reanimator/bee
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 1 Wiki Activity

refactor: modularize audit and harden build validation

Browse Source
This commit is contained in:
Mikhail Chusavitin
2026-08-31 21:22:16 +03:00
parent bb22ccfafe
commit ac4bc0b2b7
78 changed files with 13598 additions and 13130 deletions
Download Patch File Download Diff File Expand all files Collapse all files
audit/internal/app/app.go
+28 -36
View File
@@ -6,7 +6,6 @@ import (
"fmt"
"log/slog"
"os"
"strconv"
"strings"
"bee/audit/internal/collector"
@@ -96,14 +95,38 @@ type installer interface {
type GPUPresenceResult struct {
Nvidia bool
AMD bool
// NvidiaInitializing / AMDInitializing report a PCI device of that vendor
// when DetectGPUVendor did not report the vendor as operational.
NvidiaInitializing bool
AMDInitializing bool
}
// DetectGPUPresence combines the existing operational vendor detection with a
// PCI display-class scan. The latter distinguishes absent hardware from a
// detected PCI function whose runtime is not operational yet.
func (a *App) DetectGPUPresence() GPUPresenceResult {
vendor := a.sat.DetectGPUVendor()
return GPUPresenceResult{
res := GPUPresenceResult{
Nvidia: vendor == "nvidia",
AMD: vendor == "amd",
}
physNvidia, physAMD := a.sat.PhysicalGPUVendors()
res.NvidiaInitializing = physNvidia && !res.Nvidia
res.AMDInitializing = physAMD && !res.AMD
return res
}
// RuntimeHealthNow collects a fresh runtime-health snapshot (driver / CUDA
// readiness, GSP state). Unlike ReadRuntimeHealth it does not read the
// boot-time JSON; callers that poll for the GPU stack to come up need
// current data.
func (a *App) RuntimeHealthNow() (schema.RuntimeHealth, error) {
return a.runtime.CollectRuntimeHealth(DefaultExportDir)
}
// TPMPresent reports whether this host has a TPM the checks can talk to.
func (a *App) TPMPresent() bool {
return a.sat.TPMPresent()
}
func (a *App) IsLiveMediaInRAM() bool {
@@ -132,13 +155,14 @@ type satRunner interface {
RunNvidiaOfficialComputePack(ctx context.Context, baseDir string, durationSec int, gpuIndices []int, staggerSec int, logFunc func(string)) (string, error)
RunNvidiaTargetedPowerPack(ctx context.Context, baseDir string, durationSec int, gpuIndices []int, logFunc func(string)) (string, error)
RunNvidiaPulseTestPack(ctx context.Context, baseDir string, durationSec int, gpuIndices []int, logFunc func(string)) (string, error)
RunNvidiaBandwidthPack(ctx context.Context, baseDir string, gpuIndices []int, logFunc func(string)) (string, error)
RunNvidiaBandwidthPack(ctx context.Context, baseDir string, gpuIndices []int, fullMatrix bool, logFunc func(string)) (string, error)
RunNvidiaStressPack(ctx context.Context, baseDir string, opts platform.NvidiaStressOptions, logFunc func(string)) (string, error)
ListNvidiaGPUStatuses() ([]platform.NvidiaGPUStatus, error)
ResetNvidiaGPU(index int) (string, error)
RunMemoryAcceptancePack(ctx context.Context, baseDir string, sizeMB, passes int, logFunc func(string)) (string, error)
RunStorageAcceptancePack(ctx context.Context, baseDir string, extended bool, logFunc func(string)) (string, error)
RunTPMValidationPack(ctx context.Context, baseDir string, logFunc func(string)) (string, error)
TPMPresent() bool
RunNvidiaConfigCheckPack(ctx context.Context, baseDir string, logFunc func(string)) (string, error)
RunPCIeLinkCheckPack(ctx context.Context, baseDir string, logFunc func(string)) (string, error)
RunNvidiaPCIeBandwidthPack(ctx context.Context, baseDir string, gpuIndices []int, logFunc func(string)) (string, error)
@@ -151,6 +175,7 @@ type satRunner interface {
SetNvidiaGPUPowerLimit(index int, watts float64) (string, error)
ResetNvidiaGPUDefaults() (string, error)
DetectGPUVendor() string
PhysicalGPUVendors() (nvidia bool, amd bool)
ListAMDGPUs() ([]platform.AMDGPUInfo, error)
RunAMDAcceptancePack(ctx context.Context, baseDir string, logFunc func(string)) (string, error)
RunAMDMemIntegrityPack(ctx context.Context, baseDir string, logFunc func(string)) (string, error)
@@ -279,15 +304,6 @@ func (a *App) RunRuntimePreflight(output string) (string, error) {
}
}
func (a *App) RunRuntimePreflightResult() (ActionResult, error) {
path, err := a.RunRuntimePreflight("file:" + DefaultRuntimeJSONPath)
body := "Runtime preflight completed."
if path != "" {
body = "Runtime health written to " + path
}
return ActionResult{Title: "Run self-check", Body: body}, err
}
func (a *App) RuntimeHealthResult() ActionResult {
health, err := ReadRuntimeHealth(DefaultRuntimeJSONPath)
if err != nil {
@@ -326,10 +342,6 @@ func (a *App) RunAuditNow(runtimeMode runtimeenv.Mode) (ActionResult, error) {
return ActionResult{Title: "Run audit", Body: body}, err
}
func (a *App) RunAuditToDefaultFile(runtimeMode runtimeenv.Mode) (string, error) {
return a.RunAudit(runtimeMode, "file:"+DefaultAuditJSONPath)
}
func (a *App) HealthSummaryResult() ActionResult {
raw, err := os.ReadFile(DefaultAuditJSONPath)
if err != nil {
@@ -399,26 +411,6 @@ func (a *App) MainBanner() string {
return strings.TrimSpace(strings.Join(lines, "\n"))
}
func (a *App) FormatToolStatuses(statuses []platform.ToolStatus) string {
var body strings.Builder
for _, tool := range statuses {
status := "MISSING"
if tool.OK {
status = "OK (" + tool.Path + ")"
}
fmt.Fprintf(&body, "- %s: %s\n", tool.Name, status)
}
return strings.TrimSpace(body.String())
}
func (a *App) ParsePrefix(raw string, fallback int) int {
value, err := strconv.Atoi(strings.TrimSpace(raw))
if err != nil || value <= 0 {
return fallback
}
return value
}
// writePSUStatusesToDB records PSU statuses collected during audit into the
// component-status DB so they are visible in the Hardware Summary card.
// PSU status is sourced from IPMI (ipmitool fru + sdr) during audit.