scsi: hisi_sas: fix the issue of losing directly attached disk when hot-plug

Hot-plugging mini-SAS wire of direct hard disk backplane may cause
disk lost. We have done this test with several types of SATA disk from
different venders, and only two models from Seagate has this problem,
ST4000NM0035-1V4107 and ST3000VM002-1ET166.

The root cause is that the disk doesn't send D2H frame after OOB
finished. SAS controller will issue phyup interrupt only when D2H frame
is received, otherwise, will be waiting there all the time.

When this issue happen, we can find the disk again with link reset.
To fix this issue, we setup an timer after OOB finished. If the PHY is
not up in 20s, do link reset. Notes: the 20s is an experience value.

Feature or Bugfix: Bugfix
Signed-off-by: Ntanxiaofei <tanxiaofei@huawei.com>
Reviewed-by: Nchenxiang <chenxiang66@hisilicon.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>

drivers/scsi/hisi_sas/hisi_sas.h

@@ -141,6 +141,7 @@ struct hisi_sas_phy {
 	struct asd_sas_phy	sas_phy;
 	struct sas_identify	identify;
 	struct completion *reset_completion;
+	struct timer_list timer;
 	spinlock_t lock;
 	u64		port_id; /* from hw */
 	u64		frame_rcvd_size;

drivers/scsi/hisi_sas/hisi_sas_main.c

@@ -920,6 +920,8 @@ static void hisi_sas_phy_init(struct hisi_hba *hisi_hba, int phy_no)
 		INIT_WORK(&phy->works[i], hisi_sas_phye_fns[i]);
 
 	spin_lock_init(&phy->lock);
+
+	timer_setup(&phy->timer, NULL, 0);
 }
 
 static void hisi_sas_port_notify_formed(struct asd_sas_phy *sas_phy)

drivers/scsi/hisi_sas/hisi_sas_v3_hw.c

@@ -1378,6 +1378,7 @@ static irqreturn_t phy_up_v3_hw(int phy_no, struct hisi_hba *hisi_hba)
 	struct device *dev = hisi_hba->dev;
 	unsigned long flags;
 
+	del_timer(&phy->timer);
 	hisi_sas_phy_write32(hisi_hba, phy_no, PHYCTRL_PHY_ENA_MSK, 1);
 
 	port_id = hisi_sas_read32(hisi_hba, PHY_PORT_NUM_MA);
@@ -1479,9 +1480,11 @@ static irqreturn_t phy_up_v3_hw(int phy_no, struct hisi_hba *hisi_hba)
 
 static irqreturn_t phy_down_v3_hw(int phy_no, struct hisi_hba *hisi_hba)
 {
+	struct hisi_sas_phy *phy = &hisi_hba->phy[phy_no];
 	u32 phy_state, sl_ctrl, txid_auto;
 	struct device *dev = hisi_hba->dev;
 
+	del_timer(&phy->timer);
 	hisi_sas_phy_write32(hisi_hba, phy_no, PHYCTRL_NOT_RDY_MSK, 1);
 
 	phy_state = hisi_sas_read32(hisi_hba, PHY_STATE);
@@ -1672,6 +1675,41 @@ static void handle_chl_int2_v3_hw(struct hisi_hba *hisi_hba, int phy_no)
 	hisi_sas_phy_write32(hisi_hba, phy_no, CHL_INT2, irq_value);
 }
 
+#define WAIT_PHYUP_TIMEOUT_V3_HW 20
+static void wait_phyup_timedout_v3_hw(struct timer_list *t)
+{
+	struct hisi_sas_phy *phy = from_timer(phy, t, timer);
+	struct hisi_hba *hisi_hba = phy->hisi_hba;
+	struct device *dev = hisi_hba->dev;
+	int phy_no = phy->sas_phy.id;
+
+	dev_warn(dev, "phy%d wait phyup timeout, issuing link reset\n", phy_no);
+	hisi_sas_notify_phy_event(phy, HISI_PHYE_LINK_RESET);
+}
+
+static void handle_chl_int0_v3_hw(struct hisi_hba *hisi_hba, int phy_no)
+{
+	u32 irq_value0 = hisi_sas_phy_read32(hisi_hba, phy_no, CHL_INT0);
+	struct device *dev = hisi_hba->dev;
+
+	if (irq_value0 & CHL_INT0_PHY_RDY_MSK) {
+		struct hisi_sas_phy *phy = &hisi_hba->phy[phy_no];
+
+		if (!timer_pending(&phy->timer)) {
+			dev_dbg(dev, "phy%d OOB ready\n", phy_no);
+			phy->timer.function = wait_phyup_timedout_v3_hw;
+			phy->timer.expires = jiffies +
+					WAIT_PHYUP_TIMEOUT_V3_HW * HZ;
+			add_timer(&phy->timer);
+		}
+	}
+
+	hisi_sas_phy_write32(hisi_hba, phy_no, CHL_INT0,
+			     irq_value0 & (~CHL_INT0_SL_RX_BCST_ACK_MSK)
+			     & (~CHL_INT0_SL_PHY_ENABLE_MSK)
+			     & (~CHL_INT0_NOT_RDY_MSK));
+}
+
 static irqreturn_t int_chnl_int_v3_hw(int irq_no, void *p)
 {
 	struct hisi_hba *hisi_hba = p;
@@ -1682,8 +1720,8 @@ static irqreturn_t int_chnl_int_v3_hw(int irq_no, void *p)
 				& 0xeeeeeeee;
 
 	while (irq_msk) {
-		u32 irq_value0 = hisi_sas_phy_read32(hisi_hba, phy_no,
-						     CHL_INT0);
+		if (irq_msk & (2 << (phy_no * 4)))
+			handle_chl_int0_v3_hw(hisi_hba, phy_no);
 
 		if (irq_msk & (4 << (phy_no * 4)))
 			handle_chl_int1_v3_hw(hisi_hba, phy_no);
@@ -1691,13 +1729,6 @@ static irqreturn_t int_chnl_int_v3_hw(int irq_no, void *p)
 		if (irq_msk & (8 << (phy_no * 4)))
 			handle_chl_int2_v3_hw(hisi_hba, phy_no);
 
-		if (irq_msk & (2 << (phy_no * 4)) && irq_value0) {
-			hisi_sas_phy_write32(hisi_hba, phy_no,
-					CHL_INT0, irq_value0
-					& (~CHL_INT0_SL_RX_BCST_ACK_MSK)
-					& (~CHL_INT0_SL_PHY_ENABLE_MSK)
-					& (~CHL_INT0_NOT_RDY_MSK));
-		}
 		irq_msk &= ~(0xe << (phy_no * 4));
 		phy_no++;
 	}