PF_MEMALLOC_NOIO | PF_IO_THREAD;
return kthread_worker_fn(worker_ptr);
}
diff --git a/fs/ext4/balloc.c b/fs/ext4/balloc.c
index 96dd3acd7204..ec2625552221 100644
--- a/fs/ext4/balloc.c
+++ b/fs/ext4/balloc.c
@@ -582,9 +582,9 @@ static int ext4_has_free_clusters(struct ext4_sb_info *sbi,
return 1;
/* Hm, nope. Are (enough) root reserved clusters available? */
- if (uid_eq(sbi->s_resuid, current_fsuid()) ||
+ if (((uid_eq(sbi->s_resuid, current_fsuid()) ||
(!gid_eq(sbi->s_resgid, GLOBAL_ROOT_GID) && in_group_p(sbi->s_resgid)) ||
- capable(CAP_SYS_RESOURCE) ||
+ capable(CAP_SYS_RESOURCE)) && !(current->flags & PF_IO_THREAD)) ||
(flags & EXT4_MB_USE_ROOT_BLOCKS)) {
if (free_clusters >= (nclusters + dirty_clusters +
diff --git a/include/linux/sched.h b/include/linux/sched.h
index f8a0e14ece41..99c786ad946e 100644
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -1445,6 +1445,7 @@ extern struct pid *cad_pid;
*/
#define PF_IDLE 0x00000002 /* I am an IDLE thread */
#define PF_EXITING 0x00000004 /* Getting shut down */
+#define PF_IO_THREAD 0x00000008 /* I'm IO thread */
#define PF_VCPU 0x00000010 /* I'm a virtual CPU */
#define PF_WQ_WORKER 0x00000020 /* I'm a workqueue worker */
#define PF_FORKNOEXEC 0x00000040 /* Forked but didn't exec */
More information about the Devel
mailing list