Commit a49049ea authored by Björn Töpel's avatar Björn Töpel Committed by Daniel Borkmann

xsk: simplified umem setup

As suggested by Daniel Borkmann, the umem setup code was a too
defensive and complex. Here, we reduce the number of checks. Also, the
memory pinning is now folded into the umem creation, and we do correct
locking.
Signed-off-by: default avatarBjörn Töpel <bjorn.topel@intel.com>
Signed-off-by: default avatarDaniel Borkmann <daniel@iogearbox.net>
parent 37b07693
......@@ -16,21 +16,10 @@
#define XDP_UMEM_MIN_FRAME_SIZE 2048
int xdp_umem_create(struct xdp_umem **umem)
{
*umem = kzalloc(sizeof(**umem), GFP_KERNEL);
if (!*umem)
return -ENOMEM;
return 0;
}
static void xdp_umem_unpin_pages(struct xdp_umem *umem)
{
unsigned int i;
if (umem->pgs) {
for (i = 0; i < umem->npgs; i++) {
struct page *page = umem->pgs[i];
......@@ -40,15 +29,12 @@ static void xdp_umem_unpin_pages(struct xdp_umem *umem)
kfree(umem->pgs);
umem->pgs = NULL;
}
}
static void xdp_umem_unaccount_pages(struct xdp_umem *umem)
{
if (umem->user) {
atomic_long_sub(umem->npgs, &umem->user->locked_vm);
free_uid(umem->user);
}
}
static void xdp_umem_release(struct xdp_umem *umem)
......@@ -66,7 +52,6 @@ static void xdp_umem_release(struct xdp_umem *umem)
umem->cq = NULL;
}
if (umem->pgs) {
xdp_umem_unpin_pages(umem);
task = get_pid_task(umem->pid, PIDTYPE_PID);
......@@ -79,9 +64,6 @@ static void xdp_umem_release(struct xdp_umem *umem)
goto out;
mmput(mm);
umem->pgs = NULL;
}
xdp_umem_unaccount_pages(umem);
out:
kfree(umem);
......@@ -167,16 +149,13 @@ static int xdp_umem_account_pages(struct xdp_umem *umem)
return 0;
}
int xdp_umem_reg(struct xdp_umem *umem, struct xdp_umem_reg *mr)
static int xdp_umem_reg(struct xdp_umem *umem, struct xdp_umem_reg *mr)
{
u32 frame_size = mr->frame_size, frame_headroom = mr->frame_headroom;
u64 addr = mr->addr, size = mr->len;
unsigned int nframes, nfpp;
int size_chk, err;
if (!umem)
return -EINVAL;
if (frame_size < XDP_UMEM_MIN_FRAME_SIZE || frame_size > PAGE_SIZE) {
/* Strictly speaking we could support this, if:
* - huge pages, or*
......@@ -245,6 +224,24 @@ int xdp_umem_reg(struct xdp_umem *umem, struct xdp_umem_reg *mr)
return err;
}
struct xdp_umem *xdp_umem_create(struct xdp_umem_reg *mr)
{
struct xdp_umem *umem;
int err;
umem = kzalloc(sizeof(*umem), GFP_KERNEL);
if (!umem)
return ERR_PTR(-ENOMEM);
err = xdp_umem_reg(umem, mr);
if (err) {
kfree(umem);
return ERR_PTR(err);
}
return umem;
}
bool xdp_umem_validate_queues(struct xdp_umem *umem)
{
return umem->fq && umem->cq;
......
......@@ -50,9 +50,8 @@ static inline char *xdp_umem_get_data_with_headroom(struct xdp_umem *umem,
}
bool xdp_umem_validate_queues(struct xdp_umem *umem);
int xdp_umem_reg(struct xdp_umem *umem, struct xdp_umem_reg *mr);
void xdp_get_umem(struct xdp_umem *umem);
void xdp_put_umem(struct xdp_umem *umem);
int xdp_umem_create(struct xdp_umem **umem);
struct xdp_umem *xdp_umem_create(struct xdp_umem_reg *mr);
#endif /* XDP_UMEM_H_ */
......@@ -406,25 +406,23 @@ static int xsk_setsockopt(struct socket *sock, int level, int optname,
struct xdp_umem_reg mr;
struct xdp_umem *umem;
if (xs->umem)
return -EBUSY;
if (copy_from_user(&mr, optval, sizeof(mr)))
return -EFAULT;
mutex_lock(&xs->mutex);
err = xdp_umem_create(&umem);
if (xs->umem) {
mutex_unlock(&xs->mutex);
return -EBUSY;
}
err = xdp_umem_reg(umem, &mr);
if (err) {
kfree(umem);
umem = xdp_umem_create(&mr);
if (IS_ERR(umem)) {
mutex_unlock(&xs->mutex);
return err;
return PTR_ERR(umem);
}
/* Make sure umem is ready before it can be seen by others */
smp_wmb();
xs->umem = umem;
mutex_unlock(&xs->mutex);
return 0;
......@@ -435,13 +433,15 @@ static int xsk_setsockopt(struct socket *sock, int level, int optname,
struct xsk_queue **q;
int entries;
if (!xs->umem)
return -EINVAL;
if (copy_from_user(&entries, optval, sizeof(entries)))
return -EFAULT;
mutex_lock(&xs->mutex);
if (!xs->umem) {
mutex_unlock(&xs->mutex);
return -EINVAL;
}
q = (optname == XDP_UMEM_FILL_RING) ? &xs->umem->fq :
&xs->umem->cq;
err = xsk_init_queue(entries, q, true);
......
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