* [PULL v2 1/1] virtio/s390: implement PM operations for virtio_ccw
From: Cornelia Huck @ 2017-12-18 16:21 UTC (permalink / raw)
To: mst; +Cc: linux-s390, kvm, Cornelia Huck, virtualization
In-Reply-To: <20171218162123.963-1-cohuck@redhat.com>
From: Christian Borntraeger <borntraeger@de.ibm.com>
Suspend/Resume to/from disk currently fails. Let us wire
up the necessary callbacks. This is mostly just forwarding
the requests to the virtio drivers. The only thing that
has to be done in virtio_ccw itself is to re-set the
virtio revision.
Suggested-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Message-Id: <20171207141102.70190-2-borntraeger@de.ibm.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
[CH: merged <20171218083706.223836-1-borntraeger@de.ibm.com> to fix
!CONFIG_PM configs]
Signed-off-by: Cornelia Huck <cohuck@redhat.com>
---
drivers/s390/virtio/virtio_ccw.c | 29 +++++++++++++++++++++++++++++
1 file changed, 29 insertions(+)
diff --git a/drivers/s390/virtio/virtio_ccw.c b/drivers/s390/virtio/virtio_ccw.c
index b18fe2014cf2..985184ebda45 100644
--- a/drivers/s390/virtio/virtio_ccw.c
+++ b/drivers/s390/virtio/virtio_ccw.c
@@ -1300,6 +1300,9 @@ static int virtio_ccw_cio_notify(struct ccw_device *cdev, int event)
vcdev->device_lost = true;
rc = NOTIFY_DONE;
break;
+ case CIO_OPER:
+ rc = NOTIFY_OK;
+ break;
default:
rc = NOTIFY_DONE;
break;
@@ -1312,6 +1315,27 @@ static struct ccw_device_id virtio_ids[] = {
{},
};
+#ifdef CONFIG_PM_SLEEP
+static int virtio_ccw_freeze(struct ccw_device *cdev)
+{
+ struct virtio_ccw_device *vcdev = dev_get_drvdata(&cdev->dev);
+
+ return virtio_device_freeze(&vcdev->vdev);
+}
+
+static int virtio_ccw_restore(struct ccw_device *cdev)
+{
+ struct virtio_ccw_device *vcdev = dev_get_drvdata(&cdev->dev);
+ int ret;
+
+ ret = virtio_ccw_set_transport_rev(vcdev);
+ if (ret)
+ return ret;
+
+ return virtio_device_restore(&vcdev->vdev);
+}
+#endif
+
static struct ccw_driver virtio_ccw_driver = {
.driver = {
.owner = THIS_MODULE,
@@ -1324,6 +1348,11 @@ static struct ccw_driver virtio_ccw_driver = {
.set_online = virtio_ccw_online,
.notify = virtio_ccw_cio_notify,
.int_class = IRQIO_VIR,
+#ifdef CONFIG_PM_SLEEP
+ .freeze = virtio_ccw_freeze,
+ .thaw = virtio_ccw_restore,
+ .restore = virtio_ccw_restore,
+#endif
};
static int __init pure_hex(char **cp, unsigned int *val, int min_digit,
--
2.13.6
^ permalink raw reply related
* [RFC PATCH] virtio_net: Extend virtio to use VF datapath when available
From: Sridhar Samudrala @ 2017-12-19 0:40 UTC (permalink / raw)
To: mst, stephen, netdev, virtualization, alexander.duyck,
sridhar.samudrala
This patch enables virtio to switch over to a VF datapath when a VF netdev
is present with the same MAC address. It allows live migration of a VM
with a direct attached VF without the need to setup a bond/team between a
VF and virtio net device in the guest.
The hypervisor needs to unplug the VF device from the guest on the source
host and reset the MAC filter of the VF to initiate failover of datapath to
virtio before starting the migration. After the migration is completed, the
destination hypervisor sets the MAC filter on the VF and plugs it back to
the guest to switch over to VF datapath.
It is entirely based on netvsc implementation and it should be possible to
make this code generic and move it to a common location that can be shared
by netvsc and virtio.
Also, i think we should make this a negotiated feature that is off by
default via a new feature bit.
This patch is based on the discussion initiated by Jesse on this thread.
https://marc.info/?l=linux-virtualization&m=151189725224231&w=2
Signed-off-by: Sridhar Samudrala <sridhar.samudrala@intel.com>
Reviewed-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
---
drivers/net/virtio_net.c | 341 ++++++++++++++++++++++++++++++++++++++++++++++-
1 file changed, 339 insertions(+), 2 deletions(-)
diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
index 559b215c0169..a34c717bb15b 100644
--- a/drivers/net/virtio_net.c
+++ b/drivers/net/virtio_net.c
@@ -31,6 +31,8 @@
#include <linux/average.h>
#include <linux/filter.h>
#include <net/route.h>
+#include <linux/netdevice.h>
+#include <linux/netpoll.h>
static int napi_weight = NAPI_POLL_WEIGHT;
module_param(napi_weight, int, 0444);
@@ -56,6 +58,8 @@ module_param(napi_tx, bool, 0644);
*/
DECLARE_EWMA(pkt_len, 0, 64)
+#define VF_TAKEOVER_INT (HZ / 10)
+
#define VIRTNET_DRIVER_VERSION "1.0.0"
static const unsigned long guest_offloads[] = {
@@ -117,6 +121,15 @@ struct receive_queue {
char name[40];
};
+struct virtnet_vf_pcpu_stats {
+ u64 rx_packets;
+ u64 rx_bytes;
+ u64 tx_packets;
+ u64 tx_bytes;
+ struct u64_stats_sync syncp;
+ u32 tx_dropped;
+};
+
struct virtnet_info {
struct virtio_device *vdev;
struct virtqueue *cvq;
@@ -179,6 +192,11 @@ struct virtnet_info {
u32 speed;
unsigned long guest_offloads;
+
+ /* State to manage the associated VF interface. */
+ struct net_device __rcu *vf_netdev;
+ struct virtnet_vf_pcpu_stats __percpu *vf_stats;
+ struct delayed_work vf_takeover;
};
struct padded_vnet_hdr {
@@ -1300,16 +1318,51 @@ static int xmit_skb(struct send_queue *sq, struct sk_buff *skb)
return virtqueue_add_outbuf(sq->vq, sq->sg, num_sg, skb, GFP_ATOMIC);
}
+/* Send skb on the slave VF device. */
+static int virtnet_vf_xmit(struct net_device *dev, struct net_device *vf_netdev,
+ struct sk_buff *skb)
+{
+ struct virtnet_info *vi = netdev_priv(dev);
+ unsigned int len = skb->len;
+ int rc;
+
+ skb->dev = vf_netdev;
+ skb->queue_mapping = qdisc_skb_cb(skb)->slave_dev_queue_mapping;
+
+ rc = dev_queue_xmit(skb);
+ if (likely(rc == NET_XMIT_SUCCESS || rc == NET_XMIT_CN)) {
+ struct virtnet_vf_pcpu_stats *pcpu_stats
+ = this_cpu_ptr(vi->vf_stats);
+
+ u64_stats_update_begin(&pcpu_stats->syncp);
+ pcpu_stats->tx_packets++;
+ pcpu_stats->tx_bytes += len;
+ u64_stats_update_end(&pcpu_stats->syncp);
+ } else {
+ this_cpu_inc(vi->vf_stats->tx_dropped);
+ }
+
+ return rc;
+}
+
static netdev_tx_t start_xmit(struct sk_buff *skb, struct net_device *dev)
{
struct virtnet_info *vi = netdev_priv(dev);
int qnum = skb_get_queue_mapping(skb);
struct send_queue *sq = &vi->sq[qnum];
+ struct net_device *vf_netdev;
int err;
struct netdev_queue *txq = netdev_get_tx_queue(dev, qnum);
bool kick = !skb->xmit_more;
bool use_napi = sq->napi.weight;
+ /* if VF is present and up then redirect packets
+ * called with rcu_read_lock_bh
+ */
+ vf_netdev = rcu_dereference_bh(vi->vf_netdev);
+ if (vf_netdev && netif_running(vf_netdev) && !netpoll_tx_running(dev))
+ return virtnet_vf_xmit(dev, vf_netdev, skb);
+
/* Free up any pending old buffers before queueing new ones. */
free_old_xmit_skbs(sq);
@@ -1456,10 +1509,41 @@ static int virtnet_set_mac_address(struct net_device *dev, void *p)
return ret;
}
+static void virtnet_get_vf_stats(struct net_device *dev,
+ struct virtnet_vf_pcpu_stats *tot)
+{
+ struct virtnet_info *vi = netdev_priv(dev);
+ int i;
+
+ memset(tot, 0, sizeof(*tot));
+
+ for_each_possible_cpu(i) {
+ const struct virtnet_vf_pcpu_stats *stats
+ = per_cpu_ptr(vi->vf_stats, i);
+ u64 rx_packets, rx_bytes, tx_packets, tx_bytes;
+ unsigned int start;
+
+ do {
+ start = u64_stats_fetch_begin_irq(&stats->syncp);
+ rx_packets = stats->rx_packets;
+ tx_packets = stats->tx_packets;
+ rx_bytes = stats->rx_bytes;
+ tx_bytes = stats->tx_bytes;
+ } while (u64_stats_fetch_retry_irq(&stats->syncp, start));
+
+ tot->rx_packets += rx_packets;
+ tot->tx_packets += tx_packets;
+ tot->rx_bytes += rx_bytes;
+ tot->tx_bytes += tx_bytes;
+ tot->tx_dropped += stats->tx_dropped;
+ }
+}
+
static void virtnet_stats(struct net_device *dev,
struct rtnl_link_stats64 *tot)
{
struct virtnet_info *vi = netdev_priv(dev);
+ struct virtnet_vf_pcpu_stats vf_stats;
int cpu;
unsigned int start;
@@ -1490,6 +1574,13 @@ static void virtnet_stats(struct net_device *dev,
tot->rx_dropped = dev->stats.rx_dropped;
tot->rx_length_errors = dev->stats.rx_length_errors;
tot->rx_frame_errors = dev->stats.rx_frame_errors;
+
+ virtnet_get_vf_stats(dev, &vf_stats);
+ tot->rx_packets += vf_stats.rx_packets;
+ tot->tx_packets += vf_stats.tx_packets;
+ tot->rx_bytes += vf_stats.rx_bytes;
+ tot->tx_bytes += vf_stats.tx_bytes;
+ tot->tx_dropped += vf_stats.tx_dropped;
}
#ifdef CONFIG_NET_POLL_CONTROLLER
@@ -2508,6 +2599,47 @@ static int virtnet_validate(struct virtio_device *vdev)
return 0;
}
+static void __virtnet_vf_setup(struct net_device *ndev,
+ struct net_device *vf_netdev)
+{
+ int ret;
+
+ /* Align MTU of VF with master */
+ ret = dev_set_mtu(vf_netdev, ndev->mtu);
+ if (ret)
+ netdev_warn(vf_netdev,
+ "unable to change mtu to %u\n", ndev->mtu);
+
+ if (netif_running(ndev)) {
+ ret = dev_open(vf_netdev);
+ if (ret)
+ netdev_warn(vf_netdev,
+ "unable to open: %d\n", ret);
+ }
+}
+
+/* Setup VF as slave of the virtio device.
+ * Runs in workqueue to avoid recursion in netlink callbacks.
+ */
+static void virtnet_vf_setup(struct work_struct *w)
+{
+ struct virtnet_info *vi
+ = container_of(w, struct virtnet_info, vf_takeover.work);
+ struct net_device *ndev = vi->dev;
+ struct net_device *vf_netdev;
+
+ if (!rtnl_trylock()) {
+ schedule_delayed_work(&vi->vf_takeover, 0);
+ return;
+ }
+
+ vf_netdev = rtnl_dereference(vi->vf_netdev);
+ if (vf_netdev)
+ __virtnet_vf_setup(ndev, vf_netdev);
+
+ rtnl_unlock();
+}
+
static int virtnet_probe(struct virtio_device *vdev)
{
int i, err;
@@ -2600,6 +2732,11 @@ static int virtnet_probe(struct virtio_device *vdev)
}
INIT_WORK(&vi->config_work, virtnet_config_changed_work);
+ INIT_DELAYED_WORK(&vi->vf_takeover, virtnet_vf_setup);
+
+ vi->vf_stats = netdev_alloc_pcpu_stats(struct virtnet_vf_pcpu_stats);
+ if (!vi->vf_stats)
+ goto free_stats;
/* If we can receive ANY GSO packets, we must allocate large ones. */
if (virtio_has_feature(vdev, VIRTIO_NET_F_GUEST_TSO4) ||
@@ -2634,7 +2771,7 @@ static int virtnet_probe(struct virtio_device *vdev)
*/
dev_err(&vdev->dev, "device MTU appears to have changed "
"it is now %d < %d", mtu, dev->min_mtu);
- goto free_stats;
+ goto free_vf_stats;
}
dev->mtu = mtu;
@@ -2658,7 +2795,7 @@ static int virtnet_probe(struct virtio_device *vdev)
/* Allocate/initialize the rx/tx queues, and invoke find_vqs */
err = init_vqs(vi);
if (err)
- goto free_stats;
+ goto free_vf_stats;
#ifdef CONFIG_SYSFS
if (vi->mergeable_rx_bufs)
@@ -2712,6 +2849,8 @@ static int virtnet_probe(struct virtio_device *vdev)
cancel_delayed_work_sync(&vi->refill);
free_receive_page_frags(vi);
virtnet_del_vqs(vi);
+free_vf_stats:
+ free_percpu(vi->vf_stats);
free_stats:
free_percpu(vi->stats);
free:
@@ -2733,19 +2872,178 @@ static void remove_vq_common(struct virtnet_info *vi)
virtnet_del_vqs(vi);
}
+static struct net_device *get_virtio_bymac(const u8 *mac)
+{
+ struct net_device *dev;
+
+ ASSERT_RTNL();
+
+ for_each_netdev(&init_net, dev) {
+ if (dev->netdev_ops != &virtnet_netdev)
+ continue; /* not a virtio_net device */
+
+ if (ether_addr_equal(mac, dev->perm_addr))
+ return dev;
+ }
+
+ return NULL;
+}
+
+static struct net_device *get_virtio_byref(struct net_device *vf_netdev)
+{
+ struct net_device *dev;
+
+ ASSERT_RTNL();
+
+ for_each_netdev(&init_net, dev) {
+ struct virtnet_info *vi;
+
+ if (dev->netdev_ops != &virtnet_netdev)
+ continue; /* not a virtio_net device */
+
+ vi = netdev_priv(dev);
+ if (rtnl_dereference(vi->vf_netdev) == vf_netdev)
+ return dev; /* a match */
+ }
+
+ return NULL;
+}
+
+/* Called when VF is injecting data into network stack.
+ * Change the associated network device from VF to virtio.
+ * note: already called with rcu_read_lock
+ */
+static rx_handler_result_t virtnet_vf_handle_frame(struct sk_buff **pskb)
+{
+ struct sk_buff *skb = *pskb;
+ struct net_device *ndev = rcu_dereference(skb->dev->rx_handler_data);
+ struct virtnet_info *vi = netdev_priv(ndev);
+ struct virtnet_vf_pcpu_stats *pcpu_stats =
+ this_cpu_ptr(vi->vf_stats);
+
+ skb->dev = ndev;
+
+ u64_stats_update_begin(&pcpu_stats->syncp);
+ pcpu_stats->rx_packets++;
+ pcpu_stats->rx_bytes += skb->len;
+ u64_stats_update_end(&pcpu_stats->syncp);
+
+ return RX_HANDLER_ANOTHER;
+}
+
+static int virtnet_vf_join(struct net_device *vf_netdev,
+ struct net_device *ndev)
+{
+ struct virtnet_info *vi = netdev_priv(ndev);
+ int ret;
+
+ ret = netdev_rx_handler_register(vf_netdev,
+ virtnet_vf_handle_frame, ndev);
+ if (ret != 0) {
+ netdev_err(vf_netdev,
+ "can not register virtio VF receive handler (err = %d)\n",
+ ret);
+ goto rx_handler_failed;
+ }
+
+ ret = netdev_upper_dev_link(vf_netdev, ndev, NULL);
+ if (ret != 0) {
+ netdev_err(vf_netdev,
+ "can not set master device %s (err = %d)\n",
+ ndev->name, ret);
+ goto upper_link_failed;
+ }
+
+ /* set slave flag before open to prevent IPv6 addrconf */
+ vf_netdev->flags |= IFF_SLAVE;
+
+ schedule_delayed_work(&vi->vf_takeover, VF_TAKEOVER_INT);
+
+ call_netdevice_notifiers(NETDEV_JOIN, vf_netdev);
+
+ netdev_info(vf_netdev, "joined to %s\n", ndev->name);
+ return 0;
+
+upper_link_failed:
+ netdev_rx_handler_unregister(vf_netdev);
+rx_handler_failed:
+ return ret;
+}
+
+static int virtnet_register_vf(struct net_device *vf_netdev)
+{
+ struct net_device *ndev;
+ struct virtnet_info *vi;
+
+ if (vf_netdev->addr_len != ETH_ALEN)
+ return NOTIFY_DONE;
+
+ /* We will use the MAC address to locate the virtio_net interface to
+ * associate with the VF interface. If we don't find a matching
+ * virtio interface, move on.
+ */
+ ndev = get_virtio_bymac(vf_netdev->perm_addr);
+ if (!ndev)
+ return NOTIFY_DONE;
+
+ vi = netdev_priv(ndev);
+ if (rtnl_dereference(vi->vf_netdev))
+ return NOTIFY_DONE;
+
+ if (virtnet_vf_join(vf_netdev, ndev) != 0)
+ return NOTIFY_DONE;
+
+ netdev_info(ndev, "VF registering %s\n", vf_netdev->name);
+
+ dev_hold(vf_netdev);
+ rcu_assign_pointer(vi->vf_netdev, vf_netdev);
+
+ return NOTIFY_OK;
+}
+
+static int virtnet_unregister_vf(struct net_device *vf_netdev)
+{
+ struct net_device *ndev;
+ struct virtnet_info *vi;
+
+ ndev = get_virtio_byref(vf_netdev);
+ if (!ndev)
+ return NOTIFY_DONE;
+
+ vi = netdev_priv(ndev);
+ cancel_delayed_work_sync(&vi->vf_takeover);
+
+ netdev_info(ndev, "VF unregistering %s\n", vf_netdev->name);
+
+ netdev_rx_handler_unregister(vf_netdev);
+ netdev_upper_dev_unlink(vf_netdev, ndev);
+ RCU_INIT_POINTER(vi->vf_netdev, NULL);
+ dev_put(vf_netdev);
+
+ return NOTIFY_OK;
+}
+
static void virtnet_remove(struct virtio_device *vdev)
{
struct virtnet_info *vi = vdev->priv;
+ struct net_device *vf_netdev;
virtnet_cpu_notif_remove(vi);
/* Make sure no work handler is accessing the device. */
flush_work(&vi->config_work);
+ rtnl_lock();
+ vf_netdev = rtnl_dereference(vi->vf_netdev);
+ if (vf_netdev)
+ virtnet_unregister_vf(vf_netdev);
+ rtnl_unlock();
+
unregister_netdev(vi->dev);
remove_vq_common(vi);
+ free_percpu(vi->vf_stats);
free_percpu(vi->stats);
free_netdev(vi->dev);
}
@@ -2823,6 +3121,42 @@ static struct virtio_driver virtio_net_driver = {
#endif
};
+static int virtio_netdev_event(struct notifier_block *this,
+ unsigned long event, void *ptr)
+{
+ struct net_device *event_dev = netdev_notifier_info_to_dev(ptr);
+
+ /* Skip our own events */
+ if (event_dev->netdev_ops == &virtnet_netdev)
+ return NOTIFY_DONE;
+
+ /* Avoid non-Ethernet type devices */
+ if (event_dev->type != ARPHRD_ETHER)
+ return NOTIFY_DONE;
+
+ /* Avoid Vlan dev with same MAC registering as VF */
+ if (is_vlan_dev(event_dev))
+ return NOTIFY_DONE;
+
+ /* Avoid Bonding master dev with same MAC registering as VF */
+ if ((event_dev->priv_flags & IFF_BONDING) &&
+ (event_dev->flags & IFF_MASTER))
+ return NOTIFY_DONE;
+
+ switch (event) {
+ case NETDEV_REGISTER:
+ return virtnet_register_vf(event_dev);
+ case NETDEV_UNREGISTER:
+ return virtnet_unregister_vf(event_dev);
+ default:
+ return NOTIFY_DONE;
+ }
+}
+
+static struct notifier_block virtio_netdev_notifier = {
+ .notifier_call = virtio_netdev_event,
+};
+
static __init int virtio_net_driver_init(void)
{
int ret;
@@ -2841,6 +3175,8 @@ static __init int virtio_net_driver_init(void)
ret = register_virtio_driver(&virtio_net_driver);
if (ret)
goto err_virtio;
+
+ register_netdevice_notifier(&virtio_netdev_notifier);
return 0;
err_virtio:
cpuhp_remove_multi_state(CPUHP_VIRT_NET_DEAD);
@@ -2853,6 +3189,7 @@ module_init(virtio_net_driver_init);
static __exit void virtio_net_driver_exit(void)
{
+ unregister_netdevice_notifier(&virtio_netdev_notifier);
unregister_virtio_driver(&virtio_net_driver);
cpuhp_remove_multi_state(CPUHP_VIRT_NET_DEAD);
cpuhp_remove_multi_state(virtionet_online);
--
2.14.3
^ permalink raw reply related
* Re: Possible unsafe usage of skb->cb in virtio-net
From: Jason Wang @ 2017-12-19 9:06 UTC (permalink / raw)
To: Michael S. Tsirkin, Ilya Lesokhin
Cc: netdev, Willem de Bruijn, David Miller, virtualization
In-Reply-To: <20171102142758-mutt-send-email-mst@kernel.org>
On 2017年11月02日 21:01, Michael S. Tsirkin wrote:
> On Thu, Nov 02, 2017 at 11:40:36AM +0000, Ilya Lesokhin wrote:
>> Hi,
>> I've noticed that the virtio-net uses skb->cb.
>>
>> I don't know all the detail by my understanding is it caused problem with the mlx5 driver
>> and was fixed here:
>> https://github.com/torvalds/linux/commit/34802a42b3528b0e18ea4517c8b23e1214a09332
>>
>> Thanks,
>> Ilya
> Thanks a lot for the pointer.
>
> I think this was in response to this:
> https://patchwork.ozlabs.org/patch/558324/
>
>>> + skb_push(skb, skb->data - skb_data_orig);
>>> sq->skb[pi] = skb;
>>>
>>> MLX5E_TX_SKB_CB(skb)->num_wqebbs = DIV_ROUND_UP(ds_cnt,
>> And in the middle of this we have:
>>
>> skb_pull_inline(skb, ihs);
>>
>> This is looks illegal.
>>
>> You must not modify the data pointers of any SKB that you receive for
>> sending via ->ndo_start_xmit() unless you know that absolutely you are
>> the one and only reference that exists to that SKB.
>>
>> And exactly for the case you are trying to "fix" here, you do not. If
>> the SKB is cloned, or has an elevated users count, someone else can be
>> looking at it exactly at the same time you are messing with the data
>> pointers.
>>
>> I bet mlx4 has this bug too.
>>
>> You must fix this properly, by keeping track of an offset or similar
>> internally to your driver, rather than changing the SKB data pointers.
> What virtio does is this:
>
> can_push = vi->any_header_sg &&
> !((unsigned long)skb->data & (__alignof__(*hdr) - 1)) &&
> !skb_header_cloned(skb) && skb_headroom(skb) >= hdr_len;
> /* Even if we can, don't push here yet as this would skew
> * csum_start offset below. */
> if (can_push)
> hdr = (struct virtio_net_hdr_mrg_rxbuf *)(skb->data - hdr_len);
> else
> hdr = skb_vnet_hdr(skb);
>
>
> This doesn't change the data pointers in a cloned skb but it does change the cb.
> Is it true that it's illegal to touch the cb in a cloned skb then?
>
I think not.
skb_clone() call __skb_copy_header() which did:
memcpy(new->cb, old->cb, sizeof(old->cb));
Thanks
_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization
^ permalink raw reply
* Re: [PATCH v3 5/5] virtio: add comments for virtio_register_device
From: Cornelia Huck @ 2017-12-19 11:11 UTC (permalink / raw)
To: weiping zhang; +Cc: virtualization, mst
In-Reply-To: <56d8ecc6f11b7580f3d38adcc0fd71ac60f72b02.1513517240.git.zhangweiping@didichuxing.com>
On Sun, 17 Dec 2017 21:48:05 +0800
weiping zhang <zwp10758@gmail.com> wrote:
> As mentioned at drivers/base/core.c:
> /*
> * NOTE: _Never_ directly free @dev after calling this function, even
> * if it returned an error! Always use put_device() to give up the
> * reference initialized in this function instead.
> */
> virtio_register_device may fail before/after call device_register, the
> caller should do a proper cleanup. Caller cann't use kfree directly,
> if virtio_register_device has already called device_register. Caller
> cann't use put_device directly, if virtio_register_device has not yet
> call device_register, because kobject_put may give a warning cause
> dev->kobj has not been initialized.
This comment makes me inclined to think that we should also rethink
register_virtio_device(). On failure, we cannot do kfree() due to
driver core interaction; but we cannot do a put_device() either, since
the refcount may not yet have been initialized -- unless we check the
device status, which triggers I/O (at least on s390).
We really want to do the same cleanup on error in every case. What
about splitting device_register() into device_initialize() and
device_add()? If we move device_initialize() before getting an index,
we should be fine with doing put_device() on error in every case.
>
> Signed-off-by: weiping zhang <zhangweiping@didichuxing.com>
> ---
> drivers/virtio/virtio.c | 13 +++++++++++++
> 1 file changed, 13 insertions(+)
>
> diff --git a/drivers/virtio/virtio.c b/drivers/virtio/virtio.c
> index c5b057bd..4f0718b 100644
> --- a/drivers/virtio/virtio.c
> +++ b/drivers/virtio/virtio.c
> @@ -309,6 +309,19 @@ void unregister_virtio_driver(struct virtio_driver *driver)
> }
> EXPORT_SYMBOL_GPL(unregister_virtio_driver);
>
> +/**
> + * register_virtio_device - register virtio device
> + *
> + * @dev : virtio device interested
> + *
> + * This funciton may fail after call device_register, as mentioned at
> + * drivers/base/core.c we must use put_device(), _never_ directly free @dev.
> + * The caller should take care of the status of @dev and determine to use
> + * put_device or kfree. If @dev in VIRTIO_CONFIG_S_ACKNOWLEDGE status that
> + * means caller should use put_device otherwise use kfree directly.
> + *
> + * Returns: 0 on success, error on failure.
> + */
> int register_virtio_device(struct virtio_device *dev)
> {
> int err;
^ permalink raw reply
* [PATCH v20 0/7] Virtio-balloon Enhancement
From: Wei Wang @ 2017-12-19 12:17 UTC (permalink / raw)
To: virtio-dev, linux-kernel, qemu-devel, virtualization, kvm,
linux-mm, mst, mhocko, akpm, mawilcox
Cc: aarcange, yang.zhang.wz, riel, quan.xu0, penguin-kernel,
liliang.opensource, willy, amit.shah, cornelia.huck, pbonzini,
nilal, mgorman
This patch series enhances the existing virtio-balloon with the following
new features:
1) fast ballooning: transfer ballooned pages between the guest and host in
chunks using sgs, instead of one array each time; and
2) free page block reporting: a new virtqueue to report guest free pages
to the host.
The second feature can be used to accelerate live migration of VMs. Here
are some details:
Live migration needs to transfer the VM's memory from the source machine
to the destination round by round. For the 1st round, all the VM's memory
is transferred. From the 2nd round, only the pieces of memory that were
written by the guest (after the 1st round) are transferred. One method
that is popularly used by the hypervisor to track which part of memory is
written is to write-protect all the guest memory.
The second feature enables the optimization of the 1st round memory
transfer - the hypervisor can skip the transfer of guest free pages in the
1st round. It is not concerned that the memory pages are used after they
are given to the hypervisor as a hint of the free pages, because they will
be tracked by the hypervisor and transferred in the next round if they are
used and written.
ChangeLog:
v19->v20:
1) patch 1: xbitmap
- add __rcu to "void **slot";
- remove the exceptional path.
2) patch 3: xbitmap
- DeveloperNotes: add an item to comment that the current bit range
related APIs operating on extremely large ranges (e.g.
[0, ULONG_MAX)) will take too long time. This can be optimized in
the future.
- remove the exceptional path;
- remove xb_preload_and_set();
- reimplement xb_clear_bit_range to make its usage close to
bitmap_clear;
- rename xb_find_next_set_bit to xb_find_set, and re-implement it
in a style close to find_next_bit;
- rename xb_find_next_zero_bit to xb_find_clear, and re-implement
it in a stytle close to find_next_zero_bit;
- separate the implementation of xb_find_set and xb_find_clear for
the convenience of future updates.
3) patch 4: virtio-balloon
- xb_set_page: change the way to call xb_ related APIs
v18->v19:
1) patch 3:
- xb_clear_bit_range and xb_find_next_bit will deal with range [start,
end), where end is changed to be exclusive of the range.
- add overflow checks at the end of xb_clear_bit_range and
xb_find_next_bit
- add overflow related test cases
2) patch 4:
- change back to the previous add_one_sg methond, which is based on the
scatterlist struct
- tell_host_sgs: use "uint64_t len" to avoid overflow
- batch_balloon_page_sg: a simpler function to implement the batching of
sgs
3) patch 6: batch_free_page_sg: batch sgs using the previous scatterlist struct
4) patch 7: add a config field, poison_val, to tell the host about the poison
value
v17->v18:
1) patch 1-2: new to solve some tools related compilation issues
2) patch 3: revert to the original xbitmap implementation from Matthew
Wilcox with some minor changes (e.g. comments added to the exported
functions)
3) patch 4: summarize the changes we want to make to patch 3
4) patch 5: add the developer notes as a reminder for users to avoid
concurrent accesses to the ida bitmap
5) patch 6: a new vring API to allow users to directly pass in a physical
address to a vring desc
6) patch 7: ballooning time is reduced from ~490ms to ~440ms with the new
implementation
- use the new API from patch 6 to send balloon pages
- xb_preload with "GFP_NOWAIT | __GFP_NOWARN" flag;
- handle the case when xb_set_page() fails to avoid memory leak;
- put xb_set_page() under the balloon lock
7) patch 9: simper implementation
- start free page reporting by sending a new cmd id from the host
- guest acks the start or stop via adding a cmd id to the free page vq
- use vb->report_free_page, instead of vb->report_free_page_stop
- use WRITE_ONCE/READ_ONCE to access vb->report_free_page
- use the new API from patch 6 to send free pages to avoid the
unnecessary use of kaddr.
8) patch 10: new patch to solve the page posioning issue reported by
Michael S. Tsirkin
v16->v17:
1) patch 1: please check the commit log there;
2) patch 3: included Michael S. Tsirkin patch to fix the potential
deadlock issue;
3) patch 4: use BUG_ON if virtqueue_add_ returns error, which is
expected never to happen;
4) patch 4: add leak_balloon_sg_oom, which is used in the oom case when
VIRTIO_BALLOON_F_SG is in use;
5) patch 6: use config registers, instead of a vq, as the command channel
between the host and guest;
6) patch 6: add the command sequence id support.
v15->v16:
1) mm: stop reporting the free pfn range if the callback returns false;
2) mm: move some implementaion of walk_free_mem_block into a function to
make the code layout looks better;
3) xbitmap: added some optimizations suggested by Matthew, please refer to
the ChangLog in the xbitmap patch for details.
4) xbitmap: added a test suite
5) virtio-balloon: bail out with a warning when virtqueue_add_inbuf returns
an error
6) virtio-balloon: some small code re-arrangement, e.g. detachinf used buf
from the vq before adding a new buf
v14->v15:
1) mm: make the report callback return a bool value - returning 1 to stop
walking through the free page list.
2) virtio-balloon: batching sgs of balloon pages till the vq is full
3) virtio-balloon: create a new workqueue, rather than using the default
system_wq, to queue the free page reporting work item.
4) virtio-balloon: add a ctrl_vq to be a central control plane which will
handle all the future control related commands between the host and guest.
Add free page report as the first feature controlled under ctrl_vq, and
the free_page_vq is a data plane vq dedicated to the transmission of free
page blocks.
v13->v14:
1) xbitmap: move the code from lib/radix-tree.c to lib/xbitmap.c.
2) xbitmap: consolidate the implementation of xb_bit_set/clear/test into
one xb_bit_ops.
3) xbitmap: add documents for the exported APIs.
4) mm: rewrite the function to walk through free page blocks.
5) virtio-balloon: when reporting a free page blcok to the device, if the
vq is full (less likey to happen in practice), just skip reporting this
block, instead of busywaiting till an entry gets released.
6) virtio-balloon: fail the probe function if adding the signal buf in
init_vqs fails.
v12->v13:
1) mm: use a callback function to handle the the free page blocks from the
report function. This avoids exposing the zone internal to a kernel
module.
2) virtio-balloon: send balloon pages or a free page block using a single
sg each time. This has the benefits of simpler implementation with no new
APIs.
3) virtio-balloon: the free_page_vq is used to report free pages only (no
multiple usages interleaving)
4) virtio-balloon: Balloon pages and free page blocks are sent via input
sgs, and the completion signal to the host is sent via an output sg.
v11->v12:
1) xbitmap: use the xbitmap from Matthew Wilcox to record ballooned pages.
2) virtio-ring: enable the driver to build up a desc chain using vring
desc.
3) virtio-ring: Add locking to the existing START_USE() and END_USE()
macro to lock/unlock the vq when a vq operation starts/ends.
4) virtio-ring: add virtqueue_kick_sync() and virtqueue_kick_async()
5) virtio-balloon: describe chunks of ballooned pages and free pages
blocks directly using one or more chains of desc from the vq.
v10->v11:
1) virtio_balloon: use vring_desc to describe a chunk;
2) virtio_ring: support to add an indirect desc table to virtqueue;
3) virtio_balloon: use cmdq to report guest memory statistics.
v9->v10:
1) mm: put report_unused_page_block() under CONFIG_VIRTIO_BALLOON;
2) virtio-balloon: add virtballoon_validate();
3) virtio-balloon: msg format change;
4) virtio-balloon: move miscq handling to a task on system_freezable_wq;
5) virtio-balloon: code cleanup.
v8->v9:
1) Split the two new features, VIRTIO_BALLOON_F_BALLOON_CHUNKS and
VIRTIO_BALLOON_F_MISC_VQ, which were mixed together in the previous
implementation;
2) Simpler function to get the free page block.
v7->v8:
1) Use only one chunk format, instead of two.
2) re-write the virtio-balloon implementation patch.
3) commit changes
4) patch re-org
Matthew Wilcox (1):
xbitmap: Introduce xbitmap
Wei Wang (6):
xbitmap: potential improvement
xbitmap: add more operations
virtio-balloon: VIRTIO_BALLOON_F_SG
mm: support reporting free page blocks
virtio-balloon: VIRTIO_BALLOON_F_FREE_PAGE_VQ
virtio-balloon: don't report free pages when page poisoning is enabled
drivers/virtio/virtio_balloon.c | 444 +++++++++++++++++++++++++++----
include/linux/mm.h | 6 +
include/linux/radix-tree.h | 2 +
include/linux/xbitmap.h | 55 ++++
include/uapi/linux/virtio_balloon.h | 7 +
lib/Makefile | 2 +-
lib/radix-tree.c | 40 ++-
lib/xbitmap.c | 330 +++++++++++++++++++++++
mm/page_alloc.c | 91 +++++++
tools/include/linux/bitmap.h | 34 +++
tools/include/linux/kernel.h | 2 +
tools/testing/radix-tree/Makefile | 12 +-
tools/testing/radix-tree/linux/xbitmap.h | 1 +
tools/testing/radix-tree/main.c | 4 +
tools/testing/radix-tree/test.h | 1 +
15 files changed, 976 insertions(+), 55 deletions(-)
create mode 100644 include/linux/xbitmap.h
create mode 100644 lib/xbitmap.c
create mode 100644 tools/testing/radix-tree/linux/xbitmap.h
--
2.7.4
^ permalink raw reply
* [PATCH v20 1/7] xbitmap: Introduce xbitmap
From: Wei Wang @ 2017-12-19 12:17 UTC (permalink / raw)
To: virtio-dev, linux-kernel, qemu-devel, virtualization, kvm,
linux-mm, mst, mhocko, akpm, mawilcox
Cc: aarcange, yang.zhang.wz, riel, quan.xu0, penguin-kernel,
liliang.opensource, willy, amit.shah, cornelia.huck, pbonzini,
nilal, mgorman
In-Reply-To: <1513685879-21823-1-git-send-email-wei.w.wang@intel.com>
From: Matthew Wilcox <mawilcox@microsoft.com>
The eXtensible Bitmap is a sparse bitmap representation which is
efficient for set bits which tend to cluster. It supports up to
'unsigned long' worth of bits, and this commit adds the bare bones --
xb_set_bit(), xb_clear_bit() and xb_test_bit().
Signed-off-by: Wei Wang <wei.w.wang@intel.com>
Cc: Matthew Wilcox <mawilcox@microsoft.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
---
include/linux/radix-tree.h | 2 +
include/linux/xbitmap.h | 49 ++++++++++++
lib/Makefile | 2 +-
lib/radix-tree.c | 25 +++++-
lib/xbitmap.c | 130 +++++++++++++++++++++++++++++++
tools/testing/radix-tree/Makefile | 12 ++-
tools/testing/radix-tree/linux/xbitmap.h | 1 +
tools/testing/radix-tree/main.c | 4 +
tools/testing/radix-tree/test.h | 1 +
9 files changed, 221 insertions(+), 5 deletions(-)
create mode 100644 include/linux/xbitmap.h
create mode 100644 lib/xbitmap.c
create mode 100644 tools/testing/radix-tree/linux/xbitmap.h
diff --git a/include/linux/radix-tree.h b/include/linux/radix-tree.h
index 23a9c89..5c16179a 100644
--- a/include/linux/radix-tree.h
+++ b/include/linux/radix-tree.h
@@ -315,6 +315,8 @@ void radix_tree_iter_delete(struct radix_tree_root *,
struct radix_tree_iter *iter, void __rcu **slot);
void *radix_tree_delete_item(struct radix_tree_root *, unsigned long, void *);
void *radix_tree_delete(struct radix_tree_root *, unsigned long);
+bool __radix_tree_delete(struct radix_tree_root *r, struct radix_tree_node *n,
+ void __rcu **slot);
void radix_tree_clear_tags(struct radix_tree_root *, struct radix_tree_node *,
void __rcu **slot);
unsigned int radix_tree_gang_lookup(const struct radix_tree_root *,
diff --git a/include/linux/xbitmap.h b/include/linux/xbitmap.h
new file mode 100644
index 0000000..4ac2b8d
--- /dev/null
+++ b/include/linux/xbitmap.h
@@ -0,0 +1,49 @@
+/*
+ * eXtensible Bitmaps
+ * Copyright (c) 2017 Microsoft Corporation <mawilcox@microsoft.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation; either version 2 of the
+ * License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * eXtensible Bitmaps provide an unlimited-size sparse bitmap facility.
+ * All bits are initially zero.
+ */
+
+#include <linux/idr.h>
+
+struct xb {
+ struct radix_tree_root xbrt;
+};
+
+#define XB_INIT { \
+ .xbrt = RADIX_TREE_INIT(IDR_RT_MARKER | GFP_NOWAIT), \
+}
+#define DEFINE_XB(name) struct xb name = XB_INIT
+
+static inline void xb_init(struct xb *xb)
+{
+ INIT_RADIX_TREE(&xb->xbrt, IDR_RT_MARKER | GFP_NOWAIT);
+}
+
+int xb_set_bit(struct xb *xb, unsigned long bit);
+bool xb_test_bit(const struct xb *xb, unsigned long bit);
+void xb_clear_bit(struct xb *xb, unsigned long bit);
+
+static inline bool xb_empty(const struct xb *xb)
+{
+ return radix_tree_empty(&xb->xbrt);
+}
+
+void xb_preload(gfp_t);
+
+static inline void xb_preload_end(void)
+{
+ preempt_enable();
+}
diff --git a/lib/Makefile b/lib/Makefile
index d11c48e..08a8183 100644
--- a/lib/Makefile
+++ b/lib/Makefile
@@ -19,7 +19,7 @@ KCOV_INSTRUMENT_dynamic_debug.o := n
lib-y := ctype.o string.o vsprintf.o cmdline.o \
rbtree.o radix-tree.o dump_stack.o timerqueue.o\
- idr.o int_sqrt.o extable.o \
+ idr.o xbitmap.o int_sqrt.o extable.o \
sha1.o chacha20.o irq_regs.o argv_split.o \
flex_proportions.o ratelimit.o show_mem.o \
is_single_threaded.o plist.o decompress.o kobject_uevent.o \
diff --git a/lib/radix-tree.c b/lib/radix-tree.c
index c8d5556..2650e9e 100644
--- a/lib/radix-tree.c
+++ b/lib/radix-tree.c
@@ -37,7 +37,7 @@
#include <linux/rcupdate.h>
#include <linux/slab.h>
#include <linux/string.h>
-
+#include <linux/xbitmap.h>
/* Number of nodes in fully populated tree of given height */
static unsigned long height_to_maxnodes[RADIX_TREE_MAX_PATH + 1] __read_mostly;
@@ -77,6 +77,11 @@ static struct kmem_cache *radix_tree_node_cachep;
RADIX_TREE_MAP_SHIFT))
#define IDA_PRELOAD_SIZE (IDA_MAX_PATH * 2 - 1)
+#define XB_INDEX_BITS (BITS_PER_LONG - ilog2(IDA_BITMAP_BITS))
+#define XB_MAX_PATH (DIV_ROUND_UP(XB_INDEX_BITS, \
+ RADIX_TREE_MAP_SHIFT))
+#define XB_PRELOAD_SIZE (XB_MAX_PATH * 2 - 1)
+
/*
* Per-cpu pool of preloaded nodes
*/
@@ -839,6 +844,8 @@ int __radix_tree_create(struct radix_tree_root *root, unsigned long index,
offset, 0, 0);
if (!child)
return -ENOMEM;
+ if (is_idr(root))
+ all_tag_set(child, IDR_FREE);
rcu_assign_pointer(*slot, node_to_entry(child));
if (node)
node->count++;
@@ -1982,7 +1989,7 @@ void __radix_tree_delete_node(struct radix_tree_root *root,
delete_node(root, node, update_node);
}
-static bool __radix_tree_delete(struct radix_tree_root *root,
+bool __radix_tree_delete(struct radix_tree_root *root,
struct radix_tree_node *node, void __rcu **slot)
{
void *old = rcu_dereference_raw(*slot);
@@ -2135,6 +2142,20 @@ int ida_pre_get(struct ida *ida, gfp_t gfp)
}
EXPORT_SYMBOL(ida_pre_get);
+void xb_preload(gfp_t gfp)
+{
+ __radix_tree_preload(gfp, XB_PRELOAD_SIZE);
+ if (!this_cpu_read(ida_bitmap)) {
+ struct ida_bitmap *bitmap = kmalloc(sizeof(*bitmap), gfp);
+
+ if (!bitmap)
+ return;
+ bitmap = this_cpu_cmpxchg(ida_bitmap, NULL, bitmap);
+ kfree(bitmap);
+ }
+}
+EXPORT_SYMBOL(xb_preload);
+
void __rcu **idr_get_free_cmn(struct radix_tree_root *root,
struct radix_tree_iter *iter, gfp_t gfp,
unsigned long max)
diff --git a/lib/xbitmap.c b/lib/xbitmap.c
new file mode 100644
index 0000000..236afa9
--- /dev/null
+++ b/lib/xbitmap.c
@@ -0,0 +1,130 @@
+#include <linux/export.h>
+#include <linux/xbitmap.h>
+#include <linux/bitmap.h>
+#include <linux/slab.h>
+
+/**
+ * xb_set_bit - set a bit in the xbitmap
+ * @xb: the xbitmap tree used to record the bit
+ * @bit: index of the bit to set
+ *
+ * This function is used to set a bit in the xbitmap. If the bitmap that @bit
+ * resides in is not there, the per-cpu ida_bitmap will be taken.
+ *
+ * Returns: 0 on success. -EAGAIN or -ENOMEM indicates that @bit is not set.
+ */
+int xb_set_bit(struct xb *xb, unsigned long bit)
+{
+ int err;
+ unsigned long index = bit / IDA_BITMAP_BITS;
+ struct radix_tree_root *root = &xb->xbrt;
+ struct radix_tree_node *node;
+ void __rcu **slot;
+ struct ida_bitmap *bitmap;
+
+ bit %= IDA_BITMAP_BITS;
+ err = __radix_tree_create(root, index, 0, &node, &slot);
+ if (err)
+ return err;
+ bitmap = rcu_dereference_raw(*slot);
+ if (!bitmap) {
+ bitmap = this_cpu_xchg(ida_bitmap, NULL);
+ if (!bitmap)
+ return -EAGAIN;
+ memset(bitmap, 0, sizeof(*bitmap));
+ __radix_tree_replace(root, node, slot, bitmap, NULL);
+ }
+
+ __set_bit(bit, bitmap->bitmap);
+ return 0;
+}
+EXPORT_SYMBOL(xb_set_bit);
+
+/**
+ * xb_clear_bit - clear a bit in the xbitmap
+ * @xb: the xbitmap tree used to record the bit
+ * @bit: index of the bit to clear
+ *
+ * This function is used to clear a bit in the xbitmap. If all the bits of the
+ * bitmap are 0, the bitmap will be freed.
+ */
+void xb_clear_bit(struct xb *xb, unsigned long bit)
+{
+ unsigned long index = bit / IDA_BITMAP_BITS;
+ struct radix_tree_root *root = &xb->xbrt;
+ struct radix_tree_node *node;
+ void __rcu **slot;
+ struct ida_bitmap *bitmap;
+
+ bit %= IDA_BITMAP_BITS;
+ bitmap = __radix_tree_lookup(root, index, &node, &slot);
+ if (!bitmap)
+ return;
+
+ __clear_bit(bit, bitmap->bitmap);
+ if (bitmap_empty(bitmap->bitmap, IDA_BITMAP_BITS)) {
+ kfree(bitmap);
+ __radix_tree_delete(root, node, slot);
+ }
+}
+EXPORT_SYMBOL(xb_clear_bit);
+
+/**
+ * xb_test_bit - test a bit in the xbitmap
+ * @xb: the xbitmap tree used to record the bit
+ * @bit: index of the bit to test
+ *
+ * This function is used to test a bit in the xbitmap.
+ *
+ * Returns: true if the bit is set, or false otherwise.
+ */
+bool xb_test_bit(const struct xb *xb, unsigned long bit)
+{
+ unsigned long index = bit / IDA_BITMAP_BITS;
+ const struct radix_tree_root *root = &xb->xbrt;
+ struct ida_bitmap *bitmap = radix_tree_lookup(root, index);
+
+ bit %= IDA_BITMAP_BITS;
+
+ if (!bitmap)
+ return false;
+ return test_bit(bit, bitmap->bitmap);
+}
+EXPORT_SYMBOL(xb_test_bit);
+
+#ifndef __KERNEL__
+
+static DEFINE_XB(xb1);
+
+void xbitmap_check_bit(unsigned long bit)
+{
+ xb_preload(GFP_KERNEL);
+ assert(!xb_test_bit(&xb1, bit));
+ assert(xb_set_bit(&xb1, bit) == 0);
+ assert(xb_test_bit(&xb1, bit));
+ assert(xb_clear_bit(&xb1, bit) == 0);
+ assert(xb_empty(&xb1));
+ assert(xb_clear_bit(&xb1, bit) == 0);
+ assert(xb_empty(&xb1));
+ xb_preload_end();
+}
+
+void xbitmap_checks(void)
+{
+ xb_init(&xb1);
+ xbitmap_check_bit(0);
+ xbitmap_check_bit(30);
+ xbitmap_check_bit(31);
+ xbitmap_check_bit(1023);
+ xbitmap_check_bit(1024);
+ xbitmap_check_bit(1025);
+ xbitmap_check_bit((1UL << 63) | (1UL << 24));
+ xbitmap_check_bit((1UL << 63) | (1UL << 24) | 70);
+}
+
+int __weak main(void)
+{
+ radix_tree_init();
+ xbitmap_checks();
+}
+#endif
diff --git a/tools/testing/radix-tree/Makefile b/tools/testing/radix-tree/Makefile
index fa7ee36..34ece78 100644
--- a/tools/testing/radix-tree/Makefile
+++ b/tools/testing/radix-tree/Makefile
@@ -6,7 +6,8 @@ LDLIBS+= -lpthread -lurcu
TARGETS = main idr-test multiorder
CORE_OFILES := radix-tree.o idr.o linux.o test.o find_bit.o
OFILES = main.o $(CORE_OFILES) regression1.o regression2.o regression3.o \
- tag_check.o multiorder.o idr-test.o iteration_check.o benchmark.o
+ tag_check.o multiorder.o idr-test.o iteration_check.o benchmark.o \
+ xbitmap.o
ifndef SHIFT
SHIFT=3
@@ -25,8 +26,11 @@ idr-test: idr-test.o $(CORE_OFILES)
multiorder: multiorder.o $(CORE_OFILES)
+xbitmap: xbitmap.o $(CORE_OFILES)
+ $(CC) $(CFLAGS) $(LDFLAGS) $^ -o xbitmap
+
clean:
- $(RM) $(TARGETS) *.o radix-tree.c idr.c generated/map-shift.h
+ $(RM) $(TARGETS) *.o radix-tree.c idr.c xbitmap.c generated/map-shift.h
vpath %.c ../../lib
@@ -34,6 +38,7 @@ $(OFILES): Makefile *.h */*.h generated/map-shift.h \
../../include/linux/*.h \
../../include/asm/*.h \
../../../include/linux/radix-tree.h \
+ ../../../include/linux/xbitmap.h \
../../../include/linux/idr.h
radix-tree.c: ../../../lib/radix-tree.c
@@ -42,6 +47,9 @@ radix-tree.c: ../../../lib/radix-tree.c
idr.c: ../../../lib/idr.c
sed -e 's/^static //' -e 's/__always_inline //' -e 's/inline //' < $< > $@
+xbitmap.c: ../../../lib/xbitmap.c
+ sed -e 's/^static //' -e 's/__always_inline //' -e 's/inline //' < $< > $@
+
.PHONY: mapshift
mapshift:
diff --git a/tools/testing/radix-tree/linux/xbitmap.h b/tools/testing/radix-tree/linux/xbitmap.h
new file mode 100644
index 0000000..61de214
--- /dev/null
+++ b/tools/testing/radix-tree/linux/xbitmap.h
@@ -0,0 +1 @@
+#include "../../../../include/linux/xbitmap.h"
diff --git a/tools/testing/radix-tree/main.c b/tools/testing/radix-tree/main.c
index 257f3f8..d112363 100644
--- a/tools/testing/radix-tree/main.c
+++ b/tools/testing/radix-tree/main.c
@@ -326,6 +326,10 @@ static void single_thread_tests(bool long_run)
rcu_barrier();
printv(2, "after idr_checks: %d allocated, preempt %d\n",
nr_allocated, preempt_count);
+ xbitmap_checks();
+ rcu_barrier();
+ printv(2, "after xbitmap_checks: %d allocated, preempt %d\n",
+ nr_allocated, preempt_count);
big_gang_check(long_run);
rcu_barrier();
printv(2, "after big_gang_check: %d allocated, preempt %d\n",
diff --git a/tools/testing/radix-tree/test.h b/tools/testing/radix-tree/test.h
index d9c031d..8175d6b 100644
--- a/tools/testing/radix-tree/test.h
+++ b/tools/testing/radix-tree/test.h
@@ -38,6 +38,7 @@ void benchmark(void);
void idr_checks(void);
void ida_checks(void);
void ida_thread_tests(void);
+void xbitmap_checks(void);
struct item *
item_tag_set(struct radix_tree_root *root, unsigned long index, int tag);
--
2.7.4
^ permalink raw reply related
* [PATCH v20 2/7] xbitmap: potential improvement
From: Wei Wang @ 2017-12-19 12:17 UTC (permalink / raw)
To: virtio-dev, linux-kernel, qemu-devel, virtualization, kvm,
linux-mm, mst, mhocko, akpm, mawilcox
Cc: aarcange, yang.zhang.wz, riel, quan.xu0, penguin-kernel,
liliang.opensource, willy, amit.shah, cornelia.huck, pbonzini,
nilal, mgorman
In-Reply-To: <1513685879-21823-1-git-send-email-wei.w.wang@intel.com>
This patch made some changes to the original xbitmap implementation from
the linux-dax tree:
- xb_set_bit: delete the new inserted radix_tree_node when failing to
get the per cpu ida bitmap, this avoids the kind of memory leak of the
unused radix tree node left in the tree.
- xb_preload: with the original implementation, the CPU that successfully
do __radix_tree_preload() may get into sleep by kmalloc(), which has a
risk of getting the caller of xb_preload() scheduled to another CPU
after waken up, and the new CPU may not have radix_tree_node
pre-allocated there, this will be a problem when inserting a node to
the tree later. This patch moves __radix_tree_preload() after kmalloc()
and returns a boolean to indicate the success or failure. Also, add the
__must_check annotation to xb_preload for prudence purpose.
Signed-off-by: Wei Wang <wei.w.wang@intel.com>
Cc: Matthew Wilcox <mawilcox@microsoft.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
---
include/linux/xbitmap.h | 2 +-
lib/radix-tree.c | 21 ++++++++++++++++++---
lib/xbitmap.c | 4 +++-
3 files changed, 22 insertions(+), 5 deletions(-)
diff --git a/include/linux/xbitmap.h b/include/linux/xbitmap.h
index 4ac2b8d..108f929 100644
--- a/include/linux/xbitmap.h
+++ b/include/linux/xbitmap.h
@@ -41,7 +41,7 @@ static inline bool xb_empty(const struct xb *xb)
return radix_tree_empty(&xb->xbrt);
}
-void xb_preload(gfp_t);
+int xb_preload(gfp_t);
static inline void xb_preload_end(void)
{
diff --git a/lib/radix-tree.c b/lib/radix-tree.c
index 2650e9e..f30347a 100644
--- a/lib/radix-tree.c
+++ b/lib/radix-tree.c
@@ -2142,17 +2142,32 @@ int ida_pre_get(struct ida *ida, gfp_t gfp)
}
EXPORT_SYMBOL(ida_pre_get);
-void xb_preload(gfp_t gfp)
+/**
+ * xb_preload - preload for xb_set_bit()
+ * @gfp_mask: allocation mask to use for preloading
+ *
+ * Preallocate memory to use for the next call to xb_set_bit(). On success,
+ * return zero, with preemption disabled. On error, return -ENOMEM with
+ * preemption not disabled.
+ */
+__must_check int xb_preload(gfp_t gfp)
{
- __radix_tree_preload(gfp, XB_PRELOAD_SIZE);
if (!this_cpu_read(ida_bitmap)) {
struct ida_bitmap *bitmap = kmalloc(sizeof(*bitmap), gfp);
if (!bitmap)
- return;
+ return -ENOMEM;
+ /*
+ * The per-CPU variable is updated with preemption enabled.
+ * If the calling task is unlucky to be scheduled to another
+ * CPU which has no ida_bitmap allocation, it will be detected
+ * when setting a bit (i.e. xb_set_bit()).
+ */
bitmap = this_cpu_cmpxchg(ida_bitmap, NULL, bitmap);
kfree(bitmap);
}
+
+ return __radix_tree_preload(gfp, XB_PRELOAD_SIZE);
}
EXPORT_SYMBOL(xb_preload);
diff --git a/lib/xbitmap.c b/lib/xbitmap.c
index 236afa9..2dcfad5 100644
--- a/lib/xbitmap.c
+++ b/lib/xbitmap.c
@@ -29,8 +29,10 @@ int xb_set_bit(struct xb *xb, unsigned long bit)
bitmap = rcu_dereference_raw(*slot);
if (!bitmap) {
bitmap = this_cpu_xchg(ida_bitmap, NULL);
- if (!bitmap)
+ if (!bitmap) {
+ __radix_tree_delete(root, node, slot);
return -EAGAIN;
+ }
memset(bitmap, 0, sizeof(*bitmap));
__radix_tree_replace(root, node, slot, bitmap, NULL);
}
--
2.7.4
^ permalink raw reply related
* [PATCH v20 3/7] xbitmap: add more operations
From: Wei Wang @ 2017-12-19 12:17 UTC (permalink / raw)
To: virtio-dev, linux-kernel, qemu-devel, virtualization, kvm,
linux-mm, mst, mhocko, akpm, mawilcox
Cc: aarcange, yang.zhang.wz, riel, quan.xu0, penguin-kernel,
liliang.opensource, willy, amit.shah, cornelia.huck, pbonzini,
nilal, mgorman
In-Reply-To: <1513685879-21823-1-git-send-email-wei.w.wang@intel.com>
This patch adds support to find next 1 or 0 bit in a xbmitmap range and
clear a range of bits.
More possible optimizations to add in the future:
1) xb_set_bit_range: set a range of bits.
2) when searching a bit, if the bit is not found in the slot, move on to
the next slot directly.
3) add tags to help searching.
Signed-off-by: Wei Wang <wei.w.wang@intel.com>
Cc: Matthew Wilcox <mawilcox@microsoft.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Suggested-by: Matthew Wilcox <mawilcox@microsoft.com>
---
include/linux/xbitmap.h | 6 ++
lib/xbitmap.c | 198 +++++++++++++++++++++++++++++++++++++++++++
tools/include/linux/bitmap.h | 34 ++++++++
tools/include/linux/kernel.h | 2 +
4 files changed, 240 insertions(+)
diff --git a/include/linux/xbitmap.h b/include/linux/xbitmap.h
index 108f929..ede1029 100644
--- a/include/linux/xbitmap.h
+++ b/include/linux/xbitmap.h
@@ -35,6 +35,12 @@ static inline void xb_init(struct xb *xb)
int xb_set_bit(struct xb *xb, unsigned long bit);
bool xb_test_bit(const struct xb *xb, unsigned long bit);
void xb_clear_bit(struct xb *xb, unsigned long bit);
+void xb_clear_bit_range(struct xb *xb, unsigned long start,
+ unsigned long nbits);
+unsigned long xb_find_set(struct xb *xb, unsigned long size,
+ unsigned long offset);
+unsigned long xb_find_zero(struct xb *xb, unsigned long size,
+ unsigned long offset);
static inline bool xb_empty(const struct xb *xb)
{
diff --git a/lib/xbitmap.c b/lib/xbitmap.c
index 2dcfad5..bb0a5b2 100644
--- a/lib/xbitmap.c
+++ b/lib/xbitmap.c
@@ -3,6 +3,16 @@
#include <linux/bitmap.h>
#include <linux/slab.h>
+/*
+ * Developer notes:
+ * - locks are required to gurantee there is no concurrent
+ * calls of xb_set_bit, xb_clear_bit, xb_clear_bit_range, xb_test_bit,
+ * xb_find_set, or xb_find_clear to operate on the same ida bitmap.
+ * - The current implementation of xb_clear_bit_range, xb_find_set and
+ * xb_find_clear may cause long latency when the bit range to operate
+ * on is super large (e.g. [0, ULONG_MAX)).
+ */
+
/**
* xb_set_bit - set a bit in the xbitmap
* @xb: the xbitmap tree used to record the bit
@@ -72,6 +82,49 @@ void xb_clear_bit(struct xb *xb, unsigned long bit)
EXPORT_SYMBOL(xb_clear_bit);
/**
+ * xb_clear_bit_range - clear a range of bits in the xbitmap
+ * @start: the start of the bit range, inclusive
+ * @nbits: number of bits to clear
+ *
+ * This function is used to clear a range of bits in the xbitmap. If all the
+ * bits in the bitmap are 0, the bitmap will be freed.
+ */
+void xb_clear_bit_range(struct xb *xb, unsigned long start,
+ unsigned long nbits)
+{
+ struct radix_tree_root *root = &xb->xbrt;
+ struct radix_tree_node *node;
+ void __rcu **slot;
+ struct ida_bitmap *bitmap;
+ unsigned long index = start / IDA_BITMAP_BITS;
+ unsigned long bit = start % IDA_BITMAP_BITS;
+
+ if (nbits > ULONG_MAX - start)
+ nbits = ULONG_MAX - start;
+
+ while (nbits) {
+ unsigned int __nbits = min(nbits,
+ (unsigned long)IDA_BITMAP_BITS - bit);
+
+ bitmap = __radix_tree_lookup(root, index, &node, &slot);
+ if (bitmap) {
+ if (__nbits != IDA_BITMAP_BITS)
+ bitmap_clear(bitmap->bitmap, bit, __nbits);
+
+ if (__nbits == IDA_BITMAP_BITS ||
+ bitmap_empty(bitmap->bitmap, IDA_BITMAP_BITS)) {
+ kfree(bitmap);
+ __radix_tree_delete(root, node, slot);
+ }
+ }
+ bit = 0;
+ index++;
+ nbits -= __nbits;
+ }
+}
+EXPORT_SYMBOL(xb_clear_bit_range);
+
+/**
* xb_test_bit - test a bit in the xbitmap
* @xb: the xbitmap tree used to record the bit
* @bit: index of the bit to test
@@ -94,6 +147,99 @@ bool xb_test_bit(const struct xb *xb, unsigned long bit)
}
EXPORT_SYMBOL(xb_test_bit);
+/**
+ * xb_find_set - find the next set bit in a range of bits
+ * @xb: the xbitmap to search from
+ * @offset: the offset in the range to start searching
+ * @size: the size of the range
+ *
+ * Returns: the found bit or, @size if no set bit is found.
+ */
+unsigned long xb_find_set(struct xb *xb, unsigned long size,
+ unsigned long offset)
+{
+ struct radix_tree_root *root = &xb->xbrt;
+ struct radix_tree_node *node;
+ void __rcu **slot;
+ struct ida_bitmap *bitmap;
+ unsigned long index = offset / IDA_BITMAP_BITS;
+ unsigned long index_end = size / IDA_BITMAP_BITS;
+ unsigned long bit = offset % IDA_BITMAP_BITS;
+
+ if (unlikely(offset >= size))
+ return size;
+
+ while (index <= index_end) {
+ unsigned long ret;
+ unsigned int nbits = size - index * IDA_BITMAP_BITS;
+
+ bitmap = __radix_tree_lookup(root, index, &node, &slot);
+ if (!node) {
+ index = (index | RADIX_TREE_MAP_MASK) + 1;
+ continue;
+ }
+
+ if (bitmap) {
+ if (nbits > IDA_BITMAP_BITS)
+ nbits = IDA_BITMAP_BITS;
+
+ ret = find_next_bit(bitmap->bitmap, nbits, bit);
+ if (ret != nbits)
+ return ret + index * IDA_BITMAP_BITS;
+ }
+ bit = 0;
+ index++;
+ }
+
+ return size;
+}
+EXPORT_SYMBOL(xb_find_set);
+
+/**
+ * xb_find_zero - find the next zero bit in a range of bits
+ * @xb: the xbitmap to search from
+ * @offset: the offset in the range to start searching
+ * @size: the size of the range
+ *
+ * Returns: the found bit or, @size if no zero bit is found.
+ */
+unsigned long xb_find_zero(struct xb *xb, unsigned long size,
+ unsigned long offset)
+{
+ struct radix_tree_root *root = &xb->xbrt;
+ struct radix_tree_node *node;
+ void __rcu **slot;
+ struct ida_bitmap *bitmap;
+ unsigned long index = offset / IDA_BITMAP_BITS;
+ unsigned long index_end = size / IDA_BITMAP_BITS;
+ unsigned long bit = offset % IDA_BITMAP_BITS;
+
+ if (unlikely(offset >= size))
+ return size;
+
+ while (index <= index_end) {
+ unsigned long ret;
+ unsigned int nbits = size - index * IDA_BITMAP_BITS;
+
+ bitmap = __radix_tree_lookup(root, index, &node, &slot);
+ if (bitmap) {
+ if (nbits > IDA_BITMAP_BITS)
+ nbits = IDA_BITMAP_BITS;
+
+ ret = find_next_zero_bit(bitmap->bitmap, nbits, bit);
+ if (ret != nbits)
+ return ret + index * IDA_BITMAP_BITS;
+ } else {
+ return bit + index * IDA_BITMAP_BITS;
+ }
+ bit = 0;
+ index++;
+ }
+
+ return size;
+}
+EXPORT_SYMBOL(xb_find_zero);
+
#ifndef __KERNEL__
static DEFINE_XB(xb1);
@@ -111,6 +257,56 @@ void xbitmap_check_bit(unsigned long bit)
xb_preload_end();
}
+static void xbitmap_check_bit_range(void)
+{
+ /*
+ * Regular tests
+ * set bit 2000, 2001, 2040
+ * Next 1 in [0, 2048) --> 2000
+ * Next 1 in [2000, 2002) --> 2000
+ * Next 1 in [2002, 2041) --> 2040
+ * Next 1 in [2002, 2040) --> none
+ * Next 0 in [2000, 2048) --> 2002
+ * Next 0 in [2048, 2060) --> 2048
+ */
+ xb_preload(GFP_KERNEL);
+ assert(!xb_set_bit(&xb1, 2000));
+ assert(!xb_set_bit(&xb1, 2001));
+ assert(!xb_set_bit(&xb1, 2040));
+ assert(xb_find_set(&xb1, 2048, 0) == 2000);
+ assert(xb_find_set(&xb1, 2002, 2000) == 2000);
+ assert(xb_find_set(&xb1, 2041, 2002) == 2040);
+ assert(xb_find_set(&xb1, 2040, 2002) == 2040);
+ assert(xb_find_zero(&xb1, 2048, 2000) == 2002);
+ assert(xb_find_zero(&xb1, 2060, 2048) == 2048);
+ xb_clear_bit_range(&xb1, 0, 2048);
+ assert(xb_find_set(&xb1, 2048, 0) == 2048);
+ xb_preload_end();
+
+ /*
+ * Overflow tests:
+ * Set bit 1 and ULONG_MAX - 4
+ * Next 1 in [ULONG_MAX - 4, ULONG_MAX) --> ULONG_MAX - 4
+ * Next 1 [ULONG_MAX - 3, ULONG_MAX + 4) --> none
+ * Next 0 [ULONG_MAX - 4, ULONG_MAX + 4) --> none
+ */
+ xb_preload(GFP_KERNEL);
+ assert(!xb_set_bit(&xb1, 1));
+ xb_preload_end();
+ xb_preload(GFP_KERNEL);
+ assert(!xb_set_bit(&xb1, ULONG_MAX - 4));
+ assert(xb_find_set(&xb1, ULONG_MAX, ULONG_MAX - 4) == ULONG_MAX - 4);
+ assert(xb_find_set(&xb1, ULONG_MAX + 4, ULONG_MAX - 3) ==
+ ULONG_MAX + 4);
+ assert(xb_find_zero(&xb1, ULONG_MAX + 4, ULONG_MAX - 4) ==
+ ULONG_MAX + 4);
+ xb_clear_bit_range(&xb1, ULONG_MAX - 4, 4);
+ assert(xb_find_set(&xb1, ULONG_MAX, ULONG_MAX - 10) == ULONG_MAX);
+ xb_clear_bit_range(&xb1, 0, 2);
+ assert(xb_find_set(&xb1, 2, 0) == 2);
+ xb_preload_end();
+}
+
void xbitmap_checks(void)
{
xb_init(&xb1);
@@ -122,6 +318,8 @@ void xbitmap_checks(void)
xbitmap_check_bit(1025);
xbitmap_check_bit((1UL << 63) | (1UL << 24));
xbitmap_check_bit((1UL << 63) | (1UL << 24) | 70);
+
+ xbitmap_check_bit_range();
}
int __weak main(void)
diff --git a/tools/include/linux/bitmap.h b/tools/include/linux/bitmap.h
index ca16027..8d0bc1b 100644
--- a/tools/include/linux/bitmap.h
+++ b/tools/include/linux/bitmap.h
@@ -37,6 +37,40 @@ static inline void bitmap_zero(unsigned long *dst, int nbits)
}
}
+static inline void __bitmap_clear(unsigned long *map, unsigned int start,
+ int len)
+{
+ unsigned long *p = map + BIT_WORD(start);
+ const unsigned int size = start + len;
+ int bits_to_clear = BITS_PER_LONG - (start % BITS_PER_LONG);
+ unsigned long mask_to_clear = BITMAP_FIRST_WORD_MASK(start);
+
+ while (len - bits_to_clear >= 0) {
+ *p &= ~mask_to_clear;
+ len -= bits_to_clear;
+ bits_to_clear = BITS_PER_LONG;
+ mask_to_clear = ~0UL;
+ p++;
+ }
+ if (len) {
+ mask_to_clear &= BITMAP_LAST_WORD_MASK(size);
+ *p &= ~mask_to_clear;
+ }
+}
+
+static inline __always_inline void bitmap_clear(unsigned long *map,
+ unsigned int start,
+ unsigned int nbits)
+{
+ if (__builtin_constant_p(nbits) && nbits == 1)
+ __clear_bit(start, map);
+ else if (__builtin_constant_p(start & 7) && IS_ALIGNED(start, 8) &&
+ __builtin_constant_p(nbits & 7) && IS_ALIGNED(nbits, 8))
+ memset((char *)map + start / 8, 0, nbits / 8);
+ else
+ __bitmap_clear(map, start, nbits);
+}
+
static inline void bitmap_fill(unsigned long *dst, unsigned int nbits)
{
unsigned int nlongs = BITS_TO_LONGS(nbits);
diff --git a/tools/include/linux/kernel.h b/tools/include/linux/kernel.h
index 0ad8844..3c992ae 100644
--- a/tools/include/linux/kernel.h
+++ b/tools/include/linux/kernel.h
@@ -13,6 +13,8 @@
#define UINT_MAX (~0U)
#endif
+#define IS_ALIGNED(x, a) (((x) & ((typeof(x))(a) - 1)) == 0)
+
#define DIV_ROUND_UP(n,d) (((n) + (d) - 1) / (d))
#define PERF_ALIGN(x, a) __PERF_ALIGN_MASK(x, (typeof(x))(a)-1)
--
2.7.4
^ permalink raw reply related
* [PATCH v20 4/7] virtio-balloon: VIRTIO_BALLOON_F_SG
From: Wei Wang @ 2017-12-19 12:17 UTC (permalink / raw)
To: virtio-dev, linux-kernel, qemu-devel, virtualization, kvm,
linux-mm, mst, mhocko, akpm, mawilcox
Cc: aarcange, yang.zhang.wz, riel, quan.xu0, penguin-kernel,
liliang.opensource, willy, amit.shah, cornelia.huck, pbonzini,
nilal, mgorman
In-Reply-To: <1513685879-21823-1-git-send-email-wei.w.wang@intel.com>
Add a new feature, VIRTIO_BALLOON_F_SG, which enables the transfer of
balloon (i.e. inflated/deflated) pages using scatter-gather lists to the
host.
The implementation of the previous virtio-balloon is not very efficient,
because the balloon pages are transferred to the host by one array each
time. Here is the breakdown of the time in percentage spent on each step
of the balloon inflating process (inflating 7GB of an 8GB idle guest).
1) allocating pages (6.5%)
2) sending PFNs to host (68.3%)
3) address translation (6.1%)
4) madvise (19%)
It takes about 4126ms for the inflating process to complete. The above
profiling shows that the bottlenecks are stage 2) and stage 4).
This patch optimizes step 2) by transferring pages to host in sgs. An sg
describes a chunk of guest physically continuous pages. With this
mechanism, step 4) can also be optimized by doing address translation and
madvise() in chunks rather than page by page.
With this new feature, the above ballooning process takes ~460ms resulting
in an improvement of ~89%.
TODO:
- optimize stage 1) by allocating/freeing a chunk of pages instead of a
single page each time.
- sort the internal balloon page queue.
Signed-off-by: Wei Wang <wei.w.wang@intel.com>
Signed-off-by: Liang Li <liang.z.li@intel.com>
Suggested-by: Michael S. Tsirkin <mst@redhat.com>
Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
---
drivers/virtio/virtio_balloon.c | 234 +++++++++++++++++++++++++++++++++---
include/uapi/linux/virtio_balloon.h | 1 +
2 files changed, 217 insertions(+), 18 deletions(-)
diff --git a/drivers/virtio/virtio_balloon.c b/drivers/virtio/virtio_balloon.c
index a1fb52c..fff0a3f 100644
--- a/drivers/virtio/virtio_balloon.c
+++ b/drivers/virtio/virtio_balloon.c
@@ -32,6 +32,8 @@
#include <linux/mm.h>
#include <linux/mount.h>
#include <linux/magic.h>
+#include <linux/xbitmap.h>
+#include <asm/page.h>
/*
* Balloon device works in 4K page units. So each page is pointed to by
@@ -79,6 +81,9 @@ struct virtio_balloon {
/* Synchronize access/update to this struct virtio_balloon elements */
struct mutex balloon_lock;
+ /* The xbitmap used to record balloon pages */
+ struct xb page_xb;
+
/* The array of pfns we tell the Host about. */
unsigned int num_pfns;
__virtio32 pfns[VIRTIO_BALLOON_ARRAY_PFNS_MAX];
@@ -141,15 +146,129 @@ static void set_page_pfns(struct virtio_balloon *vb,
page_to_balloon_pfn(page) + i);
}
+static void kick_and_wait(struct virtqueue *vq, wait_queue_head_t wq_head)
+{
+ unsigned int len;
+
+ virtqueue_kick(vq);
+ wait_event(wq_head, virtqueue_get_buf(vq, &len));
+}
+
+static void add_one_sg(struct virtqueue *vq, unsigned long pfn, uint32_t len)
+{
+ struct scatterlist sg;
+ unsigned int unused;
+ int err;
+
+ sg_init_table(&sg, 1);
+ sg_set_page(&sg, pfn_to_page(pfn), len, 0);
+
+ /* Detach all the used buffers from the vq */
+ while (virtqueue_get_buf(vq, &unused))
+ ;
+
+ err = virtqueue_add_inbuf(vq, &sg, 1, vq, GFP_KERNEL);
+ /*
+ * This is expected to never fail: there is always at least 1 entry
+ * available on the vq, because when the vq is full the worker thread
+ * that adds the sg will be put into sleep until at least 1 entry is
+ * available to use.
+ */
+ BUG_ON(err);
+}
+
+static void batch_balloon_page_sg(struct virtio_balloon *vb,
+ struct virtqueue *vq,
+ unsigned long pfn,
+ uint32_t len)
+{
+ add_one_sg(vq, pfn, len);
+
+ /* Batch till the vq is full */
+ if (!vq->num_free)
+ kick_and_wait(vq, vb->acked);
+}
+
+/*
+ * Send balloon pages in sgs to host. The balloon pages are recorded in the
+ * page xbitmap. Each bit in the bitmap corresponds to a page of PAGE_SIZE.
+ * The page xbitmap is searched for continuous "1" bits, which correspond
+ * to continuous pages, to chunk into sgs.
+ *
+ * @page_xb_start and @page_xb_end form the range of bits in the xbitmap that
+ * need to be searched.
+ */
+static void tell_host_sgs(struct virtio_balloon *vb,
+ struct virtqueue *vq,
+ unsigned long page_xb_start,
+ unsigned long page_xb_end)
+{
+ unsigned long pfn_start, pfn_end;
+ uint32_t max_len = round_down(UINT_MAX, PAGE_SIZE);
+ uint64_t len;
+
+ pfn_start = page_xb_start;
+ while (pfn_start < page_xb_end) {
+ pfn_start = xb_find_set(&vb->page_xb, page_xb_end + 1,
+ pfn_start);
+ if (pfn_start == page_xb_end + 1)
+ break;
+ pfn_end = xb_find_zero(&vb->page_xb, page_xb_end + 1,
+ pfn_start);
+ len = (pfn_end - pfn_start) << PAGE_SHIFT;
+ while (len > max_len) {
+ batch_balloon_page_sg(vb, vq, pfn_start, max_len);
+ pfn_start += max_len >> PAGE_SHIFT;
+ len -= max_len;
+ }
+ batch_balloon_page_sg(vb, vq, pfn_start, (uint32_t)len);
+ pfn_start = pfn_end + 1;
+ }
+
+ /*
+ * The last few sgs may not reach the batch size, but need a kick to
+ * notify the device to handle them.
+ */
+ if (vq->num_free != virtqueue_get_vring_size(vq))
+ kick_and_wait(vq, vb->acked);
+
+ xb_clear_bit_range(&vb->page_xb, page_xb_start, page_xb_end);
+}
+
+static inline int xb_set_page(struct virtio_balloon *vb,
+ struct page *page,
+ unsigned long *pfn_min,
+ unsigned long *pfn_max)
+{
+ unsigned long pfn = page_to_pfn(page);
+ int ret;
+
+ *pfn_min = min(pfn, *pfn_min);
+ *pfn_max = max(pfn, *pfn_max);
+
+ do {
+ if (xb_preload(GFP_NOWAIT | __GFP_NOWARN) < 0)
+ return -ENOMEM;
+
+ ret = xb_set_bit(&vb->page_xb, pfn);
+ xb_preload_end();
+ } while (unlikely(ret == -EAGAIN));
+
+ return ret;
+}
+
static unsigned fill_balloon(struct virtio_balloon *vb, size_t num)
{
unsigned num_allocated_pages;
unsigned num_pfns;
struct page *page;
LIST_HEAD(pages);
+ bool use_sg = virtio_has_feature(vb->vdev, VIRTIO_BALLOON_F_SG);
+ unsigned long pfn_max = 0, pfn_min = ULONG_MAX;
/* We can only do one array worth at a time. */
- num = min(num, ARRAY_SIZE(vb->pfns));
+ if (!use_sg)
+ num = min(num, ARRAY_SIZE(vb->pfns));
for (num_pfns = 0; num_pfns < num;
num_pfns += VIRTIO_BALLOON_PAGES_PER_PAGE) {
@@ -173,8 +292,15 @@ static unsigned fill_balloon(struct virtio_balloon *vb, size_t num)
while ((page = balloon_page_pop(&pages))) {
balloon_page_enqueue(&vb->vb_dev_info, page);
+ if (use_sg) {
+ if (xb_set_page(vb, page, &pfn_min, &pfn_max) < 0) {
+ __free_page(page);
+ continue;
+ }
+ } else {
+ set_page_pfns(vb, vb->pfns + vb->num_pfns, page);
+ }
- set_page_pfns(vb, vb->pfns + vb->num_pfns, page);
vb->num_pages += VIRTIO_BALLOON_PAGES_PER_PAGE;
if (!virtio_has_feature(vb->vdev,
VIRTIO_BALLOON_F_DEFLATE_ON_OOM))
@@ -184,8 +310,12 @@ static unsigned fill_balloon(struct virtio_balloon *vb, size_t num)
num_allocated_pages = vb->num_pfns;
/* Did we get any? */
- if (vb->num_pfns != 0)
- tell_host(vb, vb->inflate_vq);
+ if (vb->num_pfns) {
+ if (use_sg)
+ tell_host_sgs(vb, vb->inflate_vq, pfn_min, pfn_max);
+ else
+ tell_host(vb, vb->inflate_vq);
+ }
mutex_unlock(&vb->balloon_lock);
return num_allocated_pages;
@@ -211,9 +341,12 @@ static unsigned leak_balloon(struct virtio_balloon *vb, size_t num)
struct page *page;
struct balloon_dev_info *vb_dev_info = &vb->vb_dev_info;
LIST_HEAD(pages);
+ bool use_sg = virtio_has_feature(vb->vdev, VIRTIO_BALLOON_F_SG);
+ unsigned long pfn_max = 0, pfn_min = ULONG_MAX;
- /* We can only do one array worth at a time. */
- num = min(num, ARRAY_SIZE(vb->pfns));
+ /* Traditionally, we can only do one array worth at a time. */
+ if (!use_sg)
+ num = min(num, ARRAY_SIZE(vb->pfns));
mutex_lock(&vb->balloon_lock);
/* We can't release more pages than taken */
@@ -223,7 +356,14 @@ static unsigned leak_balloon(struct virtio_balloon *vb, size_t num)
page = balloon_page_dequeue(vb_dev_info);
if (!page)
break;
- set_page_pfns(vb, vb->pfns + vb->num_pfns, page);
+ if (use_sg) {
+ if (xb_set_page(vb, page, &pfn_min, &pfn_max) < 0) {
+ balloon_page_enqueue(&vb->vb_dev_info, page);
+ break;
+ }
+ } else {
+ set_page_pfns(vb, vb->pfns + vb->num_pfns, page);
+ }
list_add(&page->lru, &pages);
vb->num_pages -= VIRTIO_BALLOON_PAGES_PER_PAGE;
}
@@ -234,13 +374,55 @@ static unsigned leak_balloon(struct virtio_balloon *vb, size_t num)
* virtio_has_feature(vdev, VIRTIO_BALLOON_F_MUST_TELL_HOST);
* is true, we *have* to do it in this order
*/
- if (vb->num_pfns != 0)
- tell_host(vb, vb->deflate_vq);
+ if (vb->num_pfns) {
+ if (use_sg)
+ tell_host_sgs(vb, vb->deflate_vq, pfn_min, pfn_max);
+ else
+ tell_host(vb, vb->deflate_vq);
+ }
release_pages_balloon(vb, &pages);
mutex_unlock(&vb->balloon_lock);
return num_freed_pages;
}
+/*
+ * The regular leak_balloon() with VIRTIO_BALLOON_F_SG needs memory allocation
+ * for xbitmap, which is not suitable for the oom case. This function does not
+ * use xbitmap to chunk pages, so it can be used by oom notifier to deflate
+ * pages when VIRTIO_BALLOON_F_SG is negotiated.
+ */
+static unsigned int leak_balloon_sg_oom(struct virtio_balloon *vb)
+{
+ unsigned int n;
+ struct page *page;
+ struct balloon_dev_info *vb_dev_info = &vb->vb_dev_info;
+ struct virtqueue *vq = vb->deflate_vq;
+ LIST_HEAD(pages);
+
+ mutex_lock(&vb->balloon_lock);
+ for (n = 0; n < oom_pages; n++) {
+ page = balloon_page_dequeue(vb_dev_info);
+ if (!page)
+ break;
+
+ list_add(&page->lru, &pages);
+ vb->num_pages -= VIRTIO_BALLOON_PAGES_PER_PAGE;
+ batch_balloon_page_sg(vb, vb->deflate_vq, page_to_pfn(page),
+ PAGE_SIZE);
+ release_pages_balloon(vb, &pages);
+ }
+
+ /*
+ * The last few sgs may not reach the batch size, but need a kick to
+ * notify the device to handle them.
+ */
+ if (vq->num_free != virtqueue_get_vring_size(vq))
+ kick_and_wait(vq, vb->acked);
+ mutex_unlock(&vb->balloon_lock);
+
+ return n;
+}
+
static inline void update_stat(struct virtio_balloon *vb, int idx,
u16 tag, u64 val)
{
@@ -380,7 +562,10 @@ static int virtballoon_oom_notify(struct notifier_block *self,
return NOTIFY_OK;
freed = parm;
- num_freed_pages = leak_balloon(vb, oom_pages);
+ if (virtio_has_feature(vb->vdev, VIRTIO_BALLOON_F_SG))
+ num_freed_pages = leak_balloon_sg_oom(vb);
+ else
+ num_freed_pages = leak_balloon(vb, oom_pages);
update_balloon_size(vb);
*freed += num_freed_pages;
@@ -477,6 +662,7 @@ static int virtballoon_migratepage(struct balloon_dev_info *vb_dev_info,
{
struct virtio_balloon *vb = container_of(vb_dev_info,
struct virtio_balloon, vb_dev_info);
+ bool use_sg = virtio_has_feature(vb->vdev, VIRTIO_BALLOON_F_SG);
unsigned long flags;
/*
@@ -498,16 +684,24 @@ static int virtballoon_migratepage(struct balloon_dev_info *vb_dev_info,
vb_dev_info->isolated_pages--;
__count_vm_event(BALLOON_MIGRATE);
spin_unlock_irqrestore(&vb_dev_info->pages_lock, flags);
- vb->num_pfns = VIRTIO_BALLOON_PAGES_PER_PAGE;
- set_page_pfns(vb, vb->pfns, newpage);
- tell_host(vb, vb->inflate_vq);
-
+ if (use_sg) {
+ add_one_sg(vb->inflate_vq, page_to_pfn(newpage), PAGE_SIZE);
+ kick_and_wait(vb->inflate_vq, vb->acked);
+ } else {
+ vb->num_pfns = VIRTIO_BALLOON_PAGES_PER_PAGE;
+ set_page_pfns(vb, vb->pfns, newpage);
+ tell_host(vb, vb->inflate_vq);
+ }
/* balloon's page migration 2nd step -- deflate "page" */
balloon_page_delete(page);
- vb->num_pfns = VIRTIO_BALLOON_PAGES_PER_PAGE;
- set_page_pfns(vb, vb->pfns, page);
- tell_host(vb, vb->deflate_vq);
-
+ if (use_sg) {
+ add_one_sg(vb->deflate_vq, page_to_pfn(page), PAGE_SIZE);
+ kick_and_wait(vb->deflate_vq, vb->acked);
+ } else {
+ vb->num_pfns = VIRTIO_BALLOON_PAGES_PER_PAGE;
+ set_page_pfns(vb, vb->pfns, page);
+ tell_host(vb, vb->deflate_vq);
+ }
mutex_unlock(&vb->balloon_lock);
put_page(page); /* balloon reference */
@@ -566,6 +760,9 @@ static int virtballoon_probe(struct virtio_device *vdev)
if (err)
goto out_free_vb;
+ if (virtio_has_feature(vdev, VIRTIO_BALLOON_F_SG))
+ xb_init(&vb->page_xb);
+
vb->nb.notifier_call = virtballoon_oom_notify;
vb->nb.priority = VIRTBALLOON_OOM_NOTIFY_PRIORITY;
err = register_oom_notifier(&vb->nb);
@@ -682,6 +879,7 @@ static unsigned int features[] = {
VIRTIO_BALLOON_F_MUST_TELL_HOST,
VIRTIO_BALLOON_F_STATS_VQ,
VIRTIO_BALLOON_F_DEFLATE_ON_OOM,
+ VIRTIO_BALLOON_F_SG,
};
static struct virtio_driver virtio_balloon_driver = {
diff --git a/include/uapi/linux/virtio_balloon.h b/include/uapi/linux/virtio_balloon.h
index 343d7dd..37780a7 100644
--- a/include/uapi/linux/virtio_balloon.h
+++ b/include/uapi/linux/virtio_balloon.h
@@ -34,6 +34,7 @@
#define VIRTIO_BALLOON_F_MUST_TELL_HOST 0 /* Tell before reclaiming pages */
#define VIRTIO_BALLOON_F_STATS_VQ 1 /* Memory Stats virtqueue */
#define VIRTIO_BALLOON_F_DEFLATE_ON_OOM 2 /* Deflate balloon on OOM */
+#define VIRTIO_BALLOON_F_SG 3 /* Use sg instead of PFN lists */
/* Size of a PFN in the balloon interface. */
#define VIRTIO_BALLOON_PFN_SHIFT 12
--
2.7.4
^ permalink raw reply related
* [PATCH v20 5/7] mm: support reporting free page blocks
From: Wei Wang @ 2017-12-19 12:17 UTC (permalink / raw)
To: virtio-dev, linux-kernel, qemu-devel, virtualization, kvm,
linux-mm, mst, mhocko, akpm, mawilcox
Cc: aarcange, yang.zhang.wz, riel, quan.xu0, penguin-kernel,
liliang.opensource, willy, amit.shah, cornelia.huck, pbonzini,
nilal, mgorman
In-Reply-To: <1513685879-21823-1-git-send-email-wei.w.wang@intel.com>
This patch adds support to walk through the free page blocks in the
system and report them via a callback function. Some page blocks may
leave the free list after zone->lock is released, so it is the caller's
responsibility to either detect or prevent the use of such pages.
One use example of this patch is to accelerate live migration by skipping
the transfer of free pages reported from the guest. A popular method used
by the hypervisor to track which part of memory is written during live
migration is to write-protect all the guest memory. So, those pages that
are reported as free pages but are written after the report function
returns will be captured by the hypervisor, and they will be added to the
next round of memory transfer.
Signed-off-by: Wei Wang <wei.w.wang@intel.com>
Signed-off-by: Liang Li <liang.z.li@intel.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Michal Hocko <mhocko@kernel.org>
---
include/linux/mm.h | 6 ++++
mm/page_alloc.c | 91 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 97 insertions(+)
diff --git a/include/linux/mm.h b/include/linux/mm.h
index ea818ff..b3077dd 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -1938,6 +1938,12 @@ extern void free_area_init_node(int nid, unsigned long * zones_size,
unsigned long zone_start_pfn, unsigned long *zholes_size);
extern void free_initmem(void);
+extern void walk_free_mem_block(void *opaque,
+ int min_order,
+ bool (*report_pfn_range)(void *opaque,
+ unsigned long pfn,
+ unsigned long num));
+
/*
* Free reserved pages within range [PAGE_ALIGN(start), end & PAGE_MASK)
* into the buddy system. The freed pages will be poisoned with pattern
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 7e5e775..f074503 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -4899,6 +4899,97 @@ void show_free_areas(unsigned int filter, nodemask_t *nodemask)
show_swap_cache_info();
}
+/*
+ * Walk through a free page list and report the found pfn range via the
+ * callback.
+ *
+ * Return false if the callback requests to stop reporting. Otherwise,
+ * return true.
+ */
+static bool walk_free_page_list(void *opaque,
+ struct zone *zone,
+ int order,
+ enum migratetype mt,
+ bool (*report_pfn_range)(void *,
+ unsigned long,
+ unsigned long))
+{
+ struct page *page;
+ struct list_head *list;
+ unsigned long pfn, flags;
+ bool ret;
+
+ spin_lock_irqsave(&zone->lock, flags);
+ list = &zone->free_area[order].free_list[mt];
+ list_for_each_entry(page, list, lru) {
+ pfn = page_to_pfn(page);
+ ret = report_pfn_range(opaque, pfn, 1 << order);
+ if (!ret)
+ break;
+ }
+ spin_unlock_irqrestore(&zone->lock, flags);
+
+ return ret;
+}
+
+/**
+ * walk_free_mem_block - Walk through the free page blocks in the system
+ * @opaque: the context passed from the caller
+ * @min_order: the minimum order of free lists to check
+ * @report_pfn_range: the callback to report the pfn range of the free pages
+ *
+ * If the callback returns false, stop iterating the list of free page blocks.
+ * Otherwise, continue to report.
+ *
+ * Please note that there are no locking guarantees for the callback and
+ * that the reported pfn range might be freed or disappear after the
+ * callback returns so the caller has to be very careful how it is used.
+ *
+ * The callback itself must not sleep or perform any operations which would
+ * require any memory allocations directly (not even GFP_NOWAIT/GFP_ATOMIC)
+ * or via any lock dependency. It is generally advisable to implement
+ * the callback as simple as possible and defer any heavy lifting to a
+ * different context.
+ *
+ * There is no guarantee that each free range will be reported only once
+ * during one walk_free_mem_block invocation.
+ *
+ * pfn_to_page on the given range is strongly discouraged and if there is
+ * an absolute need for that make sure to contact MM people to discuss
+ * potential problems.
+ *
+ * The function itself might sleep so it cannot be called from atomic
+ * contexts.
+ *
+ * In general low orders tend to be very volatile and so it makes more
+ * sense to query larger ones first for various optimizations which like
+ * ballooning etc... This will reduce the overhead as well.
+ */
+void walk_free_mem_block(void *opaque,
+ int min_order,
+ bool (*report_pfn_range)(void *opaque,
+ unsigned long pfn,
+ unsigned long num))
+{
+ struct zone *zone;
+ int order;
+ enum migratetype mt;
+ bool ret;
+
+ for_each_populated_zone(zone) {
+ for (order = MAX_ORDER - 1; order >= min_order; order--) {
+ for (mt = 0; mt < MIGRATE_TYPES; mt++) {
+ ret = walk_free_page_list(opaque, zone,
+ order, mt,
+ report_pfn_range);
+ if (!ret)
+ return;
+ }
+ }
+ }
+}
+EXPORT_SYMBOL_GPL(walk_free_mem_block);
+
static void zoneref_set_zone(struct zone *zone, struct zoneref *zoneref)
{
zoneref->zone = zone;
--
2.7.4
^ permalink raw reply related
* [PATCH v20 6/7] virtio-balloon: VIRTIO_BALLOON_F_FREE_PAGE_VQ
From: Wei Wang @ 2017-12-19 12:17 UTC (permalink / raw)
To: virtio-dev, linux-kernel, qemu-devel, virtualization, kvm,
linux-mm, mst, mhocko, akpm, mawilcox
Cc: aarcange, yang.zhang.wz, riel, quan.xu0, penguin-kernel,
liliang.opensource, willy, amit.shah, cornelia.huck, pbonzini,
nilal, mgorman
In-Reply-To: <1513685879-21823-1-git-send-email-wei.w.wang@intel.com>
Negotiation of the VIRTIO_BALLOON_F_FREE_PAGE_VQ feature indicates the
support of reporting hints of guest free pages to host via virtio-balloon.
Host requests the guest to report free pages by sending a new cmd
id to the guest via the free_page_report_cmd_id configuration register.
When the guest starts to report, the first element added to the free page
vq is the cmd id given by host. When the guest finishes the reporting
of all the free pages, VIRTIO_BALLOON_FREE_PAGE_REPORT_STOP_ID is added
to the vq to tell host that the reporting is done. Host may also requests
the guest to stop the reporting in advance by sending the stop cmd id to
the guest via the configuration register.
Signed-off-by: Wei Wang <wei.w.wang@intel.com>
Signed-off-by: Liang Li <liang.z.li@intel.com>
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: Michal Hocko <mhocko@kernel.org>
---
drivers/virtio/virtio_balloon.c | 202 ++++++++++++++++++++++++++++++------
include/uapi/linux/virtio_balloon.h | 4 +
2 files changed, 174 insertions(+), 32 deletions(-)
diff --git a/drivers/virtio/virtio_balloon.c b/drivers/virtio/virtio_balloon.c
index fff0a3f..eae65c1 100644
--- a/drivers/virtio/virtio_balloon.c
+++ b/drivers/virtio/virtio_balloon.c
@@ -55,7 +55,12 @@ static struct vfsmount *balloon_mnt;
struct virtio_balloon {
struct virtio_device *vdev;
- struct virtqueue *inflate_vq, *deflate_vq, *stats_vq;
+ struct virtqueue *inflate_vq, *deflate_vq, *stats_vq, *free_page_vq;
+
+ /* Balloon's own wq for cpu-intensive work items */
+ struct workqueue_struct *balloon_wq;
+ /* The free page reporting work item submitted to the balloon wq */
+ struct work_struct report_free_page_work;
/* The balloon servicing is delegated to a freezable workqueue. */
struct work_struct update_balloon_stats_work;
@@ -65,6 +70,13 @@ struct virtio_balloon {
spinlock_t stop_update_lock;
bool stop_update;
+ /* Start to report free pages */
+ bool report_free_page;
+ /* Stores the cmd id given by host to start the free page reporting */
+ uint32_t start_cmd_id;
+ /* Stores STOP_ID as a sign to tell host that the reporting is done */
+ uint32_t stop_cmd_id;
+
/* Waiting for host to ack the pages we released. */
wait_queue_head_t acked;
@@ -189,6 +201,28 @@ static void batch_balloon_page_sg(struct virtio_balloon *vb,
kick_and_wait(vq, vb->acked);
}
+static void batch_free_page_sg(struct virtqueue *vq,
+ unsigned long pfn,
+ uint32_t len)
+{
+ add_one_sg(vq, pfn, len);
+
+ /* Batch till the vq is full */
+ if (!vq->num_free)
+ virtqueue_kick(vq);
+}
+
+static void send_cmd_id(struct virtio_balloon *vb, void *addr)
+{
+ struct scatterlist sg;
+ int err;
+
+ sg_init_one(&sg, addr, sizeof(uint32_t));
+ err = virtqueue_add_outbuf(vb->free_page_vq, &sg, 1, vb, GFP_KERNEL);
+ BUG_ON(err);
+ virtqueue_kick(vb->free_page_vq);
+}
+
/*
* Send balloon pages in sgs to host. The balloon pages are recorded in the
* page xbitmap. Each bit in the bitmap corresponds to a page of PAGE_SIZE.
@@ -498,17 +532,6 @@ static void stats_handle_request(struct virtio_balloon *vb)
virtqueue_kick(vq);
}
-static void virtballoon_changed(struct virtio_device *vdev)
-{
- struct virtio_balloon *vb = vdev->priv;
- unsigned long flags;
-
- spin_lock_irqsave(&vb->stop_update_lock, flags);
- if (!vb->stop_update)
- queue_work(system_freezable_wq, &vb->update_balloon_size_work);
- spin_unlock_irqrestore(&vb->stop_update_lock, flags);
-}
-
static inline s64 towards_target(struct virtio_balloon *vb)
{
s64 target;
@@ -525,6 +548,36 @@ static inline s64 towards_target(struct virtio_balloon *vb)
return target - vb->num_pages;
}
+static void virtballoon_changed(struct virtio_device *vdev)
+{
+ struct virtio_balloon *vb = vdev->priv;
+ unsigned long flags;
+ __u32 cmd_id;
+ s64 diff = towards_target(vb);
+
+ if (diff) {
+ spin_lock_irqsave(&vb->stop_update_lock, flags);
+ if (!vb->stop_update)
+ queue_work(system_freezable_wq,
+ &vb->update_balloon_size_work);
+ spin_unlock_irqrestore(&vb->stop_update_lock, flags);
+ }
+
+ virtio_cread(vb->vdev, struct virtio_balloon_config,
+ free_page_report_cmd_id, &cmd_id);
+ if (cmd_id == VIRTIO_BALLOON_FREE_PAGE_REPORT_STOP_ID) {
+ WRITE_ONCE(vb->report_free_page, false);
+ } else if (cmd_id != vb->start_cmd_id) {
+ /*
+ * Host requests to start the reporting by sending a new cmd
+ * id.
+ */
+ WRITE_ONCE(vb->report_free_page, true);
+ vb->start_cmd_id = cmd_id;
+ queue_work(vb->balloon_wq, &vb->report_free_page_work);
+ }
+}
+
static void update_balloon_size(struct virtio_balloon *vb)
{
u32 actual = vb->num_pages;
@@ -602,40 +655,116 @@ static void update_balloon_size_func(struct work_struct *work)
static int init_vqs(struct virtio_balloon *vb)
{
- struct virtqueue *vqs[3];
- vq_callback_t *callbacks[] = { balloon_ack, balloon_ack, stats_request };
- static const char * const names[] = { "inflate", "deflate", "stats" };
- int err, nvqs;
+ struct virtqueue **vqs;
+ vq_callback_t **callbacks;
+ const char **names;
+ struct scatterlist sg;
+ int i, nvqs, err = -ENOMEM;
+
+ /* Inflateq and deflateq are used unconditionally */
+ nvqs = 2;
+ if (virtio_has_feature(vb->vdev, VIRTIO_BALLOON_F_STATS_VQ))
+ nvqs++;
+ if (virtio_has_feature(vb->vdev, VIRTIO_BALLOON_F_FREE_PAGE_VQ))
+ nvqs++;
+
+ /* Allocate space for find_vqs parameters */
+ vqs = kcalloc(nvqs, sizeof(*vqs), GFP_KERNEL);
+ if (!vqs)
+ goto err_vq;
+ callbacks = kmalloc_array(nvqs, sizeof(*callbacks), GFP_KERNEL);
+ if (!callbacks)
+ goto err_callback;
+ names = kmalloc_array(nvqs, sizeof(*names), GFP_KERNEL);
+ if (!names)
+ goto err_names;
+
+ callbacks[0] = balloon_ack;
+ names[0] = "inflate";
+ callbacks[1] = balloon_ack;
+ names[1] = "deflate";
+
+ i = 2;
+ if (virtio_has_feature(vb->vdev, VIRTIO_BALLOON_F_STATS_VQ)) {
+ callbacks[i] = stats_request;
+ names[i] = "stats";
+ i++;
+ }
- /*
- * We expect two virtqueues: inflate and deflate, and
- * optionally stat.
- */
- nvqs = virtio_has_feature(vb->vdev, VIRTIO_BALLOON_F_STATS_VQ) ? 3 : 2;
- err = virtio_find_vqs(vb->vdev, nvqs, vqs, callbacks, names, NULL);
+ if (virtio_has_feature(vb->vdev, VIRTIO_BALLOON_F_FREE_PAGE_VQ)) {
+ callbacks[i] = NULL;
+ names[i] = "free_page_vq";
+ }
+
+ err = vb->vdev->config->find_vqs(vb->vdev, nvqs, vqs, callbacks, names,
+ NULL, NULL);
if (err)
- return err;
+ goto err_find;
vb->inflate_vq = vqs[0];
vb->deflate_vq = vqs[1];
+ i = 2;
if (virtio_has_feature(vb->vdev, VIRTIO_BALLOON_F_STATS_VQ)) {
- struct scatterlist sg;
- unsigned int num_stats;
- vb->stats_vq = vqs[2];
-
+ vb->stats_vq = vqs[i++];
/*
* Prime this virtqueue with one buffer so the hypervisor can
* use it to signal us later (it can't be broken yet!).
*/
- num_stats = update_balloon_stats(vb);
-
- sg_init_one(&sg, vb->stats, sizeof(vb->stats[0]) * num_stats);
+ sg_init_one(&sg, vb->stats, sizeof(vb->stats));
if (virtqueue_add_outbuf(vb->stats_vq, &sg, 1, vb, GFP_KERNEL)
- < 0)
- BUG();
+ < 0) {
+ dev_warn(&vb->vdev->dev, "%s: add stat_vq failed\n",
+ __func__);
+ goto err_find;
+ }
virtqueue_kick(vb->stats_vq);
}
+
+ if (virtio_has_feature(vb->vdev, VIRTIO_BALLOON_F_FREE_PAGE_VQ))
+ vb->free_page_vq = vqs[i];
+
+ kfree(names);
+ kfree(callbacks);
+ kfree(vqs);
return 0;
+
+err_find:
+ kfree(names);
+err_names:
+ kfree(callbacks);
+err_callback:
+ kfree(vqs);
+err_vq:
+ return err;
+}
+
+static bool virtio_balloon_send_free_pages(void *opaque, unsigned long pfn,
+ unsigned long nr_pages)
+{
+ struct virtio_balloon *vb = (struct virtio_balloon *)opaque;
+ uint32_t len = nr_pages << PAGE_SHIFT;
+
+ if (!READ_ONCE(vb->report_free_page))
+ return false;
+
+ batch_free_page_sg(vb->free_page_vq, pfn, len);
+
+ return true;
+}
+
+static void report_free_page(struct work_struct *work)
+{
+ struct virtio_balloon *vb;
+
+ vb = container_of(work, struct virtio_balloon, report_free_page_work);
+ /* Start by sending the obtained cmd id to the host with an outbuf */
+ send_cmd_id(vb, &vb->start_cmd_id);
+ walk_free_mem_block(vb, 0, &virtio_balloon_send_free_pages);
+ /*
+ * End by sending the stop id to the host with an outbuf. Use the
+ * non-batching mode here to trigger a kick after adding the stop id.
+ */
+ send_cmd_id(vb, &vb->stop_cmd_id);
}
#ifdef CONFIG_BALLOON_COMPACTION
@@ -763,6 +892,13 @@ static int virtballoon_probe(struct virtio_device *vdev)
if (virtio_has_feature(vdev, VIRTIO_BALLOON_F_SG))
xb_init(&vb->page_xb);
+ if (virtio_has_feature(vdev, VIRTIO_BALLOON_F_FREE_PAGE_VQ)) {
+ vb->balloon_wq = alloc_workqueue("balloon-wq",
+ WQ_FREEZABLE | WQ_CPU_INTENSIVE, 0);
+ INIT_WORK(&vb->report_free_page_work, report_free_page);
+ vb->stop_cmd_id = VIRTIO_BALLOON_FREE_PAGE_REPORT_STOP_ID;
+ }
+
vb->nb.notifier_call = virtballoon_oom_notify;
vb->nb.priority = VIRTBALLOON_OOM_NOTIFY_PRIORITY;
err = register_oom_notifier(&vb->nb);
@@ -827,6 +963,7 @@ static void virtballoon_remove(struct virtio_device *vdev)
spin_unlock_irq(&vb->stop_update_lock);
cancel_work_sync(&vb->update_balloon_size_work);
cancel_work_sync(&vb->update_balloon_stats_work);
+ cancel_work_sync(&vb->report_free_page_work);
remove_common(vb);
#ifdef CONFIG_BALLOON_COMPACTION
@@ -880,6 +1017,7 @@ static unsigned int features[] = {
VIRTIO_BALLOON_F_STATS_VQ,
VIRTIO_BALLOON_F_DEFLATE_ON_OOM,
VIRTIO_BALLOON_F_SG,
+ VIRTIO_BALLOON_F_FREE_PAGE_VQ,
};
static struct virtio_driver virtio_balloon_driver = {
diff --git a/include/uapi/linux/virtio_balloon.h b/include/uapi/linux/virtio_balloon.h
index 37780a7..58f1274 100644
--- a/include/uapi/linux/virtio_balloon.h
+++ b/include/uapi/linux/virtio_balloon.h
@@ -35,15 +35,19 @@
#define VIRTIO_BALLOON_F_STATS_VQ 1 /* Memory Stats virtqueue */
#define VIRTIO_BALLOON_F_DEFLATE_ON_OOM 2 /* Deflate balloon on OOM */
#define VIRTIO_BALLOON_F_SG 3 /* Use sg instead of PFN lists */
+#define VIRTIO_BALLOON_F_FREE_PAGE_VQ 4 /* VQ to report free pages */
/* Size of a PFN in the balloon interface. */
#define VIRTIO_BALLOON_PFN_SHIFT 12
+#define VIRTIO_BALLOON_FREE_PAGE_REPORT_STOP_ID 0
struct virtio_balloon_config {
/* Number of pages host wants Guest to give up. */
__u32 num_pages;
/* Number of pages we've actually got in balloon. */
__u32 actual;
+ /* Free page report command id, readonly by guest */
+ __u32 free_page_report_cmd_id;
};
#define VIRTIO_BALLOON_S_SWAP_IN 0 /* Amount of memory swapped in */
--
2.7.4
^ permalink raw reply related
* [PATCH v20 7/7] virtio-balloon: don't report free pages when page poisoning is enabled
From: Wei Wang @ 2017-12-19 12:17 UTC (permalink / raw)
To: virtio-dev, linux-kernel, qemu-devel, virtualization, kvm,
linux-mm, mst, mhocko, akpm, mawilcox
Cc: aarcange, yang.zhang.wz, riel, quan.xu0, penguin-kernel,
liliang.opensource, willy, amit.shah, cornelia.huck, pbonzini,
nilal, mgorman
In-Reply-To: <1513685879-21823-1-git-send-email-wei.w.wang@intel.com>
The guest free pages should not be discarded by the live migration thread
when page poisoning is enabled with PAGE_POISONING_NO_SANITY=n, because
skipping the transfer of such poisoned free pages will trigger false
positive when new pages are allocated and checked on the destination.
This patch adds a config field, poison_val. Guest writes to the config
field to tell the host about the poisoning value. The value will be 0 in
the following cases:
1) PAGE_POISONING_NO_SANITY is enabled;
2) page poisoning is disabled; or
3) PAGE_POISONING_ZERO is enabled.
Signed-off-by: Wei Wang <wei.w.wang@intel.com>
Suggested-by: Michael S. Tsirkin <mst@redhat.com>
Cc: Michal Hocko <mhocko@suse.com>
---
drivers/virtio/virtio_balloon.c | 8 ++++++++
include/uapi/linux/virtio_balloon.h | 2 ++
2 files changed, 10 insertions(+)
diff --git a/drivers/virtio/virtio_balloon.c b/drivers/virtio/virtio_balloon.c
index eae65c1..1fa8598 100644
--- a/drivers/virtio/virtio_balloon.c
+++ b/drivers/virtio/virtio_balloon.c
@@ -860,6 +860,7 @@ static struct file_system_type balloon_fs = {
static int virtballoon_probe(struct virtio_device *vdev)
{
struct virtio_balloon *vb;
+ __u32 poison_val;
int err;
if (!vdev->config->get) {
@@ -897,6 +898,13 @@ static int virtballoon_probe(struct virtio_device *vdev)
WQ_FREEZABLE | WQ_CPU_INTENSIVE, 0);
INIT_WORK(&vb->report_free_page_work, report_free_page);
vb->stop_cmd_id = VIRTIO_BALLOON_FREE_PAGE_REPORT_STOP_ID;
+ if (IS_ENABLED(CONFIG_PAGE_POISONING_NO_SANITY) ||
+ !page_poisoning_enabled())
+ poison_val = 0;
+ else
+ poison_val = PAGE_POISON;
+ virtio_cwrite(vb->vdev, struct virtio_balloon_config,
+ poison_val, &poison_val);
}
vb->nb.notifier_call = virtballoon_oom_notify;
diff --git a/include/uapi/linux/virtio_balloon.h b/include/uapi/linux/virtio_balloon.h
index 58f1274..f270e9e 100644
--- a/include/uapi/linux/virtio_balloon.h
+++ b/include/uapi/linux/virtio_balloon.h
@@ -48,6 +48,8 @@ struct virtio_balloon_config {
__u32 actual;
/* Free page report command id, readonly by guest */
__u32 free_page_report_cmd_id;
+ /* Stores PAGE_POISON if page poisoning with sanity check is in use */
+ __u32 poison_val;
};
#define VIRTIO_BALLOON_S_SWAP_IN 0 /* Amount of memory swapped in */
--
2.7.4
^ permalink raw reply related
* Re: [PATCH v20 0/7] Virtio-balloon Enhancement
From: Matthew Wilcox @ 2017-12-19 14:40 UTC (permalink / raw)
To: Tetsuo Handa
Cc: yang.zhang.wz, kvm, mst, liliang.opensource, qemu-devel,
virtualization, linux-mm, aarcange, virtio-dev, mawilcox, nilal,
riel, cornelia.huck, mhocko, quan.xu0, linux-kernel, amit.shah,
pbonzini, akpm, mgorman
In-Reply-To: <201712192305.AAE21882.MtQHJOFFSFVOLO@I-love.SAKURA.ne.jp>
On Tue, Dec 19, 2017 at 11:05:11PM +0900, Tetsuo Handa wrote:
> Removing exceptional path made this patch easier to read.
> But what I meant is
>
> Can you eliminate exception path and fold all xbitmap patches into one, and
> post only one xbitmap patch without virtio-balloon changes?
>
> .
>
> I still think we don't need xb_preload()/xb_preload_end().
> I think xb_find_set() has a bug in !node path.
Don't think. Write a test-case. Please. If it shows a bug, then great,
Wei has an example of what the bug is to fix. If it doesn't show a bug,
then we can add it to the test suite anyway, to ensure that case continues
to work in the future.
> Also, please avoid unconditionally adding to builtin modules.
> There are users who want to save even few KB.
>
> --
> To unsubscribe, send a message with 'unsubscribe linux-mm' in
> the body to majordomo@kvack.org. For more info on Linux MM,
> see: http://www.linux-mm.org/ .
> Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply
* Re: [PATCH v3 5/5] virtio: add comments for virtio_register_device
From: weiping zhang @ 2017-12-19 15:23 UTC (permalink / raw)
To: Cornelia Huck; +Cc: virtualization, Michael S . Tsirkin
In-Reply-To: <20171219121159.17b8c060.cohuck@redhat.com>
2017-12-19 19:11 GMT+08:00 Cornelia Huck <cohuck@redhat.com>:
> On Sun, 17 Dec 2017 21:48:05 +0800
> weiping zhang <zwp10758@gmail.com> wrote:
>
>> As mentioned at drivers/base/core.c:
>> /*
>> * NOTE: _Never_ directly free @dev after calling this function, even
>> * if it returned an error! Always use put_device() to give up the
>> * reference initialized in this function instead.
>> */
>> virtio_register_device may fail before/after call device_register, the
>> caller should do a proper cleanup. Caller cann't use kfree directly,
>> if virtio_register_device has already called device_register. Caller
>> cann't use put_device directly, if virtio_register_device has not yet
>> call device_register, because kobject_put may give a warning cause
>> dev->kobj has not been initialized.
>
> This comment makes me inclined to think that we should also rethink
> register_virtio_device(). On failure, we cannot do kfree() due to
> driver core interaction; but we cannot do a put_device() either, since
> the refcount may not yet have been initialized -- unless we check the
> device status, which triggers I/O (at least on s390).
>
> We really want to do the same cleanup on error in every case. What
> about splitting device_register() into device_initialize() and
> device_add()? If we move device_initialize() before getting an index,
> we should be fine with doing put_device() on error in every case.
>
Good idea (^_^), I''ll apply it at V4.
^ permalink raw reply
* Re: [RFC PATCH] virtio_net: Extend virtio to use VF datapath when available
From: Michael S. Tsirkin @ 2017-12-19 15:47 UTC (permalink / raw)
To: Sridhar Samudrala; +Cc: netdev, alexander.duyck, virtualization
In-Reply-To: <1513644036-45230-1-git-send-email-sridhar.samudrala@intel.com>
I'll need to look at this more, in particular the feature
bit is missing here. For now one question:
On Mon, Dec 18, 2017 at 04:40:36PM -0800, Sridhar Samudrala wrote:
> @@ -56,6 +58,8 @@ module_param(napi_tx, bool, 0644);
> */
> DECLARE_EWMA(pkt_len, 0, 64)
>
> +#define VF_TAKEOVER_INT (HZ / 10)
> +
> #define VIRTNET_DRIVER_VERSION "1.0.0"
>
> static const unsigned long guest_offloads[] = {
Why is this delay necessary? And why by 100ms?
^ permalink raw reply
* Re: [PATCH v20 1/7] xbitmap: Introduce xbitmap
From: Philippe Ombredanne @ 2017-12-19 15:58 UTC (permalink / raw)
To: mawilcox
Cc: yang.zhang.wz, kvm, mst, penguin-kernel, liliang.opensource,
qemu-devel, virtualization, linux-mm, aarcange, virtio-dev, willy,
nilal, riel, cornelia.huck, mhocko, quan.xu0, LKML, amit.shah,
Paolo Bonzini, Andrew Morton, mgorman
In-Reply-To: <1513685879-21823-2-git-send-email-wei.w.wang@intel.com>
Matthew,
On Tue, Dec 19, 2017 at 1:17 PM, Wei Wang <wei.w.wang@intel.com> wrote:
> From: Matthew Wilcox <mawilcox@microsoft.com>
>
> The eXtensible Bitmap is a sparse bitmap representation which is
> efficient for set bits which tend to cluster. It supports up to
> 'unsigned long' worth of bits, and this commit adds the bare bones --
> xb_set_bit(), xb_clear_bit() and xb_test_bit().
<snip>
> --- /dev/null
> +++ b/include/linux/xbitmap.h
> @@ -0,0 +1,49 @@
> +/*
> + * eXtensible Bitmaps
> + * Copyright (c) 2017 Microsoft Corporation <mawilcox@microsoft.com>
> + *
> + * This program is free software; you can redistribute it and/or
> + * modify it under the terms of the GNU General Public License as
> + * published by the Free Software Foundation; either version 2 of the
> + * License, or (at your option) any later version.
> + *
> + * This program is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
> + * GNU General Public License for more details.
> + *
> + * eXtensible Bitmaps provide an unlimited-size sparse bitmap facility.
> + * All bits are initially zero.
> + */
Would you mind using the new SPDX tags documented in Thomas patch set
[1] rather than this fine but longer legalese?
And if you could spread the word to others in your team this would be very nice.
Thank you!
[1] https://lkml.org/lkml/2017/12/4/934
--
Cordially
Philippe Ombredanne
^ permalink raw reply
* Re: [RFC PATCH] virtio_net: Extend virtio to use VF datapath when available
From: Samudrala, Sridhar @ 2017-12-19 17:41 UTC (permalink / raw)
To: Michael S. Tsirkin; +Cc: netdev, alexander.duyck, virtualization
In-Reply-To: <20171219174259-mutt-send-email-mst@kernel.org>
On 12/19/2017 7:47 AM, Michael S. Tsirkin wrote:
> I'll need to look at this more, in particular the feature
> bit is missing here. For now one question:
>
> On Mon, Dec 18, 2017 at 04:40:36PM -0800, Sridhar Samudrala wrote:
>> @@ -56,6 +58,8 @@ module_param(napi_tx, bool, 0644);
>> */
>> DECLARE_EWMA(pkt_len, 0, 64)
>>
>> +#define VF_TAKEOVER_INT (HZ / 10)
>> +
>> #define VIRTNET_DRIVER_VERSION "1.0.0"
>>
>> static const unsigned long guest_offloads[] = {
> Why is this delay necessary? And why by 100ms?
This is based on netvsc implementation and here is the commit that
added this delay. Not sure if this needs to be 100ms.
commit 6123c66854c174e4982f98195100c1d990f9e5e6
Author: stephen hemminger <stephen@networkplumber.org>
Date: Wed Aug 9 17:46:03 2017 -0700
netvsc: delay setup of VF device
When VF device is discovered, delay bring it automatically up in
order to allow userspace to some simple changes (like renaming).
_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization
^ permalink raw reply
* Re: [RFC PATCH] virtio_net: Extend virtio to use VF datapath when available
From: Stephen Hemminger @ 2017-12-19 17:55 UTC (permalink / raw)
To: Samudrala, Sridhar
Cc: netdev, virtualization, alexander.duyck, Michael S. Tsirkin
In-Reply-To: <ca709329-a47a-2e76-b1fc-38fc665d7c0b@intel.com>
On Tue, 19 Dec 2017 09:41:39 -0800
"Samudrala, Sridhar" <sridhar.samudrala@intel.com> wrote:
> On 12/19/2017 7:47 AM, Michael S. Tsirkin wrote:
> > I'll need to look at this more, in particular the feature
> > bit is missing here. For now one question:
> >
> > On Mon, Dec 18, 2017 at 04:40:36PM -0800, Sridhar Samudrala wrote:
> >> @@ -56,6 +58,8 @@ module_param(napi_tx, bool, 0644);
> >> */
> >> DECLARE_EWMA(pkt_len, 0, 64)
> >>
> >> +#define VF_TAKEOVER_INT (HZ / 10)
> >> +
> >> #define VIRTNET_DRIVER_VERSION "1.0.0"
> >>
> >> static const unsigned long guest_offloads[] = {
> > Why is this delay necessary? And why by 100ms?
>
> This is based on netvsc implementation and here is the commit that
> added this delay. Not sure if this needs to be 100ms.
>
> commit 6123c66854c174e4982f98195100c1d990f9e5e6
> Author: stephen hemminger <stephen@networkplumber.org>
> Date: Wed Aug 9 17:46:03 2017 -0700
>
> netvsc: delay setup of VF device
>
> When VF device is discovered, delay bring it automatically up in
> order to allow userspace to some simple changes (like renaming).
>
>
>
could be 10ms, just enough to let udev do its renaming
_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization
^ permalink raw reply
* Re: [RFC PATCH] virtio_net: Extend virtio to use VF datapath when available
From: Michael S. Tsirkin @ 2017-12-19 18:07 UTC (permalink / raw)
To: Stephen Hemminger
Cc: Samudrala, Sridhar, virtualization, alexander.duyck, netdev
In-Reply-To: <20171219095548.61d70d10@xeon-e3>
On Tue, Dec 19, 2017 at 09:55:48AM -0800, Stephen Hemminger wrote:
> On Tue, 19 Dec 2017 09:41:39 -0800
> "Samudrala, Sridhar" <sridhar.samudrala@intel.com> wrote:
>
> > On 12/19/2017 7:47 AM, Michael S. Tsirkin wrote:
> > > I'll need to look at this more, in particular the feature
> > > bit is missing here. For now one question:
> > >
> > > On Mon, Dec 18, 2017 at 04:40:36PM -0800, Sridhar Samudrala wrote:
> > >> @@ -56,6 +58,8 @@ module_param(napi_tx, bool, 0644);
> > >> */
> > >> DECLARE_EWMA(pkt_len, 0, 64)
> > >>
> > >> +#define VF_TAKEOVER_INT (HZ / 10)
> > >> +
> > >> #define VIRTNET_DRIVER_VERSION "1.0.0"
> > >>
> > >> static const unsigned long guest_offloads[] = {
> > > Why is this delay necessary? And why by 100ms?
> >
> > This is based on netvsc implementation and here is the commit that
> > added this delay. Not sure if this needs to be 100ms.
> >
> > commit 6123c66854c174e4982f98195100c1d990f9e5e6
> > Author: stephen hemminger <stephen@networkplumber.org>
> > Date: Wed Aug 9 17:46:03 2017 -0700
> >
> > netvsc: delay setup of VF device
> >
> > When VF device is discovered, delay bring it automatically up in
> > order to allow userspace to some simple changes (like renaming).
> >
> >
> >
>
> could be 10ms, just enough to let udev do its renaming
Isn't there a way not to depend on udev completing its thing within a given timeframe?
^ permalink raw reply
* Re: [PATCH v20 0/7] Virtio-balloon Enhancement
From: Michael S. Tsirkin @ 2017-12-19 18:08 UTC (permalink / raw)
To: Tetsuo Handa
Cc: yang.zhang.wz, kvm, liliang.opensource, qemu-devel,
virtualization, linux-mm, aarcange, virtio-dev, mawilcox, willy,
nilal, riel, cornelia.huck, mhocko, quan.xu0, linux-kernel,
amit.shah, pbonzini, akpm, mgorman
In-Reply-To: <201712192305.AAE21882.MtQHJOFFSFVOLO@I-love.SAKURA.ne.jp>
On Tue, Dec 19, 2017 at 11:05:11PM +0900, Tetsuo Handa wrote:
> Wei Wang wrote:
> > ChangeLog:
> > v19->v20:
> > 1) patch 1: xbitmap
> > - add __rcu to "void **slot";
> > - remove the exceptional path.
> > 2) patch 3: xbitmap
> > - DeveloperNotes: add an item to comment that the current bit range
> > related APIs operating on extremely large ranges (e.g.
> > [0, ULONG_MAX)) will take too long time. This can be optimized in
> > the future.
> > - remove the exceptional path;
> > - remove xb_preload_and_set();
> > - reimplement xb_clear_bit_range to make its usage close to
> > bitmap_clear;
> > - rename xb_find_next_set_bit to xb_find_set, and re-implement it
> > in a style close to find_next_bit;
> > - rename xb_find_next_zero_bit to xb_find_clear, and re-implement
> > it in a stytle close to find_next_zero_bit;
> > - separate the implementation of xb_find_set and xb_find_clear for
> > the convenience of future updates.
>
> Removing exceptional path made this patch easier to read.
> But what I meant is
>
> Can you eliminate exception path and fold all xbitmap patches into one, and
> post only one xbitmap patch without virtio-balloon changes?
And then people will complain that patch is too big
and hard to understand without any users.
As long as patches don't change the same lines of code,
it's fine to split up imho. In this case it's done
to attribute code better, seems like a reasonable thing to do
to me.
> .
>
> I still think we don't need xb_preload()/xb_preload_end().
> I think xb_find_set() has a bug in !node path.
>
> Also, please avoid unconditionally adding to builtin modules.
> There are users who want to save even few KB.
^ permalink raw reply
* Re: [RFC PATCH] virtio_net: Extend virtio to use VF datapath when available
From: Stephen Hemminger @ 2017-12-19 18:13 UTC (permalink / raw)
To: Michael S. Tsirkin
Cc: Samudrala, Sridhar, virtualization, alexander.duyck, netdev
In-Reply-To: <20171219200611-mutt-send-email-mst@kernel.org>
On Tue, 19 Dec 2017 20:07:01 +0200
"Michael S. Tsirkin" <mst@redhat.com> wrote:
> On Tue, Dec 19, 2017 at 09:55:48AM -0800, Stephen Hemminger wrote:
> > On Tue, 19 Dec 2017 09:41:39 -0800
> > "Samudrala, Sridhar" <sridhar.samudrala@intel.com> wrote:
> >
> > > On 12/19/2017 7:47 AM, Michael S. Tsirkin wrote:
> > > > I'll need to look at this more, in particular the feature
> > > > bit is missing here. For now one question:
> > > >
> > > > On Mon, Dec 18, 2017 at 04:40:36PM -0800, Sridhar Samudrala wrote:
> > > >> @@ -56,6 +58,8 @@ module_param(napi_tx, bool, 0644);
> > > >> */
> > > >> DECLARE_EWMA(pkt_len, 0, 64)
> > > >>
> > > >> +#define VF_TAKEOVER_INT (HZ / 10)
> > > >> +
> > > >> #define VIRTNET_DRIVER_VERSION "1.0.0"
> > > >>
> > > >> static const unsigned long guest_offloads[] = {
> > > > Why is this delay necessary? And why by 100ms?
> > >
> > > This is based on netvsc implementation and here is the commit that
> > > added this delay. Not sure if this needs to be 100ms.
> > >
> > > commit 6123c66854c174e4982f98195100c1d990f9e5e6
> > > Author: stephen hemminger <stephen@networkplumber.org>
> > > Date: Wed Aug 9 17:46:03 2017 -0700
> > >
> > > netvsc: delay setup of VF device
> > >
> > > When VF device is discovered, delay bring it automatically up in
> > > order to allow userspace to some simple changes (like renaming).
> > >
> > >
> > >
> >
> > could be 10ms, just enough to let udev do its renaming
>
> Isn't there a way not to depend on udev completing its thing within a given timeframe?
Not that I know. the path is quite indirect.
_______________________________________________
Virtualization mailing list
Virtualization@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/virtualization
^ permalink raw reply
* Re: [RFC PATCH] virtio_net: Extend virtio to use VF datapath when available
From: David Miller @ 2017-12-19 18:20 UTC (permalink / raw)
To: sridhar.samudrala; +Cc: mst, netdev, alexander.duyck, virtualization
In-Reply-To: <ca709329-a47a-2e76-b1fc-38fc665d7c0b@intel.com>
From: "Samudrala, Sridhar" <sridhar.samudrala@intel.com>
Date: Tue, 19 Dec 2017 09:41:39 -0800
> This is based on netvsc implementation and here is the commit that
> added this delay. Not sure if this needs to be 100ms.
>
> commit 6123c66854c174e4982f98195100c1d990f9e5e6
> Author: stephen hemminger <stephen@networkplumber.org>
> Date: Wed Aug 9 17:46:03 2017 -0700
>
> netvsc: delay setup of VF device
>
> When VF device is discovered, delay bring it automatically up in
> order to allow userspace to some simple changes (like renaming).
This is kind of bogus, I should have called this out when the patch
was posted.
Any delay is wrong, there needs to be tight synchronization if a
userspace operation must occur before proceeding. If something
happens and userspace is delayed, this whole thing doesn't work.
^ permalink raw reply
* Re: [RFC PATCH] virtio_net: Extend virtio to use VF datapath when available
From: David Miller @ 2017-12-19 18:21 UTC (permalink / raw)
To: stephen; +Cc: mst, sridhar.samudrala, alexander.duyck, virtualization, netdev
In-Reply-To: <20171219095548.61d70d10@xeon-e3>
From: Stephen Hemminger <stephen@networkplumber.org>
Date: Tue, 19 Dec 2017 09:55:48 -0800
> could be 10ms, just enough to let udev do its renaming
Please, move to some kind of notification or event based handling of
this problem.
No delay is safe, what if userspace gets swapped out or whatever
else might make userspace stall unexpectedly?
^ permalink raw reply
* Re: [RFC PATCH] virtio_net: Extend virtio to use VF datapath when available
From: Stephen Hemminger @ 2017-12-19 18:41 UTC (permalink / raw)
To: David Miller
Cc: mst, sridhar.samudrala, alexander.duyck, virtualization, netdev
In-Reply-To: <20171219.132117.666798483874609477.davem@davemloft.net>
On Tue, 19 Dec 2017 13:21:17 -0500 (EST)
David Miller <davem@davemloft.net> wrote:
> From: Stephen Hemminger <stephen@networkplumber.org>
> Date: Tue, 19 Dec 2017 09:55:48 -0800
>
> > could be 10ms, just enough to let udev do its renaming
>
> Please, move to some kind of notification or event based handling of
> this problem.
>
> No delay is safe, what if userspace gets swapped out or whatever
> else might make userspace stall unexpectedly?
>
The plan is to remove the delay and do the naming in the kernel.
This was suggested by Lennart since udev is only doing naming policy
because kernel names were not repeatable.
This makes the VF show up as "ethN_vf" on Hyper-V which is user friendly.
Patch is pending.
^ permalink raw reply
* Re: [RFC PATCH] virtio_net: Extend virtio to use VF datapath when available
From: Samudrala, Sridhar @ 2017-12-19 19:42 UTC (permalink / raw)
To: Stephen Hemminger, David Miller
Cc: netdev, virtualization, alexander.duyck, mst
In-Reply-To: <20171219104159.30d9caaf@xeon-e3>
On 12/19/2017 10:41 AM, Stephen Hemminger wrote:
> On Tue, 19 Dec 2017 13:21:17 -0500 (EST)
> David Miller <davem@davemloft.net> wrote:
>
>> From: Stephen Hemminger <stephen@networkplumber.org>
>> Date: Tue, 19 Dec 2017 09:55:48 -0800
>>
>>> could be 10ms, just enough to let udev do its renaming
>> Please, move to some kind of notification or event based handling of
>> this problem.
>>
>> No delay is safe, what if userspace gets swapped out or whatever
>> else might make userspace stall unexpectedly?
>>
> The plan is to remove the delay and do the naming in the kernel.
> This was suggested by Lennart since udev is only doing naming policy
> because kernel names were not repeatable.
>
> This makes the VF show up as "ethN_vf" on Hyper-V which is user friendly.
>
> Patch is pending.
Do we really need to delay the setup until the name is changed?
Can't we call dev_set_mtu() and dev_open() until dev_change_name() is done?
Thanks
Sridhar
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox