[dpdk-dev,v3,1/6] ethdev: fix port data mismatched in multiple process model
Checks
Commit Message
Assume we have two virtio ports, 00:03.0 and 00:04.0. The first one is
managed by the kernel driver, while the later one is managed by DPDK.
Now we start the primary process. 00:03.0 will be skipped by DPDK virtio
PMD driver (since it's being used by the kernel). 00:04.0 would be
successfully initiated by DPDK virtio PMD (if nothing abnormal happens).
After that, we would get a port id 0, and all the related info needed
by virtio (virtio_hw) is stored at rte_eth_dev_data[0].
Then we start the secondary process. As usual, 00:03.0 will be firstly
probed. It firstly tries to get a local eth_dev structure for it (by
rte_eth_dev_allocate):
port_id = rte_eth_dev_find_free_port();
...
eth_dev = &rte_eth_devices[port_id];
eth_dev->data = &rte_eth_dev_data[port_id];
...
return eth_dev;
Since it's a first PCI device, port_id will be 0. eth_dev->data would
then point to rte_eth_dev_data[0]. And here things start going wrong,
as rte_eth_dev_data[0] actually stores the virtio_hw for 00:04.0.
That said, in the secondary process, DPDK will continue to drive PCI
device 00.03.0 (despite the fact it's been managed by kernel), with
the info from PCI device 00:04.0. Which is wrong.
The fix is to attach the port already registered by the primary process:
iterate the rte_eth_dev_data[], and get the port id who's PCI ID matches
the current PCI device.
This would let us maintain same port ID for the same PCI device, keeping
the chance of referencing to wrong data minimal.
Fixes: af75078fece3 ("first public release")
Cc: stable@dpdk.org
Cc: Thomas Monjalon <thomas.monjalon@6wind.com>
Cc: Bruce Richardson <bruce.richardson@intel.com>
Cc: Ferruh Yigit <ferruh.yigit@intel.com>
Signed-off-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
---
v3: - do not move rte_eth_dev_data_alloc to pci_probe
- rename eth_dev_attach to eth_dev_attach_secondary
- introduce eth_dev_init() for common eth_dev struct initiation
- move comment block inside the "if" block
---
lib/librte_ether/rte_ethdev.c | 77 ++++++++++++++++++++++++++++++++++++++-----
1 file changed, 68 insertions(+), 9 deletions(-)
Comments
2017-01-06 18:16, Yuanhan Liu:
> +static void
> +eth_dev_init(struct rte_eth_dev *eth_dev, uint8_t port_id, const char *name)
> +{
> + eth_dev->data = &rte_eth_dev_data[port_id];
> + eth_dev->attached = DEV_ATTACHED;
> + eth_dev_last_created_port = port_id;
> + nb_ports++;
> +
> + if (rte_eal_process_type() == RTE_PROC_PRIMARY) {
> + snprintf(eth_dev->data->name, sizeof(eth_dev->data->name),
> + "%s", name);
> + eth_dev->data->port_id = port_id;
> + }
Why not keeping eth_dev->data filling in rte_eth_dev_allocate?
> +}
> +
> struct rte_eth_dev *
> rte_eth_dev_allocate(const char *name)
> {
> @@ -211,12 +226,41 @@ struct rte_eth_dev *
> }
>
> eth_dev = &rte_eth_devices[port_id];
Why not moving this line in eth_dev_init?
> - eth_dev->data = &rte_eth_dev_data[port_id];
> - snprintf(eth_dev->data->name, sizeof(eth_dev->data->name), "%s", name);
> - eth_dev->data->port_id = port_id;
> - eth_dev->attached = DEV_ATTACHED;
> - eth_dev_last_created_port = port_id;
> - nb_ports++;
> + eth_dev_init(eth_dev, port_id, name);
> +
> + return eth_dev;
> +}
[...]
> +/*
> + * Attach to a port already registered by the primary process, which
> + * makes sure that the same device would have the same port id both
> + * in the primary and secondary process.
> + */
> +static struct rte_eth_dev *
> +eth_dev_attach_secondary(const char *name)
OK, good description
[...]
> - eth_dev = rte_eth_dev_allocate(ethdev_name);
> - if (eth_dev == NULL)
> - return -ENOMEM;
> + if (rte_eal_process_type() == RTE_PROC_PRIMARY) {
> + eth_dev = rte_eth_dev_allocate(ethdev_name);
> + if (eth_dev == NULL)
> + return -ENOMEM;
You could merge here the rest of primary init below.
> + } else {
> + eth_dev = eth_dev_attach_secondary(ethdev_name);
> + if (eth_dev == NULL) {
> + /*
> + * if we failed to attach a device, it means
> + * the device is skipped, due to some errors.
> + * Take virtio-net device as example, it could
> + * due to the device is managed by virtio-net
> + * kernel driver. For such case, we return a
> + * positive value, to let EAL skip it as well.
> + */
I'm not sure we need an example here.
Is the virtio case special?
nit: "it could due" looks to be a typo
> + return 1;
> + }
> + }
>
> if (rte_eal_process_type() == RTE_PROC_PRIMARY) {
> eth_dev->data->dev_private = rte_zmalloc("ethdev private structure",
On Fri, Jan 06, 2017 at 02:12:48PM +0100, Thomas Monjalon wrote:
> 2017-01-06 18:16, Yuanhan Liu:
> > +static void
> > +eth_dev_init(struct rte_eth_dev *eth_dev, uint8_t port_id, const char *name)
> > +{
> > + eth_dev->data = &rte_eth_dev_data[port_id];
> > + eth_dev->attached = DEV_ATTACHED;
> > + eth_dev_last_created_port = port_id;
> > + nb_ports++;
> > +
> > + if (rte_eal_process_type() == RTE_PROC_PRIMARY) {
> > + snprintf(eth_dev->data->name, sizeof(eth_dev->data->name),
> > + "%s", name);
> > + eth_dev->data->port_id = port_id;
> > + }
>
> Why not keeping eth_dev->data filling in rte_eth_dev_allocate?
I could do that.
>
> > +}
> > +
> > struct rte_eth_dev *
> > rte_eth_dev_allocate(const char *name)
> > {
> > @@ -211,12 +226,41 @@ struct rte_eth_dev *
> > }
> >
> > eth_dev = &rte_eth_devices[port_id];
>
> Why not moving this line in eth_dev_init?
Ditto.
>
> > - eth_dev->data = &rte_eth_dev_data[port_id];
> > - snprintf(eth_dev->data->name, sizeof(eth_dev->data->name), "%s", name);
> > - eth_dev->data->port_id = port_id;
> > - eth_dev->attached = DEV_ATTACHED;
> > - eth_dev_last_created_port = port_id;
> > - nb_ports++;
> > + eth_dev_init(eth_dev, port_id, name);
> > +
> > + return eth_dev;
> > +}
>
> [...]
> > +/*
> > + * Attach to a port already registered by the primary process, which
> > + * makes sure that the same device would have the same port id both
> > + * in the primary and secondary process.
> > + */
> > +static struct rte_eth_dev *
> > +eth_dev_attach_secondary(const char *name)
>
> OK, good description
>
> [...]
> > - eth_dev = rte_eth_dev_allocate(ethdev_name);
> > - if (eth_dev == NULL)
> > - return -ENOMEM;
> > + if (rte_eal_process_type() == RTE_PROC_PRIMARY) {
> > + eth_dev = rte_eth_dev_allocate(ethdev_name);
> > + if (eth_dev == NULL)
> > + return -ENOMEM;
>
> You could merge here the rest of primary init below.
Oh, right.
>
> > + } else {
> > + eth_dev = eth_dev_attach_secondary(ethdev_name);
> > + if (eth_dev == NULL) {
> > + /*
> > + * if we failed to attach a device, it means
> > + * the device is skipped, due to some errors.
> > + * Take virtio-net device as example, it could
> > + * due to the device is managed by virtio-net
> > + * kernel driver. For such case, we return a
> > + * positive value, to let EAL skip it as well.
> > + */
>
> I'm not sure we need an example here.
> Is the virtio case special?
Not quite sure, likely not. I was thinking a detailed example helps
people to understand better. I could remove it if you think it's
not necessary.
> nit: "it could due" looks to be a typo
Oops.
--yliu
@@ -189,6 +189,21 @@ struct rte_eth_dev *
return RTE_MAX_ETHPORTS;
}
+static void
+eth_dev_init(struct rte_eth_dev *eth_dev, uint8_t port_id, const char *name)
+{
+ eth_dev->data = &rte_eth_dev_data[port_id];
+ eth_dev->attached = DEV_ATTACHED;
+ eth_dev_last_created_port = port_id;
+ nb_ports++;
+
+ if (rte_eal_process_type() == RTE_PROC_PRIMARY) {
+ snprintf(eth_dev->data->name, sizeof(eth_dev->data->name),
+ "%s", name);
+ eth_dev->data->port_id = port_id;
+ }
+}
+
struct rte_eth_dev *
rte_eth_dev_allocate(const char *name)
{
@@ -211,12 +226,41 @@ struct rte_eth_dev *
}
eth_dev = &rte_eth_devices[port_id];
- eth_dev->data = &rte_eth_dev_data[port_id];
- snprintf(eth_dev->data->name, sizeof(eth_dev->data->name), "%s", name);
- eth_dev->data->port_id = port_id;
- eth_dev->attached = DEV_ATTACHED;
- eth_dev_last_created_port = port_id;
- nb_ports++;
+ eth_dev_init(eth_dev, port_id, name);
+
+ return eth_dev;
+}
+
+/*
+ * Attach to a port already registered by the primary process, which
+ * makes sure that the same device would have the same port id both
+ * in the primary and secondary process.
+ */
+static struct rte_eth_dev *
+eth_dev_attach_secondary(const char *name)
+{
+ uint8_t i;
+ struct rte_eth_dev *eth_dev;
+
+ if (rte_eth_dev_data == NULL)
+ rte_eth_dev_data_alloc();
+
+ for (i = 0; i < RTE_MAX_ETHPORTS; i++) {
+ if (strcmp(rte_eth_dev_data[i].name, name) == 0)
+ break;
+ }
+ if (i == RTE_MAX_ETHPORTS) {
+ RTE_PMD_DEBUG_TRACE(
+ "device %s is not driven by the primary process\n",
+ name);
+ return NULL;
+ }
+
+ RTE_ASSERT(eth_dev->data->port_id == i);
+
+ eth_dev = &rte_eth_devices[i];
+ eth_dev_init(eth_dev, i, NULL);
+
return eth_dev;
}
@@ -246,9 +290,24 @@ struct rte_eth_dev *
rte_eal_pci_device_name(&pci_dev->addr, ethdev_name,
sizeof(ethdev_name));
- eth_dev = rte_eth_dev_allocate(ethdev_name);
- if (eth_dev == NULL)
- return -ENOMEM;
+ if (rte_eal_process_type() == RTE_PROC_PRIMARY) {
+ eth_dev = rte_eth_dev_allocate(ethdev_name);
+ if (eth_dev == NULL)
+ return -ENOMEM;
+ } else {
+ eth_dev = eth_dev_attach_secondary(ethdev_name);
+ if (eth_dev == NULL) {
+ /*
+ * if we failed to attach a device, it means
+ * the device is skipped, due to some errors.
+ * Take virtio-net device as example, it could
+ * due to the device is managed by virtio-net
+ * kernel driver. For such case, we return a
+ * positive value, to let EAL skip it as well.
+ */
+ return 1;
+ }
+ }
if (rte_eal_process_type() == RTE_PROC_PRIMARY) {
eth_dev->data->dev_private = rte_zmalloc("ethdev private structure",