linux_dsm_epyc7002/net/xdp/xsk_queue.h

265 lines
5.7 KiB
C
Raw Normal View History

/* SPDX-License-Identifier: GPL-2.0 */
/* XDP user-space ring structure
* Copyright(c) 2018 Intel Corporation.
*/
#ifndef _LINUX_XSK_QUEUE_H
#define _LINUX_XSK_QUEUE_H
#include <linux/types.h>
#include <linux/if_xdp.h>
#include <net/xdp_sock.h>
#define RX_BATCH_SIZE 16
#define LAZY_UPDATE_THRESHOLD 128
struct xdp_ring {
u32 producer ____cacheline_aligned_in_smp;
u32 consumer ____cacheline_aligned_in_smp;
};
/* Used for the RX and TX queues for packets */
struct xdp_rxtx_ring {
struct xdp_ring ptrs;
struct xdp_desc desc[0] ____cacheline_aligned_in_smp;
};
/* Used for the fill and completion queues for buffers */
struct xdp_umem_ring {
struct xdp_ring ptrs;
xsk: new descriptor addressing scheme Currently, AF_XDP only supports a fixed frame-size memory scheme where each frame is referenced via an index (idx). A user passes the frame index to the kernel, and the kernel acts upon the data. Some NICs, however, do not have a fixed frame-size model, instead they have a model where a memory window is passed to the hardware and multiple frames are filled into that window (referred to as the "type-writer" model). By changing the descriptor format from the current frame index addressing scheme, AF_XDP can in the future be extended to support these kinds of NICs. In the index-based model, an idx refers to a frame of size frame_size. Addressing a frame in the UMEM is done by offseting the UMEM starting address by a global offset, idx * frame_size + offset. Communicating via the fill- and completion-rings are done by means of idx. In this commit, the idx is removed in favor of an address (addr), which is a relative address ranging over the UMEM. To convert an idx-based address to the new addr is simply: addr = idx * frame_size + offset. We also stop referring to the UMEM "frame" as a frame. Instead it is simply called a chunk. To transfer ownership of a chunk to the kernel, the addr of the chunk is passed in the fill-ring. Note, that the kernel will mask addr to make it chunk aligned, so there is no need for userspace to do that. E.g., for a chunk size of 2k, passing an addr of 2048, 2050 or 3000 to the fill-ring will refer to the same chunk. On the completion-ring, the addr will match that of the Tx descriptor, passed to the kernel. Changing the descriptor format to use chunks/addr will allow for future changes to move to a type-writer based model, where multiple frames can reside in one chunk. In this model passing one single chunk into the fill-ring, would potentially result in multiple Rx descriptors. This commit changes the uapi of AF_XDP sockets, and updates the documentation. Signed-off-by: Björn Töpel <bjorn.topel@intel.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-06-04 18:57:13 +07:00
u64 desc[0] ____cacheline_aligned_in_smp;
};
struct xsk_queue {
u64 chunk_mask;
u64 size;
u32 ring_mask;
u32 nentries;
u32 prod_head;
u32 prod_tail;
u32 cons_head;
u32 cons_tail;
struct xdp_ring *ring;
u64 invalid_descs;
};
/* Common functions operating for both RXTX and umem queues */
static inline u64 xskq_nb_invalid_descs(struct xsk_queue *q)
{
return q ? q->invalid_descs : 0;
}
static inline u32 xskq_nb_avail(struct xsk_queue *q, u32 dcnt)
{
u32 entries = q->prod_tail - q->cons_tail;
if (entries == 0) {
/* Refresh the local pointer */
q->prod_tail = READ_ONCE(q->ring->producer);
entries = q->prod_tail - q->cons_tail;
}
return (entries > dcnt) ? dcnt : entries;
}
static inline u32 xskq_nb_free(struct xsk_queue *q, u32 producer, u32 dcnt)
{
u32 free_entries = q->nentries - (producer - q->cons_tail);
if (free_entries >= dcnt)
return free_entries;
/* Refresh the local tail pointer */
q->cons_tail = READ_ONCE(q->ring->consumer);
return q->nentries - (producer - q->cons_tail);
}
/* UMEM queue */
xsk: new descriptor addressing scheme Currently, AF_XDP only supports a fixed frame-size memory scheme where each frame is referenced via an index (idx). A user passes the frame index to the kernel, and the kernel acts upon the data. Some NICs, however, do not have a fixed frame-size model, instead they have a model where a memory window is passed to the hardware and multiple frames are filled into that window (referred to as the "type-writer" model). By changing the descriptor format from the current frame index addressing scheme, AF_XDP can in the future be extended to support these kinds of NICs. In the index-based model, an idx refers to a frame of size frame_size. Addressing a frame in the UMEM is done by offseting the UMEM starting address by a global offset, idx * frame_size + offset. Communicating via the fill- and completion-rings are done by means of idx. In this commit, the idx is removed in favor of an address (addr), which is a relative address ranging over the UMEM. To convert an idx-based address to the new addr is simply: addr = idx * frame_size + offset. We also stop referring to the UMEM "frame" as a frame. Instead it is simply called a chunk. To transfer ownership of a chunk to the kernel, the addr of the chunk is passed in the fill-ring. Note, that the kernel will mask addr to make it chunk aligned, so there is no need for userspace to do that. E.g., for a chunk size of 2k, passing an addr of 2048, 2050 or 3000 to the fill-ring will refer to the same chunk. On the completion-ring, the addr will match that of the Tx descriptor, passed to the kernel. Changing the descriptor format to use chunks/addr will allow for future changes to move to a type-writer based model, where multiple frames can reside in one chunk. In this model passing one single chunk into the fill-ring, would potentially result in multiple Rx descriptors. This commit changes the uapi of AF_XDP sockets, and updates the documentation. Signed-off-by: Björn Töpel <bjorn.topel@intel.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-06-04 18:57:13 +07:00
static inline bool xskq_is_valid_addr(struct xsk_queue *q, u64 addr)
{
if (addr >= q->size) {
q->invalid_descs++;
return false;
}
xsk: new descriptor addressing scheme Currently, AF_XDP only supports a fixed frame-size memory scheme where each frame is referenced via an index (idx). A user passes the frame index to the kernel, and the kernel acts upon the data. Some NICs, however, do not have a fixed frame-size model, instead they have a model where a memory window is passed to the hardware and multiple frames are filled into that window (referred to as the "type-writer" model). By changing the descriptor format from the current frame index addressing scheme, AF_XDP can in the future be extended to support these kinds of NICs. In the index-based model, an idx refers to a frame of size frame_size. Addressing a frame in the UMEM is done by offseting the UMEM starting address by a global offset, idx * frame_size + offset. Communicating via the fill- and completion-rings are done by means of idx. In this commit, the idx is removed in favor of an address (addr), which is a relative address ranging over the UMEM. To convert an idx-based address to the new addr is simply: addr = idx * frame_size + offset. We also stop referring to the UMEM "frame" as a frame. Instead it is simply called a chunk. To transfer ownership of a chunk to the kernel, the addr of the chunk is passed in the fill-ring. Note, that the kernel will mask addr to make it chunk aligned, so there is no need for userspace to do that. E.g., for a chunk size of 2k, passing an addr of 2048, 2050 or 3000 to the fill-ring will refer to the same chunk. On the completion-ring, the addr will match that of the Tx descriptor, passed to the kernel. Changing the descriptor format to use chunks/addr will allow for future changes to move to a type-writer based model, where multiple frames can reside in one chunk. In this model passing one single chunk into the fill-ring, would potentially result in multiple Rx descriptors. This commit changes the uapi of AF_XDP sockets, and updates the documentation. Signed-off-by: Björn Töpel <bjorn.topel@intel.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-06-04 18:57:13 +07:00
return true;
}
xsk: new descriptor addressing scheme Currently, AF_XDP only supports a fixed frame-size memory scheme where each frame is referenced via an index (idx). A user passes the frame index to the kernel, and the kernel acts upon the data. Some NICs, however, do not have a fixed frame-size model, instead they have a model where a memory window is passed to the hardware and multiple frames are filled into that window (referred to as the "type-writer" model). By changing the descriptor format from the current frame index addressing scheme, AF_XDP can in the future be extended to support these kinds of NICs. In the index-based model, an idx refers to a frame of size frame_size. Addressing a frame in the UMEM is done by offseting the UMEM starting address by a global offset, idx * frame_size + offset. Communicating via the fill- and completion-rings are done by means of idx. In this commit, the idx is removed in favor of an address (addr), which is a relative address ranging over the UMEM. To convert an idx-based address to the new addr is simply: addr = idx * frame_size + offset. We also stop referring to the UMEM "frame" as a frame. Instead it is simply called a chunk. To transfer ownership of a chunk to the kernel, the addr of the chunk is passed in the fill-ring. Note, that the kernel will mask addr to make it chunk aligned, so there is no need for userspace to do that. E.g., for a chunk size of 2k, passing an addr of 2048, 2050 or 3000 to the fill-ring will refer to the same chunk. On the completion-ring, the addr will match that of the Tx descriptor, passed to the kernel. Changing the descriptor format to use chunks/addr will allow for future changes to move to a type-writer based model, where multiple frames can reside in one chunk. In this model passing one single chunk into the fill-ring, would potentially result in multiple Rx descriptors. This commit changes the uapi of AF_XDP sockets, and updates the documentation. Signed-off-by: Björn Töpel <bjorn.topel@intel.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-06-04 18:57:13 +07:00
static inline u64 *xskq_validate_addr(struct xsk_queue *q, u64 *addr)
{
while (q->cons_tail != q->cons_head) {
struct xdp_umem_ring *ring = (struct xdp_umem_ring *)q->ring;
unsigned int idx = q->cons_tail & q->ring_mask;
*addr = READ_ONCE(ring->desc[idx]) & q->chunk_mask;
xsk: new descriptor addressing scheme Currently, AF_XDP only supports a fixed frame-size memory scheme where each frame is referenced via an index (idx). A user passes the frame index to the kernel, and the kernel acts upon the data. Some NICs, however, do not have a fixed frame-size model, instead they have a model where a memory window is passed to the hardware and multiple frames are filled into that window (referred to as the "type-writer" model). By changing the descriptor format from the current frame index addressing scheme, AF_XDP can in the future be extended to support these kinds of NICs. In the index-based model, an idx refers to a frame of size frame_size. Addressing a frame in the UMEM is done by offseting the UMEM starting address by a global offset, idx * frame_size + offset. Communicating via the fill- and completion-rings are done by means of idx. In this commit, the idx is removed in favor of an address (addr), which is a relative address ranging over the UMEM. To convert an idx-based address to the new addr is simply: addr = idx * frame_size + offset. We also stop referring to the UMEM "frame" as a frame. Instead it is simply called a chunk. To transfer ownership of a chunk to the kernel, the addr of the chunk is passed in the fill-ring. Note, that the kernel will mask addr to make it chunk aligned, so there is no need for userspace to do that. E.g., for a chunk size of 2k, passing an addr of 2048, 2050 or 3000 to the fill-ring will refer to the same chunk. On the completion-ring, the addr will match that of the Tx descriptor, passed to the kernel. Changing the descriptor format to use chunks/addr will allow for future changes to move to a type-writer based model, where multiple frames can reside in one chunk. In this model passing one single chunk into the fill-ring, would potentially result in multiple Rx descriptors. This commit changes the uapi of AF_XDP sockets, and updates the documentation. Signed-off-by: Björn Töpel <bjorn.topel@intel.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-06-04 18:57:13 +07:00
if (xskq_is_valid_addr(q, *addr))
return addr;
q->cons_tail++;
}
return NULL;
}
xsk: new descriptor addressing scheme Currently, AF_XDP only supports a fixed frame-size memory scheme where each frame is referenced via an index (idx). A user passes the frame index to the kernel, and the kernel acts upon the data. Some NICs, however, do not have a fixed frame-size model, instead they have a model where a memory window is passed to the hardware and multiple frames are filled into that window (referred to as the "type-writer" model). By changing the descriptor format from the current frame index addressing scheme, AF_XDP can in the future be extended to support these kinds of NICs. In the index-based model, an idx refers to a frame of size frame_size. Addressing a frame in the UMEM is done by offseting the UMEM starting address by a global offset, idx * frame_size + offset. Communicating via the fill- and completion-rings are done by means of idx. In this commit, the idx is removed in favor of an address (addr), which is a relative address ranging over the UMEM. To convert an idx-based address to the new addr is simply: addr = idx * frame_size + offset. We also stop referring to the UMEM "frame" as a frame. Instead it is simply called a chunk. To transfer ownership of a chunk to the kernel, the addr of the chunk is passed in the fill-ring. Note, that the kernel will mask addr to make it chunk aligned, so there is no need for userspace to do that. E.g., for a chunk size of 2k, passing an addr of 2048, 2050 or 3000 to the fill-ring will refer to the same chunk. On the completion-ring, the addr will match that of the Tx descriptor, passed to the kernel. Changing the descriptor format to use chunks/addr will allow for future changes to move to a type-writer based model, where multiple frames can reside in one chunk. In this model passing one single chunk into the fill-ring, would potentially result in multiple Rx descriptors. This commit changes the uapi of AF_XDP sockets, and updates the documentation. Signed-off-by: Björn Töpel <bjorn.topel@intel.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-06-04 18:57:13 +07:00
static inline u64 *xskq_peek_addr(struct xsk_queue *q, u64 *addr)
{
if (q->cons_tail == q->cons_head) {
WRITE_ONCE(q->ring->consumer, q->cons_tail);
q->cons_head = q->cons_tail + xskq_nb_avail(q, RX_BATCH_SIZE);
/* Order consumer and data */
smp_rmb();
}
xsk: new descriptor addressing scheme Currently, AF_XDP only supports a fixed frame-size memory scheme where each frame is referenced via an index (idx). A user passes the frame index to the kernel, and the kernel acts upon the data. Some NICs, however, do not have a fixed frame-size model, instead they have a model where a memory window is passed to the hardware and multiple frames are filled into that window (referred to as the "type-writer" model). By changing the descriptor format from the current frame index addressing scheme, AF_XDP can in the future be extended to support these kinds of NICs. In the index-based model, an idx refers to a frame of size frame_size. Addressing a frame in the UMEM is done by offseting the UMEM starting address by a global offset, idx * frame_size + offset. Communicating via the fill- and completion-rings are done by means of idx. In this commit, the idx is removed in favor of an address (addr), which is a relative address ranging over the UMEM. To convert an idx-based address to the new addr is simply: addr = idx * frame_size + offset. We also stop referring to the UMEM "frame" as a frame. Instead it is simply called a chunk. To transfer ownership of a chunk to the kernel, the addr of the chunk is passed in the fill-ring. Note, that the kernel will mask addr to make it chunk aligned, so there is no need for userspace to do that. E.g., for a chunk size of 2k, passing an addr of 2048, 2050 or 3000 to the fill-ring will refer to the same chunk. On the completion-ring, the addr will match that of the Tx descriptor, passed to the kernel. Changing the descriptor format to use chunks/addr will allow for future changes to move to a type-writer based model, where multiple frames can reside in one chunk. In this model passing one single chunk into the fill-ring, would potentially result in multiple Rx descriptors. This commit changes the uapi of AF_XDP sockets, and updates the documentation. Signed-off-by: Björn Töpel <bjorn.topel@intel.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-06-04 18:57:13 +07:00
return xskq_validate_addr(q, addr);
}
xsk: new descriptor addressing scheme Currently, AF_XDP only supports a fixed frame-size memory scheme where each frame is referenced via an index (idx). A user passes the frame index to the kernel, and the kernel acts upon the data. Some NICs, however, do not have a fixed frame-size model, instead they have a model where a memory window is passed to the hardware and multiple frames are filled into that window (referred to as the "type-writer" model). By changing the descriptor format from the current frame index addressing scheme, AF_XDP can in the future be extended to support these kinds of NICs. In the index-based model, an idx refers to a frame of size frame_size. Addressing a frame in the UMEM is done by offseting the UMEM starting address by a global offset, idx * frame_size + offset. Communicating via the fill- and completion-rings are done by means of idx. In this commit, the idx is removed in favor of an address (addr), which is a relative address ranging over the UMEM. To convert an idx-based address to the new addr is simply: addr = idx * frame_size + offset. We also stop referring to the UMEM "frame" as a frame. Instead it is simply called a chunk. To transfer ownership of a chunk to the kernel, the addr of the chunk is passed in the fill-ring. Note, that the kernel will mask addr to make it chunk aligned, so there is no need for userspace to do that. E.g., for a chunk size of 2k, passing an addr of 2048, 2050 or 3000 to the fill-ring will refer to the same chunk. On the completion-ring, the addr will match that of the Tx descriptor, passed to the kernel. Changing the descriptor format to use chunks/addr will allow for future changes to move to a type-writer based model, where multiple frames can reside in one chunk. In this model passing one single chunk into the fill-ring, would potentially result in multiple Rx descriptors. This commit changes the uapi of AF_XDP sockets, and updates the documentation. Signed-off-by: Björn Töpel <bjorn.topel@intel.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-06-04 18:57:13 +07:00
static inline void xskq_discard_addr(struct xsk_queue *q)
{
q->cons_tail++;
}
xsk: new descriptor addressing scheme Currently, AF_XDP only supports a fixed frame-size memory scheme where each frame is referenced via an index (idx). A user passes the frame index to the kernel, and the kernel acts upon the data. Some NICs, however, do not have a fixed frame-size model, instead they have a model where a memory window is passed to the hardware and multiple frames are filled into that window (referred to as the "type-writer" model). By changing the descriptor format from the current frame index addressing scheme, AF_XDP can in the future be extended to support these kinds of NICs. In the index-based model, an idx refers to a frame of size frame_size. Addressing a frame in the UMEM is done by offseting the UMEM starting address by a global offset, idx * frame_size + offset. Communicating via the fill- and completion-rings are done by means of idx. In this commit, the idx is removed in favor of an address (addr), which is a relative address ranging over the UMEM. To convert an idx-based address to the new addr is simply: addr = idx * frame_size + offset. We also stop referring to the UMEM "frame" as a frame. Instead it is simply called a chunk. To transfer ownership of a chunk to the kernel, the addr of the chunk is passed in the fill-ring. Note, that the kernel will mask addr to make it chunk aligned, so there is no need for userspace to do that. E.g., for a chunk size of 2k, passing an addr of 2048, 2050 or 3000 to the fill-ring will refer to the same chunk. On the completion-ring, the addr will match that of the Tx descriptor, passed to the kernel. Changing the descriptor format to use chunks/addr will allow for future changes to move to a type-writer based model, where multiple frames can reside in one chunk. In this model passing one single chunk into the fill-ring, would potentially result in multiple Rx descriptors. This commit changes the uapi of AF_XDP sockets, and updates the documentation. Signed-off-by: Björn Töpel <bjorn.topel@intel.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-06-04 18:57:13 +07:00
static inline int xskq_produce_addr(struct xsk_queue *q, u64 addr)
{
struct xdp_umem_ring *ring = (struct xdp_umem_ring *)q->ring;
if (xskq_nb_free(q, q->prod_tail, 1) == 0)
return -ENOSPC;
xsk: new descriptor addressing scheme Currently, AF_XDP only supports a fixed frame-size memory scheme where each frame is referenced via an index (idx). A user passes the frame index to the kernel, and the kernel acts upon the data. Some NICs, however, do not have a fixed frame-size model, instead they have a model where a memory window is passed to the hardware and multiple frames are filled into that window (referred to as the "type-writer" model). By changing the descriptor format from the current frame index addressing scheme, AF_XDP can in the future be extended to support these kinds of NICs. In the index-based model, an idx refers to a frame of size frame_size. Addressing a frame in the UMEM is done by offseting the UMEM starting address by a global offset, idx * frame_size + offset. Communicating via the fill- and completion-rings are done by means of idx. In this commit, the idx is removed in favor of an address (addr), which is a relative address ranging over the UMEM. To convert an idx-based address to the new addr is simply: addr = idx * frame_size + offset. We also stop referring to the UMEM "frame" as a frame. Instead it is simply called a chunk. To transfer ownership of a chunk to the kernel, the addr of the chunk is passed in the fill-ring. Note, that the kernel will mask addr to make it chunk aligned, so there is no need for userspace to do that. E.g., for a chunk size of 2k, passing an addr of 2048, 2050 or 3000 to the fill-ring will refer to the same chunk. On the completion-ring, the addr will match that of the Tx descriptor, passed to the kernel. Changing the descriptor format to use chunks/addr will allow for future changes to move to a type-writer based model, where multiple frames can reside in one chunk. In this model passing one single chunk into the fill-ring, would potentially result in multiple Rx descriptors. This commit changes the uapi of AF_XDP sockets, and updates the documentation. Signed-off-by: Björn Töpel <bjorn.topel@intel.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-06-04 18:57:13 +07:00
ring->desc[q->prod_tail++ & q->ring_mask] = addr;
/* Order producer and data */
smp_wmb();
WRITE_ONCE(q->ring->producer, q->prod_tail);
return 0;
}
static inline int xskq_produce_addr_lazy(struct xsk_queue *q, u64 addr)
{
struct xdp_umem_ring *ring = (struct xdp_umem_ring *)q->ring;
if (xskq_nb_free(q, q->prod_head, LAZY_UPDATE_THRESHOLD) == 0)
return -ENOSPC;
ring->desc[q->prod_head++ & q->ring_mask] = addr;
return 0;
}
static inline void xskq_produce_flush_addr_n(struct xsk_queue *q,
u32 nb_entries)
{
/* Order producer and data */
smp_wmb();
q->prod_tail += nb_entries;
WRITE_ONCE(q->ring->producer, q->prod_tail);
}
xsk: new descriptor addressing scheme Currently, AF_XDP only supports a fixed frame-size memory scheme where each frame is referenced via an index (idx). A user passes the frame index to the kernel, and the kernel acts upon the data. Some NICs, however, do not have a fixed frame-size model, instead they have a model where a memory window is passed to the hardware and multiple frames are filled into that window (referred to as the "type-writer" model). By changing the descriptor format from the current frame index addressing scheme, AF_XDP can in the future be extended to support these kinds of NICs. In the index-based model, an idx refers to a frame of size frame_size. Addressing a frame in the UMEM is done by offseting the UMEM starting address by a global offset, idx * frame_size + offset. Communicating via the fill- and completion-rings are done by means of idx. In this commit, the idx is removed in favor of an address (addr), which is a relative address ranging over the UMEM. To convert an idx-based address to the new addr is simply: addr = idx * frame_size + offset. We also stop referring to the UMEM "frame" as a frame. Instead it is simply called a chunk. To transfer ownership of a chunk to the kernel, the addr of the chunk is passed in the fill-ring. Note, that the kernel will mask addr to make it chunk aligned, so there is no need for userspace to do that. E.g., for a chunk size of 2k, passing an addr of 2048, 2050 or 3000 to the fill-ring will refer to the same chunk. On the completion-ring, the addr will match that of the Tx descriptor, passed to the kernel. Changing the descriptor format to use chunks/addr will allow for future changes to move to a type-writer based model, where multiple frames can reside in one chunk. In this model passing one single chunk into the fill-ring, would potentially result in multiple Rx descriptors. This commit changes the uapi of AF_XDP sockets, and updates the documentation. Signed-off-by: Björn Töpel <bjorn.topel@intel.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-06-04 18:57:13 +07:00
static inline int xskq_reserve_addr(struct xsk_queue *q)
{
if (xskq_nb_free(q, q->prod_head, 1) == 0)
return -ENOSPC;
q->prod_head++;
return 0;
}
/* Rx/Tx queue */
static inline bool xskq_is_valid_desc(struct xsk_queue *q, struct xdp_desc *d)
{
xsk: new descriptor addressing scheme Currently, AF_XDP only supports a fixed frame-size memory scheme where each frame is referenced via an index (idx). A user passes the frame index to the kernel, and the kernel acts upon the data. Some NICs, however, do not have a fixed frame-size model, instead they have a model where a memory window is passed to the hardware and multiple frames are filled into that window (referred to as the "type-writer" model). By changing the descriptor format from the current frame index addressing scheme, AF_XDP can in the future be extended to support these kinds of NICs. In the index-based model, an idx refers to a frame of size frame_size. Addressing a frame in the UMEM is done by offseting the UMEM starting address by a global offset, idx * frame_size + offset. Communicating via the fill- and completion-rings are done by means of idx. In this commit, the idx is removed in favor of an address (addr), which is a relative address ranging over the UMEM. To convert an idx-based address to the new addr is simply: addr = idx * frame_size + offset. We also stop referring to the UMEM "frame" as a frame. Instead it is simply called a chunk. To transfer ownership of a chunk to the kernel, the addr of the chunk is passed in the fill-ring. Note, that the kernel will mask addr to make it chunk aligned, so there is no need for userspace to do that. E.g., for a chunk size of 2k, passing an addr of 2048, 2050 or 3000 to the fill-ring will refer to the same chunk. On the completion-ring, the addr will match that of the Tx descriptor, passed to the kernel. Changing the descriptor format to use chunks/addr will allow for future changes to move to a type-writer based model, where multiple frames can reside in one chunk. In this model passing one single chunk into the fill-ring, would potentially result in multiple Rx descriptors. This commit changes the uapi of AF_XDP sockets, and updates the documentation. Signed-off-by: Björn Töpel <bjorn.topel@intel.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-06-04 18:57:13 +07:00
if (!xskq_is_valid_addr(q, d->addr))
return false;
if (((d->addr + d->len) & q->chunk_mask) != (d->addr & q->chunk_mask) ||
d->options) {
q->invalid_descs++;
return false;
}
return true;
}
static inline struct xdp_desc *xskq_validate_desc(struct xsk_queue *q,
struct xdp_desc *desc)
{
while (q->cons_tail != q->cons_head) {
struct xdp_rxtx_ring *ring = (struct xdp_rxtx_ring *)q->ring;
unsigned int idx = q->cons_tail & q->ring_mask;
*desc = READ_ONCE(ring->desc[idx]);
if (xskq_is_valid_desc(q, desc))
return desc;
q->cons_tail++;
}
return NULL;
}
static inline struct xdp_desc *xskq_peek_desc(struct xsk_queue *q,
struct xdp_desc *desc)
{
if (q->cons_tail == q->cons_head) {
WRITE_ONCE(q->ring->consumer, q->cons_tail);
q->cons_head = q->cons_tail + xskq_nb_avail(q, RX_BATCH_SIZE);
/* Order consumer and data */
smp_rmb();
}
return xskq_validate_desc(q, desc);
}
static inline void xskq_discard_desc(struct xsk_queue *q)
{
q->cons_tail++;
}
static inline int xskq_produce_batch_desc(struct xsk_queue *q,
xsk: new descriptor addressing scheme Currently, AF_XDP only supports a fixed frame-size memory scheme where each frame is referenced via an index (idx). A user passes the frame index to the kernel, and the kernel acts upon the data. Some NICs, however, do not have a fixed frame-size model, instead they have a model where a memory window is passed to the hardware and multiple frames are filled into that window (referred to as the "type-writer" model). By changing the descriptor format from the current frame index addressing scheme, AF_XDP can in the future be extended to support these kinds of NICs. In the index-based model, an idx refers to a frame of size frame_size. Addressing a frame in the UMEM is done by offseting the UMEM starting address by a global offset, idx * frame_size + offset. Communicating via the fill- and completion-rings are done by means of idx. In this commit, the idx is removed in favor of an address (addr), which is a relative address ranging over the UMEM. To convert an idx-based address to the new addr is simply: addr = idx * frame_size + offset. We also stop referring to the UMEM "frame" as a frame. Instead it is simply called a chunk. To transfer ownership of a chunk to the kernel, the addr of the chunk is passed in the fill-ring. Note, that the kernel will mask addr to make it chunk aligned, so there is no need for userspace to do that. E.g., for a chunk size of 2k, passing an addr of 2048, 2050 or 3000 to the fill-ring will refer to the same chunk. On the completion-ring, the addr will match that of the Tx descriptor, passed to the kernel. Changing the descriptor format to use chunks/addr will allow for future changes to move to a type-writer based model, where multiple frames can reside in one chunk. In this model passing one single chunk into the fill-ring, would potentially result in multiple Rx descriptors. This commit changes the uapi of AF_XDP sockets, and updates the documentation. Signed-off-by: Björn Töpel <bjorn.topel@intel.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-06-04 18:57:13 +07:00
u64 addr, u32 len)
{
struct xdp_rxtx_ring *ring = (struct xdp_rxtx_ring *)q->ring;
unsigned int idx;
if (xskq_nb_free(q, q->prod_head, 1) == 0)
return -ENOSPC;
idx = (q->prod_head++) & q->ring_mask;
xsk: new descriptor addressing scheme Currently, AF_XDP only supports a fixed frame-size memory scheme where each frame is referenced via an index (idx). A user passes the frame index to the kernel, and the kernel acts upon the data. Some NICs, however, do not have a fixed frame-size model, instead they have a model where a memory window is passed to the hardware and multiple frames are filled into that window (referred to as the "type-writer" model). By changing the descriptor format from the current frame index addressing scheme, AF_XDP can in the future be extended to support these kinds of NICs. In the index-based model, an idx refers to a frame of size frame_size. Addressing a frame in the UMEM is done by offseting the UMEM starting address by a global offset, idx * frame_size + offset. Communicating via the fill- and completion-rings are done by means of idx. In this commit, the idx is removed in favor of an address (addr), which is a relative address ranging over the UMEM. To convert an idx-based address to the new addr is simply: addr = idx * frame_size + offset. We also stop referring to the UMEM "frame" as a frame. Instead it is simply called a chunk. To transfer ownership of a chunk to the kernel, the addr of the chunk is passed in the fill-ring. Note, that the kernel will mask addr to make it chunk aligned, so there is no need for userspace to do that. E.g., for a chunk size of 2k, passing an addr of 2048, 2050 or 3000 to the fill-ring will refer to the same chunk. On the completion-ring, the addr will match that of the Tx descriptor, passed to the kernel. Changing the descriptor format to use chunks/addr will allow for future changes to move to a type-writer based model, where multiple frames can reside in one chunk. In this model passing one single chunk into the fill-ring, would potentially result in multiple Rx descriptors. This commit changes the uapi of AF_XDP sockets, and updates the documentation. Signed-off-by: Björn Töpel <bjorn.topel@intel.com> Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
2018-06-04 18:57:13 +07:00
ring->desc[idx].addr = addr;
ring->desc[idx].len = len;
return 0;
}
static inline void xskq_produce_flush_desc(struct xsk_queue *q)
{
/* Order producer and data */
smp_wmb();
q->prod_tail = q->prod_head,
WRITE_ONCE(q->ring->producer, q->prod_tail);
}
static inline bool xskq_full_desc(struct xsk_queue *q)
{
return xskq_nb_avail(q, q->nentries) == q->nentries;
}
static inline bool xskq_empty_desc(struct xsk_queue *q)
{
return xskq_nb_free(q, q->prod_tail, q->nentries) == q->nentries;
}
void xskq_set_umem(struct xsk_queue *q, u64 size, u64 chunk_mask);
struct xsk_queue *xskq_create(u32 nentries, bool umem_queue);
void xskq_destroy(struct xsk_queue *q_ops);
/* Executed by the core when the entire UMEM gets freed */
void xsk_reuseq_destroy(struct xdp_umem *umem);
#endif /* _LINUX_XSK_QUEUE_H */