refactor: modularize audit and harden build validation
This commit is contained in:
@@ -196,11 +196,11 @@ func (a *App) RunNvidiaPulseTestPack(ctx context.Context, baseDir string, durati
|
||||
return a.sat.RunNvidiaPulseTestPack(ctx, baseDir, durationSec, gpuIndices, logFunc)
|
||||
}
|
||||
|
||||
func (a *App) RunNvidiaBandwidthPack(ctx context.Context, baseDir string, gpuIndices []int, logFunc func(string)) (string, error) {
|
||||
func (a *App) RunNvidiaBandwidthPack(ctx context.Context, baseDir string, gpuIndices []int, fullMatrix bool, logFunc func(string)) (string, error) {
|
||||
if strings.TrimSpace(baseDir) == "" {
|
||||
baseDir = DefaultSATBaseDir
|
||||
}
|
||||
return a.sat.RunNvidiaBandwidthPack(ctx, baseDir, gpuIndices, logFunc)
|
||||
return a.sat.RunNvidiaBandwidthPack(ctx, baseDir, gpuIndices, fullMatrix, logFunc)
|
||||
}
|
||||
|
||||
func (a *App) RunNvidiaStressPackCtx(ctx context.Context, baseDir string, opts platform.NvidiaStressOptions, logFunc func(string)) (string, error) {
|
||||
@@ -237,11 +237,6 @@ func (a *App) RunCPUAcceptancePackCtx(ctx context.Context, baseDir string, durat
|
||||
return a.sat.RunCPUAcceptancePack(ctx, baseDir, durationSec, logFunc)
|
||||
}
|
||||
|
||||
func (a *App) RunCPUAcceptancePackResult(baseDir string, durationSec int) (ActionResult, error) {
|
||||
path, err := a.RunCPUAcceptancePack(baseDir, durationSec, nil)
|
||||
return ActionResult{Title: "CPU SAT", Body: satResultBody(path)}, err
|
||||
}
|
||||
|
||||
func (a *App) RunStorageAcceptancePack(baseDir string, logFunc func(string)) (string, error) {
|
||||
return a.RunStorageAcceptancePackCtx(context.Background(), baseDir, false, logFunc)
|
||||
}
|
||||
@@ -276,11 +271,6 @@ func (a *App) RunNvidiaConfigCheckPack(baseDir string, logFunc func(string)) (st
|
||||
return a.RunNvidiaConfigCheckPackCtx(context.Background(), baseDir, logFunc)
|
||||
}
|
||||
|
||||
func (a *App) RunNvidiaConfigCheckPackResult(baseDir string) (ActionResult, error) {
|
||||
path, err := a.RunNvidiaConfigCheckPack(baseDir, nil)
|
||||
return ActionResult{Title: "GPU Config & NVLink Check", Body: satResultBody(path)}, err
|
||||
}
|
||||
|
||||
func (a *App) RunPCIeLinkCheckPackCtx(ctx context.Context, baseDir string, logFunc func(string)) (string, error) {
|
||||
if strings.TrimSpace(baseDir) == "" {
|
||||
baseDir = DefaultSATBaseDir
|
||||
@@ -292,11 +282,6 @@ func (a *App) RunPCIeLinkCheckPack(baseDir string, logFunc func(string)) (string
|
||||
return a.RunPCIeLinkCheckPackCtx(context.Background(), baseDir, logFunc)
|
||||
}
|
||||
|
||||
func (a *App) RunPCIeLinkCheckPackResult(baseDir string) (ActionResult, error) {
|
||||
path, err := a.RunPCIeLinkCheckPack(baseDir, nil)
|
||||
return ActionResult{Title: "PCIe Link Check", Body: satResultBody(path)}, err
|
||||
}
|
||||
|
||||
func (a *App) RunNvidiaPCIeBandwidthPackCtx(ctx context.Context, baseDir string, gpuIndices []int, logFunc func(string)) (string, error) {
|
||||
if strings.TrimSpace(baseDir) == "" {
|
||||
baseDir = DefaultSATBaseDir
|
||||
@@ -308,11 +293,6 @@ func (a *App) RunNvidiaPCIeBandwidthPack(baseDir string, gpuIndices []int, logFu
|
||||
return a.RunNvidiaPCIeBandwidthPackCtx(context.Background(), baseDir, gpuIndices, logFunc)
|
||||
}
|
||||
|
||||
func (a *App) RunNvidiaPCIeBandwidthPackResult(baseDir string) (ActionResult, error) {
|
||||
path, err := a.RunNvidiaPCIeBandwidthPack(baseDir, nil, nil)
|
||||
return ActionResult{Title: "NVIDIA GPU PCIe Bandwidth Check", Body: satResultBody(path)}, err
|
||||
}
|
||||
|
||||
func (a *App) DetectGPUVendor() string {
|
||||
return a.sat.DetectGPUVendor()
|
||||
}
|
||||
@@ -332,11 +312,6 @@ func (a *App) RunAMDAcceptancePackCtx(ctx context.Context, baseDir string, logFu
|
||||
return a.sat.RunAMDAcceptancePack(ctx, baseDir, logFunc)
|
||||
}
|
||||
|
||||
func (a *App) RunAMDAcceptancePackResult(baseDir string) (ActionResult, error) {
|
||||
path, err := a.RunAMDAcceptancePack(baseDir, nil)
|
||||
return ActionResult{Title: "AMD GPU SAT", Body: satResultBody(path)}, err
|
||||
}
|
||||
|
||||
func (a *App) RunAMDMemIntegrityPackCtx(ctx context.Context, baseDir string, logFunc func(string)) (string, error) {
|
||||
if strings.TrimSpace(baseDir) == "" {
|
||||
baseDir = DefaultSATBaseDir
|
||||
@@ -399,69 +374,6 @@ func (a *App) RunNCCLTests(ctx context.Context, baseDir string, gpuIndices []int
|
||||
return a.sat.RunNCCLTests(ctx, baseDir, gpuIndices, logFunc)
|
||||
}
|
||||
|
||||
func (a *App) RunNCCLTestsResult(ctx context.Context) (ActionResult, error) {
|
||||
path, err := a.RunNCCLTests(ctx, DefaultSATBaseDir, nil, nil)
|
||||
body := "Results: " + path
|
||||
if err != nil && err != context.Canceled {
|
||||
body += "\nERROR: " + err.Error()
|
||||
}
|
||||
return ActionResult{Title: "NCCL bandwidth test", Body: body}, err
|
||||
}
|
||||
|
||||
func (a *App) RunFanStressTestResult(ctx context.Context, opts platform.FanStressOptions) (ActionResult, error) {
|
||||
path, err := a.RunFanStressTest(ctx, "", opts)
|
||||
body := formatFanStressResult(path)
|
||||
if err != nil && err != context.Canceled {
|
||||
body += "\nERROR: " + err.Error()
|
||||
}
|
||||
return ActionResult{Title: "GPU Platform Stress Test", Body: body}, err
|
||||
}
|
||||
|
||||
// formatFanStressResult formats the summary.txt from a fan-stress run, including
|
||||
// the per-step pass/fail display and the analysis section (throttling, max temps, fan response).
|
||||
func formatFanStressResult(archivePath string) string {
|
||||
if archivePath == "" {
|
||||
return "No output produced."
|
||||
}
|
||||
runDir := strings.TrimSuffix(archivePath, ".tar.gz")
|
||||
raw, err := os.ReadFile(filepath.Join(runDir, "summary.txt"))
|
||||
if err != nil {
|
||||
return "Archive written to " + archivePath
|
||||
}
|
||||
content := strings.TrimSpace(string(raw))
|
||||
kv := parseKeyValueSummary(content)
|
||||
|
||||
var b strings.Builder
|
||||
b.WriteString(formatSATDetail(content))
|
||||
|
||||
// Append analysis section.
|
||||
var analysis []string
|
||||
if v, ok := kv["throttling_detected"]; ok {
|
||||
label := "NO"
|
||||
if v == "true" {
|
||||
label = "YES ← throttling detected during load"
|
||||
}
|
||||
analysis = append(analysis, "Throttling: "+label)
|
||||
}
|
||||
if v, ok := kv["max_gpu_temp_c"]; ok && v != "0.0" {
|
||||
analysis = append(analysis, "Max GPU temp: "+v+"°C")
|
||||
}
|
||||
if v, ok := kv["max_cpu_temp_c"]; ok && v != "0.0" {
|
||||
analysis = append(analysis, "Max CPU temp: "+v+"°C")
|
||||
}
|
||||
if v, ok := kv["fan_response_sec"]; ok && v != "N/A" && v != "-1.0" {
|
||||
analysis = append(analysis, "Fan response: "+v+"s")
|
||||
}
|
||||
|
||||
if len(analysis) > 0 {
|
||||
b.WriteString("\n\n=== Analysis ===\n")
|
||||
for _, line := range analysis {
|
||||
b.WriteString(line + "\n")
|
||||
}
|
||||
}
|
||||
return strings.TrimSpace(b.String())
|
||||
}
|
||||
|
||||
// satResultBody reads summary.txt from the SAT run directory (archive path without .tar.gz)
|
||||
// and returns a formatted human-readable result. Falls back to a plain message if unreadable.
|
||||
func satResultBody(archivePath string) string {
|
||||
|
||||
Reference in New Issue
Block a user