diff --git a/Documentation/devicetree/bindings/usb/snps,dwc3.yaml b/Documentation/devicetree/bindings/usb/snps,dwc3.yaml
index d516979b0681fa..25f2cf049ca99d 100644
--- a/Documentation/devicetree/bindings/usb/snps,dwc3.yaml
+++ b/Documentation/devicetree/bindings/usb/snps,dwc3.yaml
@@ -231,6 +231,16 @@ properties:
     description: When set, disable u2mac linestate check during HS transmit
     type: boolean
 
+  snps,enhanced-nak-fs-quirk:
+    description:
+      When set, the controller schedules many more handshakes to Async FS
+      endpoints, improving throughput when they frequently respond with NAKs.
+
+  snps,enhanced-nak-hs-quirk:
+    description:
+      When set, the controller schedules many more handshakes to Async HS
+      endpoints, improving throughput when they frequently respond with NAKs.
+
   snps,parkmode-disable-ss-quirk:
     description:
       When set, disable park mode for all Superspeed bus instances.
diff --git a/arch/arm64/boot/dts/broadcom/rp1.dtsi b/arch/arm64/boot/dts/broadcom/rp1.dtsi
index 1f60f3f03a1004..920a3a00d3384c 100644
--- a/arch/arm64/boot/dts/broadcom/rp1.dtsi
+++ b/arch/arm64/boot/dts/broadcom/rp1.dtsi
@@ -1077,6 +1077,7 @@
 			usb3-lpm-capable;
 			snps,axi-pipe-limit = /bits/ 8 <8>;
 			snps,dis_rxdet_inp3_quirk;
+			snps,enhanced-nak-fs-quirk;
 			snps,parkmode-disable-ss-quirk;
 			snps,parkmode-disable-hs-quirk;
 			snps,parkmode-disable-fsls-quirk;
@@ -1093,6 +1094,7 @@
 			usb3-lpm-capable;
 			snps,axi-pipe-limit = /bits/ 8 <8>;
 			snps,dis_rxdet_inp3_quirk;
+			snps,enhanced-nak-fs-quirk;
 			snps,parkmode-disable-ss-quirk;
 			snps,parkmode-disable-hs-quirk;
 			snps,parkmode-disable-fsls-quirk;
diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c
index c3c5711316d2e3..545cc4a9b440b1 100644
--- a/drivers/usb/dwc3/core.c
+++ b/drivers/usb/dwc3/core.c
@@ -1366,6 +1366,12 @@ static int dwc3_core_init(struct dwc3 *dwc)
 		if (dwc->dis_tx_ipgap_linecheck_quirk)
 			reg |= DWC3_GUCTL1_TX_IPGAP_LINECHECK_DIS;
 
+		if (dwc->enh_nak_fs_quirk)
+			reg |= DWC3_GUCTL1_NAK_PER_ENH_FS;
+
+		if (dwc->enh_nak_hs_quirk)
+			reg |= DWC3_GUCTL1_NAK_PER_ENH_HS;
+
 		if (dwc->parkmode_disable_ss_quirk)
 			reg |= DWC3_GUCTL1_PARKMODE_DISABLE_SS;
 
@@ -1669,6 +1675,10 @@ static void dwc3_get_properties(struct dwc3 *dwc)
 				"snps,resume-hs-terminations");
 	dwc->ulpi_ext_vbus_drv = device_property_read_bool(dev,
 				"snps,ulpi-ext-vbus-drv");
+	dwc->enh_nak_fs_quirk = device_property_read_bool(dev,
+				"snps,enhanced-nak-fs-quirk");
+	dwc->enh_nak_hs_quirk = device_property_read_bool(dev,
+				"snps,enhanced-nak-hs-quirk");
 	dwc->parkmode_disable_ss_quirk = device_property_read_bool(dev,
 				"snps,parkmode-disable-ss-quirk");
 	dwc->parkmode_disable_hs_quirk = device_property_read_bool(dev,
diff --git a/drivers/usb/dwc3/core.h b/drivers/usb/dwc3/core.h
index 1338b30466af74..0bc9d4a544f882 100644
--- a/drivers/usb/dwc3/core.h
+++ b/drivers/usb/dwc3/core.h
@@ -269,6 +269,8 @@
 #define DWC3_GUCTL1_TX_IPGAP_LINECHECK_DIS	BIT(28)
 #define DWC3_GUCTL1_DEV_FORCE_20_CLK_FOR_30_CLK	BIT(26)
 #define DWC3_GUCTL1_DEV_L1_EXIT_BY_HW		BIT(24)
+#define DWC3_GUCTL1_NAK_PER_ENH_FS		BIT(19)
+#define DWC3_GUCTL1_NAK_PER_ENH_HS		BIT(18)
 #define DWC3_GUCTL1_PARKMODE_DISABLE_SS		BIT(17)
 #define DWC3_GUCTL1_PARKMODE_DISABLE_HS		BIT(16)
 #define DWC3_GUCTL1_PARKMODE_DISABLE_FSLS	BIT(15)
@@ -1117,6 +1119,8 @@ struct dwc3_scratchpad_array {
  *			generation after resume from suspend.
  * @ulpi_ext_vbus_drv: Set to confiure the upli chip to drives CPEN pin
  *			VBUS with an external supply.
+ * @enh_nak_fs_quirk: Set to schedule more handshakes to Async FS endpoints.
+ * @enh_nak_hs_quirk: Set to schedule more handshakes to Async HS endpoints.
  * @parkmode_disable_ss_quirk: If set, disable park mode feature for all
  *			Superspeed instances.
  * @parkmode_disable_hs_quirk: If set, disable park mode feature for all
@@ -1347,6 +1351,8 @@ struct dwc3 {
 	unsigned		dis_tx_ipgap_linecheck_quirk:1;
 	unsigned		resume_hs_terminations:1;
 	unsigned		ulpi_ext_vbus_drv:1;
+	unsigned		enh_nak_fs_quirk:1;
+	unsigned		enh_nak_hs_quirk:1;
 	unsigned		parkmode_disable_ss_quirk:1;
 	unsigned		parkmode_disable_hs_quirk:1;
 	unsigned		parkmode_disable_fsls_quirk:1;
diff --git a/drivers/usb/host/xhci-ring.c b/drivers/usb/host/xhci-ring.c
index 46268b208e88fa..a9af54cf36d6f1 100644
--- a/drivers/usb/host/xhci-ring.c
+++ b/drivers/usb/host/xhci-ring.c
@@ -505,6 +505,19 @@ void xhci_ring_ep_doorbell(struct xhci_hcd *xhci,
 
 	trace_xhci_ring_ep_doorbell(slot_id, DB_VALUE(ep_index, stream_id));
 
+	/*
+	 * For non-coherent systems with PCIe DMA (such as Pi 4, Pi 5) there
+	 * is a theoretical race between the TRB write and barrier, which
+	 * is reported complete as soon as the write leaves the CPU domain,
+	 * the doorbell write, which may be reported as complete by the RC
+	 * at some arbitrary point, and the visibility of new TRBs in system
+	 * RAM by the endpoint DMA engine.
+	 *
+	 * This read before the write positively serialises the CPU state
+	 * by incurring a round-trip across the link.
+	 */
+	readl(db_addr);
+
 	writel(DB_VALUE(ep_index, stream_id), db_addr);
 	/* flush the write */
 	readl(db_addr);