Commit eab668a6 authored by Haggai Eran's avatar Haggai Eran Committed by Roland Dreier

IB/mlx5: Add support for RDMA read/write responder page faults

Signed-off-by: default avatarShachar Raindel <raindel@mellanox.com>
Signed-off-by: default avatarHaggai Eran <haggaie@mellanox.com>
Signed-off-by: default avatarRoland Dreier <roland@purestorage.com>
parent 7bdf65d4
...@@ -35,6 +35,8 @@ ...@@ -35,6 +35,8 @@
#include "mlx5_ib.h" #include "mlx5_ib.h"
#define MAX_PREFETCH_LEN (4*1024*1024U)
struct workqueue_struct *mlx5_ib_page_fault_wq; struct workqueue_struct *mlx5_ib_page_fault_wq;
#define COPY_ODP_BIT_MLX_TO_IB(reg, ib_caps, field_name, bit_name) do { \ #define COPY_ODP_BIT_MLX_TO_IB(reg, ib_caps, field_name, bit_name) do { \
...@@ -490,6 +492,80 @@ static void mlx5_ib_mr_wqe_pfault_handler(struct mlx5_ib_qp *qp, ...@@ -490,6 +492,80 @@ static void mlx5_ib_mr_wqe_pfault_handler(struct mlx5_ib_qp *qp,
free_page((unsigned long)buffer); free_page((unsigned long)buffer);
} }
static int pages_in_range(u64 address, u32 length)
{
return (ALIGN(address + length, PAGE_SIZE) -
(address & PAGE_MASK)) >> PAGE_SHIFT;
}
static void mlx5_ib_mr_rdma_pfault_handler(struct mlx5_ib_qp *qp,
struct mlx5_ib_pfault *pfault)
{
struct mlx5_pagefault *mpfault = &pfault->mpfault;
u64 address;
u32 length;
u32 prefetch_len = mpfault->bytes_committed;
int prefetch_activated = 0;
u32 rkey = mpfault->rdma.r_key;
int ret;
/* The RDMA responder handler handles the page fault in two parts.
* First it brings the necessary pages for the current packet
* (and uses the pfault context), and then (after resuming the QP)
* prefetches more pages. The second operation cannot use the pfault
* context and therefore uses the dummy_pfault context allocated on
* the stack */
struct mlx5_ib_pfault dummy_pfault = {};
dummy_pfault.mpfault.bytes_committed = 0;
mpfault->rdma.rdma_va += mpfault->bytes_committed;
mpfault->rdma.rdma_op_len -= min(mpfault->bytes_committed,
mpfault->rdma.rdma_op_len);
mpfault->bytes_committed = 0;
address = mpfault->rdma.rdma_va;
length = mpfault->rdma.rdma_op_len;
/* For some operations, the hardware cannot tell the exact message
* length, and in those cases it reports zero. Use prefetch
* logic. */
if (length == 0) {
prefetch_activated = 1;
length = mpfault->rdma.packet_size;
prefetch_len = min(MAX_PREFETCH_LEN, prefetch_len);
}
ret = pagefault_single_data_segment(qp, pfault, rkey, address, length,
NULL);
if (ret == -EAGAIN) {
/* We're racing with an invalidation, don't prefetch */
prefetch_activated = 0;
} else if (ret < 0 || pages_in_range(address, length) > ret) {
mlx5_ib_page_fault_resume(qp, pfault, 1);
return;
}
mlx5_ib_page_fault_resume(qp, pfault, 0);
/* At this point, there might be a new pagefault already arriving in
* the eq, switch to the dummy pagefault for the rest of the
* processing. We're still OK with the objects being alive as the
* work-queue is being fenced. */
if (prefetch_activated) {
ret = pagefault_single_data_segment(qp, &dummy_pfault, rkey,
address,
prefetch_len,
NULL);
if (ret < 0) {
pr_warn("Prefetch failed (ret = %d, prefetch_activated = %d) for QPN %d, address: 0x%.16llx, length = 0x%.16x\n",
ret, prefetch_activated,
qp->ibqp.qp_num, address, prefetch_len);
}
}
}
void mlx5_ib_mr_pfault_handler(struct mlx5_ib_qp *qp, void mlx5_ib_mr_pfault_handler(struct mlx5_ib_qp *qp,
struct mlx5_ib_pfault *pfault) struct mlx5_ib_pfault *pfault)
{ {
...@@ -499,6 +575,9 @@ void mlx5_ib_mr_pfault_handler(struct mlx5_ib_qp *qp, ...@@ -499,6 +575,9 @@ void mlx5_ib_mr_pfault_handler(struct mlx5_ib_qp *qp,
case MLX5_PFAULT_SUBTYPE_WQE: case MLX5_PFAULT_SUBTYPE_WQE:
mlx5_ib_mr_wqe_pfault_handler(qp, pfault); mlx5_ib_mr_wqe_pfault_handler(qp, pfault);
break; break;
case MLX5_PFAULT_SUBTYPE_RDMA:
mlx5_ib_mr_rdma_pfault_handler(qp, pfault);
break;
default: default:
pr_warn("Invalid page fault event subtype: 0x%x\n", pr_warn("Invalid page fault event subtype: 0x%x\n",
event_subtype); event_subtype);
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment