diff options
author | Jens Axboe <axboe@kernel.dk> | 2020-07-08 08:02:13 -0600 |
---|---|---|
committer | Jens Axboe <axboe@kernel.dk> | 2020-07-08 08:02:13 -0600 |
commit | 482c6b614a4750f71ed9c928bb5b2007a05dd694 (patch) | |
tree | 60f9140b0c24fb4b2b1059afcce8ac77ecc3d3c1 /drivers/misc/habanalabs/gaudi/gaudi.c | |
parent | 0e6e255e7a58cdf4ee4163f83deeb5ce4946051e (diff) | |
parent | dcb7fd82c75ee2d6e6f9d8cc71c52519ed52e258 (diff) |
Merge tag 'v5.8-rc4' into for-5.9/drivers
Merge in 5.8-rc4 for-5.9/block to setup for-5.9/drivers, to provide
a clean base and making the life for the NVMe changes easier.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
* tag 'v5.8-rc4': (732 commits)
Linux 5.8-rc4
x86/ldt: use "pr_info_once()" instead of open-coding it badly
MIPS: Do not use smp_processor_id() in preemptible code
MIPS: Add missing EHB in mtc0 -> mfc0 sequence for DSPen
.gitignore: Do not track `defconfig` from `make savedefconfig`
io_uring: fix regression with always ignoring signals in io_cqring_wait()
x86/ldt: Disable 16-bit segments on Xen PV
x86/entry/32: Fix #MC and #DB wiring on x86_32
x86/entry/xen: Route #DB correctly on Xen PV
x86/entry, selftests: Further improve user entry sanity checks
x86/entry/compat: Clear RAX high bits on Xen PV SYSENTER
i2c: mlxcpld: check correct size of maximum RECV_LEN packet
i2c: add Kconfig help text for slave mode
i2c: slave-eeprom: update documentation
i2c: eg20t: Load module automatically if ID matches
i2c: designware: platdrv: Set class based on DMI
i2c: algo-pca: Add 0x78 as SCL stuck low status for PCA9665
mm/page_alloc: fix documentation error
vmalloc: fix the owner argument for the new __vmalloc_node_range callers
mm/cma.c: use exact_nid true to fix possible per-numa cma leak
...
Diffstat (limited to 'drivers/misc/habanalabs/gaudi/gaudi.c')
-rw-r--r-- | drivers/misc/habanalabs/gaudi/gaudi.c | 37 |
1 files changed, 35 insertions, 2 deletions
diff --git a/drivers/misc/habanalabs/gaudi/gaudi.c b/drivers/misc/habanalabs/gaudi/gaudi.c index 61f88e9884ce..834470d10b46 100644 --- a/drivers/misc/habanalabs/gaudi/gaudi.c +++ b/drivers/misc/habanalabs/gaudi/gaudi.c @@ -96,7 +96,7 @@ #define GAUDI_NUM_OF_QM_ARB_ERR_CAUSE 3 -#define GAUDI_ARB_WDT_TIMEOUT 0x400000 +#define GAUDI_ARB_WDT_TIMEOUT 0x1000000 static const char gaudi_irq_name[GAUDI_MSI_ENTRIES][GAUDI_MAX_STRING_LEN] = { "gaudi cq 0_0", "gaudi cq 0_1", "gaudi cq 0_2", "gaudi cq 0_3", @@ -1893,6 +1893,8 @@ static void gaudi_init_pci_dma_qman(struct hl_device *hdev, int dma_id, WREG32(mmDMA0_QM_CP_MSG_BASE3_ADDR_LO_0 + q_off, so_base_ws_lo); WREG32(mmDMA0_QM_CP_MSG_BASE3_ADDR_HI_0 + q_off, so_base_ws_hi); + WREG32(mmDMA0_QM_CP_BARRIER_CFG_0 + q_off, 0x100); + /* The following configuration is needed only once per QMAN */ if (qman_id == 0) { /* Configure RAZWI IRQ */ @@ -2725,6 +2727,12 @@ static int gaudi_mmu_init(struct hl_device *hdev) WREG32(mmSTLB_HOP_CONFIGURATION, hdev->mmu_huge_page_opt ? 0x30440 : 0x40440); + /* + * The H/W expects the first PI after init to be 1. After wraparound + * we'll write 0. + */ + gaudi->mmu_cache_inv_pi = 1; + gaudi->hw_cap_initialized |= HW_CAP_MMU; return 0; @@ -3790,6 +3798,25 @@ static int gaudi_validate_dma_pkt_no_mmu(struct hl_device *hdev, src_in_host); } +static int gaudi_validate_load_and_exe_pkt(struct hl_device *hdev, + struct hl_cs_parser *parser, + struct packet_load_and_exe *user_pkt) +{ + u32 cfg; + + cfg = le32_to_cpu(user_pkt->cfg); + + if (cfg & GAUDI_PKT_LOAD_AND_EXE_CFG_DST_MASK) { + dev_err(hdev->dev, + "User not allowed to use Load and Execute\n"); + return -EPERM; + } + + parser->patched_cb_size += sizeof(struct packet_load_and_exe); + + return 0; +} + static int gaudi_validate_cb(struct hl_device *hdev, struct hl_cs_parser *parser, bool is_mmu) { @@ -3838,6 +3865,11 @@ static int gaudi_validate_cb(struct hl_device *hdev, rc = -EPERM; break; + case PACKET_LOAD_AND_EXE: + rc = gaudi_validate_load_and_exe_pkt(hdev, parser, + (struct packet_load_and_exe *) user_pkt); + break; + case PACKET_LIN_DMA: parser->contains_dma_pkt = true; if (is_mmu) @@ -3855,7 +3887,6 @@ static int gaudi_validate_cb(struct hl_device *hdev, case PACKET_FENCE: case PACKET_NOP: case PACKET_ARB_POINT: - case PACKET_LOAD_AND_EXE: parser->patched_cb_size += pkt_size; break; @@ -5994,6 +6025,8 @@ static int gaudi_mmu_invalidate_cache(struct hl_device *hdev, bool is_hard, mutex_lock(&hdev->mmu_cache_lock); /* L0 & L1 invalidation */ + WREG32(mmSTLB_INV_PS, 3); + WREG32(mmSTLB_CACHE_INV, gaudi->mmu_cache_inv_pi++); WREG32(mmSTLB_INV_PS, 2); rc = hl_poll_timeout( |