From f419b81bdfcde0bc30e7bf8084efefa65a7c8960 Mon Sep 17 00:00:00 2001
From: Joseph Greathouse <Joseph.Greathouse@amd.com>
Date: Fri, 29 Oct 2021 16:25:10 -0500
Subject: [PATCH] Add capability to pull gfx architecture from KFD topology

New versions of amdkfd include the gfx architecture version number
for all GPUs surfaced in the HSA topology. This patch adds this as
the preferred way for rocm_agent_enumerator to check for supported
gfx architecture numbers.

Kernels that are missing this feature will not have the value in
the topology. rocm_agent_enumerator will fall back to checking
against the PCI IDs in this case. If PCI IDs fail, we fall back
to the heavyweight rocminfo method.

Change-Id: I5cf22e1069114675092e97ae52331b829cfafb04
---
 rocm_agent_enumerator | 34 ++++++++++++++++++++++++++++++++--
 1 file changed, 32 insertions(+), 2 deletions(-)

diff --git a/rocm_agent_enumerator b/rocm_agent_enumerator
index 3698f8c2b6..4ed6a3c17a 100755
--- a/rocm_agent_enumerator
+++ b/rocm_agent_enumerator
@@ -169,6 +169,31 @@ def readFromLSPCI():
 
   return target_list
 
+def readFromKFD():
+  target_list = []
+
+  topology_dir = '/sys/class/kfd/kfd/topology/nodes/'
+  for node in sorted(os.listdir(topology_dir)):
+    node_path = os.path.join(topology_dir, node)
+    if os.path.isdir(node_path):
+      prop_path = node_path + '/properties'
+      if os.path.isfile(prop_path):
+        target_search_term = re.compile("gfx_target_version.+")
+        with open(prop_path) as f:
+          line = f.readline()
+          while line != '' :
+            search_result = target_search_term.search(line)
+            if search_result is not None:
+              device_id = int(search_result.group(0).split(' ')[1], 10)
+              if device_id != 0:
+                major_ver = int((device_id / 10000) % 100)
+                minor_ver = int((device_id / 100) % 100)
+                stepping_ver = int(device_id % 100)
+                target_list.append("gfx" + format(major_ver, 'x') + format(minor_ver, 'x') + format(stepping_ver, 'x'))
+            line = f.readline()
+
+  return target_list
+
 def main():
   if len(sys.argv) == 2 and sys.argv[1] == '-name' :
     """ Prints the list of available AMD GCN target names extracted from rocminfo, a tool
@@ -186,13 +211,18 @@ def main():
                        execution of "rocminfo" is not possible.
     2. target.lst : user-supplied text file. This is used in a container setting
                   where ROCm stack may usually not available.
-    3. lspci : enumerate PCI bus and locate supported devices from a hard-coded
+    3. HSA topology : gathers the information from the HSA node topology in
+                      /sys/class/kfd/kfd/topology/nodes/
+    4. lspci : enumerate PCI bus and locate supported devices from a hard-coded
                lookup table.
-    4. rocminfo : a tool shipped with this script to enumerate GPU agents
+    5. rocminfo : a tool shipped with this script to enumerate GPU agents
                 available on a working ROCm stack.
     """
     target_list = readFromTargetLstFile()
 
+    if len(target_list) == 0:
+      target_list = readFromKFD()
+
     if len(target_list) == 0:
       target_list = readFromLSPCI()