Re: [PATCH 24/94] radix tree test suite: Add keme_cache_alloc_bulk() support

From: Liam Howlett
Date: Fri May 28 2021 - 15:29:45 EST


* Suren Baghdasaryan <surenb@xxxxxxxxxx> [210528 14:17]:
> On Wed, Apr 28, 2021 at 8:36 AM Liam Howlett <liam.howlett@xxxxxxxxxx> wrote:
> >
> > Signed-off-by: Liam R. Howlett <Liam.Howlett@xxxxxxxxxx>
> > ---
> > tools/testing/radix-tree/linux.c | 51 +++++++++++++++++++++++++++
> > tools/testing/radix-tree/linux/slab.h | 1 +
> > 2 files changed, 52 insertions(+)
> >
> > diff --git a/tools/testing/radix-tree/linux.c b/tools/testing/radix-tree/linux.c
> > index 380bbc0a48d6..fb19a40ebb46 100644
> > --- a/tools/testing/radix-tree/linux.c
> > +++ b/tools/testing/radix-tree/linux.c
> > @@ -99,6 +99,57 @@ void kmem_cache_free_bulk(struct kmem_cache *cachep, size_t size, void **list)
> > for (int i = 0; i < size; i++)
> > kmem_cache_free(cachep, list[i]);
> > }
> > +int kmem_cache_alloc_bulk(struct kmem_cache *cachep, gfp_t gfp, size_t size,
> > + void **p)
> > +{
> > + size_t i;
> > +
> > + if (kmalloc_verbose)
> > + printk("Bulk alloc %lu\n", size);
> > +
> > + if (!(gfp & __GFP_DIRECT_RECLAIM) && cachep->non_kernel < size)
> > + return 0;
> > +
> > + if (!(gfp & __GFP_DIRECT_RECLAIM))
> > + cachep->non_kernel -= size;
> > +
> > + pthread_mutex_lock(&cachep->lock);
> > + if (cachep->nr_objs >= size) {
> > + struct radix_tree_node *node = cachep->objs;
> > +
>
> I don't think the loop below is correct because "node" is not being
> changed on each iteration:
>
> > + for (i = 0; i < size; i++) {
> > + cachep->nr_objs--;
> > + cachep->objs = node->parent;
>
> In the above assignment cachep->objs will be assigned the same value
> on all iterations.
>
> > + p[i] = cachep->objs;
>
> p[0] should point to the node, however here it would point to the node->parent.
>
> > + }
> > + pthread_mutex_unlock(&cachep->lock);
> > + node->parent = NULL;
>
> here you terminated the original cachep->objs which is not even inside
> the "p" list at this point (it was skipped).

I just verified that this code wasn't hit in my current test code. I
will test and fix this. Good catch.

>
> > + } else {
> > + pthread_mutex_unlock(&cachep->lock);
> > + for (i = 0; i < size; i++) {
> > + if (cachep->align) {
> > + posix_memalign(&p[i], cachep->align,
> > + cachep->size * size);
> > + } else {
> > + p[i] = malloc(cachep->size * size);
> > + }
> > + if (cachep->ctor)
> > + cachep->ctor(p[i]);
> > + else if (gfp & __GFP_ZERO)
> > + memset(p[i], 0, cachep->size);
> > + }
> > + }
> > +
> > + for (i = 0; i < size; i++) {
> > + uatomic_inc(&nr_allocated);
> > + uatomic_inc(&nr_tallocated);
>
> I don't see nr_tallocated even in linux-next branch. Was it introduced
> in one of the previous patches and I missed it?

It was introduced with the maple tree itself. I will spin this off as
its own patch with the same edits as nr_allocated.

>
> > + if (kmalloc_verbose)
> > + printf("Allocating %p from slab\n", p[i]);
> > + }
> > +
> > + return size;
> > +}
> > +
> >
> > void *kmalloc(size_t size, gfp_t gfp)
> > {
> > diff --git a/tools/testing/radix-tree/linux/slab.h b/tools/testing/radix-tree/linux/slab.h
> > index 53b79c15b3a2..ba42b8cc11d0 100644
> > --- a/tools/testing/radix-tree/linux/slab.h
> > +++ b/tools/testing/radix-tree/linux/slab.h
> > @@ -25,4 +25,5 @@ struct kmem_cache *kmem_cache_create(const char *name, unsigned int size,
> > void (*ctor)(void *));
> >
> > void kmem_cache_free_bulk(struct kmem_cache *cachep, size_t, void **);
> > +int kmem_cache_alloc_bulk(struct kmem_cache *cachep, gfp_t, size_t, void **);
> > #endif /* SLAB_H */
> > --
> > 2.30.2