[PATCH v2 0/3] virtio support cache indirect desc

[PATCH v2 0/3] virtio support cache indirect desc

[Xuan Zhuo]

If the VIRTIO_RING_F_INDIRECT_DESC negotiation succeeds, and the number
of sgs used for sending packets is greater than 1. We must constantly
call __kmalloc/kfree to allocate/release desc.

In the case of extremely fast package delivery, the overhead cannot be
ignored:

  27.46%  [kernel]  [k] virtqueue_add
  16.66%  [kernel]  [k] detach_buf_split
  16.51%  [kernel]  [k] virtnet_xsk_xmit
  14.04%  [kernel]  [k] virtqueue_add_outbuf
   5.18%  [kernel]  [k] __kmalloc
   4.08%  [kernel]  [k] kfree
   2.80%  [kernel]  [k] virtqueue_get_buf_ctx
   2.22%  [kernel]  [k] xsk_tx_peek_desc
   2.08%  [kernel]  [k] memset_erms
   0.83%  [kernel]  [k] virtqueue_kick_prepare
   0.76%  [kernel]  [k] virtnet_xsk_run
   0.62%  [kernel]  [k] __free_old_xmit_ptr
   0.60%  [kernel]  [k] vring_map_one_sg
   0.53%  [kernel]  [k] native_apic_mem_write
   0.46%  [kernel]  [k] sg_next
   0.43%  [kernel]  [k] sg_init_table
   0.41%  [kernel]  [k] kmalloc_slab

This patch adds a cache function to virtio to cache these allocated indirect
desc instead of constantly allocating and releasing desc.

v2:
  use struct list_head to cache the desc

Xuan Zhuo (3):
  virtio: cache indirect desc for split
  virtio: cache indirect desc for packed
  virtio-net: enable virtio indirect cache

 drivers/net/virtio_net.c     |  12 ++++
 drivers/virtio/virtio.c      |   6 ++
 drivers/virtio/virtio_ring.c | 104 +++++++++++++++++++++++++++++------
 include/linux/virtio.h       |  14 +++++
 4 files changed, 119 insertions(+), 17 deletions(-)

--
2.31.0

[PATCH v2 1/3] virtio: cache indirect desc for split

[Xuan Zhuo]

In the case of using indirect, indirect desc must be allocated and
released each time, which increases a lot of cpu overhead.

Here, a cache is added for indirect. If the number of indirect desc to be
applied for is less than VIRT_QUEUE_CACHE_DESC_NUM, the desc array with
the size of VIRT_QUEUE_CACHE_DESC_NUM is fixed and cached for reuse.

Signed-off-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
---
 drivers/virtio/virtio.c      |  6 +++
 drivers/virtio/virtio_ring.c | 77 ++++++++++++++++++++++++++++++++----
 include/linux/virtio.h       | 14 +++++++
 3 files changed, 89 insertions(+), 8 deletions(-)

diff --git a/drivers/virtio/virtio.c b/drivers/virtio/virtio.c
index 0a5b54034d4b..1047149ac2a4 100644
--- a/drivers/virtio/virtio.c
+++ b/drivers/virtio/virtio.c
@@ -431,6 +431,12 @@ bool is_virtio_device(struct device *dev)
 }
 EXPORT_SYMBOL_GPL(is_virtio_device);
 
+void virtio_set_desc_cache(struct virtio_device *dev, u32 thr)
+{
+	dev->desc_cache_thr = thr;
+}
+EXPORT_SYMBOL_GPL(virtio_set_desc_cache);
+
 void unregister_virtio_device(struct virtio_device *dev)
 {
 	int index = dev->index; /* save for after device release */
diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c
index dd95dfd85e98..0ebcd4f12d3b 100644
--- a/drivers/virtio/virtio_ring.c
+++ b/drivers/virtio/virtio_ring.c
@@ -117,6 +117,15 @@ struct vring_virtqueue {
 	/* Hint for event idx: already triggered no need to disable. */
 	bool event_triggered;
 
+	/* desc cache threshold
+	 *    0   - disable desc cache
+	 *    > 0 - enable desc cache. As the threshold of the desc cache.
+	 */
+	u32 desc_cache_thr;
+
+	/* desc cache chain */
+	struct list_head desc_cache;
+
 	union {
 		/* Available for split ring */
 		struct {
@@ -423,7 +432,53 @@ static unsigned int vring_unmap_one_split(const struct vring_virtqueue *vq,
 	return extra[i].next;
 }
 
-static struct vring_desc *alloc_indirect_split(struct virtqueue *_vq,
+static void desc_cache_free(struct list_head *head)
+{
+	struct list_head *n, *pos;
+
+	BUILD_BUG_ON(sizeof(struct list_head) > sizeof(struct vring_desc));
+	BUILD_BUG_ON(sizeof(struct list_head) > sizeof(struct vring_packed_desc));
+
+	list_for_each_prev_safe(pos, n, head)
+		kfree(pos);
+}
+
+static void __desc_cache_put(struct vring_virtqueue *vq,
+			     struct list_head *node, int n)
+{
+	if (n <= vq->desc_cache_thr)
+		list_add(node, &vq->desc_cache);
+	else
+		kfree(node);
+}
+
+#define desc_cache_put(vq, desc, n) \
+	__desc_cache_put(vq, (struct list_head *)desc, n)
+
+static void *desc_cache_get(struct vring_virtqueue *vq,
+			    int size, int n, gfp_t gfp)
+{
+	struct list_head *node;
+
+	if (n > vq->desc_cache_thr)
+		return kmalloc_array(n, size, gfp);
+
+	if (!list_empty(&vq->desc_cache)) {
+		node = vq->desc_cache.next;
+		list_del(node);
+		return node;
+	}
+
+	return kmalloc_array(vq->desc_cache_thr, size, gfp);
+}
+
+#define _desc_cache_get(vq, n, gfp, tp) \
+	((tp *)desc_cache_get(vq, (sizeof(tp)), n, gfp))
+
+#define desc_cache_get_split(vq, n, gfp) \
+	_desc_cache_get(vq, n, gfp, struct vring_desc)
+
+static struct vring_desc *alloc_indirect_split(struct vring_virtqueue *vq,
 					       unsigned int total_sg,
 					       gfp_t gfp)
 {
@@ -437,12 +492,12 @@ static struct vring_desc *alloc_indirect_split(struct virtqueue *_vq,
 	 */
 	gfp &= ~__GFP_HIGHMEM;
 
-	desc = kmalloc_array(total_sg, sizeof(struct vring_desc), gfp);
+	desc = desc_cache_get_split(vq, total_sg, gfp);
 	if (!desc)
 		return NULL;
 
 	for (i = 0; i < total_sg; i++)
-		desc[i].next = cpu_to_virtio16(_vq->vdev, i + 1);
+		desc[i].next = cpu_to_virtio16(vq->vq.vdev, i + 1);
 	return desc;
 }
 
@@ -508,7 +563,7 @@ static inline int virtqueue_add_split(struct virtqueue *_vq,
 	head = vq->free_head;
 
 	if (virtqueue_use_indirect(_vq, total_sg))
-		desc = alloc_indirect_split(_vq, total_sg, gfp);
+		desc = alloc_indirect_split(vq, total_sg, gfp);
 	else {
 		desc = NULL;
 		WARN_ON_ONCE(total_sg > vq->split.vring.num && !vq->indirect);
@@ -652,7 +707,7 @@ static inline int virtqueue_add_split(struct virtqueue *_vq,
 	}
 
 	if (indirect)
-		kfree(desc);
+		desc_cache_put(vq, desc, total_sg);
 
 	END_USE(vq);
 	return -ENOMEM;
@@ -717,7 +772,7 @@ static void detach_buf_split(struct vring_virtqueue *vq, unsigned int head,
 	if (vq->indirect) {
 		struct vring_desc *indir_desc =
 				vq->split.desc_state[head].indir_desc;
-		u32 len;
+		u32 len, n;
 
 		/* Free the indirect table, if any, now that it's unmapped. */
 		if (!indir_desc)
@@ -729,10 +784,12 @@ static void detach_buf_split(struct vring_virtqueue *vq, unsigned int head,
 				VRING_DESC_F_INDIRECT));
 		BUG_ON(len == 0 || len % sizeof(struct vring_desc));
 
-		for (j = 0; j < len / sizeof(struct vring_desc); j++)
+		n = len / sizeof(struct vring_desc);
+
+		for (j = 0; j < n; j++)
 			vring_unmap_one_split_indirect(vq, &indir_desc[j]);
 
-		kfree(indir_desc);
+		desc_cache_put(vq, indir_desc, n);
 		vq->split.desc_state[head].indir_desc = NULL;
 	} else if (ctx) {
 		*ctx = vq->split.desc_state[head].indir_desc;
@@ -2199,6 +2256,9 @@ struct virtqueue *__vring_new_virtqueue(unsigned int index,
 	vq->indirect = virtio_has_feature(vdev, VIRTIO_RING_F_INDIRECT_DESC) &&
 		!context;
 	vq->event = virtio_has_feature(vdev, VIRTIO_RING_F_EVENT_IDX);
+	vq->desc_cache_thr = vdev->desc_cache_thr;
+
+	INIT_LIST_HEAD(&vq->desc_cache);
 
 	if (virtio_has_feature(vdev, VIRTIO_F_ORDER_PLATFORM))
 		vq->weak_barriers = false;
@@ -2329,6 +2389,7 @@ void vring_del_virtqueue(struct virtqueue *_vq)
 	if (!vq->packed_ring) {
 		kfree(vq->split.desc_state);
 		kfree(vq->split.desc_extra);
+		desc_cache_free(&vq->desc_cache);
 	}
 	kfree(vq);
 }
diff --git a/include/linux/virtio.h b/include/linux/virtio.h
index 41edbc01ffa4..bda6f9853e97 100644
--- a/include/linux/virtio.h
+++ b/include/linux/virtio.h
@@ -118,6 +118,7 @@ struct virtio_device {
 	struct list_head vqs;
 	u64 features;
 	void *priv;
+	u32 desc_cache_thr;
 };
 
 static inline struct virtio_device *dev_to_virtio(struct device *_dev)
@@ -130,6 +131,19 @@ int register_virtio_device(struct virtio_device *dev);
 void unregister_virtio_device(struct virtio_device *dev);
 bool is_virtio_device(struct device *dev);
 
+/**
+ * virtio_set_desc_cache - set virtio ring desc cache threshold
+ *
+ * virtio will cache the allocated indirect desc.
+ *
+ * This function must be called before find_vqs.
+ *
+ * @thr:
+ *    0   - disable desc cache
+ *    > 0 - enable desc cache. As the threshold of the desc cache.
+ */
+void virtio_set_desc_cache(struct virtio_device *dev, u32 thr);
+
 void virtio_break_device(struct virtio_device *dev);
 
 void virtio_config_changed(struct virtio_device *dev);
-- 
2.31.0

[PATCH v2 2/3] virtio: cache indirect desc for packed

[Xuan Zhuo]

In the case of using indirect, indirect desc must be allocated and
released each time, which increases a lot of cpu overhead.

Here, a cache is added for indirect. If the number of indirect desc to be
applied for is less than VIRT_QUEUE_CACHE_DESC_NUM, the desc array with
the size of VIRT_QUEUE_CACHE_DESC_NUM is fixed and cached for reuse.

Signed-off-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
---
 drivers/virtio/virtio_ring.c | 29 +++++++++++++++++++----------
 1 file changed, 19 insertions(+), 10 deletions(-)

diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c
index 0ebcd4f12d3b..e6d1985a87a8 100644
--- a/drivers/virtio/virtio_ring.c
+++ b/drivers/virtio/virtio_ring.c
@@ -1089,7 +1089,11 @@ static void vring_unmap_desc_packed(const struct vring_virtqueue *vq,
 	}
 }
 
-static struct vring_packed_desc *alloc_indirect_packed(unsigned int total_sg,
+#define desc_cache_get_packed(vq, n, gfp) \
+	_desc_cache_get(vq, n, gfp, struct vring_packed_desc)
+
+static struct vring_packed_desc *alloc_indirect_packed(struct vring_virtqueue *vq,
+						       unsigned int total_sg,
 						       gfp_t gfp)
 {
 	struct vring_packed_desc *desc;
@@ -1101,7 +1105,7 @@ static struct vring_packed_desc *alloc_indirect_packed(unsigned int total_sg,
 	 */
 	gfp &= ~__GFP_HIGHMEM;
 
-	desc = kmalloc_array(total_sg, sizeof(struct vring_packed_desc), gfp);
+	desc = desc_cache_get_packed(vq, total_sg, gfp);
 
 	return desc;
 }
@@ -1121,7 +1125,7 @@ static int virtqueue_add_indirect_packed(struct vring_virtqueue *vq,
 	dma_addr_t addr;
 
 	head = vq->packed.next_avail_idx;
-	desc = alloc_indirect_packed(total_sg, gfp);
+	desc = alloc_indirect_packed(vq, total_sg, gfp);
 
 	if (unlikely(vq->vq.num_free < 1)) {
 		pr_debug("Can't add buf len 1 - avail = 0\n");
@@ -1212,7 +1216,7 @@ static int virtqueue_add_indirect_packed(struct vring_virtqueue *vq,
 	for (i = 0; i < err_idx; i++)
 		vring_unmap_desc_packed(vq, &desc[i]);
 
-	kfree(desc);
+	desc_cache_put(vq, desc, total_sg);
 
 	END_USE(vq);
 	return -ENOMEM;
@@ -1437,20 +1441,22 @@ static void detach_buf_packed(struct vring_virtqueue *vq,
 	}
 
 	if (vq->indirect) {
-		u32 len;
+		u32 len, n;
 
 		/* Free the indirect table, if any, now that it's unmapped. */
 		desc = state->indir_desc;
 		if (!desc)
 			return;
 
+		len = vq->packed.desc_extra[id].len;
+		n = len / sizeof(struct vring_packed_desc);
+
 		if (vq->use_dma_api) {
-			len = vq->packed.desc_extra[id].len;
-			for (i = 0; i < len / sizeof(struct vring_packed_desc);
-					i++)
+			for (i = 0; i < n; i++)
 				vring_unmap_desc_packed(vq, &desc[i]);
 		}
-		kfree(desc);
+
+		desc_cache_put(vq, desc, n);
 		state->indir_desc = NULL;
 	} else if (ctx) {
 		*ctx = state->indir_desc;
@@ -1768,6 +1774,9 @@ static struct virtqueue *vring_create_virtqueue_packed(
 	vq->indirect = virtio_has_feature(vdev, VIRTIO_RING_F_INDIRECT_DESC) &&
 		!context;
 	vq->event = virtio_has_feature(vdev, VIRTIO_RING_F_EVENT_IDX);
+	vq->desc_cache_thr = vdev->desc_cache_thr;
+
+	INIT_LIST_HEAD(&vq->desc_cache);
 
 	if (virtio_has_feature(vdev, VIRTIO_F_ORDER_PLATFORM))
 		vq->weak_barriers = false;
@@ -2389,8 +2398,8 @@ void vring_del_virtqueue(struct virtqueue *_vq)
 	if (!vq->packed_ring) {
 		kfree(vq->split.desc_state);
 		kfree(vq->split.desc_extra);
-		desc_cache_free(&vq->desc_cache);
 	}
+	desc_cache_free(&vq->desc_cache);
 	kfree(vq);
 }
 EXPORT_SYMBOL_GPL(vring_del_virtqueue);
-- 
2.31.0

[PATCH v2 3/3] virtio-net: enable virtio indirect cache

[Xuan Zhuo]

If the VIRTIO_RING_F_INDIRECT_DESC negotiation succeeds, and the number
of sgs used for sending packets is greater than 1. We must constantly
call __kmalloc/kfree to allocate/release desc.

In the case of extremely fast package delivery, the overhead cannot be
ignored:

  27.46%  [kernel]  [k] virtqueue_add
  16.66%  [kernel]  [k] detach_buf_split
  16.51%  [kernel]  [k] virtnet_xsk_xmit
  14.04%  [kernel]  [k] virtqueue_add_outbuf
   5.18%  [kernel]  [k] __kmalloc
   4.08%  [kernel]  [k] kfree
   2.80%  [kernel]  [k] virtqueue_get_buf_ctx
   2.22%  [kernel]  [k] xsk_tx_peek_desc
   2.08%  [kernel]  [k] memset_erms
   0.83%  [kernel]  [k] virtqueue_kick_prepare
   0.76%  [kernel]  [k] virtnet_xsk_run
   0.62%  [kernel]  [k] __free_old_xmit_ptr
   0.60%  [kernel]  [k] vring_map_one_sg
   0.53%  [kernel]  [k] native_apic_mem_write
   0.46%  [kernel]  [k] sg_next
   0.43%  [kernel]  [k] sg_init_table
   0.41%  [kernel]  [k] kmalloc_slab

Compared to not using virtio indirect cache, virtio-net can get a 16%
performance improvement when using indirect desc cache.

In the test case, the CPU where the package is sent has reached 100%.
The following are the PPS in two cases:

    indirect desc cache  | no cache
    3074658              | 2685132
    3111866              | 2666118
    3152527              | 2653632
    3125867              | 2669820
    3027147              | 2644464
    3069211              | 2669777
    3038522              | 2675645
    3034507              | 2671302
    3102257              | 2685504
    3083712              | 2692800
    3051771              | 2676928
    3080684              | 2695040
    3147816              | 2720876
    3123887              | 2705492
    3180963              | 2699520
    3191579              | 2676480
    3161670              | 2686272
    3189768              | 2692588
    3174272              | 2686692
    3143434              | 2682416

Signed-off-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
---
 drivers/net/virtio_net.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index 4ad25a8b0870..e1ade176ab46 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -31,6 +31,13 @@ module_param(csum, bool, 0444);
 module_param(gso, bool, 0444);
 module_param(napi_tx, bool, 0644);
 
+/**
+ * Because virtio desc cache will increase memory overhead, users can turn it
+ * off or select an acceptable value. The maximum value is 2 + MAX_SKB_FRAGS.
+ */
+static u32 virtio_desc_cache_thr = 4;
+module_param(virtio_desc_cache_thr, uint, 0644);
+
 /* FIXME: MTU in config. */
 #define GOOD_PACKET_LEN (ETH_HLEN + VLAN_HLEN + ETH_DATA_LEN)
 #define GOOD_COPY_LEN	128
@@ -3214,6 +3221,11 @@ static int virtnet_probe(struct virtio_device *vdev)
 		vi->curr_queue_pairs = num_online_cpus();
 	vi->max_queue_pairs = max_queue_pairs;
 
+	if (virtio_desc_cache_thr > 2 + MAX_SKB_FRAGS)
+		virtio_set_desc_cache(vdev, 2 + MAX_SKB_FRAGS);
+	else
+		virtio_set_desc_cache(vdev, virtio_desc_cache_thr);
+
 	/* Allocate/initialize the rx/tx queues, and invoke find_vqs */
 	err = init_vqs(vi);
 	if (err)
-- 
2.31.0

Re: [PATCH v2 3/3] virtio-net: enable virtio indirect cache

[kernel test robot]

[-- Attachment #1: Type: text/plain, Size: 1976 bytes --]

Hi Xuan,

Thank you for the patch! Perhaps something to improve:

[auto build test WARNING on horms-ipvs/master]
[also build test WARNING on linus/master v5.15-rc7]
[cannot apply to mst-vhost/linux-next next-20211028]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch]

url:    https://github.com/0day-ci/linux/commits/Xuan-Zhuo/virtio-support-cache-indirect-desc/20211028-185145
base:   https://git.kernel.org/pub/scm/linux/kernel/git/horms/ipvs.git master
config: i386-randconfig-a004-20211028 (attached as .config)
compiler: gcc-9 (Debian 9.3.0-22) 9.3.0
reproduce (this is a W=1 build):
        # https://github.com/0day-ci/linux/commit/e8418946355cc294b006c6692990dae15a22d85f
        git remote add linux-review https://github.com/0day-ci/linux
        git fetch --no-tags linux-review Xuan-Zhuo/virtio-support-cache-indirect-desc/20211028-185145
        git checkout e8418946355cc294b006c6692990dae15a22d85f
        # save the attached .config to linux build tree
        make W=1 ARCH=i386 

If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>

All warnings (new ones prefixed by >>):

>> drivers/net/virtio_net.c:35: warning: This comment starts with '/**', but isn't a kernel-doc comment. Refer Documentation/doc-guide/kernel-doc.rst
    * Because virtio desc cache will increase memory overhead, users can turn it


vim +35 drivers/net/virtio_net.c

    33	
    34	/**
  > 35	 * Because virtio desc cache will increase memory overhead, users can turn it
    36	 * off or select an acceptable value. The maximum value is 2 + MAX_SKB_FRAGS.
    37	 */
    38	static u32 virtio_desc_cache_thr = 4;
    39	module_param(virtio_desc_cache_thr, uint, 0644);
    40	

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org

[-- Attachment #2: .config.gz --]
[-- Type: application/gzip, Size: 32532 bytes --]

Re: [PATCH v2 1/3] virtio: cache indirect desc for split

[Jason Wang]

On Thu, Oct 28, 2021 at 6:49 PM Xuan Zhuo <xuanzhuo@linux.alibaba.com> wrote:
>
> In the case of using indirect, indirect desc must be allocated and
> released each time, which increases a lot of cpu overhead.
>
> Here, a cache is added for indirect. If the number of indirect desc to be
> applied for is less than VIRT_QUEUE_CACHE_DESC_NUM, the desc array with
> the size of VIRT_QUEUE_CACHE_DESC_NUM is fixed and cached for reuse.

The commit log looks out of date? It's actually desc_cache_thr now.

>
> Signed-off-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
> ---
>  drivers/virtio/virtio.c      |  6 +++
>  drivers/virtio/virtio_ring.c | 77 ++++++++++++++++++++++++++++++++----
>  include/linux/virtio.h       | 14 +++++++
>  3 files changed, 89 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/virtio/virtio.c b/drivers/virtio/virtio.c
> index 0a5b54034d4b..1047149ac2a4 100644
> --- a/drivers/virtio/virtio.c
> +++ b/drivers/virtio/virtio.c
> @@ -431,6 +431,12 @@ bool is_virtio_device(struct device *dev)
>  }
>  EXPORT_SYMBOL_GPL(is_virtio_device);
>
> +void virtio_set_desc_cache(struct virtio_device *dev, u32 thr)
> +{
> +       dev->desc_cache_thr = thr;
> +}
> +EXPORT_SYMBOL_GPL(virtio_set_desc_cache);
> +
>  void unregister_virtio_device(struct virtio_device *dev)
>  {
>         int index = dev->index; /* save for after device release */
> diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c
> index dd95dfd85e98..0ebcd4f12d3b 100644
> --- a/drivers/virtio/virtio_ring.c
> +++ b/drivers/virtio/virtio_ring.c
> @@ -117,6 +117,15 @@ struct vring_virtqueue {
>         /* Hint for event idx: already triggered no need to disable. */
>         bool event_triggered;
>
> +       /* desc cache threshold
> +        *    0   - disable desc cache
> +        *    > 0 - enable desc cache. As the threshold of the desc cache.
> +        */
> +       u32 desc_cache_thr;
> +
> +       /* desc cache chain */
> +       struct list_head desc_cache;
> +
>         union {
>                 /* Available for split ring */
>                 struct {
> @@ -423,7 +432,53 @@ static unsigned int vring_unmap_one_split(const struct vring_virtqueue *vq,
>         return extra[i].next;
>  }
>
> -static struct vring_desc *alloc_indirect_split(struct virtqueue *_vq,
> +static void desc_cache_free(struct list_head *head)
> +{
> +       struct list_head *n, *pos;
> +
> +       BUILD_BUG_ON(sizeof(struct list_head) > sizeof(struct vring_desc));
> +       BUILD_BUG_ON(sizeof(struct list_head) > sizeof(struct vring_packed_desc));
> +
> +       list_for_each_prev_safe(pos, n, head)
> +               kfree(pos);
> +}
> +
> +static void __desc_cache_put(struct vring_virtqueue *vq,
> +                            struct list_head *node, int n)
> +{
> +       if (n <= vq->desc_cache_thr)
> +               list_add(node, &vq->desc_cache);
> +       else
> +               kfree(node);
> +}
> +
> +#define desc_cache_put(vq, desc, n) \
> +       __desc_cache_put(vq, (struct list_head *)desc, n)
> +
> +static void *desc_cache_get(struct vring_virtqueue *vq,
> +                           int size, int n, gfp_t gfp)
> +{
> +       struct list_head *node;
> +
> +       if (n > vq->desc_cache_thr)
> +               return kmalloc_array(n, size, gfp);
> +
> +       if (!list_empty(&vq->desc_cache)) {
> +               node = vq->desc_cache.next;
> +               list_del(node);
> +               return node;
> +       }
> +
> +       return kmalloc_array(vq->desc_cache_thr, size, gfp);

Instead of doing the if .. kmalloc_array here. I wonder if we can
simply pre-allocate per buffer indirect descriptors array.

E.g if we use MAX_SKB_FRAGS + 1, it will be ~64KB for 256 queue size
and ~272KB for 1024 queue size.

Then we avoid list manipulation and kmalloc in datapath with much more
simpler codes.

Thanks

> +}
> +
> +#define _desc_cache_get(vq, n, gfp, tp) \
> +       ((tp *)desc_cache_get(vq, (sizeof(tp)), n, gfp))
> +
> +#define desc_cache_get_split(vq, n, gfp) \
> +       _desc_cache_get(vq, n, gfp, struct vring_desc)
> +
> +static struct vring_desc *alloc_indirect_split(struct vring_virtqueue *vq,
>                                                unsigned int total_sg,
>                                                gfp_t gfp)
>  {
> @@ -437,12 +492,12 @@ static struct vring_desc *alloc_indirect_split(struct virtqueue *_vq,
>          */
>         gfp &= ~__GFP_HIGHMEM;
>
> -       desc = kmalloc_array(total_sg, sizeof(struct vring_desc), gfp);
> +       desc = desc_cache_get_split(vq, total_sg, gfp);
>         if (!desc)
>                 return NULL;
>
>         for (i = 0; i < total_sg; i++)
> -               desc[i].next = cpu_to_virtio16(_vq->vdev, i + 1);
> +               desc[i].next = cpu_to_virtio16(vq->vq.vdev, i + 1);
>         return desc;
>  }
>
> @@ -508,7 +563,7 @@ static inline int virtqueue_add_split(struct virtqueue *_vq,
>         head = vq->free_head;
>
>         if (virtqueue_use_indirect(_vq, total_sg))
> -               desc = alloc_indirect_split(_vq, total_sg, gfp);
> +               desc = alloc_indirect_split(vq, total_sg, gfp);
>         else {
>                 desc = NULL;
>                 WARN_ON_ONCE(total_sg > vq->split.vring.num && !vq->indirect);
> @@ -652,7 +707,7 @@ static inline int virtqueue_add_split(struct virtqueue *_vq,
>         }
>
>         if (indirect)
> -               kfree(desc);
> +               desc_cache_put(vq, desc, total_sg);
>
>         END_USE(vq);
>         return -ENOMEM;
> @@ -717,7 +772,7 @@ static void detach_buf_split(struct vring_virtqueue *vq, unsigned int head,
>         if (vq->indirect) {
>                 struct vring_desc *indir_desc =
>                                 vq->split.desc_state[head].indir_desc;
> -               u32 len;
> +               u32 len, n;
>
>                 /* Free the indirect table, if any, now that it's unmapped. */
>                 if (!indir_desc)
> @@ -729,10 +784,12 @@ static void detach_buf_split(struct vring_virtqueue *vq, unsigned int head,
>                                 VRING_DESC_F_INDIRECT));
>                 BUG_ON(len == 0 || len % sizeof(struct vring_desc));
>
> -               for (j = 0; j < len / sizeof(struct vring_desc); j++)
> +               n = len / sizeof(struct vring_desc);
> +
> +               for (j = 0; j < n; j++)
>                         vring_unmap_one_split_indirect(vq, &indir_desc[j]);
>
> -               kfree(indir_desc);
> +               desc_cache_put(vq, indir_desc, n);
>                 vq->split.desc_state[head].indir_desc = NULL;
>         } else if (ctx) {
>                 *ctx = vq->split.desc_state[head].indir_desc;
> @@ -2199,6 +2256,9 @@ struct virtqueue *__vring_new_virtqueue(unsigned int index,
>         vq->indirect = virtio_has_feature(vdev, VIRTIO_RING_F_INDIRECT_DESC) &&
>                 !context;
>         vq->event = virtio_has_feature(vdev, VIRTIO_RING_F_EVENT_IDX);
> +       vq->desc_cache_thr = vdev->desc_cache_thr;
> +
> +       INIT_LIST_HEAD(&vq->desc_cache);
>
>         if (virtio_has_feature(vdev, VIRTIO_F_ORDER_PLATFORM))
>                 vq->weak_barriers = false;
> @@ -2329,6 +2389,7 @@ void vring_del_virtqueue(struct virtqueue *_vq)
>         if (!vq->packed_ring) {
>                 kfree(vq->split.desc_state);
>                 kfree(vq->split.desc_extra);
> +               desc_cache_free(&vq->desc_cache);
>         }
>         kfree(vq);
>  }
> diff --git a/include/linux/virtio.h b/include/linux/virtio.h
> index 41edbc01ffa4..bda6f9853e97 100644
> --- a/include/linux/virtio.h
> +++ b/include/linux/virtio.h
> @@ -118,6 +118,7 @@ struct virtio_device {
>         struct list_head vqs;
>         u64 features;
>         void *priv;
> +       u32 desc_cache_thr;
>  };
>
>  static inline struct virtio_device *dev_to_virtio(struct device *_dev)
> @@ -130,6 +131,19 @@ int register_virtio_device(struct virtio_device *dev);
>  void unregister_virtio_device(struct virtio_device *dev);
>  bool is_virtio_device(struct device *dev);
>
> +/**
> + * virtio_set_desc_cache - set virtio ring desc cache threshold
> + *
> + * virtio will cache the allocated indirect desc.
> + *
> + * This function must be called before find_vqs.
> + *
> + * @thr:
> + *    0   - disable desc cache
> + *    > 0 - enable desc cache. As the threshold of the desc cache.
> + */
> +void virtio_set_desc_cache(struct virtio_device *dev, u32 thr);
> +
>  void virtio_break_device(struct virtio_device *dev);
>
>  void virtio_config_changed(struct virtio_device *dev);
> --
> 2.31.0
>

Re: [PATCH v2 1/3] virtio: cache indirect desc for split

[Michael S. Tsirkin]

On Thu, Oct 28, 2021 at 06:49:17PM +0800, Xuan Zhuo wrote:
> In the case of using indirect, indirect desc must be allocated and
> released each time, which increases a lot of cpu overhead.
> 
> Here, a cache is added for indirect. If the number of indirect desc to be
> applied for is less than VIRT_QUEUE_CACHE_DESC_NUM, the desc array with
> the size of VIRT_QUEUE_CACHE_DESC_NUM is fixed and cached for reuse.
> 
> Signed-off-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>

What bothers me here is what happens if cache gets
filled on one numa node, then used on another?


> ---
>  drivers/virtio/virtio.c      |  6 +++
>  drivers/virtio/virtio_ring.c | 77 ++++++++++++++++++++++++++++++++----
>  include/linux/virtio.h       | 14 +++++++
>  3 files changed, 89 insertions(+), 8 deletions(-)
> 
> diff --git a/drivers/virtio/virtio.c b/drivers/virtio/virtio.c
> index 0a5b54034d4b..1047149ac2a4 100644
> --- a/drivers/virtio/virtio.c
> +++ b/drivers/virtio/virtio.c
> @@ -431,6 +431,12 @@ bool is_virtio_device(struct device *dev)
>  }
>  EXPORT_SYMBOL_GPL(is_virtio_device);
>  
> +void virtio_set_desc_cache(struct virtio_device *dev, u32 thr)
> +{
> +	dev->desc_cache_thr = thr;
> +}
> +EXPORT_SYMBOL_GPL(virtio_set_desc_cache);
> +
>  void unregister_virtio_device(struct virtio_device *dev)
>  {
>  	int index = dev->index; /* save for after device release */
> diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c
> index dd95dfd85e98..0ebcd4f12d3b 100644
> --- a/drivers/virtio/virtio_ring.c
> +++ b/drivers/virtio/virtio_ring.c
> @@ -117,6 +117,15 @@ struct vring_virtqueue {
>  	/* Hint for event idx: already triggered no need to disable. */
>  	bool event_triggered;
>  
> +	/* desc cache threshold
> +	 *    0   - disable desc cache
> +	 *    > 0 - enable desc cache. As the threshold of the desc cache.
> +	 */
> +	u32 desc_cache_thr;

not really descriptive. also pls eschew abbreviation.

> +
> +	/* desc cache chain */
> +	struct list_head desc_cache;

hmm this puts extra pressure on cache. you never need to drop
things in the middle. llist_head would be better I
think ... no?


> +
>  	union {
>  		/* Available for split ring */
>  		struct {
> @@ -423,7 +432,53 @@ static unsigned int vring_unmap_one_split(const struct vring_virtqueue *vq,
>  	return extra[i].next;
>  }
>  
> -static struct vring_desc *alloc_indirect_split(struct virtqueue *_vq,
> +static void desc_cache_free(struct list_head *head)
> +{
> +	struct list_head *n, *pos;
> +
> +	BUILD_BUG_ON(sizeof(struct list_head) > sizeof(struct vring_desc));
> +	BUILD_BUG_ON(sizeof(struct list_head) > sizeof(struct vring_packed_desc));
> +
> +	list_for_each_prev_safe(pos, n, head)
> +		kfree(pos);
> +}
> +
> +static void __desc_cache_put(struct vring_virtqueue *vq,
> +			     struct list_head *node, int n)
> +{
> +	if (n <= vq->desc_cache_thr)
> +		list_add(node, &vq->desc_cache);
> +	else
> +		kfree(node);

this bothers me. Do we really need a full VQ's worth of
indirect descriptors? Can't we set a limit on how many
are used?


> +}
> +
> +#define desc_cache_put(vq, desc, n) \
> +	__desc_cache_put(vq, (struct list_head *)desc, n)

replace with an inline function pls. in fact we dont need
__desc_cache_put at all.


> +
> +static void *desc_cache_get(struct vring_virtqueue *vq,
> +			    int size, int n, gfp_t gfp)
> +{
> +	struct list_head *node;
> +
> +	if (n > vq->desc_cache_thr)
> +		return kmalloc_array(n, size, gfp);
> +
> +	if (!list_empty(&vq->desc_cache)) {
> +		node = vq->desc_cache.next;
> +		list_del(node);
> +		return node;
> +	}
> +
> +	return kmalloc_array(vq->desc_cache_thr, size, gfp);
> +}
> +
> +#define _desc_cache_get(vq, n, gfp, tp) \
> +	((tp *)desc_cache_get(vq, (sizeof(tp)), n, gfp))
> +
> +#define desc_cache_get_split(vq, n, gfp) \
> +	_desc_cache_get(vq, n, gfp, struct vring_desc)
> +

same thing here.

> +static struct vring_desc *alloc_indirect_split(struct vring_virtqueue *vq,
>  					       unsigned int total_sg,
>  					       gfp_t gfp)
>  {
> @@ -437,12 +492,12 @@ static struct vring_desc *alloc_indirect_split(struct virtqueue *_vq,
>  	 */
>  	gfp &= ~__GFP_HIGHMEM;
>  
> -	desc = kmalloc_array(total_sg, sizeof(struct vring_desc), gfp);
> +	desc = desc_cache_get_split(vq, total_sg, gfp);
>  	if (!desc)
>  		return NULL;
>  
>  	for (i = 0; i < total_sg; i++)
> -		desc[i].next = cpu_to_virtio16(_vq->vdev, i + 1);
> +		desc[i].next = cpu_to_virtio16(vq->vq.vdev, i + 1);
>  	return desc;
>  }
>  
> @@ -508,7 +563,7 @@ static inline int virtqueue_add_split(struct virtqueue *_vq,
>  	head = vq->free_head;
>  
>  	if (virtqueue_use_indirect(_vq, total_sg))
> -		desc = alloc_indirect_split(_vq, total_sg, gfp);
> +		desc = alloc_indirect_split(vq, total_sg, gfp);
>  	else {
>  		desc = NULL;
>  		WARN_ON_ONCE(total_sg > vq->split.vring.num && !vq->indirect);
> @@ -652,7 +707,7 @@ static inline int virtqueue_add_split(struct virtqueue *_vq,
>  	}
>  
>  	if (indirect)
> -		kfree(desc);
> +		desc_cache_put(vq, desc, total_sg);
>  
>  	END_USE(vq);
>  	return -ENOMEM;
> @@ -717,7 +772,7 @@ static void detach_buf_split(struct vring_virtqueue *vq, unsigned int head,
>  	if (vq->indirect) {
>  		struct vring_desc *indir_desc =
>  				vq->split.desc_state[head].indir_desc;
> -		u32 len;
> +		u32 len, n;
>  
>  		/* Free the indirect table, if any, now that it's unmapped. */
>  		if (!indir_desc)
> @@ -729,10 +784,12 @@ static void detach_buf_split(struct vring_virtqueue *vq, unsigned int head,
>  				VRING_DESC_F_INDIRECT));
>  		BUG_ON(len == 0 || len % sizeof(struct vring_desc));
>  
> -		for (j = 0; j < len / sizeof(struct vring_desc); j++)
> +		n = len / sizeof(struct vring_desc);
> +
> +		for (j = 0; j < n; j++)
>  			vring_unmap_one_split_indirect(vq, &indir_desc[j]);
>  
> -		kfree(indir_desc);
> +		desc_cache_put(vq, indir_desc, n);
>  		vq->split.desc_state[head].indir_desc = NULL;
>  	} else if (ctx) {
>  		*ctx = vq->split.desc_state[head].indir_desc;
> @@ -2199,6 +2256,9 @@ struct virtqueue *__vring_new_virtqueue(unsigned int index,
>  	vq->indirect = virtio_has_feature(vdev, VIRTIO_RING_F_INDIRECT_DESC) &&
>  		!context;
>  	vq->event = virtio_has_feature(vdev, VIRTIO_RING_F_EVENT_IDX);
> +	vq->desc_cache_thr = vdev->desc_cache_thr;
> +
> +	INIT_LIST_HEAD(&vq->desc_cache);
>  
>  	if (virtio_has_feature(vdev, VIRTIO_F_ORDER_PLATFORM))
>  		vq->weak_barriers = false;

So e.g. for rx, we are wasting memory since indirect isn't used.


> @@ -2329,6 +2389,7 @@ void vring_del_virtqueue(struct virtqueue *_vq)
>  	if (!vq->packed_ring) {
>  		kfree(vq->split.desc_state);
>  		kfree(vq->split.desc_extra);
> +		desc_cache_free(&vq->desc_cache);
>  	}
>  	kfree(vq);
>  }
> diff --git a/include/linux/virtio.h b/include/linux/virtio.h
> index 41edbc01ffa4..bda6f9853e97 100644
> --- a/include/linux/virtio.h
> +++ b/include/linux/virtio.h
> @@ -118,6 +118,7 @@ struct virtio_device {
>  	struct list_head vqs;
>  	u64 features;
>  	void *priv;
> +	u32 desc_cache_thr;
>  };
>  
>  static inline struct virtio_device *dev_to_virtio(struct device *_dev)
> @@ -130,6 +131,19 @@ int register_virtio_device(struct virtio_device *dev);
>  void unregister_virtio_device(struct virtio_device *dev);
>  bool is_virtio_device(struct device *dev);
>  
> +/**
> + * virtio_set_desc_cache - set virtio ring desc cache threshold
> + *
> + * virtio will cache the allocated indirect desc.
> + *
> + * This function must be called before find_vqs.
> + *
> + * @thr:
> + *    0   - disable desc cache
> + *    > 0 - enable desc cache. As the threshold of the desc cache.
> + */
> +void virtio_set_desc_cache(struct virtio_device *dev, u32 thr);
> +
>  void virtio_break_device(struct virtio_device *dev);
>  
>  void virtio_config_changed(struct virtio_device *dev);
> -- 
> 2.31.0

Re: [PATCH v2 3/3] virtio-net: enable virtio indirect cache

[Michael S. Tsirkin]

On Thu, Oct 28, 2021 at 06:49:19PM +0800, Xuan Zhuo wrote:
> If the VIRTIO_RING_F_INDIRECT_DESC negotiation succeeds, and the number
> of sgs used for sending packets is greater than 1. We must constantly
> call __kmalloc/kfree to allocate/release desc.
> 
> In the case of extremely fast package delivery, the overhead cannot be
> ignored:
> 
>   27.46%  [kernel]  [k] virtqueue_add
>   16.66%  [kernel]  [k] detach_buf_split
>   16.51%  [kernel]  [k] virtnet_xsk_xmit
>   14.04%  [kernel]  [k] virtqueue_add_outbuf
>    5.18%  [kernel]  [k] __kmalloc
>    4.08%  [kernel]  [k] kfree
>    2.80%  [kernel]  [k] virtqueue_get_buf_ctx
>    2.22%  [kernel]  [k] xsk_tx_peek_desc
>    2.08%  [kernel]  [k] memset_erms
>    0.83%  [kernel]  [k] virtqueue_kick_prepare
>    0.76%  [kernel]  [k] virtnet_xsk_run
>    0.62%  [kernel]  [k] __free_old_xmit_ptr
>    0.60%  [kernel]  [k] vring_map_one_sg
>    0.53%  [kernel]  [k] native_apic_mem_write
>    0.46%  [kernel]  [k] sg_next
>    0.43%  [kernel]  [k] sg_init_table
>    0.41%  [kernel]  [k] kmalloc_slab
> 
> Compared to not using virtio indirect cache, virtio-net can get a 16%
> performance improvement when using indirect desc cache.
> 
> In the test case, the CPU where the package is sent has reached 100%.
> The following are the PPS in two cases:
> 
>     indirect desc cache  | no cache
>     3074658              | 2685132
>     3111866              | 2666118
>     3152527              | 2653632
>     3125867              | 2669820
>     3027147              | 2644464
>     3069211              | 2669777
>     3038522              | 2675645
>     3034507              | 2671302
>     3102257              | 2685504
>     3083712              | 2692800
>     3051771              | 2676928
>     3080684              | 2695040
>     3147816              | 2720876
>     3123887              | 2705492
>     3180963              | 2699520
>     3191579              | 2676480
>     3161670              | 2686272
>     3189768              | 2692588
>     3174272              | 2686692
>     3143434              | 2682416
> 
> Signed-off-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
> ---
>  drivers/net/virtio_net.c | 12 ++++++++++++
>  1 file changed, 12 insertions(+)
> 
> diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
> index 4ad25a8b0870..e1ade176ab46 100644
> --- a/drivers/net/virtio_net.c
> +++ b/drivers/net/virtio_net.c
> @@ -31,6 +31,13 @@ module_param(csum, bool, 0444);
>  module_param(gso, bool, 0444);
>  module_param(napi_tx, bool, 0644);
>  
> +/**
> + * Because virtio desc cache will increase memory overhead, users can turn it
> + * off or select an acceptable value. The maximum value is 2 + MAX_SKB_FRAGS.
> + */

Maybe add code to validate it and cap it at acceptable values then.

> +static u32 virtio_desc_cache_thr = 4;

Wouldn't something like CACHE_LINE_SIZE make more sense here?

> +module_param(virtio_desc_cache_thr, uint, 0644);
> +
>  /* FIXME: MTU in config. */
>  #define GOOD_PACKET_LEN (ETH_HLEN + VLAN_HLEN + ETH_DATA_LEN)
>  #define GOOD_COPY_LEN	128


> @@ -3214,6 +3221,11 @@ static int virtnet_probe(struct virtio_device *vdev)
>  		vi->curr_queue_pairs = num_online_cpus();
>  	vi->max_queue_pairs = max_queue_pairs;
>  
> +	if (virtio_desc_cache_thr > 2 + MAX_SKB_FRAGS)
> +		virtio_set_desc_cache(vdev, 2 + MAX_SKB_FRAGS);
> +	else
> +		virtio_set_desc_cache(vdev, virtio_desc_cache_thr);
> +
>  	/* Allocate/initialize the rx/tx queues, and invoke find_vqs */
>  	err = init_vqs(vi);
>  	if (err)
> -- 
> 2.31.0

Re: [PATCH v2 1/3] virtio: cache indirect desc for split

[Michael S. Tsirkin]

On Thu, Oct 28, 2021 at 06:49:17PM +0800, Xuan Zhuo wrote:
> In the case of using indirect, indirect desc must be allocated and
> released each time, which increases a lot of cpu overhead.
> 
> Here, a cache is added for indirect. If the number of indirect desc to be
> applied for is less than VIRT_QUEUE_CACHE_DESC_NUM, the desc array with
> the size of VIRT_QUEUE_CACHE_DESC_NUM is fixed and cached for reuse.
> 
> Signed-off-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
> ---
>  drivers/virtio/virtio.c      |  6 +++
>  drivers/virtio/virtio_ring.c | 77 ++++++++++++++++++++++++++++++++----
>  include/linux/virtio.h       | 14 +++++++
>  3 files changed, 89 insertions(+), 8 deletions(-)
> 
> diff --git a/drivers/virtio/virtio.c b/drivers/virtio/virtio.c
> index 0a5b54034d4b..1047149ac2a4 100644
> --- a/drivers/virtio/virtio.c
> +++ b/drivers/virtio/virtio.c
> @@ -431,6 +431,12 @@ bool is_virtio_device(struct device *dev)
>  }
>  EXPORT_SYMBOL_GPL(is_virtio_device);
>  
> +void virtio_set_desc_cache(struct virtio_device *dev, u32 thr)
> +{
> +	dev->desc_cache_thr = thr;
> +}
> +EXPORT_SYMBOL_GPL(virtio_set_desc_cache);
> +
>  void unregister_virtio_device(struct virtio_device *dev)
>  {
>  	int index = dev->index; /* save for after device release */
> diff --git a/drivers/virtio/virtio_ring.c b/drivers/virtio/virtio_ring.c
> index dd95dfd85e98..0ebcd4f12d3b 100644
> --- a/drivers/virtio/virtio_ring.c
> +++ b/drivers/virtio/virtio_ring.c
> @@ -117,6 +117,15 @@ struct vring_virtqueue {
>  	/* Hint for event idx: already triggered no need to disable. */
>  	bool event_triggered;
>  
> +	/* desc cache threshold
> +	 *    0   - disable desc cache
> +	 *    > 0 - enable desc cache. As the threshold of the desc cache.
> +	 */
> +	u32 desc_cache_thr;
> +
> +	/* desc cache chain */
> +	struct list_head desc_cache;
> +
>  	union {
>  		/* Available for split ring */
>  		struct {
> @@ -423,7 +432,53 @@ static unsigned int vring_unmap_one_split(const struct vring_virtqueue *vq,
>  	return extra[i].next;
>  }
>  
> -static struct vring_desc *alloc_indirect_split(struct virtqueue *_vq,
> +static void desc_cache_free(struct list_head *head)
> +{
> +	struct list_head *n, *pos;
> +
> +	BUILD_BUG_ON(sizeof(struct list_head) > sizeof(struct vring_desc));
> +	BUILD_BUG_ON(sizeof(struct list_head) > sizeof(struct vring_packed_desc));
> +
> +	list_for_each_prev_safe(pos, n, head)
> +		kfree(pos);
> +}
> +
> +static void __desc_cache_put(struct vring_virtqueue *vq,
> +			     struct list_head *node, int n)
> +{
> +	if (n <= vq->desc_cache_thr)
> +		list_add(node, &vq->desc_cache);
> +	else
> +		kfree(node);
> +}
> +
> +#define desc_cache_put(vq, desc, n) \
> +	__desc_cache_put(vq, (struct list_head *)desc, n)
> +
> +static void *desc_cache_get(struct vring_virtqueue *vq,
> +			    int size, int n, gfp_t gfp)
> +{
> +	struct list_head *node;
> +
> +	if (n > vq->desc_cache_thr)
> +		return kmalloc_array(n, size, gfp);
> +
> +	if (!list_empty(&vq->desc_cache)) {
> +		node = vq->desc_cache.next;
> +		list_del(node);
> +		return node;
> +	}
> +
> +	return kmalloc_array(vq->desc_cache_thr, size, gfp);
> +}
> +
> +#define _desc_cache_get(vq, n, gfp, tp) \
> +	((tp *)desc_cache_get(vq, (sizeof(tp)), n, gfp))
> +
> +#define desc_cache_get_split(vq, n, gfp) \
> +	_desc_cache_get(vq, n, gfp, struct vring_desc)
> +
> +static struct vring_desc *alloc_indirect_split(struct vring_virtqueue *vq,
>  					       unsigned int total_sg,
>  					       gfp_t gfp)
>  {
> @@ -437,12 +492,12 @@ static struct vring_desc *alloc_indirect_split(struct virtqueue *_vq,
>  	 */
>  	gfp &= ~__GFP_HIGHMEM;
>  
> -	desc = kmalloc_array(total_sg, sizeof(struct vring_desc), gfp);
> +	desc = desc_cache_get_split(vq, total_sg, gfp);
>  	if (!desc)
>  		return NULL;
>  
>  	for (i = 0; i < total_sg; i++)
> -		desc[i].next = cpu_to_virtio16(_vq->vdev, i + 1);
> +		desc[i].next = cpu_to_virtio16(vq->vq.vdev, i + 1);
>  	return desc;
>  }
>  
> @@ -508,7 +563,7 @@ static inline int virtqueue_add_split(struct virtqueue *_vq,
>  	head = vq->free_head;
>  
>  	if (virtqueue_use_indirect(_vq, total_sg))
> -		desc = alloc_indirect_split(_vq, total_sg, gfp);
> +		desc = alloc_indirect_split(vq, total_sg, gfp);
>  	else {
>  		desc = NULL;
>  		WARN_ON_ONCE(total_sg > vq->split.vring.num && !vq->indirect);
> @@ -652,7 +707,7 @@ static inline int virtqueue_add_split(struct virtqueue *_vq,
>  	}
>  
>  	if (indirect)
> -		kfree(desc);
> +		desc_cache_put(vq, desc, total_sg);
>  
>  	END_USE(vq);
>  	return -ENOMEM;
> @@ -717,7 +772,7 @@ static void detach_buf_split(struct vring_virtqueue *vq, unsigned int head,
>  	if (vq->indirect) {
>  		struct vring_desc *indir_desc =
>  				vq->split.desc_state[head].indir_desc;
> -		u32 len;
> +		u32 len, n;
>  
>  		/* Free the indirect table, if any, now that it's unmapped. */
>  		if (!indir_desc)
> @@ -729,10 +784,12 @@ static void detach_buf_split(struct vring_virtqueue *vq, unsigned int head,
>  				VRING_DESC_F_INDIRECT));
>  		BUG_ON(len == 0 || len % sizeof(struct vring_desc));
>  
> -		for (j = 0; j < len / sizeof(struct vring_desc); j++)
> +		n = len / sizeof(struct vring_desc);
> +
> +		for (j = 0; j < n; j++)
>  			vring_unmap_one_split_indirect(vq, &indir_desc[j]);
>  
> -		kfree(indir_desc);
> +		desc_cache_put(vq, indir_desc, n);
>  		vq->split.desc_state[head].indir_desc = NULL;
>  	} else if (ctx) {
>  		*ctx = vq->split.desc_state[head].indir_desc;
> @@ -2199,6 +2256,9 @@ struct virtqueue *__vring_new_virtqueue(unsigned int index,
>  	vq->indirect = virtio_has_feature(vdev, VIRTIO_RING_F_INDIRECT_DESC) &&
>  		!context;
>  	vq->event = virtio_has_feature(vdev, VIRTIO_RING_F_EVENT_IDX);
> +	vq->desc_cache_thr = vdev->desc_cache_thr;
> +
> +	INIT_LIST_HEAD(&vq->desc_cache);
>  
>  	if (virtio_has_feature(vdev, VIRTIO_F_ORDER_PLATFORM))
>  		vq->weak_barriers = false;
> @@ -2329,6 +2389,7 @@ void vring_del_virtqueue(struct virtqueue *_vq)
>  	if (!vq->packed_ring) {
>  		kfree(vq->split.desc_state);
>  		kfree(vq->split.desc_extra);
> +		desc_cache_free(&vq->desc_cache);
>  	}
>  	kfree(vq);
>  }
> diff --git a/include/linux/virtio.h b/include/linux/virtio.h
> index 41edbc01ffa4..bda6f9853e97 100644
> --- a/include/linux/virtio.h
> +++ b/include/linux/virtio.h
> @@ -118,6 +118,7 @@ struct virtio_device {
>  	struct list_head vqs;
>  	u64 features;
>  	void *priv;
> +	u32 desc_cache_thr;
>  };
>  
>  static inline struct virtio_device *dev_to_virtio(struct device *_dev)
> @@ -130,6 +131,19 @@ int register_virtio_device(struct virtio_device *dev);
>  void unregister_virtio_device(struct virtio_device *dev);
>  bool is_virtio_device(struct device *dev);
>  
> +/**
> + * virtio_set_desc_cache - set virtio ring desc cache threshold
> + *
> + * virtio will cache the allocated indirect desc.
> + *
> + * This function must be called before find_vqs.
> + *
> + * @thr:
> + *    0   - disable desc cache
> + *    > 0 - enable desc cache. As the threshold of the desc cache.
> + */
> +void virtio_set_desc_cache(struct virtio_device *dev, u32 thr);
> +

Would a better API be a cache size in bytes? This controls how much
memory is spent after all.

>  void virtio_break_device(struct virtio_device *dev);
>  
>  void virtio_config_changed(struct virtio_device *dev);
> -- 
> 2.31.0