fix: resolve SHM hang, DRM device mismatch, and duplicate VAAPI context

BUG-2 (HIGH): SHM Buffer event caused permanent hang
  In the ZwlrScreencopyFrameV1 dispatcher, receiving a SHM Buffer event
  left in_flight_surface stuck at AllocQueued forever, preventing
  queue_alloc_frame() from requesting new frames.
  Fix: treat Buffer as a metadata offer (v3 protocol), wait for
  BufferDone to decide failure, and add AllocQueued state guard to
  LinuxDmabuf handler.

BUG-3 (MEDIUM): Portal backend picked wrong GPU on multi-GPU systems
  state_portal.rs hardcoded /dev/dri/renderD128 then renderD129, which
  selects the wrong GPU when PipeWire uses a different device.
  Fix: extract find_drm_render_nodes() as shared utility; defer DRM
  device selection to first PipeWire frame; test each candidate with
  av_hwframe_transfer_data to find the GPU that can actually import
  the DMA-BUF frame.

BUG-4 (LOW): VAAPI device context created twice unnecessarily
  try_finalize_output() created an AvHwDevCtx stored in EverythingButFmt,
  but negotiate_format() discarded it (_hw_device_ctx) and EncState::new
  created a new one.
  Fix: thread the existing hw_device_ctx through negotiate_format() and
  create_encoder() to EncState::new() which reuses it when provided.

src/state_portal.rs

@@ -1,5 +1,6 @@
 // 采集门户状态模块 —— 通过 PipeWire/DMA-BUF 进行屏幕采集并编码
 use std::mem;
+use std::os::fd::AsRawFd;
 use std::path::PathBuf;
 
 use anyhow::{bail, Result};
@@ -39,8 +40,8 @@ pub struct StatePortal {
     errored: bool,
     /// 是否为第一帧（首帧跳过帧率限制）
     first_frame: bool,
-    /// DRM 渲染设备路径（如 /dev/dri/renderD128）
-    drm_device: PathBuf,
+    /// DRM 渲染设备路径（如 /dev/dri/renderD128）；None 表示首帧自动检测
+    drm_device: Option<PathBuf>,
     /// 第一帧的时间戳（纳秒），用于计算相对 PTS
     first_pts_ns: Option<i64>,
 }
@@ -51,7 +52,11 @@ impl StatePortal {
     /// 初始化 DRM 设备路径和 PipeWire 采集端点，编码器延迟到第一帧到达时创建。
     pub fn new(args: Args) -> Result<Self> {
         let drm_device = resolve_drm_device(&args)?;
-        tracing::info!("Using DRM device: {}", drm_device.display());
+        if let Some(ref drm_device) = drm_device {
+            tracing::info!("Using DRM device: {}", drm_device.display());
+        } else {
+            tracing::info!("DRM device auto-detection enabled");
+        }
 
         let cap = CapPortal::new(&args)?;
 
@@ -92,8 +97,9 @@ impl StatePortal {
                             frame.modifier
                         );
 
+                        let drm_path = self.resolve_drm_device_for_frame(&frame)?;
                         let enc = avhw::create_encoder(
-                            &self.drm_device,
+                            &drm_path,
                             self.args.output.as_ref(),
                             frame.width,
                             frame.height,
@@ -101,6 +107,7 @@ impl StatePortal {
                             Transform::Normal,
                             self.args.bitrate,
                             self.args.gop_size,
+                            None,
                         )?;
 
                         self.enc = Some(enc);
@@ -128,6 +135,44 @@ impl StatePortal {
         Ok(true)
     }
 
+    fn resolve_drm_device_for_frame(&mut self, frame: &PwDmaBufFrame) -> Result<PathBuf> {
+        if let Some(ref drm) = self.drm_device {
+            return Ok(drm.clone());
+        }
+
+        let candidates = crate::state::find_drm_render_nodes();
+        if candidates.is_empty() {
+            bail!("No DRM render device found. Specify --drm-device.");
+        }
+
+        let mut failures = Vec::new();
+        for candidate in &candidates {
+            match crate::avhw::test_dma_buf_import(candidate, frame) {
+                Ok(()) => {
+                    tracing::info!(
+                        "Auto-selected DRM device: {} (can import PipeWire DMA-BUF)",
+                        candidate.display()
+                    );
+                    self.drm_device = Some(candidate.clone());
+                    return Ok(candidate.clone());
+                }
+                Err(err) => {
+                    tracing::debug!(
+                        "DRM device {} cannot import frame: {err:#}",
+                        candidate.display()
+                    );
+                    failures.push(format!("{}: {err:#}", candidate.display()));
+                }
+            }
+        }
+
+        bail!(
+            "No DRM render device can import the PipeWire DMA-BUF frame. \
+             Specify --drm-device. Tried: {}",
+            failures.join("; ")
+        )
+    }
+
     /// 处理单帧 DMA-BUF 数据
     ///
     /// 完整的帧处理流水线：
@@ -311,25 +356,14 @@ fn build_drm_descriptor(frame: &PwDmaBufFrame) -> ffi::AVDRMFrameDescriptor {
     desc
 }
 
-use std::os::fd::AsRawFd;
-
 /// 解析 DRM 渲染设备路径
 ///
-/// 优先使用命令行指定的设备路径，否则依次尝试
-/// `/dev/dri/renderD128` 和 `/dev/dri/renderD129`。
-fn resolve_drm_device(args: &Args) -> Result<PathBuf> {
+/// 仅使用命令行指定的设备路径；未指定则在首帧到达时自动检测。
+fn resolve_drm_device(args: &Args) -> Result<Option<PathBuf>> {
     if let Some(ref drm) = args.drm_device {
-        return Ok(PathBuf::from(drm));
+        return Ok(Some(PathBuf::from(drm)));
     }
-
-    for render in &["/dev/dri/renderD128", "/dev/dri/renderD129"] {
-        let path = PathBuf::from(render);
-        if path.exists() {
-            return Ok(path);
-        }
-    }
-
-    bail!("No DRM render device found. Specify --drm-device.")
+    Ok(None)
 }
 
 #[cfg(test)]