Commit ba2b062f authored by Greg Kroah-Hartman's avatar Greg Kroah-Hartman
Browse files

Merge tag 'misc-habanalabs-fixes-2021-05-08' of... 

Merge tag 'misc-habanalabs-fixes-2021-05-08' of https://git.kernel.org/pub/scm/linux/kernel/git/ogabbay/linux into char-misc-linus

Oded writes:

This tag contains the following fixes for 5.13-rc2:

- Expose PLL information per ASIC. This also fixes some casting warnings.
- Skip reading further firmware errors in case PCI link is down.
- Security firmware error should be handled as error and not warning.
- Allow user to ignore firmware errors.
- Fix bug in timeout calculation when waiting for interrupt of CS.
- Fix bug of potential use-after-free.

* tag 'misc-habanalabs-fixes-2021-05-08' of https://git.kernel.org/pub/scm/linux/kernel/git/ogabbay/linux:
  habanalabs/gaudi: Fix a potential use after free in gaudi_memset_device_memory
  habanalabs: wait for interrupt wrong timeout calculation
  habanalabs: ignore f/w status error
  habanalabs: change error level of security not ready
  habanalabs: skip reading f/w errors on bad status
  habanalabs: expose ASIC specific PLL index
parents 6efb943b 115726c5

drivers/misc/habanalabs/common/command_submission.c

+1 −1
Original line number Diff line number Diff line
@@ -2017,7 +2017,7 @@ static int _hl_interrupt_wait_ioctl(struct hl_device *hdev, struct hl_ctx *ctx, 
		if (completion_value >= target_value) {

			*status = CS_WAIT_STATUS_COMPLETED;

		} else {

			timeout -= jiffies_to_usecs(completion_rc);

			timeout = completion_rc;

			goto wait_again;

		}

	} else {

drivers/misc/habanalabs/common/firmware_if.c

+31 −22
Original line number Diff line number Diff line
@@ -362,12 +362,9 @@ static int fw_read_errors(struct hl_device *hdev, u32 boot_err0_reg, 
	}



	if (err_val & CPU_BOOT_ERR0_SECURITY_NOT_RDY) {

		dev_warn(hdev->dev,

		dev_err(hdev->dev,

			"Device boot warning - security not ready\n");

		/* This is a warning so we don't want it to disable the

		 * device

		 */

		err_val &= ~CPU_BOOT_ERR0_SECURITY_NOT_RDY;

		err_exists = true;

	}



	if (err_val & CPU_BOOT_ERR0_SECURITY_FAIL) {
@@ -403,7 +400,8 @@ static int fw_read_errors(struct hl_device *hdev, u32 boot_err0_reg, 
		err_exists = true;

	}



	if (err_exists)

	if (err_exists && ((err_val & ~CPU_BOOT_ERR0_ENABLED) &

				lower_32_bits(hdev->boot_error_status_mask)))

		return -EIO;



	return 0;
@@ -661,18 +659,13 @@ int hl_fw_cpucp_total_energy_get(struct hl_device *hdev, u64 *total_energy) 
	return rc;

}



int get_used_pll_index(struct hl_device *hdev, enum pll_index input_pll_index,

int get_used_pll_index(struct hl_device *hdev, u32 input_pll_index,

						enum pll_index *pll_index)

{

	struct asic_fixed_properties *prop = &hdev->asic_prop;

	u8 pll_byte, pll_bit_off;

	bool dynamic_pll;



	if (input_pll_index >= PLL_MAX) {

		dev_err(hdev->dev, "PLL index %d is out of range\n",

							input_pll_index);

		return -EINVAL;

	}

	int fw_pll_idx;



	dynamic_pll = prop->fw_security_status_valid &&

		(prop->fw_app_security_map & CPU_BOOT_DEV_STS0_DYN_PLL_EN);
@@ -680,28 +673,39 @@ int get_used_pll_index(struct hl_device *hdev, enum pll_index input_pll_index, 
	if (!dynamic_pll) {

		/*

		 * in case we are working with legacy FW (each asic has unique

		 * PLL numbering) extract the legacy numbering

		 * PLL numbering) use the driver based index as they are

		 * aligned with fw legacy numbering

		 */

		*pll_index = hdev->legacy_pll_map[input_pll_index];

		*pll_index = input_pll_index;

		return 0;

	}



	/* retrieve a FW compatible PLL index based on

	 * ASIC specific user request

	 */

	fw_pll_idx = hdev->asic_funcs->map_pll_idx_to_fw_idx(input_pll_index);

	if (fw_pll_idx < 0) {

		dev_err(hdev->dev, "Invalid PLL index (%u) error %d\n",

			input_pll_index, fw_pll_idx);

		return -EINVAL;

	}



	/* PLL map is a u8 array */

	pll_byte = prop->cpucp_info.pll_map[input_pll_index >> 3];

	pll_bit_off = input_pll_index & 0x7;

	pll_byte = prop->cpucp_info.pll_map[fw_pll_idx >> 3];

	pll_bit_off = fw_pll_idx & 0x7;



	if (!(pll_byte & BIT(pll_bit_off))) {

		dev_err(hdev->dev, "PLL index %d is not supported\n",

							input_pll_index);

			fw_pll_idx);

		return -EINVAL;

	}



	*pll_index = input_pll_index;

	*pll_index = fw_pll_idx;



	return 0;

}



int hl_fw_cpucp_pll_info_get(struct hl_device *hdev, enum pll_index pll_index,

int hl_fw_cpucp_pll_info_get(struct hl_device *hdev, u32 pll_index,

		u16 *pll_freq_arr)

{

	struct cpucp_packet pkt;
@@ -844,6 +848,11 @@ int hl_fw_read_preboot_status(struct hl_device *hdev, u32 cpu_boot_status_reg, 
	if (rc) {

		dev_err(hdev->dev, "Failed to read preboot version\n");

		detect_cpu_boot_status(hdev, status);



		/* If we read all FF, then something is totally wrong, no point

		 * of reading specific errors

		 */

		if (status != -1)

			fw_read_errors(hdev, boot_err0_reg,

					cpu_security_boot_status_reg);

		return -EIO;

drivers/misc/habanalabs/common/habanalabs.h

+15 −8
Original line number Diff line number Diff line
@@ -930,6 +930,9 @@ enum div_select_defs { 
 *                         driver is ready to receive asynchronous events. This

 *                         function should be called during the first init and

 *                         after every hard-reset of the device

 * @get_msi_info: Retrieve asic-specific MSI ID of the f/w async event

 * @map_pll_idx_to_fw_idx: convert driver specific per asic PLL index to

 *                         generic f/w compatible PLL Indexes

 */

struct hl_asic_funcs {

	int (*early_init)(struct hl_device *hdev);
@@ -1054,6 +1057,7 @@ struct hl_asic_funcs { 
			u32 block_id, u32 block_size);

	void (*enable_events_from_fw)(struct hl_device *hdev);

	void (*get_msi_info)(u32 *table);

	int (*map_pll_idx_to_fw_idx)(u32 pll_idx);

};
@@ -1950,8 +1954,6 @@ struct hl_mmu_funcs { 
 * @aggregated_cs_counters: aggregated cs counters among all contexts

 * @mmu_priv: device-specific MMU data.

 * @mmu_func: device-related MMU functions.

 * @legacy_pll_map: map holding map between dynamic (common) PLL indexes and

 *                  static (asic specific) PLL indexes.

 * @dram_used_mem: current DRAM memory consumption.

 * @timeout_jiffies: device CS timeout value.

 * @max_power: the max power of the device, as configured by the sysadmin. This
@@ -1960,6 +1962,12 @@ struct hl_mmu_funcs { 
 * @clock_gating_mask: is clock gating enabled. bitmask that represents the

 *                     different engines. See debugfs-driver-habanalabs for

 *                     details.

 * @boot_error_status_mask: contains a mask of the device boot error status.

 *                          Each bit represents a different error, according to

 *                          the defines in hl_boot_if.h. If the bit is cleared,

 *                          the error will be ignored by the driver during

 *                          device initialization. Mainly used to debug and

 *                          workaround firmware bugs

 * @in_reset: is device in reset flow.

 * @curr_pll_profile: current PLL profile.

 * @card_type: Various ASICs have several card types. This indicates the card
@@ -2071,12 +2079,11 @@ struct hl_device { 
	struct hl_mmu_priv		mmu_priv;

	struct hl_mmu_funcs		mmu_func[MMU_NUM_PGT_LOCATIONS];



	enum pll_index			*legacy_pll_map;



	atomic64_t			dram_used_mem;

	u64				timeout_jiffies;

	u64				max_power;

	u64				clock_gating_mask;

	u64				boot_error_status_mask;

	atomic_t			in_reset;

	enum hl_pll_frequency		curr_pll_profile;

	enum cpucp_card_types		card_type;
@@ -2387,9 +2394,9 @@ int hl_fw_cpucp_pci_counters_get(struct hl_device *hdev, 
		struct hl_info_pci_counters *counters);

int hl_fw_cpucp_total_energy_get(struct hl_device *hdev,

			u64 *total_energy);

int get_used_pll_index(struct hl_device *hdev, enum pll_index input_pll_index,

int get_used_pll_index(struct hl_device *hdev, u32 input_pll_index,

						enum pll_index *pll_index);

int hl_fw_cpucp_pll_info_get(struct hl_device *hdev, enum pll_index pll_index,

int hl_fw_cpucp_pll_info_get(struct hl_device *hdev, u32 pll_index,

		u16 *pll_freq_arr);

int hl_fw_cpucp_power_get(struct hl_device *hdev, u64 *power);

int hl_fw_init_cpu(struct hl_device *hdev, u32 cpu_boot_status_reg,
@@ -2411,9 +2418,9 @@ int hl_pci_set_outbound_region(struct hl_device *hdev, 
int hl_pci_init(struct hl_device *hdev);

void hl_pci_fini(struct hl_device *hdev);



long hl_get_frequency(struct hl_device *hdev, enum pll_index pll_index,

long hl_get_frequency(struct hl_device *hdev, u32 pll_index,

								bool curr);

void hl_set_frequency(struct hl_device *hdev, enum pll_index pll_index,

void hl_set_frequency(struct hl_device *hdev, u32 pll_index,

								u64 freq);

int hl_get_temperature(struct hl_device *hdev,

		       int sensor_index, u32 attr, long *value);

drivers/misc/habanalabs/common/habanalabs_drv.c

+7 −0
Original line number Diff line number Diff line
@@ -30,6 +30,7 @@ static DEFINE_MUTEX(hl_devs_idr_lock); 
static int timeout_locked = 30;

static int reset_on_lockup = 1;

static int memory_scrub = 1;

static ulong boot_error_status_mask = ULONG_MAX;



module_param(timeout_locked, int, 0444);

MODULE_PARM_DESC(timeout_locked,
@@ -43,6 +44,10 @@ module_param(memory_scrub, int, 0444); 
MODULE_PARM_DESC(memory_scrub,

	"Scrub device memory in various states (0 = no, 1 = yes, default yes)");



module_param(boot_error_status_mask, ulong, 0444);

MODULE_PARM_DESC(boot_error_status_mask,

	"Mask of the error status during device CPU boot (If bitX is cleared then error X is masked. Default all 1's)");



#define PCI_VENDOR_ID_HABANALABS	0x1da3



#define PCI_IDS_GOYA			0x0001
@@ -319,6 +324,8 @@ int create_hdev(struct hl_device **dev, struct pci_dev *pdev, 
	hdev->major = hl_major;

	hdev->reset_on_lockup = reset_on_lockup;

	hdev->memory_scrub = memory_scrub;

	hdev->boot_error_status_mask = boot_error_status_mask;



	hdev->pldm = 0;



	set_driver_behavior_per_device(hdev);

drivers/misc/habanalabs/common/sysfs.c

+2 −2
Original line number Diff line number Diff line
@@ -9,7 +9,7 @@ 


#include <linux/pci.h>



long hl_get_frequency(struct hl_device *hdev, enum pll_index pll_index,

long hl_get_frequency(struct hl_device *hdev, u32 pll_index,

								bool curr)

{

	struct cpucp_packet pkt;
@@ -44,7 +44,7 @@ long hl_get_frequency(struct hl_device *hdev, enum pll_index pll_index, 
	return (long) result;

}



void hl_set_frequency(struct hl_device *hdev, enum pll_index pll_index,

void hl_set_frequency(struct hl_device *hdev, u32 pll_index,

								u64 freq)

{

	struct cpucp_packet pkt;