On Sat 10-03-18 11:52:52, Alexey Dobriyan wrote: > All it takes to open a file and read 1 byte from it. > > seq_file will be allocated along with any private allocations, > and more importantly seq file buffer which is 1 page by default. OOM badness heuristic will not know about this memory when choosing an oom victim so we might kill a wrong victim which is not all that great but at least the consumption will be contained within a memcg which sounds like the sufficient protection for now. This is the case basically for all kmem charges...
The seq file life time should be bound to a process life time so this is OK. > Signed-off-by: Alexey Dobriyan <[email protected]> Acked-by: Michal Hocko <[email protected]> > --- > > fs/seq_file.c | 8 ++++---- > 1 file changed, 4 insertions(+), 4 deletions(-) > > --- a/fs/seq_file.c > +++ b/fs/seq_file.c > @@ -29,7 +29,7 @@ static void seq_set_overflow(struct seq_file *m) > > static void *seq_buf_alloc(unsigned long size) > { > - return kvmalloc(size, GFP_KERNEL); > + return kvmalloc(size, GFP_KERNEL_ACCOUNT); > } > > /** > @@ -566,7 +566,7 @@ static void single_stop(struct seq_file *p, void *v) > int single_open(struct file *file, int (*show)(struct seq_file *, void *), > void *data) > { > - struct seq_operations *op = kmalloc(sizeof(*op), GFP_KERNEL); > + struct seq_operations *op = kmalloc(sizeof(*op), GFP_KERNEL_ACCOUNT); > int res = -ENOMEM; > > if (op) { > @@ -628,7 +628,7 @@ void *__seq_open_private(struct file *f, const struct > seq_operations *ops, > void *private; > struct seq_file *seq; > > - private = kzalloc(psize, GFP_KERNEL); > + private = kzalloc(psize, GFP_KERNEL_ACCOUNT); > if (private == NULL) > goto out; > > @@ -1112,5 +1112,5 @@ EXPORT_SYMBOL(seq_hlist_next_percpu); > > void __init seq_file_init(void) > { > - seq_file_cache = KMEM_CACHE(seq_file, SLAB_PANIC); > + seq_file_cache = KMEM_CACHE(seq_file, SLAB_ACCOUNT|SLAB_PANIC); > } -- Michal Hocko SUSE Labs
