[PATCH 1/3] habanalabs: make set_pci_regions asic function

[PATCH 1/3] habanalabs: make set_pci_regions asic function

[Oded Gabbay]

From: Ohad Sharabi <osharabi@habana.ai>

In order to better support variants of the same ASIC
the set_pci_regions function is now an ASIC function which
allows each ASIC to implement it internally, thus keeping
all definitions static to the file.

Signed-off-by: Ohad Sharabi <osharabi@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
---
 drivers/misc/habanalabs/common/habanalabs.h | 2 ++
 drivers/misc/habanalabs/gaudi/gaudi.c       | 5 +++--
 drivers/misc/habanalabs/goya/goya.c         | 5 +++--
 3 files changed, 8 insertions(+), 4 deletions(-)

diff --git a/drivers/misc/habanalabs/common/habanalabs.h b/drivers/misc/habanalabs/common/habanalabs.h
index 38617d4a34d5..75be25d47c98 100644
--- a/drivers/misc/habanalabs/common/habanalabs.h
+++ b/drivers/misc/habanalabs/common/habanalabs.h
@@ -1160,6 +1160,7 @@ struct fw_load_mgr {
  * @init_cpu_scrambler_dram: Enable CPU specific DRAM scrambling
  * @state_dump_init: initialize constants required for state dump
  * @get_sob_addr: get SOB base address offset.
+ * @set_pci_memory_regions: setting properties of PCI memory regions
  */
 struct hl_asic_funcs {
 	int (*early_init)(struct hl_device *hdev);
@@ -1288,6 +1289,7 @@ struct hl_asic_funcs {
 	void (*init_cpu_scrambler_dram)(struct hl_device *hdev);
 	void (*state_dump_init)(struct hl_device *hdev);
 	u32 (*get_sob_addr)(struct hl_device *hdev, u32 sob_id);
+	void (*set_pci_memory_regions)(struct hl_device *hdev);
 };
 
 
diff --git a/drivers/misc/habanalabs/gaudi/gaudi.c b/drivers/misc/habanalabs/gaudi/gaudi.c
index 2fe14a138f37..0cb21440b5ff 100644
--- a/drivers/misc/habanalabs/gaudi/gaudi.c
+++ b/drivers/misc/habanalabs/gaudi/gaudi.c
@@ -1867,7 +1867,7 @@ static int gaudi_sw_init(struct hl_device *hdev)
 	hdev->supports_staged_submission = true;
 	hdev->supports_wait_for_multi_cs = true;
 
-	gaudi_set_pci_memory_regions(hdev);
+	hdev->asic_funcs->set_pci_memory_regions(hdev);
 
 	return 0;
 
@@ -9463,7 +9463,8 @@ static const struct hl_asic_funcs gaudi_funcs = {
 	.init_firmware_loader = gaudi_init_firmware_loader,
 	.init_cpu_scrambler_dram = gaudi_init_scrambler_hbm,
 	.state_dump_init = gaudi_state_dump_init,
-	.get_sob_addr = gaudi_get_sob_addr
+	.get_sob_addr = gaudi_get_sob_addr,
+	.set_pci_memory_regions = gaudi_set_pci_memory_regions
 };
 
 /**
diff --git a/drivers/misc/habanalabs/goya/goya.c b/drivers/misc/habanalabs/goya/goya.c
index fad130f6a118..9d66bdda8789 100644
--- a/drivers/misc/habanalabs/goya/goya.c
+++ b/drivers/misc/habanalabs/goya/goya.c
@@ -963,7 +963,7 @@ static int goya_sw_init(struct hl_device *hdev)
 	hdev->allow_external_soft_reset = true;
 	hdev->supports_wait_for_multi_cs = false;
 
-	goya_set_pci_memory_regions(hdev);
+	hdev->asic_funcs->set_pci_memory_regions(hdev);
 
 	return 0;
 
@@ -5670,7 +5670,8 @@ static const struct hl_asic_funcs goya_funcs = {
 	.init_firmware_loader = goya_init_firmware_loader,
 	.init_cpu_scrambler_dram = goya_cpu_init_scrambler_dram,
 	.state_dump_init = goya_state_dump_init,
-	.get_sob_addr = &goya_get_sob_addr
+	.get_sob_addr = &goya_get_sob_addr,
+	.set_pci_memory_regions = goya_set_pci_memory_regions,
 };
 
 /*
-- 
2.25.1

[PATCH 2/3] habanalabs: clear msg_to_cpu_reg to avoid misread after reset

[Oded Gabbay]

From: Koby Elbaz <kelbaz@habana.ai>

For some ASICs, the f/w reads the msg_to_cpu_reg value after
reset, and for some it doesn't.
Therefore, to be sure f/w doesn't read a wrong value after reset, we
need to clear this register before the reset occurs.

Signed-off-by: Koby Elbaz <kelbaz@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
---
 drivers/misc/habanalabs/common/firmware_if.c | 28 +++++++++-----------
 1 file changed, 12 insertions(+), 16 deletions(-)

diff --git a/drivers/misc/habanalabs/common/firmware_if.c b/drivers/misc/habanalabs/common/firmware_if.c
index 0be3f5414f0b..c232d197b57a 100644
--- a/drivers/misc/habanalabs/common/firmware_if.c
+++ b/drivers/misc/habanalabs/common/firmware_if.c
@@ -667,17 +667,15 @@ int hl_fw_cpucp_info_get(struct hl_device *hdev,
 	hdev->event_queue.check_eqe_index = false;
 
 	/* Read FW application security bits again */
-	if (hdev->asic_prop.fw_cpu_boot_dev_sts0_valid) {
-		hdev->asic_prop.fw_app_cpu_boot_dev_sts0 =
-						RREG32(sts_boot_dev_sts0_reg);
-		if (hdev->asic_prop.fw_app_cpu_boot_dev_sts0 &
+	if (prop->fw_cpu_boot_dev_sts0_valid) {
+		prop->fw_app_cpu_boot_dev_sts0 = RREG32(sts_boot_dev_sts0_reg);
+		if (prop->fw_app_cpu_boot_dev_sts0 &
 				CPU_BOOT_DEV_STS0_EQ_INDEX_EN)
 			hdev->event_queue.check_eqe_index = true;
 	}
 
-	if (hdev->asic_prop.fw_cpu_boot_dev_sts1_valid)
-		hdev->asic_prop.fw_app_cpu_boot_dev_sts1 =
-						RREG32(sts_boot_dev_sts1_reg);
+	if (prop->fw_cpu_boot_dev_sts1_valid)
+		prop->fw_app_cpu_boot_dev_sts1 = RREG32(sts_boot_dev_sts1_reg);
 
 out:
 	hdev->asic_funcs->cpu_accessible_dma_pool_free(hdev,
@@ -1012,6 +1010,11 @@ void hl_fw_ask_halt_machine_without_linux(struct hl_device *hdev)
 	} else {
 		WREG32(static_loader->kmd_msg_to_cpu_reg, KMD_MSG_GOTO_WFE);
 		msleep(static_loader->cpu_reset_wait_msec);
+
+		/* Must clear this register in order to prevent preboot
+		 * from reading WFE after reboot
+		 */
+		WREG32(static_loader->kmd_msg_to_cpu_reg, KMD_MSG_NA);
 	}
 
 	hdev->device_cpu_is_halted = true;
@@ -1242,11 +1245,6 @@ static void hl_fw_preboot_update_state(struct hl_device *hdev)
 	 *               b. Check whether hard reset is done by boot cpu
 	 * 3. FW application - a. Fetch fw application security status
 	 *                     b. Check whether hard reset is done by fw app
-	 *
-	 * Preboot:
-	 * Check security status bit (CPU_BOOT_DEV_STS0_ENABLED). If set, then-
-	 * check security enabled bit (CPU_BOOT_DEV_STS0_SECURITY_EN)
-	 * If set, then mark GIC controller to be disabled.
 	 */
 	prop->hard_reset_done_by_fw =
 		!!(cpu_boot_dev_sts0 & CPU_BOOT_DEV_STS0_FW_HARD_RST_EN);
@@ -2126,8 +2124,7 @@ static void hl_fw_linux_update_state(struct hl_device *hdev,
 
 	/* Read FW application security bits */
 	if (prop->fw_cpu_boot_dev_sts0_valid) {
-		prop->fw_app_cpu_boot_dev_sts0 =
-				RREG32(cpu_boot_dev_sts0_reg);
+		prop->fw_app_cpu_boot_dev_sts0 = RREG32(cpu_boot_dev_sts0_reg);
 
 		if (prop->fw_app_cpu_boot_dev_sts0 &
 				CPU_BOOT_DEV_STS0_FW_HARD_RST_EN)
@@ -2147,8 +2144,7 @@ static void hl_fw_linux_update_state(struct hl_device *hdev,
 	}
 
 	if (prop->fw_cpu_boot_dev_sts1_valid) {
-		prop->fw_app_cpu_boot_dev_sts1 =
-				RREG32(cpu_boot_dev_sts1_reg);
+		prop->fw_app_cpu_boot_dev_sts1 = RREG32(cpu_boot_dev_sts1_reg);
 
 		dev_dbg(hdev->dev,
 			"Firmware application CPU status1 %#x\n",
-- 
2.25.1

[PATCH 3/3] habanalabs: add validity check for event ID received from F/W

[Oded Gabbay]

From: Ofir Bitton <obitton@habana.ai>

Currently there is no validity check for event ID received from F/W,
Thus exposing driver to memory overrun.

Signed-off-by: Ofir Bitton <obitton@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
---
 drivers/misc/habanalabs/gaudi/gaudi.c | 6 ++++++
 drivers/misc/habanalabs/goya/goya.c   | 6 ++++++
 2 files changed, 12 insertions(+)

diff --git a/drivers/misc/habanalabs/gaudi/gaudi.c b/drivers/misc/habanalabs/gaudi/gaudi.c
index 0cb21440b5ff..655278e98618 100644
--- a/drivers/misc/habanalabs/gaudi/gaudi.c
+++ b/drivers/misc/habanalabs/gaudi/gaudi.c
@@ -7962,6 +7962,12 @@ static void gaudi_handle_eqe(struct hl_device *hdev,
 	u8 cause;
 	int rc;
 
+	if (event_type >= GAUDI_EVENT_SIZE) {
+		dev_err(hdev->dev, "Event type %u exceeds maximum of %u",
+				event_type, GAUDI_EVENT_SIZE - 1);
+		return;
+	}
+
 	gaudi->events_stat[event_type]++;
 	gaudi->events_stat_aggregate[event_type]++;
 
diff --git a/drivers/misc/habanalabs/goya/goya.c b/drivers/misc/habanalabs/goya/goya.c
index 9d66bdda8789..d54c700c31cd 100644
--- a/drivers/misc/habanalabs/goya/goya.c
+++ b/drivers/misc/habanalabs/goya/goya.c
@@ -4805,6 +4805,12 @@ void goya_handle_eqe(struct hl_device *hdev, struct hl_eq_entry *eq_entry)
 				>> EQ_CTL_EVENT_TYPE_SHIFT);
 	struct goya_device *goya = hdev->asic_specific;
 
+	if (event_type >= GOYA_ASYNC_EVENT_ID_SIZE) {
+		dev_err(hdev->dev, "Event type %u exceeds maximum of %u",
+				event_type, GOYA_ASYNC_EVENT_ID_SIZE - 1);
+		return;
+	}
+
 	goya->events_stat[event_type]++;
 	goya->events_stat_aggregate[event_type]++;
 
-- 
2.25.1