On 2018/11/19 3:34, Yuval Shaia wrote:
On Sun, Nov 18, 2018 at 09:37:09PM +0800, Yanjun Zhu wrote:
On 2018/11/18 21:35, Yuval Shaia wrote:
On Sun, Nov 18, 2018 at 08:21:24AM -0500, Yanjun Zhu wrote:
Attempt to allocate physically contiguous memory, but upon
failure, fall back to non-contiguous (vmalloc) allocation.
Signed-off-by: Zhu Yanjun <yanjun.zhu@xxxxxxxxxx>
---
drivers/infiniband/sw/rxe/rxe_cq.c | 2 +-
drivers/infiniband/sw/rxe/rxe_qp.c | 4 ++--
drivers/infiniband/sw/rxe/rxe_queue.c | 6 +++---
drivers/infiniband/sw/rxe/rxe_srq.c | 2 +-
4 files changed, 7 insertions(+), 7 deletions(-)
diff --git a/drivers/infiniband/sw/rxe/rxe_cq.c b/drivers/infiniband/sw/rxe/rxe_cq.c
index a57276f2cb84..b85ca024fa58 100644
--- a/drivers/infiniband/sw/rxe/rxe_cq.c
+++ b/drivers/infiniband/sw/rxe/rxe_cq.c
@@ -98,7 +98,7 @@ int rxe_cq_from_init(struct rxe_dev *rxe, struct rxe_cq *cq, int cqe,
cq->queue->buf, cq->queue->buf_size, &cq->queue->ip);
if (err) {
vfree(cq->queue->buf);
- kfree(cq->queue);
+ kvfree(cq->queue);
But queue is created with kmalloc, what am i missing here?
Now I use kvmalloc to create queue.
Zhu Yanjun
I see, my bad, i was misled by commit message.
Anyway, I do not see here an attempt to allocated with kmalloc and switch
to kvmalloc on failure, i see that you just decided to use kvmalloc
*instead* of kmalloc, right?
Yes. kvmalloc is to allocate physically contiguous memory. If failure,
fall back to non-contiguous (vmalloc) allocation.
Zhu Yanjun
Is this patch based on some other patch that is not yet in for-next?
return err;
}
diff --git a/drivers/infiniband/sw/rxe/rxe_qp.c b/drivers/infiniband/sw/rxe/rxe_qp.c
index b9710907dac2..fd233f93ac1d 100644
--- a/drivers/infiniband/sw/rxe/rxe_qp.c
+++ b/drivers/infiniband/sw/rxe/rxe_qp.c
@@ -259,7 +259,7 @@ static int rxe_qp_init_req(struct rxe_dev *rxe, struct rxe_qp *qp,
if (err) {
vfree(qp->sq.queue->buf);
- kfree(qp->sq.queue);
+ kvfree(qp->sq.queue);
return err;
}
@@ -312,7 +312,7 @@ static int rxe_qp_init_resp(struct rxe_dev *rxe, struct rxe_qp *qp,
&qp->rq.queue->ip);
if (err) {
vfree(qp->rq.queue->buf);
- kfree(qp->rq.queue);
+ kvfree(qp->rq.queue);
return err;
}
}
diff --git a/drivers/infiniband/sw/rxe/rxe_queue.c b/drivers/infiniband/sw/rxe/rxe_queue.c
index f84ab4469261..e811a655b286 100644
--- a/drivers/infiniband/sw/rxe/rxe_queue.c
+++ b/drivers/infiniband/sw/rxe/rxe_queue.c
@@ -91,7 +91,7 @@ struct rxe_queue *rxe_queue_init(struct rxe_dev *rxe,
if (*num_elem < 0)
goto err1;
- q = kmalloc(sizeof(*q), GFP_KERNEL);
+ q = kvmalloc(sizeof(*q), GFP_KERNEL);
Where is the "fall back" logic?
if (!q)
goto err1;
@@ -174,7 +174,7 @@ int rxe_queue_resize(struct rxe_queue *q,
new_q->buf_size, &new_q->ip);
if (err) {
vfree(new_q->buf);
- kfree(new_q);
+ kvfree(new_q);
goto err1;
}
@@ -208,5 +208,5 @@ void rxe_queue_cleanup(struct rxe_queue *q)
else
vfree(q->buf);
- kfree(q);
+ kvfree(q);
}
diff --git a/drivers/infiniband/sw/rxe/rxe_srq.c b/drivers/infiniband/sw/rxe/rxe_srq.c
index c41a5fee81f7..c9969d157b87 100644
--- a/drivers/infiniband/sw/rxe/rxe_srq.c
+++ b/drivers/infiniband/sw/rxe/rxe_srq.c
@@ -132,7 +132,7 @@ int rxe_srq_from_init(struct rxe_dev *rxe, struct rxe_srq *srq,
q->buf_size, &q->ip);
if (err) {
vfree(q->buf);
- kfree(q);
+ kvfree(q);
return err;
}
--
2.17.1
