*
* - HL_RESET_DEVICE_RELEASE
* Set if reset is due to device release
+ *
+ * - HL_RESET_FW
+ * F/W will perform the reset. No need to ask it to reset the device. This is relevant
+ * only when running with secured f/w
*/
#define HL_RESET_HARD (1 << 0)
#define HL_RESET_FROM_RESET_THREAD (1 << 1)
#define HL_RESET_HEARTBEAT (1 << 2)
#define HL_RESET_TDR (1 << 3)
#define HL_RESET_DEVICE_RELEASE (1 << 4)
+#define HL_RESET_FW (1 << 5)
#define HL_MAX_SOBS_PER_MONITOR 8
* @completion: fence is implemented using completion
* @refcount: refcount for this fence
* @cs_sequence: sequence of the corresponding command submission
+ * @stream_master_qid_map: streams masters QID bitmap to represent all streams
+ * masters QIDs that multi cs is waiting on
* @error: mark this fence with error
* @timestamp: timestamp upon completion
- * @stream_map: streams bitmap to represent all streams that multi cs is
- * waiting on
*/
struct hl_fence {
struct completion completion;
struct kref refcount;
u64 cs_sequence;
+ u32 stream_master_qid_map;
int error;
ktime_t timestamp;
- u8 stream_map;
};
/**
* @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
+ * @get_stream_master_qid_arr: get pointer to stream masters QID array
*/
struct hl_asic_funcs {
int (*early_init)(struct hl_device *hdev);
int (*sw_init)(struct hl_device *hdev);
int (*sw_fini)(struct hl_device *hdev);
int (*hw_init)(struct hl_device *hdev);
- void (*hw_fini)(struct hl_device *hdev, bool hard_reset);
- void (*halt_engines)(struct hl_device *hdev, bool hard_reset);
+ void (*hw_fini)(struct hl_device *hdev, bool hard_reset, bool fw_reset);
+ void (*halt_engines)(struct hl_device *hdev, bool hard_reset, bool fw_reset);
int (*suspend)(struct hl_device *hdev);
int (*resume)(struct hl_device *hdev);
int (*mmap)(struct hl_device *hdev, struct vm_area_struct *vma,
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);
+ u32* (*get_stream_master_qid_arr)(void);
};
* @sgt: pointer to the scatter-gather table that holds the pages.
* @dir: for DMA unmapping, the direction must be supplied, so save it.
* @debugfs_list: node in debugfs list of command submissions.
+ * @pid: the pid of the user process owning the memory
* @addr: user-space virtual address of the start of the memory area.
* @size: size of the memory area to pin & map.
* @dma_mapped: true if the SG was mapped to DMA addresses, false otherwise.
struct sg_table *sgt;
enum dma_data_direction dir;
struct list_head debugfs_list;
+ pid_t pid;
u64 addr;
u64 size;
u8 dma_mapped;
* @state_dump_sem: protects state_dump.
* @addr: next address to read/write from/to in read/write32.
* @mmu_addr: next virtual address to translate to physical address in mmu_show.
+ * @userptr_lookup: the target user ptr to look up for on demand.
* @mmu_asid: ASID to use while translating in mmu_show.
* @state_dump_head: index of the latest state dump
* @i2c_bus: generic u8 debugfs file for bus value to use in i2c_data_read.
struct rw_semaphore state_dump_sem;
u64 addr;
u64 mmu_addr;
+ u64 userptr_lookup;
u32 mmu_asid;
u32 state_dump_head;
u8 i2c_bus;
#define HL_STR_MAX 32
-#define HL_DEV_STS_MAX (HL_DEVICE_STATUS_NEEDS_RESET + 1)
+#define HL_DEV_STS_MAX (HL_DEVICE_STATUS_LAST + 1)
/* Theoretical limit only. A single host can only contain up to 4 or 8 PCIe
* x16 cards. In extreme cases, there are hosts that can accommodate 16 cards.
* @wq: work queue for device reset procedure.
* @reset_work: reset work to be done.
* @hdev: habanalabs device structure.
+ * @fw_reset: whether f/w will do the reset without us sending them a message to do it.
*/
struct hl_device_reset_work {
struct workqueue_struct *wq;
struct delayed_work reset_work;
struct hl_device *hdev;
+ bool fw_reset;
};
/**
* @completion: completion of any of the CS in the list
* @lock: spinlock for the completion structure
* @timestamp: timestamp for the multi-CS completion
+ * @stream_master_qid_map: bitmap of all stream masters on which the multi-CS
+ * is waiting
* @used: 1 if in use, otherwise 0
- * @stream_map: bitmap of all HW/external queues streams on which the multi-CS
- * is waiting
*/
struct multi_cs_completion {
struct completion completion;
spinlock_t lock;
s64 timestamp;
+ u32 stream_master_qid_map;
u8 used;
- u8 stream_map;
};
/**
* @timestamp: timestamp of first completed CS
* @wait_status: wait for CS status
* @completion_bitmap: bitmap of completed CSs (1- completed, otherwise 0)
+ * @stream_master_qid_map: bitmap of all stream master QIDs on which the
+ * multi-CS is waiting
* @arr_len: fence_arr and seq_arr array length
- * @stream_map: bitmap of all HW/external queues streams on which the multi-CS
- * is waiting
* @gone_cs: indication of gone CS (1- there was gone CS, otherwise 0)
* @update_ts: update timestamp. 1- update the timestamp, otherwise 0.
*/
s64 timestamp;
long wait_status;
u32 completion_bitmap;
+ u32 stream_master_qid_map;
u8 arr_len;
- u8 stream_map;
u8 gone_cs;
u8 update_ts;
};
struct multi_cs_completion multi_cs_completion[
MULTI_CS_MAX_USER_CTX];
+ u32 *stream_master_qid_arr;
atomic64_t dram_used_mem;
u64 timeout_jiffies;
u64 max_power;
u8 skip_reset_on_timeout;
u8 device_cpu_is_halted;
u8 supports_wait_for_multi_cs;
+ u8 stream_master_qid_arr_size;
/* Parameters for bring-up */
u64 nic_ports_mask;
* Kernel module functions that can be accessed by entire module
*/
+/**
+ * hl_get_sg_info() - get number of pages and the DMA address from SG list.
+ * @sg: the SG list.
+ * @dma_addr: pointer to DMA address to return.
+ *
+ * Calculate the number of consecutive pages described by the SG list. Take the
+ * offset of the address in the first page, add to it the length and round it up
+ * to the number of needed pages.
+ */
+static inline u32 hl_get_sg_info(struct scatterlist *sg, dma_addr_t *dma_addr)
+{
+ *dma_addr = sg_dma_address(sg);
+
+ return ((((*dma_addr) & (PAGE_SIZE - 1)) + sg_dma_len(sg)) +
+ (PAGE_SIZE - 1)) >> PAGE_SHIFT;
+}
+
/**
* hl_mem_area_inside_range() - Checks whether address+size are inside a range.
* @address: The start address of the area we want to validate.