From c3a0ec4f9a9f711a71d05752750fb27ce4e1207c Mon Sep 17 00:00:00 2001
From: "Liu, Shuzhou (Bill)" <Shuzhou.Liu@amd.com>
Date: Thu, 27 Mar 2025 16:36:43 -0400
Subject: [PATCH] [SWDEV-524147] Patch for handling new ras filenames (#205)

The code is changed to handle both original and ACA based ECC counters
for backward compatibilities.

Signed-off-by: Maisam Arif <Maisam.Arif@amd.com>
Co-authored-by: Maisam Arif <Maisam.Arif@amd.com>

[ROCm/amdsmi commit: 9b6e0432b2bd7e442a884500815048ab70f28c64]
---
 .../amdsmi/rocm_smi/src/rocm_smi_device.cc    | 38 ++++++++++++++++++-
 projects/amdsmi/src/amd_smi/amd_smi_utils.cc  |  7 +++-
 2 files changed, 42 insertions(+), 3 deletions(-)

diff --git a/projects/amdsmi/rocm_smi/src/rocm_smi_device.cc b/projects/amdsmi/rocm_smi/src/rocm_smi_device.cc
index 80f89f08e3..ab81a76659 100644
--- a/projects/amdsmi/rocm_smi/src/rocm_smi_device.cc
+++ b/projects/amdsmi/rocm_smi/src/rocm_smi_device.cc
@@ -736,6 +736,17 @@ std::string Device::get_sys_file_path_by_type(DevInfoTypes type) const {
   return sysfs_path;
 }
 
+// The fallback sysfs to handle backward compatibilities
+static const std::map<DevInfoTypes, std::string> kDevFallbackFile = {
+  {kDevErrCntGFX, "ras/aca_gfx"},
+  {kDevErrCntSDMA, "ras/aca_sdma"},
+  {kDevErrCntUMC, "ras/aca_umc"},
+  {kDevErrCntMMHUB, "ras/aca_mmhub"},
+  {kDevErrCntPCIEBIF, "ras/aca_pcie_bif"},
+  {kDevErrCntHDP, "ras/aca_hdp"},
+  {kDevErrCntXGMIWAFL, "ras/aca_xgmi_wafl"},
+};
+
 template <typename T>
 int Device::openSysfsFileStream(DevInfoTypes type, T *fs, const char *str) {
   auto sysfs_path = path_;
@@ -769,6 +780,28 @@ int Device::openSysfsFileStream(DevInfoTypes type, T *fs, const char *str) {
   bool reg_file;
 
   int ret = isRegularFile(sysfs_path, &reg_file);
+
+  if (ret != 0 || !reg_file) {
+    // Handle specific types if the file does not exist
+    if (kDevFallbackFile.find(type) != kDevFallbackFile.end()) {
+
+      sysfs_path = path_ + "/device/" + kDevFallbackFile.at(type);
+      DBG_FILE_ERROR(sysfs_path, str);
+
+      // Recheck the adjusted path
+      ret = isRegularFile(sysfs_path, &reg_file);
+      if (ret != 0 || !reg_file) {
+        ss << __PRETTY_FUNCTION__
+           << " | Adjusted file path also does not exist - SYSFS file ("
+           << sysfs_path 
+           << ") for DevInfoInfoType (" << get_type_string(type)
+           << "), returning " << std::to_string(ret);
+        LOG_ERROR(ss);
+        return ret;
+      }
+    }
+  }
+
   if (ret != 0) {
     ss << __PRETTY_FUNCTION__ << " | Issue: File did not exist - SYSFS file ("
        << sysfs_path
@@ -777,6 +810,7 @@ int Device::openSysfsFileStream(DevInfoTypes type, T *fs, const char *str) {
     LOG_ERROR(ss);
     return ret;
   }
+
   if (!reg_file) {
     ss << __PRETTY_FUNCTION__
        << " | Issue: File is not a regular file - SYSFS file ("
@@ -792,7 +826,7 @@ int Device::openSysfsFileStream(DevInfoTypes type, T *fs, const char *str) {
   if (!fs->is_open()) {
     ss << __PRETTY_FUNCTION__
        << " | Issue: Could not open - SYSFS file (" << sysfs_path << ") for "
-       << "DevInfoInfoType (" << get_type_string(type) << "), "
+       << "DevInfoTypes (" << get_type_string(type) << "), "
        << ", returning " << std::to_string(errno) << " ("
        << std::strerror(errno) << ")";
     LOG_ERROR(ss);
@@ -801,7 +835,7 @@ int Device::openSysfsFileStream(DevInfoTypes type, T *fs, const char *str) {
 
   ss << __PRETTY_FUNCTION__ << " | Successfully opened SYSFS file ("
      << sysfs_path
-     << ") for DevInfoInfoType (" << get_type_string(type)
+     << ") for DevInfoTypes (" << get_type_string(type)
      << ")";
   LOG_INFO(ss);
   return 0;
diff --git a/projects/amdsmi/src/amd_smi/amd_smi_utils.cc b/projects/amdsmi/src/amd_smi/amd_smi_utils.cc
index fa9f809d95..c1f42ebf8d 100644
--- a/projects/amdsmi/src/amd_smi/amd_smi_utils.cc
+++ b/projects/amdsmi/src/amd_smi/amd_smi_utils.cc
@@ -503,7 +503,12 @@ amdsmi_status_t smi_amdgpu_get_ecc_error_count(amd::smi::AMDSmiGPUDevice* device
     std::ifstream f(fullpath.c_str());
 
     if (f.fail()) {
-        return AMDSMI_STATUS_NOT_SUPPORTED;
+        //fall back to aca file
+        fullpath = "/sys/class/drm/" + device->get_gpu_path() + std::string("/device/ras/aca_umc");
+        f.open(fullpath.c_str());
+        if (f.fail()) {
+            return AMDSMI_STATUS_NOT_SUPPORTED;
+        }
     }
 
     std::string line;