From: Adrien Mazarguil <adrien.mazarguil@6wind.com>
To: Shahaf Shuler <shahafs@mellanox.com>
Cc: dev@dpdk.org
Subject: [dpdk-dev] [PATCH 5/7] net/mlx5: add port representor awareness
Date: Fri, 25 May 2018 18:35:23 +0200 [thread overview]
Message-ID: <20180525161814.13873-6-adrien.mazarguil@6wind.com> (raw)
In-Reply-To: <20180525161814.13873-1-adrien.mazarguil@6wind.com>
The current PCI probing method is not aware of Verbs port representors,
which appear as standard Verbs devices bound to the same PCI address and
cannot be distinguished.
Problem is that more often than not, the wrong Verbs device is used,
resulting in unexpected traffic.
This patch adds necessary heuristics to bind affected driver instances to
the intended (i.e. non-representor) device.
(Patch based on prior work from Yuanhan Liu)
Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
drivers/net/mlx5/mlx5.c | 61 +++++++++++++++++++++++++++++++++++++++++---
1 file changed, 57 insertions(+), 4 deletions(-)
diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index d57e8118c..d3a298332 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -1155,6 +1155,32 @@ mlx5_dev_spawn(struct rte_device *dpdk_dev,
}
/**
+ * Comparison callback to sort Verbs device names.
+ *
+ * This is meant to be used with qsort().
+ *
+ * @param a[in]
+ * Pointer to pointer to first Verbs device.
+ * @param b[in]
+ * Pointer to pointer to second Verbs device.
+ *
+ * @return
+ * 0 if both names are equal, less than 0 if the first argument is less
+ * than the second, greater than 0 otherwise.
+ */
+static int
+mlx5_cmp_ibv_name(const void *a, const void *b)
+{
+ const char *name_a = (*(const struct ibv_device *const *)a)->name;
+ const char *name_b = (*(const struct ibv_device *const *)b)->name;
+ size_t i = 0;
+
+ while (name_a[i] && name_a[i] == name_b[i])
+ ++i;
+ return atoi(name_a + i) - atoi(name_b + i);
+}
+
+/**
* DPDK callback to register a PCI device.
*
* This function creates an Ethernet device for each port of a given
@@ -1174,6 +1200,7 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
{
struct ibv_device **ibv_list;
struct rte_eth_dev **eth_list = NULL;
+ int n = 0;
int vf;
int ret;
@@ -1195,6 +1222,9 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
DRV_LOG(ERR, "cannot list devices, is ib_uverbs loaded?");
return -rte_errno;
}
+
+ struct ibv_device *ibv_match[ret + 1];
+
while (ret-- > 0) {
struct rte_pci_addr pci_addr;
@@ -1206,12 +1236,35 @@ mlx5_pci_probe(struct rte_pci_driver *pci_drv __rte_unused,
pci_dev->addr.devid != pci_addr.devid ||
pci_dev->addr.function != pci_addr.function)
continue;
- DRV_LOG(INFO, "PCI information matches, using device \"%s\"",
+ DRV_LOG(INFO, "PCI information matches for device \"%s\"",
ibv_list[ret]->name);
- break;
+ ibv_match[n++] = ibv_list[ret];
+ }
+ ibv_match[n] = NULL;
+ if (n > 1) {
+ /*
+ * The existence of several matching entries means port
+ * representors have been instantiated. No existing Verbs
+ * call nor /sys entries can tell them apart at this point.
+ *
+ * While definitely hackish, assume their names are numbered
+ * based on order of creation with master device first,
+ * followed by first port representor, followed by the
+ * second one and so on.
+ */
+ DRV_LOG(WARNING,
+ "probing device with port representors involves"
+ " heuristics with uncertain outcome");
+ qsort(ibv_match, n, sizeof(*ibv_match), mlx5_cmp_ibv_name);
+ DRV_LOG(WARNING, "assuming \"%s\" is the master device",
+ ibv_match[0]->name);
+ for (ret = 1; ret < n; ++ret)
+ DRV_LOG(WARNING,
+ "assuming \"%s\" is port representor #%d",
+ ibv_match[ret]->name, ret - 1);
}
- if (ret >= 0)
- eth_list = mlx5_dev_spawn(&pci_dev->device, ibv_list[ret], vf);
+ if (n)
+ eth_list = mlx5_dev_spawn(&pci_dev->device, ibv_match[0], vf);
mlx5_glue->free_device_list(ibv_list);
if (!eth_list || !*eth_list) {
DRV_LOG(WARNING,
--
2.11.0
next prev parent reply other threads:[~2018-05-25 16:35 UTC|newest]
Thread overview: 100+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-05-25 16:35 [dpdk-dev] [PATCH 0/7] net/mlx5: add port representor support Adrien Mazarguil
2018-05-25 16:35 ` [dpdk-dev] [PATCH 1/7] net/mlx5: rename confusing object in probe code Adrien Mazarguil
2018-06-10 11:00 ` Xueming(Steven) Li
2018-06-12 13:19 ` Adrien Mazarguil
2018-05-25 16:35 ` [dpdk-dev] [PATCH 2/7] net/mlx5: remove redundant objects " Adrien Mazarguil
2018-06-10 11:00 ` Xueming(Steven) Li
2018-06-12 13:19 ` Adrien Mazarguil
2018-05-25 16:35 ` [dpdk-dev] [PATCH 3/7] net/mlx5: split PCI from generic probing code Adrien Mazarguil
2018-06-10 12:59 ` Xueming(Steven) Li
2018-06-12 13:20 ` Adrien Mazarguil
2018-05-25 16:35 ` [dpdk-dev] [PATCH 4/7] net/mlx5: re-indent generic probing function Adrien Mazarguil
2018-06-11 11:42 ` Xueming(Steven) Li
2018-05-25 16:35 ` Adrien Mazarguil [this message]
2018-05-25 16:35 ` [dpdk-dev] [PATCH 6/7] net/mlx5: probe all port representors Adrien Mazarguil
2018-06-12 6:42 ` Xueming(Steven) Li
2018-06-12 13:20 ` Adrien Mazarguil
2018-05-25 16:35 ` [dpdk-dev] [PATCH 7/7] net/mlx5: add parameter for " Adrien Mazarguil
2018-06-12 8:02 ` Xueming(Steven) Li
2018-06-12 13:20 ` Adrien Mazarguil
2018-06-12 13:43 ` Xueming(Steven) Li
2018-06-14 8:01 ` Adrien Mazarguil
2018-06-12 14:44 ` Xueming(Steven) Li
2018-06-13 13:11 ` Adrien Mazarguil
2018-06-14 8:34 ` [dpdk-dev] [PATCH v2 0/7] net/mlx5: add port representor support Adrien Mazarguil
2018-06-14 8:34 ` [dpdk-dev] [PATCH v2 1/7] net/mlx5: rename confusing object in probe code Adrien Mazarguil
2018-06-16 8:24 ` Xueming(Steven) Li
2018-06-14 8:34 ` [dpdk-dev] [PATCH v2 2/7] net/mlx5: remove redundant objects " Adrien Mazarguil
2018-06-16 8:27 ` Xueming(Steven) Li
2018-06-17 10:14 ` Shahaf Shuler
2018-06-27 13:30 ` Adrien Mazarguil
2018-06-28 5:35 ` Shahaf Shuler
2018-06-14 8:34 ` [dpdk-dev] [PATCH v2 3/7] net/mlx5: split PCI from generic probing code Adrien Mazarguil
2018-06-16 8:29 ` Xueming(Steven) Li
2018-06-17 10:14 ` Shahaf Shuler
2018-06-27 13:31 ` Adrien Mazarguil
2018-06-14 8:34 ` [dpdk-dev] [PATCH v2 4/7] net/mlx5: re-indent generic probing function Adrien Mazarguil
2018-06-14 8:34 ` [dpdk-dev] [PATCH v2 5/7] net/mlx5: add port representor awareness Adrien Mazarguil
2018-06-16 8:37 ` Xueming(Steven) Li
2018-06-27 13:32 ` Adrien Mazarguil
2018-06-14 8:35 ` [dpdk-dev] [PATCH v2 6/7] net/mlx5: probe all port representors Adrien Mazarguil
2018-06-16 8:57 ` Xueming(Steven) Li
2018-06-17 10:15 ` Shahaf Shuler
2018-06-24 13:33 ` Shahaf Shuler
2018-06-27 13:32 ` Adrien Mazarguil
2018-06-28 5:57 ` Shahaf Shuler
2018-06-28 9:13 ` Adrien Mazarguil
2018-06-27 13:32 ` Adrien Mazarguil
2018-06-27 17:30 ` Xueming(Steven) Li
2018-06-28 6:01 ` Shahaf Shuler
2018-06-28 8:45 ` Adrien Mazarguil
2018-06-28 9:06 ` Shahaf Shuler
2018-06-27 13:32 ` Adrien Mazarguil
2018-06-14 8:35 ` [dpdk-dev] [PATCH v2 7/7] net/mlx5: add parameter for " Adrien Mazarguil
2018-06-16 8:59 ` Xueming(Steven) Li
2018-07-04 17:27 ` [dpdk-dev] [PATCH v3 00/10] net/mlx5: add port representor support Adrien Mazarguil
2018-07-04 17:27 ` [dpdk-dev] [PATCH v3 01/10] net/mlx5: rename confusing object in probe code Adrien Mazarguil
2018-07-04 17:27 ` [dpdk-dev] [PATCH v3 02/10] net/mlx5: remove redundant objects " Adrien Mazarguil
2018-07-04 17:27 ` [dpdk-dev] [PATCH v3 03/10] net/mlx5: drop useless support for several Verbs ports Adrien Mazarguil
2018-07-04 17:27 ` [dpdk-dev] [PATCH v3 04/10] net/mlx5: split PCI from generic probing code Adrien Mazarguil
2018-07-04 17:27 ` [dpdk-dev] [PATCH v3 05/10] net/mlx5: re-indent generic probing function Adrien Mazarguil
2018-07-04 17:27 ` [dpdk-dev] [PATCH v3 06/10] net/mlx5: add port representor awareness Adrien Mazarguil
2018-07-04 17:27 ` [dpdk-dev] [PATCH v3 07/10] net/mlx5: probe all port representors Adrien Mazarguil
2018-07-04 17:27 ` [dpdk-dev] [PATCH v3 08/10] net/mlx5: probe port representors in natural order Adrien Mazarguil
2018-07-04 17:27 ` [dpdk-dev] [PATCH v3 09/10] net/mlx5: add parameter for port representors Adrien Mazarguil
2018-07-04 17:27 ` [dpdk-dev] [PATCH v3 10/10] net/mlx5: support negative identifiers " Adrien Mazarguil
2018-07-05 8:45 ` [dpdk-dev] [PATCH v4 00/10] net/mlx5: add port representor support Adrien Mazarguil
2018-07-05 8:45 ` [dpdk-dev] [PATCH v4 01/10] net/mlx5: rename confusing object in probe code Adrien Mazarguil
2018-07-05 8:45 ` [dpdk-dev] [PATCH v4 02/10] net/mlx5: remove redundant objects " Adrien Mazarguil
2018-07-05 8:45 ` [dpdk-dev] [PATCH v4 03/10] net/mlx5: drop useless support for several Verbs ports Adrien Mazarguil
2018-07-05 8:45 ` [dpdk-dev] [PATCH v4 04/10] net/mlx5: split PCI from generic probing code Adrien Mazarguil
2018-07-05 8:45 ` [dpdk-dev] [PATCH v4 05/10] net/mlx5: re-indent generic probing function Adrien Mazarguil
2018-07-05 8:45 ` [dpdk-dev] [PATCH v4 06/10] net/mlx5: add port representor awareness Adrien Mazarguil
2018-07-05 8:45 ` [dpdk-dev] [PATCH v4 07/10] net/mlx5: probe all port representors Adrien Mazarguil
2018-07-09 11:57 ` Shahaf Shuler
2018-07-10 9:37 ` Adrien Mazarguil
2018-07-10 10:13 ` Shahaf Shuler
2018-07-10 10:58 ` Adrien Mazarguil
2018-07-10 11:17 ` Shahaf Shuler
2018-07-05 8:45 ` [dpdk-dev] [PATCH v4 08/10] net/mlx5: probe port representors in natural order Adrien Mazarguil
2018-07-05 8:45 ` [dpdk-dev] [PATCH v4 09/10] net/mlx5: add parameter for port representors Adrien Mazarguil
2018-07-09 11:57 ` Shahaf Shuler
2018-07-10 9:37 ` Adrien Mazarguil
2018-07-10 10:16 ` Shahaf Shuler
2018-07-10 10:58 ` Adrien Mazarguil
2018-07-10 11:15 ` Shahaf Shuler
2018-07-05 8:45 ` [dpdk-dev] [PATCH v4 10/10] net/mlx5: support negative identifiers " Adrien Mazarguil
2018-07-09 11:58 ` Shahaf Shuler
2018-07-10 9:37 ` Adrien Mazarguil
2018-07-10 16:04 ` [dpdk-dev] [PATCH v5 00/10] net/mlx5: add port representor support Adrien Mazarguil
2018-07-10 16:04 ` [dpdk-dev] [PATCH v5 01/10] net/mlx5: rename confusing object in probe code Adrien Mazarguil
2018-07-10 16:04 ` [dpdk-dev] [PATCH v5 02/10] net/mlx5: remove redundant objects " Adrien Mazarguil
2018-07-10 16:04 ` [dpdk-dev] [PATCH v5 03/10] net/mlx5: drop useless support for several Verbs ports Adrien Mazarguil
2018-07-10 16:04 ` [dpdk-dev] [PATCH v5 04/10] net/mlx5: split PCI from generic probing code Adrien Mazarguil
2018-07-10 16:04 ` [dpdk-dev] [PATCH v5 05/10] net/mlx5: re-indent generic probing function Adrien Mazarguil
2018-07-10 16:04 ` [dpdk-dev] [PATCH v5 06/10] net/mlx5: add port representor awareness Adrien Mazarguil
2018-07-10 16:04 ` [dpdk-dev] [PATCH v5 07/10] net/mlx5: probe all port representors Adrien Mazarguil
2018-07-10 16:04 ` [dpdk-dev] [PATCH v5 08/10] net/mlx5: probe port representors in natural order Adrien Mazarguil
2018-07-10 16:04 ` [dpdk-dev] [PATCH v5 09/10] net/mlx5: add parameter for port representors Adrien Mazarguil
2018-07-10 16:05 ` [dpdk-dev] [PATCH v5 10/10] net/mlx5: support negative identifiers " Adrien Mazarguil
2018-07-12 7:51 ` [dpdk-dev] [PATCH v5 00/10] net/mlx5: add port representor support Shahaf Shuler
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20180525161814.13873-6-adrien.mazarguil@6wind.com \
--to=adrien.mazarguil@6wind.com \
--cc=dev@dpdk.org \
--cc=shahafs@mellanox.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).