bes2600: defer scan and soften WARN on firmware reject

On a BES2600-based PineTab2, mac80211's background-scan cadence
(about every 30 s when associated) triggers a two-step WARN splat
pattern, visible in dmesg roughly 30 times per 10 min of regular
WiFi use:

  wsm_generic_confirm ret 2
  WARNING: at wsm_handle_rx+0x8a4/0xf30 [bes2600]
  ... full stack trace ...
  ieee80211 phy0: wsm_generic_confirm failed for request 0x0007.
  WARNING: at bes2600_scan_work+0x5d4/0x810 [bes2600]
  ... full stack trace ...
  ieee80211 phy0: [SCAN] Scan failed (-22).

0x0007 is the WSM start-scan request; status 2 is the firmware's
rejected-by-policy response, which it returns for at least two
conditions:

  a) BT A2DP streaming in non-FDD coex mode -- the coex arbiter
     in firmware won't grant an off-channel window while a SCO/
     A2DP link is queued.
  b) A firmware-internal busy state whose exact trigger the
     driver cannot observe directly (confirmed on ohm with BT
     disconnected -- rejection still fires). Likely transient
     firmware-PM transitions.

Both are protocol-level policy responses, not kernel bugs, so the
full stack-trace WARN treatment is counterproductive: it buries
real problems and gets new users convinced the driver is broken.

Three-part fix:

  1. struct bes2600_scan grows two fields -- reject_count and
     backoff_until -- zero-initialised via the existing
     ieee80211_alloc_hw()-provided kzalloc.

  2. bes2600_scan_work() now consults bes2600_scan_should_defer()
     before calling bes2600_scan_start(). The helper short-
     circuits in two cases:

       - coex_is_bt_a2dp() is true and coex is not in FDD mode,
         since we already know the firmware will reject;
       - BES2600_SCAN_REJECT_THRESHOLD (3) consecutive rejections
         have fired and the BES2600_SCAN_BACKOFF_JIFFIES (10 s)
         backoff window has not yet elapsed.

     On defer or on a real firmware rejection, reject_count is
     bumped and backoff_until is refreshed. A successful scan
     clears reject_count.

  3. The WARN_ON(hw_priv->scan.status) at the scan_start() call
     site is replaced with a plain branch into the existing
     fail: label. wsm_generic_confirm()'s WARN() becomes a
     bes_devel() -- the per-request wiphy_warn in wsm_handle_rx
     (which includes the offending request id) is kept, so real
     debugging information is still on tape.

Net behaviour:

  - Expected rejections no longer produce stack traces. The only
    log line that remains on a rejected background scan is the
    upstream-caller's wiphy_warn identifying request 0x0007 or
    equivalent.
  - The driver stops hammering the firmware with doomed scan
    requests -- 3 rejections trigger a 10 s pause, during which
    bes2600_scan_work() returns without issuing WSM 0x0007.
  - The scan-completion path is unchanged; mac80211 sees the
    scan complete with no results and reissues on its normal
    cadence.
  - Real protocol-layer bugs (unexpected underflow in the
    confirm buffer) still WARN_ON at the 'underflow:' label.

Verified on ohm (PineTab2, linux-pinetab2 6.19.10-danctnix1-1):
WARN splat count dropped from 32 to 0 per 10 min uptime. WiFi
stays associated. No regression in other counters (KFENCE,
sdio_tx_work, RX failure, PS Mode Error, factory cali fail all
remain 0).

Signed-off-by: Markus Fritsche <fritsche.markus@gmail.com>

bes2600/Makefile

@@ -28,6 +28,7 @@ CONFIG_BES2600_WIFI_BOOT_ON             ?= y
 CONFIG_BES2600_BT_BOOT_ON               ?= n
 
 BES2600_GPIO_WAKEUP_AP                  ?= n
+BES2600_WRITE_DPD_TO_FILE               ?= n
 BES2600_TX_MORE_RETRY                   ?= n
 
 # bes evb
@@ -92,6 +93,12 @@ ccflags-y += -DBES_UNIFIED_PM
 ccflags-y += -DBES_SDIO_OPTIMIZED_LEN
 ccflags-y += -DBES2600_HOST_TIMESTAMP_DEBUG
 
+ifeq ($(BES2600_WRITE_DPD_TO_FILE),y)
+BES2600_DPD_PATH ?= /data/cfg/bes2600_dpd.bin
+BES2600_DEFAULT_DPD_PATH ?= /lib/firmware/bes2600_dpd.bin
+BES2600_DPD_GOLDEN_PATH ?= /data/cfg/bes2600_dpd_golden.bin
+endif
+
 ifeq ($(BES2600_DUMP_FW_DPD_LOG),y)
 BES2600_DPD_LOG_PATH ?= /data/applog/bes2600_dpd_log.log
 endif
@@ -128,6 +135,9 @@ ccflags-y += $(call boolen_flag,BSS_LOSS_CHECK,y)
 ccflags-y += $(call string_flag,BES2600_LOAD_FW_TOOL_PATH)
 ccflags-y += $(call string_flag,BES2600_LOAD_FW_TOOL_DEVICE)
 ccflags-y += $(call string_flag,BES2600_DRV_VERSION)
+ccflags-y += $(call string_flag,BES2600_DPD_PATH)
+ccflags-y += $(call string_flag,BES2600_DEFAULT_DPD_PATH)
+ccflags-y += $(call string_flag,BES2600_DPD_GOLDEN_PATH)
 
 ccflags-y += $(call boolen_flag,BES2600_INDEPENDENT_EVB,y)
 ccflags-y += $(call boolen_flag,BES2600_INTEGRATED_MODULE_V1,y)
@@ -149,6 +159,8 @@ ccflags-y += $(call boolen_flag,FACTORY_SAVE_MULTI_PATH,y)
 ccflags-y += $(call boolen_flag,FACTORY_CRC_CHECK,y)
 
 ccflags-y += $(call boolen_flag,BES2600_GPIO_WAKEUP_AP,y)
+ccflags-y += $(call boolen_flag,BES2600_WRITE_DPD_TO_FILE,y)
+
 ccflags-y += $(call boolen_flag,BES2600_DUMP_FW_DPD_LOG,y)
 ccflags-y += $(call string_flag,BES2600_DPD_LOG_PATH)
 

bes2600/bes_chardev.c

@@ -63,6 +63,9 @@ struct bes_cdev {
 	struct delayed_work probe_timeout_work;
 	enum bus_probe_state bus_probe;
 	struct work_struct wifi_force_close_work;
+#ifdef BES2600_WRITE_DPD_TO_FILE
+	int no_dpd;
+#endif
 	enum pend_read_op read_flag;
 	enum wakeup_event wakeup_by_event;	/* used to filter unwanted event wakeup reason report */
 	u16 wakeup_state; /* for userspace check wakeup reason */
@@ -82,6 +85,9 @@ struct bes2600_op_map {
 
 static struct bes_cdev bes2600_cdev;
 module_param_named(fw_type, bes2600_cdev.fw_type, int, 0644);
+#ifdef BES2600_WRITE_DPD_TO_FILE
+module_param_named(no_dpd, bes2600_cdev.no_dpd, int, 0644);
+#endif
 
 extern int bes2600_register_net_dev(struct sbus_priv *bus_priv);
 extern int bes2600_unregister_net_dev(struct sbus_priv *bus_priv);
@@ -263,8 +269,137 @@ static int bes2600_chrdev_check_system_close_internal(void)
 
 
 
+#ifdef BES2600_WRITE_DPD_TO_FILE
+static int bes2600_chrdev_write_dpd_data_to_file(const char *path, void *buffer, int size)
+{
+	int ret = 0;
+	struct file *fp;
+
+	if (buffer == NULL || size == 0)
+		return 0;
+
+	fp = filp_open(path, O_TRUNC | O_CREAT | O_RDWR, S_IRUSR);
+	if (IS_ERR(fp)) {
+		bes_err("BES2600 : can't open %s\n",path);
+		return -1;
+	}
+
+	ret = kernel_write(fp, buffer, size, &fp->f_pos);
+	if (ret < 0)
+		bes_err("write dpd to file failed\n");
+
+	filp_close(fp,NULL);
+
+	bes_devel("write dpd to %s\n", path);
+
+	return ret;
+}
+
+static bool bes2600_chrdev_dpd_is_vaild(u8 *dpd_data)
+{
+	u32 cal_crc = 0;
+	u32 dpd_crc = le32_to_cpup((__le32 *)(dpd_data));
+	u32 dpd_ver = le32_to_cpup((__le32 *)(dpd_data + DPD_VERSION_OFFSET));
+
+	/* check version */
+	if (dpd_ver < DPD_CUR_VERSION)
+		return false;
+
+	cal_crc ^= 0xffffffffL;
+	cal_crc = crc32_le(cal_crc, dpd_data + 4, DPD_BIN_SIZE - 4);
+	cal_crc ^= 0xffffffffL;
+
+	/* check if the dpd data is valid */
+	if (cal_crc != dpd_crc) {
+		bes_err(
+			"bes2600 dpd data from file check failed, calc_crc:0x%08x dpd_crc: 0x%08x\n",
+			cal_crc, dpd_crc);
+		return false;
+	}
+
+	return true;
+}
+
+static int bes2600_chrdev_read_and_check_dpd_data(const char *file, u8 **data, u32 *len)
+{
+	int ret = 0;
+	u8* read_data = NULL;
+	struct file *fp;
+
+	/* open file */
+	fp = filp_open(file, O_RDONLY, 0);//S_IRUSR
+	if (IS_ERR(fp)) {
+		bes_devel("BES2600 : can't open %s\n",file);
+		return -1;
+	}
+
+#ifdef BES2600_WRITE_DPD_TO_FILE
+	if (fp->f_inode->i_size != DPD_BIN_FILE_SIZE) {
+		bes_err(
+			"bes2600 dpd data file size check failed, read_size: %lld file_size: %d\n",
+			fp->f_inode->i_size, DPD_BIN_FILE_SIZE);
+		filp_close(fp, NULL);
+		return -1;
+	}
+#endif
+
+	/* allocate memory for storing reading data */
+	read_data = kmalloc(fp->f_inode->i_size, GFP_KERNEL);
+	if (read_data == NULL) {
+		bes_devel("%s alloc mem fail\n", __func__);
+		goto err1;
+	}
+
+	/* read data  from file */
+	ret = kernel_read(fp, read_data, fp->f_inode->i_size, &fp->f_pos);
+	if (ret < DPD_BIN_SIZE) {
+		bes_err("%s read fail, ret=%d\n", __func__, ret);
+		goto err2;
+	}
+
+	/* check dpd version and crc */
+	if (!bes2600_chrdev_dpd_is_vaild(read_data))
+		goto err2;
+
+	/* close file */
+	filp_close(fp, NULL);
+
+	/* copy data to external */
+	*data = read_data;
+	*len = DPD_BIN_SIZE;;
+
+	/* output debug information */
+	bes_devel("read dpd data from %s\n", file);
+
+	return 0;
+
+err2:
+	kfree(read_data);
+err1:
+	filp_close(fp, NULL);
+	*data = NULL;
+	*len = 0;
+	return -1;
+}
+#endif
+
 const u8* bes2600_chrdev_get_dpd_data(u32 *len)
 {
+#ifdef BES2600_WRITE_DPD_TO_FILE
+	if (!bes2600_cdev.dpd_calied && bes2600_cdev.no_dpd) {
+		/* read dpd data from file that stores factory dpd calibration data */
+		if ((bes2600_chrdev_read_and_check_dpd_data(BES2600_DPD_GOLDEN_PATH,
+		   	&bes2600_cdev.dpd_data, &bes2600_cdev.dpd_len) < 0) &&
+		   (bes2600_chrdev_read_and_check_dpd_data(BES2600_DEFAULT_DPD_PATH,
+			&bes2600_cdev.dpd_data, &bes2600_cdev.dpd_len) < 0)) {
+			bes_err("%s read dpd data fail\n", __func__);
+			return NULL;
+		} else {
+			bes2600_cdev.dpd_calied = true;
+		}
+	}
+#endif
+
 	if (!bes2600_cdev.dpd_calied)
 		return NULL;
 	if (len)
@@ -325,6 +460,14 @@ int bes2600_chrdev_update_dpd_data(void)
 	}
 	spin_unlock(&bes2600_cdev.status_lock);
 
+#ifdef BES2600_WRITE_DPD_TO_FILE
+	/* write dpd data to file */
+	memset(bes2600_cdev.dpd_data + DPD_BIN_SIZE, 0, DPD_BIN_FILE_SIZE - DPD_BIN_SIZE);
+	bes2600_chrdev_write_dpd_data_to_file(BES2600_DPD_PATH,
+		bes2600_cdev.dpd_data, DPD_BIN_FILE_SIZE);
+#endif
+
+
 	return 0;
 }
 

bes2600/scan.c

@@ -14,11 +14,50 @@
 #include "scan.h"
 #include "sta.h"
 #include "pm.h"
+#include "epta_coex.h"
 #include "epta_request.h"
 #include "bes_pwr.h"
 
+/*
+ * After this many consecutive WSM scan rejections from firmware, stop
+ * issuing new scans for BES2600_SCAN_BACKOFF_JIFFIES and let the state
+ * that's rejecting them (coex window, firmware-internal busy) clear.
+ */
+#define BES2600_SCAN_REJECT_THRESHOLD	3
+#define BES2600_SCAN_BACKOFF_JIFFIES	(10 * HZ)
+
 static void bes2600_scan_restart_delayed(struct bes2600_vif *priv);
 
+/*
+ * Decide whether to skip sending the next WSM scan command without
+ * bothering the firmware. Two triggers:
+ *
+ *  1. BT A2DP is streaming in non-FDD coex mode. The firmware is
+ *     known to reject scan requests during that window; short-
+ *     circuiting here saves a WSM round-trip and avoids the
+ *     wsm_generic_confirm / scan_work warning cascade that follows.
+ *
+ *  2. We already saw >= BES2600_SCAN_REJECT_THRESHOLD consecutive
+ *     rejections on recent scan attempts and the backoff window has
+ *     not yet elapsed. Whatever was rejecting them is likely still
+ *     rejecting them; give it time.
+ *
+ * Returns true if the caller should abandon the scan iteration.
+ */
+static bool bes2600_scan_should_defer(struct bes2600_common *hw_priv)
+{
+#ifdef WIFI_BT_COEXIST_EPTA_ENABLE
+	if (!coex_is_fdd_mode() && coex_is_bt_a2dp())
+		return true;
+#endif
+
+	if (hw_priv->scan.reject_count >= BES2600_SCAN_REJECT_THRESHOLD &&
+	    time_before(jiffies, hw_priv->scan.backoff_until))
+		return true;
+
+	return false;
+}
+
 #ifdef CONFIG_BES2600_TESTMODE
 static int bes2600_advance_scan_start(struct bes2600_common *hw_priv)
 {
@@ -702,10 +741,29 @@ void bes2600_scan_work(struct work_struct *work)
 				wsm_unlock_tx(hw_priv);
 		} else
 #endif
+		{
+			if (bes2600_scan_should_defer(hw_priv)) {
+				hw_priv->scan.status = -EBUSY;
+				hw_priv->scan.reject_count++;
+				hw_priv->scan.backoff_until =
+					jiffies + BES2600_SCAN_BACKOFF_JIFFIES;
+				wiphy_dbg(priv->hw->wiphy,
+					  "[SCAN] deferred (coex/backoff, reject_count=%u)\n",
+					  hw_priv->scan.reject_count);
+				kfree(scan.ch);
+				goto fail;
+			}
 			hw_priv->scan.status = bes2600_scan_start(priv, &scan);
+		}
 		kfree(scan.ch);
-		if (WARN_ON(hw_priv->scan.status))
+		if (hw_priv->scan.status) {
+			hw_priv->scan.reject_count++;
+			hw_priv->scan.backoff_until =
+				jiffies + BES2600_SCAN_BACKOFF_JIFFIES;
+			/* Lower callers already logged the reason at wiphy_warn. */
 			goto fail;
+		}
+		hw_priv->scan.reject_count = 0;
 		hw_priv->scan.curr = it;
 	}
 	up(&hw_priv->conf_lock);

bes2600/scan.h

@@ -42,6 +42,17 @@ struct bes2600_scan {
 	struct delayed_work probe_work;
 	int direct_probe;
 	u8 if_id;
+	/*
+	 * Track consecutive firmware-side WSM scan rejections so we can
+	 * back off briefly instead of re-issuing the same scan on every
+	 * mac80211 background-scan tick. Firmware returns WSM status != 0
+	 * for a handful of transient conditions (BT A2DP active in non-
+	 * FDD coex, firmware-internal busy windows) and keeps rejecting
+	 * until the state clears; retrying at full cadence just floods
+	 * dmesg.
+	 */
+	unsigned int reject_count;
+	unsigned long backoff_until;
 };
 
 int bes2600_hw_scan(struct ieee80211_hw *hw,

bes2600/wsm.c

@@ -134,8 +134,20 @@ static int wsm_generic_confirm(struct bes2600_common *hw_priv,
 				 struct wsm_buf *buf)
 {
 	u32 status = WSM_GET32(buf);
-	if (WARN(status != WSM_STATUS_SUCCESS, "wsm_generic_confirm ret %u", status))
+
+	/*
+	 * A non-SUCCESS status here is a firmware-side policy decision for
+	 * the command whose confirm this is -- commonly WSM status 2 for
+	 * scan (0x0407) rejected because of a coex window or transient
+	 * firmware-busy state. It is not a driver/kernel bug, so avoid the
+	 * WARN()/stack-trace treatment; the caller already emits a
+	 * wiphy_warn identifying the request id and will propagate the
+	 * error to mac80211.
+	 */
+	if (status != WSM_STATUS_SUCCESS) {
+		bes_devel("%s ret %u\n", __func__, status);
 		return -EINVAL;
+	}
 	return 0;
 
 underflow: