Merge pull request #4302 from ollama/mxyng/forward-env

only forward some env vars

llm/server.go

@@ -292,31 +292,25 @@ func NewLlamaServer(gpus gpu.GpuInfoList, model string, ggml *GGML, adapters, pr
 			done:           make(chan error, 1),
 		}
 
-		s.cmd.Env = os.Environ()
 		s.cmd.Stdout = os.Stdout
 		s.cmd.Stderr = s.status
 
-		visibleDevicesEnv, visibleDevicesEnvVal := gpu.GpuInfoList(gpus).GetVisibleDevicesEnv()
+		if v := strings.Join(libraryPaths, string(filepath.ListSeparator)); v != "" {
-		pathEnvVal := strings.Join(libraryPaths, string(filepath.ListSeparator))
+			s.cmd.Env = append(s.cmd.Env, pathEnv+"="+v)
+		}
 
-		// Update or add the path and visible devices variable with our adjusted version
+		if k, v := gpu.GpuInfoList(gpus).GetVisibleDevicesEnv(); k != "" {
-		pathNeeded := true
+			s.cmd.Env = append(s.cmd.Env, k+"="+v)
-		devicesNeeded := visibleDevicesEnv != ""
-		for i := range s.cmd.Env {
-			cmp := strings.SplitN(s.cmd.Env[i], "=", 2)
-			if strings.EqualFold(cmp[0], pathEnv) {
-				s.cmd.Env[i] = pathEnv + "=" + pathEnvVal
-				pathNeeded = false
-			} else if devicesNeeded && strings.EqualFold(cmp[0], visibleDevicesEnv) {
-				s.cmd.Env[i] = visibleDevicesEnv + "=" + visibleDevicesEnvVal
-				devicesNeeded = false
 		}
+
+		for _, ev := range os.Environ() {
+			if strings.HasPrefix(ev, "CUDA_") ||
+				strings.HasPrefix(ev, "ROCM_") ||
+				strings.HasPrefix(ev, "HIP_") ||
+				strings.HasPrefix(ev, "HSA_") ||
+				strings.HasPrefix(ev, "GGML_") {
+				s.cmd.Env = append(s.cmd.Env, ev)
 			}
-		if pathNeeded {
-			s.cmd.Env = append(s.cmd.Env, pathEnv+"="+pathEnvVal)
-		}
-		if devicesNeeded {
-			s.cmd.Env = append(s.cmd.Env, visibleDevicesEnv+"="+visibleDevicesEnvVal)
 		}
 
 		slog.Info("starting llama server", "cmd", s.cmd.String())