devmap/cpumap: Use flush list instead of bitmap

The socket map uses a linked list instead of a bitmap to keep track of
which entries to flush. Do the same for devmap and cpumap, as this means we
don't have to care about the map index when enqueueing things into the
map (and so we can cache the map lookup).

Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Acked-by: Jonathan Lemon <jonathan.lemon@gmail.com>
Acked-by: Andrii Nakryiko <andriin@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
This commit is contained in:
Toke Høiland-Jørgensen 2019-06-28 11:12:34 +02:00 committed by Daniel Borkmann
parent c8af5cd75e
commit d5df2830ca
3 changed files with 95 additions and 119 deletions

View File

@ -32,14 +32,19 @@
/* General idea: XDP packets getting XDP redirected to another CPU, /* General idea: XDP packets getting XDP redirected to another CPU,
* will maximum be stored/queued for one driver ->poll() call. It is * will maximum be stored/queued for one driver ->poll() call. It is
* guaranteed that setting flush bit and flush operation happen on * guaranteed that queueing the frame and the flush operation happen on
* same CPU. Thus, cpu_map_flush operation can deduct via this_cpu_ptr() * same CPU. Thus, cpu_map_flush operation can deduct via this_cpu_ptr()
* which queue in bpf_cpu_map_entry contains packets. * which queue in bpf_cpu_map_entry contains packets.
*/ */
#define CPU_MAP_BULK_SIZE 8 /* 8 == one cacheline on 64-bit archs */ #define CPU_MAP_BULK_SIZE 8 /* 8 == one cacheline on 64-bit archs */
struct bpf_cpu_map_entry;
struct bpf_cpu_map;
struct xdp_bulk_queue { struct xdp_bulk_queue {
void *q[CPU_MAP_BULK_SIZE]; void *q[CPU_MAP_BULK_SIZE];
struct list_head flush_node;
struct bpf_cpu_map_entry *obj;
unsigned int count; unsigned int count;
}; };
@ -52,6 +57,8 @@ struct bpf_cpu_map_entry {
/* XDP can run multiple RX-ring queues, need __percpu enqueue store */ /* XDP can run multiple RX-ring queues, need __percpu enqueue store */
struct xdp_bulk_queue __percpu *bulkq; struct xdp_bulk_queue __percpu *bulkq;
struct bpf_cpu_map *cmap;
/* Queue with potential multi-producers, and single-consumer kthread */ /* Queue with potential multi-producers, and single-consumer kthread */
struct ptr_ring *queue; struct ptr_ring *queue;
struct task_struct *kthread; struct task_struct *kthread;
@ -65,23 +72,17 @@ struct bpf_cpu_map {
struct bpf_map map; struct bpf_map map;
/* Below members specific for map type */ /* Below members specific for map type */
struct bpf_cpu_map_entry **cpu_map; struct bpf_cpu_map_entry **cpu_map;
unsigned long __percpu *flush_needed; struct list_head __percpu *flush_list;
}; };
static int bq_flush_to_queue(struct bpf_cpu_map_entry *rcpu, static int bq_flush_to_queue(struct xdp_bulk_queue *bq, bool in_napi_ctx);
struct xdp_bulk_queue *bq, bool in_napi_ctx);
static u64 cpu_map_bitmap_size(const union bpf_attr *attr)
{
return BITS_TO_LONGS(attr->max_entries) * sizeof(unsigned long);
}
static struct bpf_map *cpu_map_alloc(union bpf_attr *attr) static struct bpf_map *cpu_map_alloc(union bpf_attr *attr)
{ {
struct bpf_cpu_map *cmap; struct bpf_cpu_map *cmap;
int err = -ENOMEM; int err = -ENOMEM;
int ret, cpu;
u64 cost; u64 cost;
int ret;
if (!capable(CAP_SYS_ADMIN)) if (!capable(CAP_SYS_ADMIN))
return ERR_PTR(-EPERM); return ERR_PTR(-EPERM);
@ -105,7 +106,7 @@ static struct bpf_map *cpu_map_alloc(union bpf_attr *attr)
/* make sure page count doesn't overflow */ /* make sure page count doesn't overflow */
cost = (u64) cmap->map.max_entries * sizeof(struct bpf_cpu_map_entry *); cost = (u64) cmap->map.max_entries * sizeof(struct bpf_cpu_map_entry *);
cost += cpu_map_bitmap_size(attr) * num_possible_cpus(); cost += sizeof(struct list_head) * num_possible_cpus();
/* Notice returns -EPERM on if map size is larger than memlock limit */ /* Notice returns -EPERM on if map size is larger than memlock limit */
ret = bpf_map_charge_init(&cmap->map.memory, cost); ret = bpf_map_charge_init(&cmap->map.memory, cost);
@ -114,12 +115,13 @@ static struct bpf_map *cpu_map_alloc(union bpf_attr *attr)
goto free_cmap; goto free_cmap;
} }
/* A per cpu bitfield with a bit per possible CPU in map */ cmap->flush_list = alloc_percpu(struct list_head);
cmap->flush_needed = __alloc_percpu(cpu_map_bitmap_size(attr), if (!cmap->flush_list)
__alignof__(unsigned long));
if (!cmap->flush_needed)
goto free_charge; goto free_charge;
for_each_possible_cpu(cpu)
INIT_LIST_HEAD(per_cpu_ptr(cmap->flush_list, cpu));
/* Alloc array for possible remote "destination" CPUs */ /* Alloc array for possible remote "destination" CPUs */
cmap->cpu_map = bpf_map_area_alloc(cmap->map.max_entries * cmap->cpu_map = bpf_map_area_alloc(cmap->map.max_entries *
sizeof(struct bpf_cpu_map_entry *), sizeof(struct bpf_cpu_map_entry *),
@ -129,7 +131,7 @@ static struct bpf_map *cpu_map_alloc(union bpf_attr *attr)
return &cmap->map; return &cmap->map;
free_percpu: free_percpu:
free_percpu(cmap->flush_needed); free_percpu(cmap->flush_list);
free_charge: free_charge:
bpf_map_charge_finish(&cmap->map.memory); bpf_map_charge_finish(&cmap->map.memory);
free_cmap: free_cmap:
@ -334,7 +336,8 @@ static struct bpf_cpu_map_entry *__cpu_map_entry_alloc(u32 qsize, u32 cpu,
{ {
gfp_t gfp = GFP_KERNEL | __GFP_NOWARN; gfp_t gfp = GFP_KERNEL | __GFP_NOWARN;
struct bpf_cpu_map_entry *rcpu; struct bpf_cpu_map_entry *rcpu;
int numa, err; struct xdp_bulk_queue *bq;
int numa, err, i;
/* Have map->numa_node, but choose node of redirect target CPU */ /* Have map->numa_node, but choose node of redirect target CPU */
numa = cpu_to_node(cpu); numa = cpu_to_node(cpu);
@ -349,6 +352,11 @@ static struct bpf_cpu_map_entry *__cpu_map_entry_alloc(u32 qsize, u32 cpu,
if (!rcpu->bulkq) if (!rcpu->bulkq)
goto free_rcu; goto free_rcu;
for_each_possible_cpu(i) {
bq = per_cpu_ptr(rcpu->bulkq, i);
bq->obj = rcpu;
}
/* Alloc queue */ /* Alloc queue */
rcpu->queue = kzalloc_node(sizeof(*rcpu->queue), gfp, numa); rcpu->queue = kzalloc_node(sizeof(*rcpu->queue), gfp, numa);
if (!rcpu->queue) if (!rcpu->queue)
@ -405,7 +413,7 @@ static void __cpu_map_entry_free(struct rcu_head *rcu)
struct xdp_bulk_queue *bq = per_cpu_ptr(rcpu->bulkq, cpu); struct xdp_bulk_queue *bq = per_cpu_ptr(rcpu->bulkq, cpu);
/* No concurrent bq_enqueue can run at this point */ /* No concurrent bq_enqueue can run at this point */
bq_flush_to_queue(rcpu, bq, false); bq_flush_to_queue(bq, false);
} }
free_percpu(rcpu->bulkq); free_percpu(rcpu->bulkq);
/* Cannot kthread_stop() here, last put free rcpu resources */ /* Cannot kthread_stop() here, last put free rcpu resources */
@ -488,6 +496,7 @@ static int cpu_map_update_elem(struct bpf_map *map, void *key, void *value,
rcpu = __cpu_map_entry_alloc(qsize, key_cpu, map->id); rcpu = __cpu_map_entry_alloc(qsize, key_cpu, map->id);
if (!rcpu) if (!rcpu)
return -ENOMEM; return -ENOMEM;
rcpu->cmap = cmap;
} }
rcu_read_lock(); rcu_read_lock();
__cpu_map_entry_replace(cmap, key_cpu, rcpu); __cpu_map_entry_replace(cmap, key_cpu, rcpu);
@ -514,14 +523,14 @@ static void cpu_map_free(struct bpf_map *map)
synchronize_rcu(); synchronize_rcu();
/* To ensure all pending flush operations have completed wait for flush /* To ensure all pending flush operations have completed wait for flush
* bitmap to indicate all flush_needed bits to be zero on _all_ cpus. * list be empty on _all_ cpus. Because the above synchronize_rcu()
* Because the above synchronize_rcu() ensures the map is disconnected * ensures the map is disconnected from the program we can assume no new
* from the program we can assume no new bits will be set. * items will be added to the list.
*/ */
for_each_online_cpu(cpu) { for_each_online_cpu(cpu) {
unsigned long *bitmap = per_cpu_ptr(cmap->flush_needed, cpu); struct list_head *flush_list = per_cpu_ptr(cmap->flush_list, cpu);
while (!bitmap_empty(bitmap, cmap->map.max_entries)) while (!list_empty(flush_list))
cond_resched(); cond_resched();
} }
@ -538,7 +547,7 @@ static void cpu_map_free(struct bpf_map *map)
/* bq flush and cleanup happens after RCU graze-period */ /* bq flush and cleanup happens after RCU graze-period */
__cpu_map_entry_replace(cmap, i, NULL); /* call_rcu */ __cpu_map_entry_replace(cmap, i, NULL); /* call_rcu */
} }
free_percpu(cmap->flush_needed); free_percpu(cmap->flush_list);
bpf_map_area_free(cmap->cpu_map); bpf_map_area_free(cmap->cpu_map);
kfree(cmap); kfree(cmap);
} }
@ -590,9 +599,9 @@ const struct bpf_map_ops cpu_map_ops = {
.map_check_btf = map_check_no_btf, .map_check_btf = map_check_no_btf,
}; };
static int bq_flush_to_queue(struct bpf_cpu_map_entry *rcpu, static int bq_flush_to_queue(struct xdp_bulk_queue *bq, bool in_napi_ctx)
struct xdp_bulk_queue *bq, bool in_napi_ctx)
{ {
struct bpf_cpu_map_entry *rcpu = bq->obj;
unsigned int processed = 0, drops = 0; unsigned int processed = 0, drops = 0;
const int to_cpu = rcpu->cpu; const int to_cpu = rcpu->cpu;
struct ptr_ring *q; struct ptr_ring *q;
@ -621,6 +630,8 @@ static int bq_flush_to_queue(struct bpf_cpu_map_entry *rcpu,
bq->count = 0; bq->count = 0;
spin_unlock(&q->producer_lock); spin_unlock(&q->producer_lock);
__list_del_clearprev(&bq->flush_node);
/* Feedback loop via tracepoints */ /* Feedback loop via tracepoints */
trace_xdp_cpumap_enqueue(rcpu->map_id, processed, drops, to_cpu); trace_xdp_cpumap_enqueue(rcpu->map_id, processed, drops, to_cpu);
return 0; return 0;
@ -631,10 +642,11 @@ static int bq_flush_to_queue(struct bpf_cpu_map_entry *rcpu,
*/ */
static int bq_enqueue(struct bpf_cpu_map_entry *rcpu, struct xdp_frame *xdpf) static int bq_enqueue(struct bpf_cpu_map_entry *rcpu, struct xdp_frame *xdpf)
{ {
struct list_head *flush_list = this_cpu_ptr(rcpu->cmap->flush_list);
struct xdp_bulk_queue *bq = this_cpu_ptr(rcpu->bulkq); struct xdp_bulk_queue *bq = this_cpu_ptr(rcpu->bulkq);
if (unlikely(bq->count == CPU_MAP_BULK_SIZE)) if (unlikely(bq->count == CPU_MAP_BULK_SIZE))
bq_flush_to_queue(rcpu, bq, true); bq_flush_to_queue(bq, true);
/* Notice, xdp_buff/page MUST be queued here, long enough for /* Notice, xdp_buff/page MUST be queued here, long enough for
* driver to code invoking us to finished, due to driver * driver to code invoking us to finished, due to driver
@ -646,6 +658,10 @@ static int bq_enqueue(struct bpf_cpu_map_entry *rcpu, struct xdp_frame *xdpf)
* operation, when completing napi->poll call. * operation, when completing napi->poll call.
*/ */
bq->q[bq->count++] = xdpf; bq->q[bq->count++] = xdpf;
if (!bq->flush_node.prev)
list_add(&bq->flush_node, flush_list);
return 0; return 0;
} }
@ -665,41 +681,16 @@ int cpu_map_enqueue(struct bpf_cpu_map_entry *rcpu, struct xdp_buff *xdp,
return 0; return 0;
} }
void __cpu_map_insert_ctx(struct bpf_map *map, u32 bit)
{
struct bpf_cpu_map *cmap = container_of(map, struct bpf_cpu_map, map);
unsigned long *bitmap = this_cpu_ptr(cmap->flush_needed);
__set_bit(bit, bitmap);
}
void __cpu_map_flush(struct bpf_map *map) void __cpu_map_flush(struct bpf_map *map)
{ {
struct bpf_cpu_map *cmap = container_of(map, struct bpf_cpu_map, map); struct bpf_cpu_map *cmap = container_of(map, struct bpf_cpu_map, map);
unsigned long *bitmap = this_cpu_ptr(cmap->flush_needed); struct list_head *flush_list = this_cpu_ptr(cmap->flush_list);
u32 bit; struct xdp_bulk_queue *bq, *tmp;
/* The napi->poll softirq makes sure __cpu_map_insert_ctx() list_for_each_entry_safe(bq, tmp, flush_list, flush_node) {
* and __cpu_map_flush() happen on same CPU. Thus, the percpu bq_flush_to_queue(bq, true);
* bitmap indicate which percpu bulkq have packets.
*/
for_each_set_bit(bit, bitmap, map->max_entries) {
struct bpf_cpu_map_entry *rcpu = READ_ONCE(cmap->cpu_map[bit]);
struct xdp_bulk_queue *bq;
/* This is possible if entry is removed by user space
* between xdp redirect and flush op.
*/
if (unlikely(!rcpu))
continue;
__clear_bit(bit, bitmap);
/* Flush all frames in bulkq to real queue */
bq = this_cpu_ptr(rcpu->bulkq);
bq_flush_to_queue(rcpu, bq, true);
/* If already running, costs spin_lock_irqsave + smb_mb */ /* If already running, costs spin_lock_irqsave + smb_mb */
wake_up_process(rcpu->kthread); wake_up_process(bq->obj->kthread);
} }
} }

View File

@ -17,9 +17,8 @@
* datapath always has a valid copy. However, the datapath does a "flush" * datapath always has a valid copy. However, the datapath does a "flush"
* operation that pushes any pending packets in the driver outside the RCU * operation that pushes any pending packets in the driver outside the RCU
* critical section. Each bpf_dtab_netdev tracks these pending operations using * critical section. Each bpf_dtab_netdev tracks these pending operations using
* an atomic per-cpu bitmap. The bpf_dtab_netdev object will not be destroyed * a per-cpu flush list. The bpf_dtab_netdev object will not be destroyed until
* until all bits are cleared indicating outstanding flush operations have * this list is empty, indicating outstanding flush operations have completed.
* completed.
* *
* BPF syscalls may race with BPF program calls on any of the update, delete * BPF syscalls may race with BPF program calls on any of the update, delete
* or lookup operations. As noted above the xchg() operation also keep the * or lookup operations. As noted above the xchg() operation also keep the
@ -48,9 +47,13 @@
(BPF_F_NUMA_NODE | BPF_F_RDONLY | BPF_F_WRONLY) (BPF_F_NUMA_NODE | BPF_F_RDONLY | BPF_F_WRONLY)
#define DEV_MAP_BULK_SIZE 16 #define DEV_MAP_BULK_SIZE 16
struct bpf_dtab_netdev;
struct xdp_bulk_queue { struct xdp_bulk_queue {
struct xdp_frame *q[DEV_MAP_BULK_SIZE]; struct xdp_frame *q[DEV_MAP_BULK_SIZE];
struct list_head flush_node;
struct net_device *dev_rx; struct net_device *dev_rx;
struct bpf_dtab_netdev *obj;
unsigned int count; unsigned int count;
}; };
@ -65,23 +68,18 @@ struct bpf_dtab_netdev {
struct bpf_dtab { struct bpf_dtab {
struct bpf_map map; struct bpf_map map;
struct bpf_dtab_netdev **netdev_map; struct bpf_dtab_netdev **netdev_map;
unsigned long __percpu *flush_needed; struct list_head __percpu *flush_list;
struct list_head list; struct list_head list;
}; };
static DEFINE_SPINLOCK(dev_map_lock); static DEFINE_SPINLOCK(dev_map_lock);
static LIST_HEAD(dev_map_list); static LIST_HEAD(dev_map_list);
static u64 dev_map_bitmap_size(const union bpf_attr *attr)
{
return BITS_TO_LONGS((u64) attr->max_entries) * sizeof(unsigned long);
}
static struct bpf_map *dev_map_alloc(union bpf_attr *attr) static struct bpf_map *dev_map_alloc(union bpf_attr *attr)
{ {
struct bpf_dtab *dtab; struct bpf_dtab *dtab;
int err, cpu;
u64 cost; u64 cost;
int err;
if (!capable(CAP_NET_ADMIN)) if (!capable(CAP_NET_ADMIN))
return ERR_PTR(-EPERM); return ERR_PTR(-EPERM);
@ -99,7 +97,7 @@ static struct bpf_map *dev_map_alloc(union bpf_attr *attr)
/* make sure page count doesn't overflow */ /* make sure page count doesn't overflow */
cost = (u64) dtab->map.max_entries * sizeof(struct bpf_dtab_netdev *); cost = (u64) dtab->map.max_entries * sizeof(struct bpf_dtab_netdev *);
cost += dev_map_bitmap_size(attr) * num_possible_cpus(); cost += sizeof(struct list_head) * num_possible_cpus();
/* if map size is larger than memlock limit, reject it */ /* if map size is larger than memlock limit, reject it */
err = bpf_map_charge_init(&dtab->map.memory, cost); err = bpf_map_charge_init(&dtab->map.memory, cost);
@ -108,28 +106,30 @@ static struct bpf_map *dev_map_alloc(union bpf_attr *attr)
err = -ENOMEM; err = -ENOMEM;
/* A per cpu bitfield with a bit per possible net device */ dtab->flush_list = alloc_percpu(struct list_head);
dtab->flush_needed = __alloc_percpu_gfp(dev_map_bitmap_size(attr), if (!dtab->flush_list)
__alignof__(unsigned long),
GFP_KERNEL | __GFP_NOWARN);
if (!dtab->flush_needed)
goto free_charge; goto free_charge;
for_each_possible_cpu(cpu)
INIT_LIST_HEAD(per_cpu_ptr(dtab->flush_list, cpu));
dtab->netdev_map = bpf_map_area_alloc(dtab->map.max_entries * dtab->netdev_map = bpf_map_area_alloc(dtab->map.max_entries *
sizeof(struct bpf_dtab_netdev *), sizeof(struct bpf_dtab_netdev *),
dtab->map.numa_node); dtab->map.numa_node);
if (!dtab->netdev_map) if (!dtab->netdev_map)
goto free_charge; goto free_percpu;
spin_lock(&dev_map_lock); spin_lock(&dev_map_lock);
list_add_tail_rcu(&dtab->list, &dev_map_list); list_add_tail_rcu(&dtab->list, &dev_map_list);
spin_unlock(&dev_map_lock); spin_unlock(&dev_map_lock);
return &dtab->map; return &dtab->map;
free_percpu:
free_percpu(dtab->flush_list);
free_charge: free_charge:
bpf_map_charge_finish(&dtab->map.memory); bpf_map_charge_finish(&dtab->map.memory);
free_dtab: free_dtab:
free_percpu(dtab->flush_needed);
kfree(dtab); kfree(dtab);
return ERR_PTR(err); return ERR_PTR(err);
} }
@ -158,14 +158,14 @@ static void dev_map_free(struct bpf_map *map)
rcu_barrier(); rcu_barrier();
/* To ensure all pending flush operations have completed wait for flush /* To ensure all pending flush operations have completed wait for flush
* bitmap to indicate all flush_needed bits to be zero on _all_ cpus. * list to empty on _all_ cpus.
* Because the above synchronize_rcu() ensures the map is disconnected * Because the above synchronize_rcu() ensures the map is disconnected
* from the program we can assume no new bits will be set. * from the program we can assume no new items will be added.
*/ */
for_each_online_cpu(cpu) { for_each_online_cpu(cpu) {
unsigned long *bitmap = per_cpu_ptr(dtab->flush_needed, cpu); struct list_head *flush_list = per_cpu_ptr(dtab->flush_list, cpu);
while (!bitmap_empty(bitmap, dtab->map.max_entries)) while (!list_empty(flush_list))
cond_resched(); cond_resched();
} }
@ -181,7 +181,7 @@ static void dev_map_free(struct bpf_map *map)
kfree(dev); kfree(dev);
} }
free_percpu(dtab->flush_needed); free_percpu(dtab->flush_list);
bpf_map_area_free(dtab->netdev_map); bpf_map_area_free(dtab->netdev_map);
kfree(dtab); kfree(dtab);
} }
@ -203,18 +203,10 @@ static int dev_map_get_next_key(struct bpf_map *map, void *key, void *next_key)
return 0; return 0;
} }
void __dev_map_insert_ctx(struct bpf_map *map, u32 bit) static int bq_xmit_all(struct xdp_bulk_queue *bq, u32 flags,
{
struct bpf_dtab *dtab = container_of(map, struct bpf_dtab, map);
unsigned long *bitmap = this_cpu_ptr(dtab->flush_needed);
__set_bit(bit, bitmap);
}
static int bq_xmit_all(struct bpf_dtab_netdev *obj,
struct xdp_bulk_queue *bq, u32 flags,
bool in_napi_ctx) bool in_napi_ctx)
{ {
struct bpf_dtab_netdev *obj = bq->obj;
struct net_device *dev = obj->dev; struct net_device *dev = obj->dev;
int sent = 0, drops = 0, err = 0; int sent = 0, drops = 0, err = 0;
int i; int i;
@ -241,6 +233,7 @@ static int bq_xmit_all(struct bpf_dtab_netdev *obj,
trace_xdp_devmap_xmit(&obj->dtab->map, obj->bit, trace_xdp_devmap_xmit(&obj->dtab->map, obj->bit,
sent, drops, bq->dev_rx, dev, err); sent, drops, bq->dev_rx, dev, err);
bq->dev_rx = NULL; bq->dev_rx = NULL;
__list_del_clearprev(&bq->flush_node);
return 0; return 0;
error: error:
/* If ndo_xdp_xmit fails with an errno, no frames have been /* If ndo_xdp_xmit fails with an errno, no frames have been
@ -263,31 +256,18 @@ static int bq_xmit_all(struct bpf_dtab_netdev *obj,
* from the driver before returning from its napi->poll() routine. The poll() * from the driver before returning from its napi->poll() routine. The poll()
* routine is called either from busy_poll context or net_rx_action signaled * routine is called either from busy_poll context or net_rx_action signaled
* from NET_RX_SOFTIRQ. Either way the poll routine must complete before the * from NET_RX_SOFTIRQ. Either way the poll routine must complete before the
* net device can be torn down. On devmap tear down we ensure the ctx bitmap * net device can be torn down. On devmap tear down we ensure the flush list
* is zeroed before completing to ensure all flush operations have completed. * is empty before completing to ensure all flush operations have completed.
*/ */
void __dev_map_flush(struct bpf_map *map) void __dev_map_flush(struct bpf_map *map)
{ {
struct bpf_dtab *dtab = container_of(map, struct bpf_dtab, map); struct bpf_dtab *dtab = container_of(map, struct bpf_dtab, map);
unsigned long *bitmap = this_cpu_ptr(dtab->flush_needed); struct list_head *flush_list = this_cpu_ptr(dtab->flush_list);
u32 bit; struct xdp_bulk_queue *bq, *tmp;
rcu_read_lock(); rcu_read_lock();
for_each_set_bit(bit, bitmap, map->max_entries) { list_for_each_entry_safe(bq, tmp, flush_list, flush_node)
struct bpf_dtab_netdev *dev = READ_ONCE(dtab->netdev_map[bit]); bq_xmit_all(bq, XDP_XMIT_FLUSH, true);
struct xdp_bulk_queue *bq;
/* This is possible if the dev entry is removed by user space
* between xdp redirect and flush op.
*/
if (unlikely(!dev))
continue;
bq = this_cpu_ptr(dev->bulkq);
bq_xmit_all(dev, bq, XDP_XMIT_FLUSH, true);
__clear_bit(bit, bitmap);
}
rcu_read_unlock(); rcu_read_unlock();
} }
@ -314,10 +294,11 @@ static int bq_enqueue(struct bpf_dtab_netdev *obj, struct xdp_frame *xdpf,
struct net_device *dev_rx) struct net_device *dev_rx)
{ {
struct list_head *flush_list = this_cpu_ptr(obj->dtab->flush_list);
struct xdp_bulk_queue *bq = this_cpu_ptr(obj->bulkq); struct xdp_bulk_queue *bq = this_cpu_ptr(obj->bulkq);
if (unlikely(bq->count == DEV_MAP_BULK_SIZE)) if (unlikely(bq->count == DEV_MAP_BULK_SIZE))
bq_xmit_all(obj, bq, 0, true); bq_xmit_all(bq, 0, true);
/* Ingress dev_rx will be the same for all xdp_frame's in /* Ingress dev_rx will be the same for all xdp_frame's in
* bulk_queue, because bq stored per-CPU and must be flushed * bulk_queue, because bq stored per-CPU and must be flushed
@ -327,6 +308,10 @@ static int bq_enqueue(struct bpf_dtab_netdev *obj, struct xdp_frame *xdpf,
bq->dev_rx = dev_rx; bq->dev_rx = dev_rx;
bq->q[bq->count++] = xdpf; bq->q[bq->count++] = xdpf;
if (!bq->flush_node.prev)
list_add(&bq->flush_node, flush_list);
return 0; return 0;
} }
@ -377,17 +362,12 @@ static void dev_map_flush_old(struct bpf_dtab_netdev *dev)
{ {
if (dev->dev->netdev_ops->ndo_xdp_xmit) { if (dev->dev->netdev_ops->ndo_xdp_xmit) {
struct xdp_bulk_queue *bq; struct xdp_bulk_queue *bq;
unsigned long *bitmap;
int cpu; int cpu;
rcu_read_lock(); rcu_read_lock();
for_each_online_cpu(cpu) { for_each_online_cpu(cpu) {
bitmap = per_cpu_ptr(dev->dtab->flush_needed, cpu);
__clear_bit(dev->bit, bitmap);
bq = per_cpu_ptr(dev->bulkq, cpu); bq = per_cpu_ptr(dev->bulkq, cpu);
bq_xmit_all(dev, bq, XDP_XMIT_FLUSH, false); bq_xmit_all(bq, XDP_XMIT_FLUSH, false);
} }
rcu_read_unlock(); rcu_read_unlock();
} }
@ -434,8 +414,10 @@ static int dev_map_update_elem(struct bpf_map *map, void *key, void *value,
struct net *net = current->nsproxy->net_ns; struct net *net = current->nsproxy->net_ns;
gfp_t gfp = GFP_ATOMIC | __GFP_NOWARN; gfp_t gfp = GFP_ATOMIC | __GFP_NOWARN;
struct bpf_dtab_netdev *dev, *old_dev; struct bpf_dtab_netdev *dev, *old_dev;
u32 i = *(u32 *)key;
u32 ifindex = *(u32 *)value; u32 ifindex = *(u32 *)value;
struct xdp_bulk_queue *bq;
u32 i = *(u32 *)key;
int cpu;
if (unlikely(map_flags > BPF_EXIST)) if (unlikely(map_flags > BPF_EXIST))
return -EINVAL; return -EINVAL;
@ -458,6 +440,11 @@ static int dev_map_update_elem(struct bpf_map *map, void *key, void *value,
return -ENOMEM; return -ENOMEM;
} }
for_each_possible_cpu(cpu) {
bq = per_cpu_ptr(dev->bulkq, cpu);
bq->obj = dev;
}
dev->dev = dev_get_by_index(net, ifindex); dev->dev = dev_get_by_index(net, ifindex);
if (!dev->dev) { if (!dev->dev) {
free_percpu(dev->bulkq); free_percpu(dev->bulkq);

View File

@ -3523,7 +3523,6 @@ static int __bpf_tx_xdp_map(struct net_device *dev_rx, void *fwd,
err = dev_map_enqueue(dst, xdp, dev_rx); err = dev_map_enqueue(dst, xdp, dev_rx);
if (unlikely(err)) if (unlikely(err))
return err; return err;
__dev_map_insert_ctx(map, index);
break; break;
} }
case BPF_MAP_TYPE_CPUMAP: { case BPF_MAP_TYPE_CPUMAP: {
@ -3532,7 +3531,6 @@ static int __bpf_tx_xdp_map(struct net_device *dev_rx, void *fwd,
err = cpu_map_enqueue(rcpu, xdp, dev_rx); err = cpu_map_enqueue(rcpu, xdp, dev_rx);
if (unlikely(err)) if (unlikely(err))
return err; return err;
__cpu_map_insert_ctx(map, index);
break; break;
} }
case BPF_MAP_TYPE_XSKMAP: { case BPF_MAP_TYPE_XSKMAP: {