[PATCH 2/2] drm/amdgpu: expose the connected port num info through sysfs

[Zhang, Hawking]

[AMD Official Use Only - General]

Series is

Reviewed-by: Hawking Zhang <Hawking.Zhang at amd.com>

Regards,
Hawking
-----Original Message-----
From: amd-gfx <amd-gfx-bounces at lists.freedesktop.org> On Behalf Of Shiwu Zhang
Sent: Tuesday, November 7, 2023 11:06
To: amd-gfx at lists.freedesktop.org
Subject: [PATCH 2/2] drm/amdgpu: expose the connected port num info through sysfs

By catting the xgmi_port_num sysfs node, it prints out the info in the format of <src node id>:<src port num> -> <dst node id>:<dst port num> for one xgmi link.

For example, in case of 4 sockets fully and evenly connected setup, it would be like as below for the first node in the hive.
01:02 -> 02:03
01:03 -> 02:02
01:07 -> 03:04
01:04 -> 03:07
01:06 -> 04:05
01:05 -> 04:06
Based on the fact that there is two xgmi links between each socket pair,
"01:02 -> 02:03" means that the current socket in question use the port 2 to connect with port 3 of the second node in the hive and so on.

v2: print out the src/dst node id for each xgmi link (lijo)
v3: replace the current_node++ with +1 to align with dst node (le)
    and use the dev_err instead of pr_err (lijo)

Signed-off-by: Shiwu Zhang <shiwu.zhang at amd.com>
Acked-by: Lijo Lazar <lijo.lazar at amd.com>
Reviewed-by: Le Ma <le.ma at amd.com>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c | 44 ++++++++++++++++++++++++
 1 file changed, 44 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c
index 2b7dc490ba6b..4bce96300e7c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c
@@ -372,6 +372,38 @@ static ssize_t amdgpu_xgmi_show_num_links(struct device *dev,
        return sysfs_emit(buf, "%s\n", buf);
 }

+static ssize_t amdgpu_xgmi_show_connected_port_num(struct device *dev,
+                                       struct device_attribute *attr,
+                                       char *buf)
+{
+       struct drm_device *ddev = dev_get_drvdata(dev);
+       struct amdgpu_device *adev = drm_to_adev(ddev);
+       struct psp_xgmi_topology_info *top = &adev->psp.xgmi_context.top_info;
+       int i, j, size = 0;
+       int current_node;
+       /*
+        * get the node id in the sysfs for the current socket and show
+        * it in the port num info output in the sysfs for easy reading.
+        * it is NOT the one retrieved from xgmi ta.
+        */
+       for (i = 0; i < top->num_nodes; i++) {
+               if (top->nodes[i].node_id == adev->gmc.xgmi.node_id) {
+                       current_node = i;
+                       break;
+               }
+       }
+
+       for (i = 0; i < top->num_nodes; i++) {
+               for (j=0; j < top->nodes[i].num_links; j++)
+                       /* node id in sysfs starts from 1 rather than 0 so +1 here */
+                       size += sysfs_emit_at(buf, size, "%02x:%02x ->  %02x:%02x\n", current_node + 1,
+                                             top->nodes[i].port_num[j].src_xgmi_port_num, i + 1,
+                                             top->nodes[i].port_num[j].dst_xgmi_port_num);
+       }
+
+       return size;
+}
+
 #define AMDGPU_XGMI_SET_FICAA(o)       ((o) | 0x456801)
 static ssize_t amdgpu_xgmi_show_error(struct device *dev,
                                      struct device_attribute *attr, @@ -411,6 +443,7 @@ static DEVICE_ATTR(xgmi_physical_id, 0444, amdgpu_xgmi_show_physical_id, NULL);  static DEVICE_ATTR(xgmi_error, S_IRUGO, amdgpu_xgmi_show_error, NULL);  static DEVICE_ATTR(xgmi_num_hops, S_IRUGO, amdgpu_xgmi_show_num_hops, NULL);  static DEVICE_ATTR(xgmi_num_links, S_IRUGO, amdgpu_xgmi_show_num_links, NULL);
+static DEVICE_ATTR(xgmi_port_num, S_IRUGO,
+amdgpu_xgmi_show_connected_port_num, NULL);

 static int amdgpu_xgmi_sysfs_add_dev_info(struct amdgpu_device *adev,
                                         struct amdgpu_hive_info *hive)
@@ -446,6 +479,13 @@ static int amdgpu_xgmi_sysfs_add_dev_info(struct amdgpu_device *adev,
        if (ret)
                pr_err("failed to create xgmi_num_links\n");

+       /* Create xgmi port num file if supported */
+       if (adev->psp.xgmi_context.xgmi_ta_caps & EXTEND_PEER_LINK_INFO_CMD_FLAG) {
+               ret = device_create_file(adev->dev, &dev_attr_xgmi_port_num);
+               if (ret)
+                       dev_err(adev->dev, "failed to create xgmi_port_num\n");
+       }
+
        /* Create sysfs link to hive info folder on the first device */
        if (hive->kobj.parent != (&adev->dev->kobj)) {
                ret = sysfs_create_link(&adev->dev->kobj, &hive->kobj, @@ -476,6 +516,8 @@ static int amdgpu_xgmi_sysfs_add_dev_info(struct amdgpu_device *adev,
        device_remove_file(adev->dev, &dev_attr_xgmi_error);
        device_remove_file(adev->dev, &dev_attr_xgmi_num_hops);
        device_remove_file(adev->dev, &dev_attr_xgmi_num_links);
+       if (adev->psp.xgmi_context.xgmi_ta_caps & EXTEND_PEER_LINK_INFO_CMD_FLAG)
+               device_remove_file(adev->dev, &dev_attr_xgmi_port_num);

 success:
        return ret;
@@ -492,6 +534,8 @@ static void amdgpu_xgmi_sysfs_rem_dev_info(struct amdgpu_device *adev,
        device_remove_file(adev->dev, &dev_attr_xgmi_error);
        device_remove_file(adev->dev, &dev_attr_xgmi_num_hops);
        device_remove_file(adev->dev, &dev_attr_xgmi_num_links);
+       if (adev->psp.xgmi_context.xgmi_ta_caps & EXTEND_PEER_LINK_INFO_CMD_FLAG)
+               device_remove_file(adev->dev, &dev_attr_xgmi_port_num);

        if (hive->kobj.parent != (&adev->dev->kobj))
                sysfs_remove_link(&adev->dev->kobj,"xgmi_hive_info");
--
2.17.1