| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-Id: <20171101022946.27114-1-jakub.kicinski@netronome.com>
Date: Tue, 31 Oct 2017 19:29:46 -0700
From: Jakub Kicinski <jakub.kicinski@...ronome.com>
To: netdev@...r.kernel.org
Cc: oss-drivers@...ronome.com, alexei.starovoitov@...il.com,
daniel@...earbox.net, Jakub Kicinski <jakub.kicinski@...ronome.com>
Subject: [RFC] net: dummy: add BPF offload callbacks for test purposes
Add callbacks for BPF offload, a handful of checks and DebugFS
files to be able to test BPF offload without any offload-
-capable HW.
Signed-off-by: Jakub Kicinski <jakub.kicinski@...ronome.com>
Reviewed-by: Quentin Monnet <quentin.monnet@...ronome.com>
---
This was a little useful for testing, I'm not sure we want
it upstream, though?
drivers/net/dummy.c | 162 +++++++++++++++++++++++++++++++++++++++++++++++++---
1 file changed, 154 insertions(+), 8 deletions(-)
diff --git a/drivers/net/dummy.c b/drivers/net/dummy.c
index 58483af80bdb..e1a1ad7902de 100644
--- a/drivers/net/dummy.c
+++ b/drivers/net/dummy.c
@@ -28,6 +28,10 @@
Alan Cox, 30th May 1994
*/
+#include <linux/bpf.h>
+#include <linux/bpf_verifier.h>
+#include <linux/debugfs.h>
+#include <linux/filter.h>
#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/netdevice.h>
@@ -47,6 +51,7 @@
static int numdummies = 1;
static int num_vfs;
+static bool debugfs;
struct vf_data_storage {
u8 vf_mac[ETH_ALEN];
@@ -61,8 +66,21 @@ struct vf_data_storage {
int link_state;
};
+struct dummy_bpf_offload_prog {
+ struct dummy_priv *dummy;
+ struct bpf_prog *prog;
+ const char *state;
+ struct list_head l;
+};
+
struct dummy_priv {
struct vf_data_storage *vfinfo;
+ struct bpf_prog *xdp_prog;
+ int xdp_prog_mode;
+ struct dentry *ddir;
+ bool bpf_offload_accept;
+ u32 bpf_offload_verifier_delay;
+ struct list_head bpf_offload_progs;
};
static int dummy_num_vf(struct device *dev)
@@ -85,6 +103,35 @@ static struct device dummy_parent = {
.release = release_dummy_parent,
};
+static int dummy_debugfs_bpf_offload_read(struct seq_file *file, void *data)
+{
+ struct dummy_priv *priv = file->private;
+ struct dummy_bpf_offload_prog *state;
+ unsigned int i = 0;
+
+ rtnl_lock();
+ list_for_each_entry(state, &priv->bpf_offload_progs, l)
+ seq_printf(file, "%u %u %s %d\n",
+ i++, state->prog->aux->id, state->state,
+ state->prog == state->dummy->xdp_prog);
+ rtnl_unlock();
+
+ return 0;
+}
+
+static int dummy_debugfs_bpf_offload_open(struct inode *inode, struct file *f)
+{
+ return single_open(f, dummy_debugfs_bpf_offload_read, inode->i_private);
+}
+
+static const struct file_operations dummy_bpf_offload_fops = {
+ .owner = THIS_MODULE,
+ .open = dummy_debugfs_bpf_offload_open,
+ .release = single_release,
+ .read = seq_read,
+ .llseek = seq_lseek
+};
+
/* fake multicast ability */
static void set_multicast_list(struct net_device *dev)
{
@@ -141,18 +188,36 @@ static int dummy_dev_init(struct net_device *dev)
priv->vfinfo = NULL;
- if (!num_vfs)
- return 0;
+ if (num_vfs) {
+ dev->dev.parent = &dummy_parent;
+ priv->vfinfo = kcalloc(num_vfs, sizeof(struct vf_data_storage),
+ GFP_KERNEL);
+ if (!priv->vfinfo)
+ goto err_free_dstats;
+ }
- dev->dev.parent = &dummy_parent;
- priv->vfinfo = kcalloc(num_vfs, sizeof(struct vf_data_storage),
- GFP_KERNEL);
- if (!priv->vfinfo) {
- free_percpu(dev->dstats);
- return -ENOMEM;
+ INIT_LIST_HEAD(&priv->bpf_offload_progs);
+ if (debugfs) {
+ priv->ddir = debugfs_create_dir(dev->name, NULL);
+ if (!priv->ddir)
+ goto err_free_vfinfo;
+
+ debugfs_create_bool("bpf_offload_accept", 0600,
+ priv->ddir, &priv->bpf_offload_accept);
+ debugfs_create_u32("bpf_offload_verifier_delay", 0600,
+ priv->ddir,
+ &priv->bpf_offload_verifier_delay);
+ debugfs_create_file("bpf_offload_list", 0600, priv->ddir,
+ priv, &dummy_bpf_offload_fops);
}
return 0;
+
+err_free_vfinfo:
+ kfree(priv->vfinfo);
+err_free_dstats:
+ free_percpu(dev->dstats);
+ return -ENOMEM;
}
static void dummy_dev_uninit(struct net_device *dev)
@@ -280,6 +345,81 @@ static int dummy_set_vf_link_state(struct net_device *dev, int vf, int state)
return 0;
}
+static int
+dummy_bpf_verify_insn(struct bpf_verifier_env *env, int insn_idx, int prev_insn)
+{
+ struct dummy_bpf_offload_prog *state;
+
+ state = env->prog->aux->offload->dev_priv;
+ if (state->dummy->bpf_offload_verifier_delay && !insn_idx)
+ msleep(state->dummy->bpf_offload_verifier_delay);
+
+ return 0;
+}
+
+static const struct bpf_ext_analyzer_ops dummy_bpf_analyzer_ops = {
+ .insn_hook = dummy_bpf_verify_insn,
+};
+
+static int dummy_bpf(struct net_device *dev, struct netdev_bpf *bpf)
+{
+ struct dummy_priv *priv = netdev_priv(dev);
+ struct dummy_bpf_offload_prog *state;
+
+ ASSERT_RTNL();
+
+ switch (bpf->command) {
+ case BPF_OFFLOAD_VERIFIER_PREP:
+ if (!priv->bpf_offload_accept)
+ return -EOPNOTSUPP;
+
+ state = kzalloc(sizeof(*state), GFP_KERNEL);
+ if (!state)
+ return -ENOMEM;
+ bpf->verifier.prog->aux->offload->dev_priv = state;
+ bpf->verifier.ops = &dummy_bpf_analyzer_ops;
+
+ state->dummy = priv;
+ state->prog = bpf->verifier.prog;
+ state->state = "verify";
+ list_add_tail(&state->l, &priv->bpf_offload_progs);
+ return 0;
+ case BPF_OFFLOAD_TRANSLATE:
+ state = bpf->offload.prog->aux->offload->dev_priv;
+
+ state->state = "xlated";
+ return 0;
+ case BPF_OFFLOAD_DESTROY:
+ state = bpf->offload.prog->aux->offload->dev_priv;
+
+ list_del(&state->l);
+ kfree(state);
+ return 0;
+ case XDP_QUERY_PROG:
+ if (priv->xdp_prog) {
+ bpf->prog_attached = priv->xdp_prog_mode;
+ bpf->prog_id = priv->xdp_prog->aux->id;
+ }
+ return 0;
+ case XDP_SETUP_PROG:
+ case XDP_SETUP_PROG_HW:
+ if (bpf->prog->aux->offload) {
+ WARN_ON(bpf->prog->aux->offload->netdev != dev);
+ state = bpf->prog->aux->offload->dev_priv;
+ WARN_ON(strcmp(state->state, "xlated"));
+ }
+
+ if (priv->xdp_prog)
+ bpf_prog_put(priv->xdp_prog);
+ priv->xdp_prog = bpf->prog;
+ priv->xdp_prog_mode = bpf->command == XDP_SETUP_PROG ?
+ XDP_ATTACHED_DRV : XDP_ATTACHED_HW;
+ return 0;
+ default:
+ return -EINVAL;
+ }
+}
+
static const struct net_device_ops dummy_netdev_ops = {
.ndo_init = dummy_dev_init,
.ndo_uninit = dummy_dev_uninit,
@@ -297,6 +437,7 @@ static const struct net_device_ops dummy_netdev_ops = {
.ndo_get_vf_config = dummy_get_vf_config,
.ndo_set_vf_link_state = dummy_set_vf_link_state,
.ndo_set_vf_rss_query_en = dummy_set_vf_rss_query_en,
+ .ndo_bpf = dummy_bpf,
};
static void dummy_get_drvinfo(struct net_device *dev,
@@ -327,6 +468,8 @@ static void dummy_free_netdev(struct net_device *dev)
{
struct dummy_priv *priv = netdev_priv(dev);
+ WARN_ON(!list_empty(&priv->bpf_offload_progs));
+ debugfs_remove_recursive(priv->ddir);
kfree(priv->vfinfo);
}
@@ -382,6 +525,9 @@ MODULE_PARM_DESC(numdummies, "Number of dummy pseudo devices");
module_param(num_vfs, int, 0);
MODULE_PARM_DESC(num_vfs, "Number of dummy VFs per dummy device");
+module_param(debugfs, bool, false);
+MODULE_PARM_DESC(debugfs, "Expose control interface in DebugFS");
+
static int __init dummy_init_one(void)
{
struct net_device *dev_dummy;
--
2.14.1
Powered by blists - more mailing lists