Re: [PATCH] mm: introduce kv[mz]alloc helpers

From: Michal Hocko
Date: Fri Jan 06 2017 - 10:10:46 EST


On Fri 06-01-17 15:36:04, Vlastimil Babka wrote:
> On 01/04/2017 03:20 PM, Michal Hocko wrote:
> > diff --git a/net/netfilter/x_tables.c b/net/netfilter/x_tables.c
> > index 2ff499680cc6..0a5cc1237afe 100644
> > --- a/net/netfilter/x_tables.c
> > +++ b/net/netfilter/x_tables.c
> > @@ -712,17 +712,8 @@ EXPORT_SYMBOL(xt_check_entry_offsets);
> > */
> > unsigned int *xt_alloc_entry_offsets(unsigned int size)
> > {
> > - unsigned int *off;
> > + return kvmalloc(size * sizeof(unsigned int), GFP_KERNEL);;
> >
> > - off = kcalloc(size, sizeof(unsigned int), GFP_KERNEL | __GFP_NOWARN);
> > -
> > - if (off)
> > - return off;
> > -
> > - if (size < (SIZE_MAX / sizeof(unsigned int)))
> > - off = vmalloc(size * sizeof(unsigned int));
> > -
> > - return off;
>
> This one seems to have tried hard to avoid the multiplication overflow
> by using kcalloc() and doing the size check before vmalloc(), so I
> wonder if it's safe to just remove the checks completely?

Tetsuo has already pointed that out and it is fixed in my loacal
version.

> > }
> > EXPORT_SYMBOL(xt_alloc_entry_offsets);
> >
> > diff --git a/net/sched/sch_fq.c b/net/sched/sch_fq.c
> > index 86309a3156a5..5678eff40f61 100644
> > --- a/net/sched/sch_fq.c
> > +++ b/net/sched/sch_fq.c
> > @@ -624,16 +624,6 @@ static void fq_rehash(struct fq_sched_data *q,
> > q->stat_gc_flows += fcnt;
> > }
> >
> > -static void *fq_alloc_node(size_t sz, int node)
> > -{
> > - void *ptr;
> > -
> > - ptr = kmalloc_node(sz, GFP_KERNEL | __GFP_REPEAT | __GFP_NOWARN, node);
>
> Another patch 3 material?

I will just drop this part for now. c3bd85495aef6 doesn't say a word about why
it needs __GFP_REPEAT. I will ask Eric.

> > - if (!ptr)
> > - ptr = vmalloc_node(sz, node);
> > - return ptr;
> > -}
> > -
> > static void fq_free(void *addr)
> > {
> > kvfree(addr);
> > @@ -650,7 +640,7 @@ static int fq_resize(struct Qdisc *sch, u32 log)
> > return 0;
> >
> > /* If XPS was setup, we can allocate memory on right NUMA node */
> > - array = fq_alloc_node(sizeof(struct rb_root) << log,
> > + array = kvmalloc_node(sizeof(struct rb_root) << log, GFP_KERNEL,
> > netdev_queue_numa_node_read(sch->dev_queue));
> > if (!array)
> > return -ENOMEM;
>
> With that fixed,
>
> Acked-by: Vlastimil Babka <vbabka@xxxxxxx>

Thanks!

--
Michal Hocko
SUSE Labs