/* * Copyright 2014 IBM Corp. * * This program is free software; you can redistribute it and/or * modify it under the terms of the GNU General Public License * as published by the Free Software Foundation; either version * 2 of the License, or (at your option) any later version. */ #include #include #include #include #include #include #include #include #include #include #include "cxl.h" #include "trace.h" static int afu_control(struct cxl_afu *afu, u64 command, u64 clear, u64 result, u64 mask, bool enabled) { u64 AFU_Cntl; unsigned long timeout = jiffies + (HZ * CXL_TIMEOUT); int rc = 0; spin_lock(&afu->afu_cntl_lock); pr_devel("AFU command starting: %llx\n", command); trace_cxl_afu_ctrl(afu, command); AFU_Cntl = cxl_p2n_read(afu, CXL_AFU_Cntl_An); cxl_p2n_write(afu, CXL_AFU_Cntl_An, (AFU_Cntl & ~clear) | command); AFU_Cntl = cxl_p2n_read(afu, CXL_AFU_Cntl_An); while ((AFU_Cntl & mask) != result) { if (time_after_eq(jiffies, timeout)) { dev_warn(&afu->dev, "WARNING: AFU control timed out!\n"); rc = -EBUSY; goto out; } if (!cxl_ops->link_ok(afu->adapter, afu)) { afu->enabled = enabled; rc = -EIO; goto out; } pr_devel_ratelimited("AFU control... (0x%016llx)\n", AFU_Cntl | command); cpu_relax(); AFU_Cntl = cxl_p2n_read(afu, CXL_AFU_Cntl_An); }; if (AFU_Cntl & CXL_AFU_Cntl_An_RA) { /* * Workaround for a bug in the XSL used in the Mellanox CX4 * that fails to clear the RA bit after an AFU reset, * preventing subsequent AFU resets from working. */ cxl_p2n_write(afu, CXL_AFU_Cntl_An, AFU_Cntl & ~CXL_AFU_Cntl_An_RA); } pr_devel("AFU command complete: %llx\n", command); afu->enabled = enabled; out: trace_cxl_afu_ctrl_done(afu, command, rc); spin_unlock(&afu->afu_cntl_lock); return rc; } static int afu_enable(struct cxl_afu *afu) { pr_devel("AFU enable request\n"); return afu_control(afu, CXL_AFU_Cntl_An_E, 0, CXL_AFU_Cntl_An_ES_Enabled, CXL_AFU_Cntl_An_ES_MASK, true); } int cxl_afu_disable(struct cxl_afu *afu) { pr_devel("AFU disable request\n"); return afu_control(afu, 0, CXL_AFU_Cntl_An_E, CXL_AFU_Cntl_An_ES_Disabled, CXL_AFU_Cntl_An_ES_MASK, false); } /* This will disable as well as reset */ static int native_afu_reset(struct cxl_afu *afu) { pr_devel("AFU reset request\n"); return afu_control(afu, CXL_AFU_Cntl_An_RA, 0, CXL_AFU_Cntl_An_RS_Complete | CXL_AFU_Cntl_An_ES_Disabled, CXL_AFU_Cntl_An_RS_MASK | CXL_AFU_Cntl_An_ES_MASK, false); } static int native_afu_check_and_enable(struct cxl_afu *afu) { if (!cxl_ops->link_ok(afu->adapter, afu)) { WARN(1, "Refusing to enable afu while link down!\n"); return -EIO; } if (afu->enabled) return 0; return afu_enable(afu); } int cxl_psl_purge(struct cxl_afu *afu) { u64 PSL_CNTL = cxl_p1n_read(afu, CXL_PSL_SCNTL_An); u64 AFU_Cntl = cxl_p2n_read(afu, CXL_AFU_Cntl_An); u64 dsisr, dar; u64 start, end; unsigned long timeout = jiffies + (HZ * CXL_TIMEOUT); int rc = 0; trace_cxl_psl_ctrl(afu, CXL_PSL_SCNTL_An_Pc); pr_devel("PSL purge request\n"); if (!cxl_ops->link_ok(afu->adapter, afu)) { dev_warn(&afu->dev, "PSL Purge called with link down, ignoring\n"); rc = -EIO; goto out; } if ((AFU_Cntl & CXL_AFU_Cntl_An_ES_MASK) != CXL_AFU_Cntl_An_ES_Disabled) { WARN(1, "psl_purge request while AFU not disabled!\n"); cxl_afu_disable(afu); } cxl_p1n_write(afu, CXL_PSL_SCNTL_An, PSL_CNTL | CXL_PSL_SCNTL_An_Pc); start = local_clock(); PSL_CNTL = cxl_p1n_read(afu, CXL_PSL_SCNTL_An); while ((PSL_CNTL & CXL_PSL_SCNTL_An_Ps_MASK) == CXL_PSL_SCNTL_An_Ps_Pending) { if (time_after_eq(jiffies, timeout)) { dev_warn(&afu->dev, "WARNING: PSL Purge timed out!\n"); rc = -EBUSY; goto out; } if (!cxl_ops->link_ok(afu->adapter, afu)) { rc = -EIO; goto out; } dsisr = cxl_p2n_read(afu, CXL_PSL_DSISR_An); pr_devel_ratelimited("PSL purging... PSL_CNTL: 0x%016llx PSL_DSISR: 0x%016llx\n", PSL_CNTL, dsisr); if (dsisr & CXL_PSL_DSISR_TRANS) { dar = cxl_p2n_read(afu, CXL_PSL_DAR_An); dev_notice(&afu->dev, "PSL purge terminating pending translation, DSISR: 0x%016llx, DAR: 0x%016llx\n", dsisr, dar); cxl_p2n_write(afu, CXL_PSL_TFC_An, CXL_PSL_TFC_An_AE); } else if (dsisr) { dev_notice(&afu->dev, "PSL purge acknowledging pending non-translation fault, DSISR: 0x%016llx\n", dsisr); cxl_p2n_write(afu, CXL_PSL_TFC_An, CXL_PSL_TFC_An_A); } else { cpu_relax(); } PSL_CNTL = cxl_p1n_read(afu, CXL_PSL_SCNTL_An); }; end = local_clock(); pr_devel("PSL purged in %lld ns\n", end - start); cxl_p1n_write(afu, CXL_PSL_SCNTL_An, PSL_CNTL & ~CXL_PSL_SCNTL_An_Pc); out: trace_cxl_psl_ctrl_done(afu, CXL_PSL_SCNTL_An_Pc, rc); return rc; } static int spa_max_procs(int spa_size) { /* * From the CAIA: * end_of_SPA_area = SPA_Base + ((n+4) * 128) + (( ((n*8) + 127) >> 7) * 128) + 255 * Most of that junk is really just an overly-complicated way of saying * the last 256 bytes are __aligned(128), so it's really: * end_of_SPA_area = end_of_PSL_queue_area + __aligned(128) 255 * and * end_of_PSL_queue_area = SPA_Base + ((n+4) * 128) + (n*8) - 1 * so * sizeof(SPA) = ((n+4) * 128) + (n*8) + __aligned(128) 256 * Ignore the alignment (which is safe in this case as long as we are * careful with our rounding) and solve for n: */ return ((spa_size / 8) - 96) / 17; } int cxl_alloc_spa(struct cxl_afu *afu) { unsigned spa_size; /* Work out how many pages to allocate */ afu->native->spa_order = -1; do { afu->native->spa_order++; spa_size = (1 << afu->native->spa_order) * PAGE_SIZE; if (spa_size > 0x100000) { dev_warn(&afu->dev, "num_of_processes too large for the SPA, limiting to %i (0x%x)\n", afu->native->spa_max_procs, afu->native->spa_size); afu->num_procs = afu->native->spa_max_procs; break; } afu->native->spa_size = spa_size; afu->native->spa_max_procs = spa_max_procs(afu->native->spa_size); } while (afu->native->spa_max_procs < afu->num_procs); if (!(afu->native->spa = (struct cxl_process_element *) __get_free_pages(GFP_KERNEL | __GFP_ZERO, afu->native->spa_order))) { pr_err("cxl_alloc_spa: Unable to allocate scheduled process area\n"); return -ENOMEM; } pr_devel("spa pages: %i afu->spa_max_procs: %i afu->num_procs: %i\n", 1<native->spa_order, afu->native->spa_max_procs, afu->num_procs); return 0; } static void attach_spa(struct cxl_afu *afu) { u64 spap; afu->native->sw_command_status = (__be64 *)((char *)afu->native->spa + ((afu->native->spa_max_procs + 3) * 128)); spap = virt_to_phys(afu->native->spa) & CXL_PSL_SPAP_Addr; spap |= ((afu->native->spa_size >> (12 - CXL_PSL_SPAP_Size_Shift)) - 1) & CXL_PSL_SPAP_Size; spap |= CXL_PSL_SPAP_V; pr_devel("cxl: SPA allocated at 0x%p. Max processes: %i, sw_command_status: 0x%p CXL_PSL_SPAP_An=0x%016llx\n", afu->native->spa, afu->native->spa_max_procs, afu->native->sw_command_status, spap); cxl_p1n_write(afu, CXL_PSL_SPAP_An, spap); } static inline void detach_spa(struct cxl_afu *afu) { cxl_p1n_write(afu, CXL_PSL_SPAP_An, 0); } void cxl_release_spa(struct cxl_afu *afu) { if (afu->native->spa) { free_pages((unsigned long) afu->native->spa, afu->native->spa_order); afu->native->spa = NULL; } } int cxl_tlb_slb_invalidate(struct cxl *adapter) { unsigned long timeout = jiffies + (HZ * CXL_TIMEOUT); pr_devel("CXL adapter wide TLBIA & SLBIA\n"); cxl_p1_write(adapter, CXL_PSL_AFUSEL, CXL_PSL_AFUSEL_A); cxl_p1_write(adapter, CXL_PSL_TLBIA, CXL_TLB_SLB_IQ_ALL); while (cxl_p1_read(adapter, CXL_PSL_TLBIA) & CXL_TLB_SLB_P) { if (time_after_eq(jiffies, timeout)) { dev_warn(&adapter->dev, "WARNING: CXL adapter wide TLBIA timed out!\n"); return -EBUSY; } if (!cxl_ops->link_ok(adapter, NULL)) return -EIO; cpu_relax(); } cxl_p1_write(adapter, CXL_PSL_SLBIA, CXL_TLB_SLB_IQ_ALL); while (cxl_p1_read(adapter, CXL_PSL_SLBIA) & CXL_TLB_SLB_P) { if (time_after_eq(jiffies, timeout)) { dev_warn(&adapter->dev, "WARNING: CXL adapter wide SLBIA timed out!\n"); return -EBUSY; } if (!cxl_ops->link_ok(adapter, NULL)) return -EIO; cpu_relax(); } return 0; } static int cxl_write_sstp(struct cxl_afu *afu, u64 sstp0, u64 sstp1) { int rc; /* 1. Disable SSTP by writing 0 to SSTP1[V] */ cxl_p2n_write(afu, CXL_SSTP1_An, 0); /* 2. Invalidate all SLB entries */ if ((rc = cxl_afu_slbia(afu))) return rc; /* 3. Set SSTP0_An */ cxl_p2n_write(afu, CXL_SSTP0_An, sstp0); /* 4. Set SSTP1_An */ cxl_p2n_write(afu, CXL_SSTP1_An, sstp1); return 0; } /* Using per slice version may improve performance here. (ie. SLBIA_An) */ static void slb_invalid(struct cxl_context *ctx) { struct cxl *adapter = ctx->afu->adapter; u64 slbia; WARN_ON(!mutex_is_locked(&ctx->afu->native->spa_mutex)); cxl_p1_write(adapter, CXL_PSL_LBISEL, ((u64)be32_to_cpu(ctx->elem->common.pid) << 32) | be32_to_cpu(ctx->elem->lpid)); cxl_p1_write(adapter, CXL_PSL_SLBIA, CXL_TLB_SLB_IQ_LPIDPID); while (1) { if (!cxl_ops->link_ok(adapter, NULL)) break; slbia = cxl_p1_read(adapter, CXL_PSL_SLBIA); if (!(slbia & CXL_TLB_SLB_P)) break; cpu_relax(); } } static int do_process_element_cmd(struct cxl_context *ctx, u64 cmd, u64 pe_state) { u64 state; unsigned long timeout = jiffies + (HZ * CXL_TIMEOUT); int rc = 0; trace_cxl_llcmd(ctx, cmd); WARN_ON(!ctx->afu->enabled); ctx->elem->software_state = cpu_to_be32(pe_state); smp_wmb(); *(ctx->afu->native->sw_command_status) = cpu_to_be64(cmd | 0 | ctx->pe); smp_mb(); cxl_p1n_write(ctx->afu, CXL_PSL_LLCMD_An, cmd | ctx->pe); while (1) { if (time_after_eq(jiffies, timeout)) { dev_warn(&ctx->afu->dev, "WARNING: Process Element Command timed out!\n"); rc = -EBUSY; goto out; } if (!cxl_ops->link_ok(ctx->afu->adapter, ctx->afu)) { dev_warn(&ctx->afu->dev, "WARNING: Device link down, aborting Process Element Command!\n"); rc = -EIO; goto out; } state = be64_to_cpup(ctx->afu->native->sw_command_status); if (state == ~0ULL) { pr_err("cxl: Error adding process element to AFU\n"); rc = -1; goto out; } if ((state & (CXL_SPA_SW_CMD_MASK | CXL_SPA_SW_STATE_MASK | CXL_SPA_SW_LINK_MASK)) == (cmd | (cmd >> 16) | ctx->pe)) break; /* * The command won't finish in the PSL if there are * outstanding DSIs. Hence we need to yield here in * case there are outstanding DSIs that we need to * service. Tuning possiblity: we could wait for a * while before sched */ schedule(); } out: trace_cxl_llcmd_done(ctx, cmd, rc); return rc; } static int add_process_element(struct cxl_context *ctx) { int rc = 0; mutex_lock(&ctx->afu->native->spa_mutex); pr_devel("%s Adding pe: %i started\n", __func__, ctx->pe); if (!(rc = do_process_element_cmd(ctx, CXL_SPA_SW_CMD_ADD, CXL_PE_SOFTWARE_STATE_V))) ctx->pe_inserted = true; pr_devel("%s Adding pe: %i finished\n", __func__, ctx->pe); mutex_unlock(&ctx->afu->native->spa_mutex); return rc; } static int terminate_process_element(struct cxl_context *ctx) { int rc = 0; /* fast path terminate if it's already invalid */ if (!(ctx->elem->software_state & cpu_to_be32(CXL_PE_SOFTWARE_STATE_V))) return rc; mutex_lock(&ctx->afu->native->spa_mutex); pr_devel("%s Terminate pe: %i started\n", __func__, ctx->pe); /* We could be asked to terminate when the hw is down. That * should always succeed: it's not running if the hw has gone * away and is being reset. */ if (cxl_ops->link_ok(ctx->afu->adapter, ctx->afu)) rc = do_process_element_cmd(ctx, CXL_SPA_SW_CMD_TERMINATE, CXL_PE_SOFTWARE_STATE_V | CXL_PE_SOFTWARE_STATE_T); ctx->elem->software_state = 0; /* Remove Valid bit */ pr_devel("%s Terminate pe: %i finished\n", __func__, ctx->pe); mutex_unlock(&ctx->afu->native->spa_mutex); return rc; } static int remove_process_element(struct cxl_context *ctx) { int rc = 0; mutex_lock(&ctx->afu->native->spa_mutex); pr_devel("%s Remove pe: %i started\n", __func__, ctx->pe); /* We could be asked to remove when the hw is down. Again, if * the hw is down, the PE is gone, so we succeed. */ if (cxl_ops->link_ok(ctx->afu->adapter, ctx->afu)) rc = do_process_element_cmd(ctx, CXL_SPA_SW_CMD_REMOVE, 0); if (!rc) ctx->pe_inserted = false; slb_invalid(ctx); pr_devel("%s Remove pe: %i finished\n", __func__, ctx->pe); mutex_unlock(&ctx->afu->native->spa_mutex); return rc; } void cxl_assign_psn_space(struct cxl_context *ctx) { if (!ctx->afu->pp_size || ctx->master) { ctx->psn_phys = ctx->afu->psn_phys; ctx->psn_size = ctx->afu->adapter->ps_size; } else { ctx->psn_phys = ctx->afu->psn_phys + (ctx->afu->native->pp_offset + ctx->afu->pp_size * ctx->pe); ctx->psn_size = ctx->afu->pp_size; } } static int activate_afu_directed(struct cxl_afu *afu) { int rc; dev_info(&afu->dev, "Activating AFU directed mode\n"); afu->num_procs = afu->max_procs_virtualised; if (afu->native->spa == NULL) { if (cxl_alloc_spa(afu)) return -ENOMEM; } attach_spa(afu); cxl_p1n_write(afu, CXL_PSL_SCNTL_An, CXL_PSL_SCNTL_An_PM_AFU); cxl_p1n_write(afu, CXL_PSL_AMOR_An, 0xFFFFFFFFFFFFFFFFULL); cxl_p1n_write(afu, CXL_PSL_ID_An, CXL_PSL_ID_An_F | CXL_PSL_ID_An_L); afu->current_mode = CXL_MODE_DIRECTED; if ((rc = cxl_chardev_m_afu_add(afu))) return rc; if ((rc = cxl_sysfs_afu_m_add(afu))) goto err; if ((rc = cxl_chardev_s_afu_add(afu))) goto err1; return 0; err1: cxl_sysfs_afu_m_remove(afu); err: cxl_chardev_afu_remove(afu); return rc; } #ifdef CONFIG_CPU_LITTLE_ENDIAN #define set_endian(sr) ((sr) |= CXL_PSL_SR_An_LE) #else #define set_endian(sr) ((sr) &= ~(CXL_PSL_SR_An_LE)) #endif static u64 calculate_sr(struct cxl_context *ctx) { u64 sr = 0; set_endian(sr); if (ctx->master) sr |= CXL_PSL_SR_An_MP; if (mfspr(SPRN_LPCR) & LPCR_TC) sr |= CXL_PSL_SR_An_TC; if (ctx->kernel) { if (!ctx->real_mode) sr |= CXL_PSL_SR_An_R; sr |= (mfmsr() & MSR_SF) | CXL_PSL_SR_An_HV; } else { sr |= CXL_PSL_SR_An_PR | CXL_PSL_SR_An_R; sr &= ~(CXL_PSL_SR_An_HV); if (!test_tsk_thread_flag(current, TIF_32BIT)) sr |= CXL_PSL_SR_An_SF; } return sr; } static void update_ivtes_directed(struct cxl_context *ctx) { bool need_update = (ctx->status == STARTED); int r; if (need_update) { WARN_ON(terminate_process_element(ctx)); WARN_ON(remove_process_element(ctx)); } for (r = 0; r < CXL_IRQ_RANGES; r++) { ctx->elem->ivte_offsets[r] = cpu_to_be16(ctx->irqs.offset[r]); ctx->elem->ivte_ranges[r] = cpu_to_be16(ctx->irqs.range[r]); } /* * Theoretically we could use the update llcmd, instead of a * terminate/remove/add (or if an atomic update was required we could * do a suspend/update/resume), however it seems there might be issues * with the update llcmd on some cards (including those using an XSL on * an ASIC) so for now it's safest to go with the commands that are * known to work. In the future if we come across a situation where the * card may be performing transactions using the same PE while we are * doing this update we might need to revisit this. */ if (need_update) WARN_ON(add_process_element(ctx)); } static int attach_afu_directed(struct cxl_context *ctx, u64 wed, u64 amr) { u32 pid; int result; cxl_assign_psn_space(ctx); ctx->elem->ctxtime = 0; /* disable */ ctx->elem->lpid = cpu_to_be32(mfspr(SPRN_LPID)); ctx->elem->haurp = 0; /* disable */ ctx->elem->sdr = cpu_to_be64(mfspr(SPRN_SDR1)); pid = current->pid; if (ctx->kernel) pid = 0; ctx->elem->common.tid = 0; ctx->elem->common.pid = cpu_to_be32(pid); ctx->elem->sr = cpu_to_be64(calculate_sr(ctx)); ctx->elem->common.csrp = 0; /* disable */ ctx->elem->common.aurp0 = 0; /* disable */ ctx->elem->common.aurp1 = 0; /* disable */ cxl_prefault(ctx, wed); ctx->elem->common.sstp0 = cpu_to_be64(ctx->sstp0); ctx->elem->common.sstp1 = cpu_to_be64(ctx->sstp1); /* * Ensure we have the multiplexed PSL interrupt set up to take faults * for kernel contexts that may not have allocated any AFU IRQs at all: */ if (ctx->irqs.range[0] == 0) { ctx->irqs.offset[0] = ctx->afu->native->psl_hwirq; ctx->irqs.range[0] = 1; } update_ivtes_directed(ctx); ctx->elem->common.amr = cpu_to_be64(amr); ctx->elem->common.wed = cpu_to_be64(wed); /* first guy needs to enable */ if ((result = cxl_ops->afu_check_and_enable(ctx->afu))) return result; return add_process_element(ctx); } static int deactivate_afu_directed(struct cxl_afu *afu) { dev_info(&afu->dev, "Deactivating AFU directed mode\n"); afu->current_mode = 0; afu->num_procs = 0; cxl_sysfs_afu_m_remove(afu); cxl_chardev_afu_remove(afu); /* * The CAIA section 2.2.1 indicates that the procedure for starting and * stopping an AFU in AFU directed mode is AFU specific, which is not * ideal since this code is generic and with one exception has no * knowledge of the AFU. This is in contrast to the procedure for * disabling a dedicated process AFU, which is documented to just * require a reset. The architecture does indicate that both an AFU * reset and an AFU disable should result in the AFU being disabled and * we do both followed by a PSL purge for safety. * * Notably we used to have some issues with the disable sequence on PSL * cards, which is why we ended up using this heavy weight procedure in * the first place, however a bug was discovered that had rendered the * disable operation ineffective, so it is conceivable that was the * sole explanation for those difficulties. Careful regression testing * is recommended if anyone attempts to remove or reorder these * operations. * * The XSL on the Mellanox CX4 behaves a little differently from the * PSL based cards and will time out an AFU reset if the AFU is still * enabled. That card is special in that we do have a means to identify * it from this code, so in that case we skip the reset and just use a * disable/purge to avoid the timeout and corresponding noise in the * kernel log. */ if (afu->adapter->native->sl_ops->needs_reset_before_disable) cxl_ops->afu_reset(afu); cxl_afu_disable(afu); cxl_psl_purge(afu); return 0; } static int activate_dedicated_process(struct cxl_afu *afu) { dev_info(&afu->dev, "Activating dedicated process mode\n"); cxl_p1n_write(afu, CXL_PSL_SCNTL_An, CXL_PSL_SCNTL_An_PM_Process); cxl_p1n_write(afu, CXL_PSL_CtxTime_An, 0); /* disable */ cxl_p1n_write(afu, CXL_PSL_SPAP_An, 0); /* disable */ cxl_p1n_write(afu, CXL_PSL_AMOR_An, 0xFFFFFFFFFFFFFFFFULL); cxl_p1n_write(afu, CXL_PSL_LPID_An, mfspr(SPRN_LPID)); cxl_p1n_write(afu, CXL_HAURP_An, 0); /* disable */ cxl_p1n_write(afu, CXL_PSL_SDR_An, mfspr(SPRN_SDR1)); cxl_p2n_write(afu, CXL_CSRP_An, 0); /* disable */ cxl_p2n_write(afu, CXL_AURP0_An, 0); /* disable */ cxl_p2n_write(afu, CXL_AURP1_An, 0); /* disable */ afu->current_mode = CXL_MODE_DEDICATED; afu->num_procs = 1; return cxl_chardev_d_afu_add(afu); } static void update_ivtes_dedicated(struct cxl_context *ctx) { struct cxl_afu *afu = ctx->afu; cxl_p1n_write(afu, CXL_PSL_IVTE_Offset_An, (((u64)ctx->irqs.offset[0] & 0xffff) << 48) | (((u64)ctx->irqs.offset[1] & 0xffff) << 32) | (((u64)ctx->irqs.offset[2] & 0xffff) << 16) | ((u64)ctx->irqs.offset[3] & 0xffff)); cxl_p1n_write(afu, CXL_PSL_IVTE_Limit_An, (u64) (((u64)ctx->irqs.range[0] & 0xffff) << 48) | (((u64)ctx->irqs.range[1] & 0xffff) << 32) | (((u64)ctx->irqs.range[2] & 0xffff) << 16) | ((u64)ctx->irqs.range[3] & 0xffff)); } static int attach_dedicated(struct cxl_context *ctx, u64 wed, u64 amr) { struct cxl_afu *afu = ctx->afu; u64 pid; int rc; pid = (u64)current->pid << 32; if (ctx->kernel) pid = 0; cxl_p2n_write(afu, CXL_PSL_PID_TID_An, pid); cxl_p1n_write(afu, CXL_PSL_SR_An, calculate_sr(ctx)); if ((rc = cxl_write_sstp(afu, ctx->sstp0, ctx->sstp1))) return rc; cxl_prefault(ctx, wed); update_ivtes_dedicated(ctx); cxl_p2n_write(afu, CXL_PSL_AMR_An, amr); /* master only context for dedicated */ cxl_assign_psn_space(ctx); if ((rc = cxl_ops->afu_reset(afu))) return rc; cxl_p2n_write(afu, CXL_PSL_WED_An, wed); return afu_enable(afu); } static int deactivate_dedicated_process(struct cxl_afu *afu) { dev_info(&afu->dev, "Deactivating dedicated process mode\n"); afu->current_mode = 0; afu->num_procs = 0; cxl_chardev_afu_remove(afu); return 0; } static int native_afu_deactivate_mode(struct cxl_afu *afu, int mode) { if (mode == CXL_MODE_DIRECTED) return deactivate_afu_directed(afu); if (mode == CXL_MODE_DEDICATED) return deactivate_dedicated_process(afu); return 0; } static int native_afu_activate_mode(struct cxl_afu *afu, int mode) { if (!mode) return 0; if (!(mode & afu->modes_supported)) return -EINVAL; if (!cxl_ops->link_ok(afu->adapter, afu)) { WARN(1, "Device link is down, refusing to activate!\n"); return -EIO; } if (mode == CXL_MODE_DIRECTED) return activate_afu_directed(afu); if (mode == CXL_MODE_DEDICATED) return activate_dedicated_process(afu); return -EINVAL; } static int native_attach_process(struct cxl_context *ctx, bool kernel, u64 wed, u64 amr) { if (!cxl_ops->link_ok(ctx->afu->adapter, ctx->afu)) { WARN(1, "Device link is down, refusing to attach process!\n"); return -EIO; } ctx->kernel = kernel; if (ctx->afu->current_mode == CXL_MODE_DIRECTED) return attach_afu_directed(ctx, wed, amr); if (ctx->afu->current_mode == CXL_MODE_DEDICATED) return attach_dedicated(ctx, wed, amr); return -EINVAL; } static inline int detach_process_native_dedicated(struct cxl_context *ctx) { /* * The CAIA section 2.1.1 indicates that we need to do an AFU reset to * stop the AFU in dedicated mode (we therefore do not make that * optional like we do in the afu directed path). It does not indicate * that we need to do an explicit disable (which should occur * implicitly as part of the reset) or purge, but we do these as well * to be on the safe side. * * Notably we used to have some issues with the disable sequence * (before the sequence was spelled out in the architecture) which is * why we were so heavy weight in the first place, however a bug was * discovered that had rendered the disable operation ineffective, so * it is conceivable that was the sole explanation for those * difficulties. Point is, we should be careful and do some regression * testing if we ever attempt to remove any part of this procedure. */ cxl_ops->afu_reset(ctx->afu); cxl_afu_disable(ctx->afu); cxl_psl_purge(ctx->afu); return 0; } static void native_update_ivtes(struct cxl_context *ctx) { if (ctx->afu->current_mode == CXL_MODE_DIRECTED) return update_ivtes_directed(ctx); if (ctx->afu->current_mode == CXL_MODE_DEDICATED) return update_ivtes_dedicated(ctx); WARN(1, "native_update_ivtes: Bad mode\n"); } static inline int detach_process_native_afu_directed(struct cxl_context *ctx) { if (!ctx->pe_inserted) return 0; if (terminate_process_element(ctx)) return -1; if (remove_process_element(ctx)) return -1; return 0; } static int native_detach_process(struct cxl_context *ctx) { trace_cxl_detach(ctx); if (ctx->afu->current_mode == CXL_MODE_DEDICATED) return detach_process_native_dedicated(ctx); return detach_process_native_afu_directed(ctx); } static int native_get_irq_info(struct cxl_afu *afu, struct cxl_irq_info *info) { u64 pidtid; /* If the adapter has gone away, we can't get any meaningful * information. */ if (!cxl_ops->link_ok(afu->adapter, afu)) return -EIO; info->dsisr = cxl_p2n_read(afu, CXL_PSL_DSISR_An); info->dar = cxl_p2n_read(afu, CXL_PSL_DAR_An); info->dsr = cxl_p2n_read(afu, CXL_PSL_DSR_An); pidtid = cxl_p2n_read(afu, CXL_PSL_PID_TID_An); info->pid = pidtid >> 32; info->tid = pidtid & 0xffffffff; info->afu_err = cxl_p2n_read(afu, CXL_AFU_ERR_An); info->errstat = cxl_p2n_read(afu, CXL_PSL_ErrStat_An); info->proc_handle = 0; return 0; } void cxl_native_psl_irq_dump_regs(struct cxl_context *ctx) { u64 fir1, fir2, fir_slice, serr, afu_debug; fir1 = cxl_p1_read(ctx->afu->adapter, CXL_PSL_FIR1); fir2 = cxl_p1_read(ctx->afu->adapter, CXL_PSL_FIR2); fir_slice = cxl_p1n_read(ctx->afu, CXL_PSL_FIR_SLICE_An); afu_debug = cxl_p1n_read(ctx->afu, CXL_AFU_DEBUG_An); dev_crit(&ctx->afu->dev, "PSL_FIR1: 0x%016llx\n", fir1); dev_crit(&ctx->afu->dev, "PSL_FIR2: 0x%016llx\n", fir2); if (ctx->afu->adapter->native->sl_ops->register_serr_irq) { serr = cxl_p1n_read(ctx->afu, CXL_PSL_SERR_An); cxl_afu_decode_psl_serr(ctx->afu, serr); } dev_crit(&ctx->afu->dev, "PSL_FIR_SLICE_An: 0x%016llx\n", fir_slice); dev_crit(&ctx->afu->dev, "CXL_PSL_AFU_DEBUG_An: 0x%016llx\n", afu_debug); } static irqreturn_t native_handle_psl_slice_error(struct cxl_context *ctx, u64 dsisr, u64 errstat) { dev_crit(&ctx->afu->dev, "PSL ERROR STATUS: 0x%016llx\n", errstat); if (ctx->afu->adapter->native->sl_ops->psl_irq_dump_registers) ctx->afu->adapter->native->sl_ops->psl_irq_dump_registers(ctx); if (ctx->afu->adapter->native->sl_ops->debugfs_stop_trace) { dev_crit(&ctx->afu->dev, "STOPPING CXL TRACE\n"); ctx->afu->adapter->native->sl_ops->debugfs_stop_trace(ctx->afu->adapter); } return cxl_ops->ack_irq(ctx, 0, errstat); } static irqreturn_t fail_psl_irq(struct cxl_afu *afu, struct cxl_irq_info *irq_info) { if (irq_info->dsisr & CXL_PSL_DSISR_TRANS) cxl_p2n_write(afu, CXL_PSL_TFC_An, CXL_PSL_TFC_An_AE); else cxl_p2n_write(afu, CXL_PSL_TFC_An, CXL_PSL_TFC_An_A); return IRQ_HANDLED; } static irqreturn_t native_irq_multiplexed(int irq, void *data) { struct cxl_afu *afu = data; struct cxl_context *ctx; struct cxl_irq_info irq_info; int ph = cxl_p2n_read(afu, CXL_PSL_PEHandle_An) & 0xffff; int ret; if ((ret = native_get_irq_info(afu, &irq_info))) { WARN(1, "Unable to get CXL IRQ Info: %i\n", ret); return fail_psl_irq(afu, &irq_info); } rcu_read_lock(); ctx = idr_find(&afu->contexts_idr, ph); if (ctx) { ret = cxl_irq(irq, ctx, &irq_info); rcu_read_unlock(); return ret; } rcu_read_unlock(); WARN(1, "Unable to demultiplex CXL PSL IRQ for PE %i DSISR %016llx DAR" " %016llx\n(Possible AFU HW issue - was a term/remove acked" " with outstanding transactions?)\n", ph, irq_info.dsisr, irq_info.dar); return fail_psl_irq(afu, &irq_info); } void native_irq_wait(struct cxl_context *ctx) { u64 dsisr; int timeout = 1000; int ph; /* * Wait until no further interrupts are presented by the PSL * for this context. */ while (timeout--) { ph = cxl_p2n_read(ctx->afu, CXL_PSL_PEHandle_An) & 0xffff; if (ph != ctx->pe) return; dsisr = cxl_p2n_read(ctx->afu, CXL_PSL_DSISR_An); if ((dsisr & CXL_PSL_DSISR_PENDING) == 0) return; /* * We are waiting for the workqueue to process our * irq, so need to let that run here. */ msleep(1); } dev_warn(&ctx->afu->dev, "WARNING: waiting on DSI for PE %i" " DSISR %016llx!\n", ph, dsisr); return; } static irqreturn_t native_slice_irq_err(int irq, void *data) { struct cxl_afu *afu = data; u64 fir_slice, errstat, serr, afu_debug, afu_error, dsisr; /* * slice err interrupt is only used with full PSL (no XSL) */ serr = cxl_p1n_read(afu, CXL_PSL_SERR_An); fir_slice = cxl_p1n_read(afu, CXL_PSL_FIR_SLICE_An); errstat = cxl_p2n_read(afu, CXL_PSL_ErrStat_An); afu_debug = cxl_p1n_read(afu, CXL_AFU_DEBUG_An); afu_error = cxl_p2n_read(afu, CXL_AFU_ERR_An); dsisr = cxl_p2n_read(afu, CXL_PSL_DSISR_An); cxl_afu_decode_psl_serr(afu, serr); dev_crit(&afu->dev, "PSL_FIR_SLICE_An: 0x%016llx\n", fir_slice); dev_crit(&afu->dev, "CXL_PSL_ErrStat_An: 0x%016llx\n", errstat); dev_crit(&afu->dev, "CXL_PSL_AFU_DEBUG_An: 0x%016llx\n", afu_debug); dev_crit(&afu->dev, "AFU_ERR_An: 0x%.16llx\n", afu_error); dev_crit(&afu->dev, "PSL_DSISR_An: 0x%.16llx\n", dsisr); cxl_p1n_write(afu, CXL_PSL_SERR_An, serr); return IRQ_HANDLED; } void cxl_native_err_irq_dump_regs(struct cxl *adapter) { u64 fir1, fir2; fir1 = cxl_p1_read(adapter, CXL_PSL_FIR1); fir2 = cxl_p1_read(adapter, CXL_PSL_FIR2); dev_crit(&adapter->dev, "PSL_FIR1: 0x%016llx\nPSL_FIR2: 0x%016llx\n", fir1, fir2); } static irqreturn_t native_irq_err(int irq, void *data) { struct cxl *adapter = data; u64 err_ivte; WARN(1, "CXL ERROR interrupt %i\n", irq); err_ivte = cxl_p1_read(adapter, CXL_PSL_ErrIVTE); dev_crit(&adapter->dev, "PSL_ErrIVTE: 0x%016llx\n", err_ivte); if (adapter->native->sl_ops->debugfs_stop_trace) { dev_crit(&adapter->dev, "STOPPING CXL TRACE\n"); adapter->native->sl_ops->debugfs_stop_trace(adapter); } if (adapter->native->sl_ops->err_irq_dump_registers) adapter->native->sl_ops->err_irq_dump_registers(adapter); return IRQ_HANDLED; } int cxl_native_register_psl_err_irq(struct cxl *adapter) { int rc; adapter->irq_name = kasprintf(GFP_KERNEL, "cxl-%s-err", dev_name(&adapter->dev)); if (!adapter->irq_name) return -ENOMEM; if ((rc = cxl_register_one_irq(adapter, native_irq_err, adapter, &adapter->native->err_hwirq, &adapter->native->err_virq, adapter->irq_name))) { kfree(adapter->irq_name); adapter->irq_name = NULL; return rc; } cxl_p1_write(adapter, CXL_PSL_ErrIVTE, adapter->native->err_hwirq & 0xffff); return 0; } void cxl_native_release_psl_err_irq(struct cxl *adapter) { if (adapter->native->err_virq != irq_find_mapping(NULL, adapter->native->err_hwirq)) return; cxl_p1_write(adapter, CXL_PSL_ErrIVTE, 0x0000000000000000); cxl_unmap_irq(adapter->native->err_virq, adapter); cxl_ops->release_one_irq(adapter, adapter->native->err_hwirq); kfree(adapter->irq_name); } int cxl_native_register_serr_irq(struct cxl_afu *afu) { u64 serr; int rc; afu->err_irq_name = kasprintf(GFP_KERNEL, "cxl-%s-err", dev_name(&afu->dev)); if (!afu->err_irq_name) return -ENOMEM; if ((rc = cxl_register_one_irq(afu->adapter, native_slice_irq_err, afu, &afu->serr_hwirq, &afu->serr_virq, afu->err_irq_name))) { kfree(afu->err_irq_name); afu->err_irq_name = NULL; return rc; } serr = cxl_p1n_read(afu, CXL_PSL_SERR_An); serr = (serr & 0x00ffffffffff0000ULL) | (afu->serr_hwirq & 0xffff); cxl_p1n_write(afu, CXL_PSL_SERR_An, serr); return 0; } void cxl_native_release_serr_irq(struct cxl_afu *afu) { if (afu->serr_virq != irq_find_mapping(NULL, afu->serr_hwirq)) return; cxl_p1n_write(afu, CXL_PSL_SERR_An, 0x0000000000000000); cxl_unmap_irq(afu->serr_virq, afu); cxl_ops->release_one_irq(afu->adapter, afu->serr_hwirq); kfree(afu->err_irq_name); } int cxl_native_register_psl_irq(struct cxl_afu *afu) { int rc; afu->psl_irq_name = kasprintf(GFP_KERNEL, "cxl-%s", dev_name(&afu->dev)); if (!afu->psl_irq_name) return -ENOMEM; if ((rc = cxl_register_one_irq(afu->adapter, native_irq_multiplexed, afu, &afu->native->psl_hwirq, &afu->native->psl_virq, afu->psl_irq_name))) { kfree(afu->psl_irq_name); afu->psl_irq_name = NULL; } return rc; } void cxl_native_release_psl_irq(struct cxl_afu *afu) { if (afu->native->psl_virq != irq_find_mapping(NULL, afu->native->psl_hwirq)) return; cxl_unmap_irq(afu->native->psl_virq, afu); cxl_ops->release_one_irq(afu->adapter, afu->native->psl_hwirq); kfree(afu->psl_irq_name); } static void recover_psl_err(struct cxl_afu *afu, u64 errstat) { u64 dsisr; pr_devel("RECOVERING FROM PSL ERROR... (0x%016llx)\n", errstat); /* Clear PSL_DSISR[PE] */ dsisr = cxl_p2n_read(afu, CXL_PSL_DSISR_An); cxl_p2n_write(afu, CXL_PSL_DSISR_An, dsisr & ~CXL_PSL_DSISR_An_PE); /* Write 1s to clear error status bits */ cxl_p2n_write(afu, CXL_PSL_ErrStat_An, errstat); } static int native_ack_irq(struct cxl_context *ctx, u64 tfc, u64 psl_reset_mask) { trace_cxl_psl_irq_ack(ctx, tfc); if (tfc) cxl_p2n_write(ctx->afu, CXL_PSL_TFC_An, tfc); if (psl_reset_mask) recover_psl_err(ctx->afu, psl_reset_mask); return 0; } int cxl_check_error(struct cxl_afu *afu) { return (cxl_p1n_read(afu, CXL_PSL_SCNTL_An) == ~0ULL); } static bool native_support_attributes(const char *attr_name, enum cxl_attrs type) { return true; } static int native_afu_cr_read64(struct cxl_afu *afu, int cr, u64 off, u64 *out) { if (unlikely(!cxl_ops->link_ok(afu->adapter, afu))) return -EIO; if (unlikely(off >= afu->crs_len)) return -ERANGE; *out = in_le64(afu->native->afu_desc_mmio + afu->crs_offset + (cr * afu->crs_len) + off); return 0; } static int native_afu_cr_read32(struct cxl_afu *afu, int cr, u64 off, u32 *out) { if (unlikely(!cxl_ops->link_ok(afu->adapter, afu))) return -EIO; if (unlikely(off >= afu->crs_len)) return -ERANGE; *out = in_le32(afu->native->afu_desc_mmio + afu->crs_offset + (cr * afu->crs_len) + off); return 0; } static int native_afu_cr_read16(struct cxl_afu *afu, int cr, u64 off, u16 *out) { u64 aligned_off = off & ~0x3L; u32 val; int rc; rc = native_afu_cr_read32(afu, cr, aligned_off, &val); if (!rc) *out = (val >> ((off & 0x3) * 8)) & 0xffff; return rc; } static int native_afu_cr_read8(struct cxl_afu *afu, int cr, u64 off, u8 *out) { u64 aligned_off = off & ~0x3L; u32 val; int rc; rc = native_afu_cr_read32(afu, cr, aligned_off, &val); if (!rc) *out = (val >> ((off & 0x3) * 8)) & 0xff; return rc; } static int native_afu_cr_write32(struct cxl_afu *afu, int cr, u64 off, u32 in) { if (unlikely(!cxl_ops->link_ok(afu->adapter, afu))) return -EIO; if (unlikely(off >= afu->crs_len)) return -ERANGE; out_le32(afu->native->afu_desc_mmio + afu->crs_offset + (cr * afu->crs_len) + off, in); return 0; } static int native_afu_cr_write16(struct cxl_afu *afu, int cr, u64 off, u16 in) { u64 aligned_off = off & ~0x3L; u32 val32, mask, shift; int rc; rc = native_afu_cr_read32(afu, cr, aligned_off, &val32); if (rc) return rc; shift = (off & 0x3) * 8; WARN_ON(shift == 24); mask = 0xffff << shift; val32 = (val32 & ~mask) | (in << shift); rc = native_afu_cr_write32(afu, cr, aligned_off, val32); return rc; } static int native_afu_cr_write8(struct cxl_afu *afu, int cr, u64 off, u8 in) { u64 aligned_off = off & ~0x3L; u32 val32, mask, shift; int rc; rc = native_afu_cr_read32(afu, cr, aligned_off, &val32); if (rc) return rc; shift = (off & 0x3) * 8; mask = 0xff << shift; val32 = (val32 & ~mask) | (in << shift); rc = native_afu_cr_write32(afu, cr, aligned_off, val32); return rc; } const struct cxl_backend_ops cxl_native_ops = { .module = THIS_MODULE, .adapter_reset = cxl_pci_reset, .alloc_one_irq = cxl_pci_alloc_one_irq, .release_one_irq = cxl_pci_release_one_irq, .alloc_irq_ranges = cxl_pci_alloc_irq_ranges, .release_irq_ranges = cxl_pci_release_irq_ranges, .setup_irq = cxl_pci_setup_irq, .handle_psl_slice_error = native_handle_psl_slice_error, .psl_interrupt = NULL, .ack_irq = native_ack_irq, .irq_wait = native_irq_wait, .attach_process = native_attach_process, .detach_process = native_detach_process, .update_ivtes = native_update_ivtes, .support_attributes = native_support_attributes, .link_ok = cxl_adapter_link_ok, .release_afu = cxl_pci_release_afu, .afu_read_err_buffer = cxl_pci_afu_read_err_buffer, .afu_check_and_enable = native_afu_check_and_enable, .afu_activate_mode = native_afu_activate_mode, .afu_deactivate_mode = native_afu_deactivate_mode, .afu_reset = native_afu_reset, .afu_cr_read8 = native_afu_cr_read8, .afu_cr_read16 = native_afu_cr_read16, .afu_cr_read32 = native_afu_cr_read32, .afu_cr_read64 = native_afu_cr_read64, .afu_cr_write8 = native_afu_cr_write8, .afu_cr_write16 = native_afu_cr_write16, .afu_cr_write32 = native_afu_cr_write32, .read_adapter_vpd = cxl_pci_read_adapter_vpd, };