habanalabs: improve communication protocol with cpucp

Current messaging communictaion protocol with cpucp can get out
of sync due to coherency issues. In order to improve the protocol
reliability, we modify the protocol to expect a different
acknowledgment for every packet sent to cpucp.

Signed-off-by: Ofir Bitton <obitton@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>

diff --git a/drivers/misc/habanalabs/common/firmware_if.c b/drivers/misc/habanalabs/common/firmware_if.c
index ba6920f2b4ab6451f1a6d6e22174e13e4580d471..31b52a223f0215d869a9d05aa21fa0f3105f1ddd 100644 (file)
--- a/drivers/misc/habanalabs/common/firmware_if.c
+++ b/drivers/misc/habanalabs/common/firmware_if.c
@@ -90,9 +90,10 @@ int hl_fw_send_pci_access_msg(struct hl_device *hdev, u32 opcode)
 int hl_fw_send_cpu_message(struct hl_device *hdev, u32 hw_queue_id, u32 *msg,
                                u16 len, u32 timeout, u64 *result)
 {
+       struct hl_hw_queue *queue = &hdev->kernel_queues[hw_queue_id];
        struct cpucp_packet *pkt;
        dma_addr_t pkt_dma_addr;
-       u32 tmp;
+       u32 tmp, expected_ack_val;
        int rc = 0;
 
        pkt = hdev->asic_funcs->cpu_accessible_dma_pool_alloc(hdev, len,
@@ -115,14 +116,22 @@ int hl_fw_send_cpu_message(struct hl_device *hdev, u32 hw_queue_id, u32 *msg,
                goto out;
        }
 
+       /* set fence to a non valid value */
+       pkt->fence = UINT_MAX;
+
        rc = hl_hw_queue_send_cb_no_cmpl(hdev, hw_queue_id, len, pkt_dma_addr);
        if (rc) {
                dev_err(hdev->dev, "Failed to send CB on CPU PQ (%d)\n", rc);
                goto out;
        }
 
+       if (hdev->asic_prop.fw_cpucp_ack_with_pi)
+               expected_ack_val = queue->pi;
+       else
+               expected_ack_val = CPUCP_PACKET_FENCE_VAL;
+
        rc = hl_poll_timeout_memory(hdev, &pkt->fence, tmp,
-                               (tmp == CPUCP_PACKET_FENCE_VAL), 1000,
+                               (tmp == expected_ack_val), 1000,
                                timeout, true);
 
        hl_hw_queue_inc_ci_kernel(hdev, hw_queue_id);
@@ -777,6 +786,10 @@ int hl_fw_init_cpu(struct hl_device *hdev, u32 cpu_boot_status_reg,
                                CPU_BOOT_DEV_STS0_FW_HARD_RST_EN)
                        prop->hard_reset_done_by_fw = true;
 
+               if (prop->fw_boot_cpu_security_map &
+                               CPU_BOOT_DEV_STS0_PKT_PI_ACK_EN)
+                       prop->fw_cpucp_ack_with_pi = true;
+
                dev_dbg(hdev->dev,
                        "Firmware boot CPU security status %#x\n",
                        prop->fw_boot_cpu_security_map);

diff --git a/drivers/misc/habanalabs/common/habanalabs.h b/drivers/misc/habanalabs/common/habanalabs.h
index 30f32f2edb8a2e1b76ef9bcc531040556b3d5115..3c54010f7ab9ce9dc5cd1a15a5709a362e4a9307 100644 (file)
--- a/drivers/misc/habanalabs/common/habanalabs.h
+++ b/drivers/misc/habanalabs/common/habanalabs.h
@@ -419,6 +419,8 @@ struct hl_mmu_properties {
  *                            from BOOT_DEV_STS0
  * @dram_supports_virtual_memory: is there an MMU towards the DRAM
  * @hard_reset_done_by_fw: true if firmware is handling hard reset flow
+ * @fw_cpucp_ack_with_pi: true if cpucp is acking messages with the PQ PI
+ *                        instead of a magic number
  * @num_functional_hbms: number of functional HBMs in each DCORE.
  */
 struct asic_fixed_properties {
@@ -479,6 +481,7 @@ struct asic_fixed_properties {
        u8                              fw_security_status_valid;
        u8                              dram_supports_virtual_memory;
        u8                              hard_reset_done_by_fw;
+       u8                              fw_cpucp_ack_with_pi;
        u8                              num_functional_hbms;
 };
 

diff --git a/drivers/misc/habanalabs/gaudi/gaudi.c b/drivers/misc/habanalabs/gaudi/gaudi.c
index 52fcaf25531a588b9960066bd734cacfb031ddab..006c34ae35c2c891429c216727c4cd742ecdd5cd 100644 (file)
--- a/drivers/misc/habanalabs/gaudi/gaudi.c
+++ b/drivers/misc/habanalabs/gaudi/gaudi.c
@@ -533,6 +533,7 @@ static int gaudi_get_fixed_properties(struct hl_device *hdev)
        prop->fw_security_disabled = true;
        prop->fw_security_status_valid = false;
        prop->hard_reset_done_by_fw = false;
+       prop->fw_cpucp_ack_with_pi = false;
 
        return 0;
 }
@@ -4438,9 +4439,12 @@ static void gaudi_ring_doorbell(struct hl_device *hdev, u32 hw_queue_id, u32 pi)
        /* ring the doorbell */
        WREG32(db_reg_offset, db_value);
 
-       if (hw_queue_id == GAUDI_QUEUE_ID_CPU_PQ)
+       if (hw_queue_id == GAUDI_QUEUE_ID_CPU_PQ) {
+               /* make sure device CPU will read latest data from host */
+               mb();
                WREG32(mmGIC_DISTRIBUTOR__5_GICD_SETSPI_NSR,
                                GAUDI_EVENT_PI_UPDATE);
+       }
 }
 
 static void gaudi_pqe_write(struct hl_device *hdev, __le64 *pqe,

diff --git a/drivers/misc/habanalabs/goya/goya.c b/drivers/misc/habanalabs/goya/goya.c
index a954e7c0237598b3bd0d8dea30613ef4197ba424..53db7e96686647ef8e4796c956ea94a9007e8cd7 100644 (file)
--- a/drivers/misc/habanalabs/goya/goya.c
+++ b/drivers/misc/habanalabs/goya/goya.c
@@ -461,6 +461,7 @@ int goya_get_fixed_properties(struct hl_device *hdev)
        prop->fw_security_disabled = true;
        prop->fw_security_status_valid = false;
        prop->hard_reset_done_by_fw = false;
+       prop->fw_cpucp_ack_with_pi = false;
 
        return 0;
 }
@@ -2806,9 +2807,12 @@ void goya_ring_doorbell(struct hl_device *hdev, u32 hw_queue_id, u32 pi)
        /* ring the doorbell */
        WREG32(db_reg_offset, db_value);
 
-       if (hw_queue_id == GOYA_QUEUE_ID_CPU_PQ)
+       if (hw_queue_id == GOYA_QUEUE_ID_CPU_PQ) {
+               /* make sure device CPU will read latest data from host */
+               mb();
                WREG32(mmGIC_DISTRIBUTOR__5_GICD_SETSPI_NSR,
                                GOYA_ASYNC_EVENT_ID_PI_UPDATE);
+       }
 }
 
 void goya_pqe_write(struct hl_device *hdev, __le64 *pqe, struct hl_bd *bd)

diff --git a/drivers/misc/habanalabs/include/common/hl_boot_if.h b/drivers/misc/habanalabs/include/common/hl_boot_if.h
index 57785478a4ef840fd14ebe95e9252d0c8d58881d..e87f5a98e19391570977c5d7bcba18eb249551fe 100644 (file)
--- a/drivers/misc/habanalabs/include/common/hl_boot_if.h
+++ b/drivers/misc/habanalabs/include/common/hl_boot_if.h
@@ -166,6 +166,10 @@
  *                                     FW handles HBM ECC indications.
  *                                     Initialized in: linux
  *
+ * CPU_BOOT_DEV_STS0_PKT_PI_ACK_EN     Packets ack value used in the armcpd
+ *                                     is set to the PI counter.
+ *                                     Initialized in: linux
+ *
  * CPU_BOOT_DEV_STS0_ENABLED           Device status register enabled.
  *                                     This is a main indication that the
  *                                     running FW populates the device status
@@ -190,6 +194,7 @@
 #define CPU_BOOT_DEV_STS0_SP_SRAM_EN                   (1 << 12)
 #define CPU_BOOT_DEV_STS0_CLK_GATE_EN                  (1 << 13)
 #define CPU_BOOT_DEV_STS0_HBM_ECC_EN                   (1 << 14)
+#define CPU_BOOT_DEV_STS0_PKT_PI_ACK_EN                        (1 << 15)
 #define CPU_BOOT_DEV_STS0_ENABLED                      (1 << 31)
 
 enum cpu_boot_status {