fault.c 8.1 KB
Newer Older
1 2 3 4 5 6 7 8 9 10
/*
 * 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 <linux/workqueue.h>
11
#include <linux/sched/signal.h>
12
#include <linux/sched/mm.h>
13 14 15 16 17 18 19 20 21 22 23
#include <linux/pid.h>
#include <linux/mm.h>
#include <linux/moduleparam.h>

#undef MODULE_PARAM_PREFIX
#define MODULE_PARAM_PREFIX "cxl" "."
#include <asm/current.h>
#include <asm/copro.h>
#include <asm/mmu.h>

#include "cxl.h"
I
Ian Munsie 已提交
24
#include "trace.h"
25

26 27 28 29 30 31 32 33 34 35
static bool sste_matches(struct cxl_sste *sste, struct copro_slb *slb)
{
	return ((sste->vsid_data == cpu_to_be64(slb->vsid)) &&
		(sste->esid_data == cpu_to_be64(slb->esid)));
}

/*
 * This finds a free SSTE for the given SLB, or returns NULL if it's already in
 * the segment table.
 */
36 37
static struct cxl_sste* find_free_sste(struct cxl_context *ctx,
				       struct copro_slb *slb)
38
{
39
	struct cxl_sste *primary, *sste, *ret = NULL;
40
	unsigned int mask = (ctx->sst_size >> 7) - 1; /* SSTP0[SegTableSize] */
41
	unsigned int entry;
42 43 44 45 46 47
	unsigned int hash;

	if (slb->vsid & SLB_VSID_B_1T)
		hash = (slb->esid >> SID_SHIFT_1T) & mask;
	else /* 256M */
		hash = (slb->esid >> SID_SHIFT) & mask;
48

49 50 51
	primary = ctx->sstp + (hash << 3);

	for (entry = 0, sste = primary; entry < 8; entry++, sste++) {
52 53 54 55
		if (!ret && !(be64_to_cpu(sste->esid_data) & SLB_ESID_V))
			ret = sste;
		if (sste_matches(sste, slb))
			return NULL;
56
	}
57 58
	if (ret)
		return ret;
59

60
	/* Nothing free, select an entry to cast out */
61
	ret = primary + ctx->sst_lru;
62
	ctx->sst_lru = (ctx->sst_lru + 1) & 0x7;
63

64
	return ret;
65 66 67 68 69 70 71 72 73
}

static void cxl_load_segment(struct cxl_context *ctx, struct copro_slb *slb)
{
	/* mask is the group index, we search primary and secondary here. */
	struct cxl_sste *sste;
	unsigned long flags;

	spin_lock_irqsave(&ctx->sste_lock, flags);
74
	sste = find_free_sste(ctx, slb);
75 76
	if (!sste)
		goto out_unlock;
77 78 79

	pr_devel("CXL Populating SST[%li]: %#llx %#llx\n",
			sste - ctx->sstp, slb->vsid, slb->esid);
I
Ian Munsie 已提交
80
	trace_cxl_ste_write(ctx, sste - ctx->sstp, slb->esid, slb->vsid);
81 82 83

	sste->vsid_data = cpu_to_be64(slb->vsid);
	sste->esid_data = cpu_to_be64(slb->esid);
84
out_unlock:
85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104
	spin_unlock_irqrestore(&ctx->sste_lock, flags);
}

static int cxl_fault_segment(struct cxl_context *ctx, struct mm_struct *mm,
			     u64 ea)
{
	struct copro_slb slb = {0,0};
	int rc;

	if (!(rc = copro_calculate_slb(mm, ea, &slb))) {
		cxl_load_segment(ctx, &slb);
	}

	return rc;
}

static void cxl_ack_ae(struct cxl_context *ctx)
{
	unsigned long flags;

105
	cxl_ops->ack_irq(ctx, CXL_PSL_TFC_An_AE, 0);
106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121

	spin_lock_irqsave(&ctx->lock, flags);
	ctx->pending_fault = true;
	ctx->fault_addr = ctx->dar;
	ctx->fault_dsisr = ctx->dsisr;
	spin_unlock_irqrestore(&ctx->lock, flags);

	wake_up_all(&ctx->wq);
}

static int cxl_handle_segment_miss(struct cxl_context *ctx,
				   struct mm_struct *mm, u64 ea)
{
	int rc;

	pr_devel("CXL interrupt: Segment fault pe: %i ea: %#llx\n", ctx->pe, ea);
I
Ian Munsie 已提交
122
	trace_cxl_ste_miss(ctx, ea);
123 124 125 126 127 128

	if ((rc = cxl_fault_segment(ctx, mm, ea)))
		cxl_ack_ae(ctx);
	else {

		mb(); /* Order seg table write to TFC MMIO write */
129
		cxl_ops->ack_irq(ctx, CXL_PSL_TFC_An_R, 0);
130 131 132 133 134 135 136 137 138 139
	}

	return IRQ_HANDLED;
}

static void cxl_handle_page_fault(struct cxl_context *ctx,
				  struct mm_struct *mm, u64 dsisr, u64 dar)
{
	unsigned flt = 0;
	int result;
140
	unsigned long access, flags, inv_flags = 0;
141

I
Ian Munsie 已提交
142 143
	trace_cxl_pte_miss(ctx, dsisr, dar);

144 145 146 147 148
	if ((result = copro_handle_mm_fault(mm, dar, dsisr, &flt))) {
		pr_devel("copro_handle_mm_fault failed: %#x\n", result);
		return cxl_ack_ae(ctx);
	}

C
Christophe Lombard 已提交
149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168
	if (!radix_enabled()) {
		/*
		 * update_mmu_cache() will not have loaded the hash since current->trap
		 * is not a 0x400 or 0x300, so just call hash_page_mm() here.
		 */
		access = _PAGE_PRESENT | _PAGE_READ;
		if (dsisr & CXL_PSL_DSISR_An_S)
			access |= _PAGE_WRITE;

		access |= _PAGE_PRIVILEGED;
		if ((!ctx->kernel) || (REGION_ID(dar) == USER_REGION_ID))
			access &= ~_PAGE_PRIVILEGED;

		if (dsisr & DSISR_NOHPTE)
			inv_flags |= HPTE_NOHPTE_UPDATE;

		local_irq_save(flags);
		hash_page_mm(mm, dar, access, 0x300, inv_flags);
		local_irq_restore(flags);
	}
169
	pr_devel("Page fault successfully handled for pe: %i!\n", ctx->pe);
170
	cxl_ops->ack_irq(ctx, CXL_PSL_TFC_An_R, 0);
171 172
}

173
/*
174 175
 * Returns the mm_struct corresponding to the context ctx.
 * mm_users == 0, the context may be in the process of being closed.
176 177 178
 */
static struct mm_struct *get_mem_context(struct cxl_context *ctx)
{
179
	if (ctx->mm == NULL)
180 181
		return NULL;

182 183
	if (!atomic_inc_not_zero(&ctx->mm->mm_users))
		return NULL;
184

185
	return ctx->mm;
186 187
}

C
Christophe Lombard 已提交
188 189
static bool cxl_is_segment_miss(struct cxl_context *ctx, u64 dsisr)
{
190
	if ((cxl_is_power8() && (dsisr & CXL_PSL_DSISR_An_DS)))
C
Christophe Lombard 已提交
191 192 193 194 195 196 197
		return true;

	return false;
}

static bool cxl_is_page_fault(struct cxl_context *ctx, u64 dsisr)
{
198
	u64 crs; /* Translation Checkout Response Status */
C
Christophe Lombard 已提交
199

200
	if ((cxl_is_power8()) && (dsisr & CXL_PSL_DSISR_An_DM))
C
Christophe Lombard 已提交
201
		return true;
202

203 204 205 206 207 208 209 210 211 212 213 214
	if (cxl_is_power9()) {
		crs = (dsisr & CXL_PSL9_DSISR_An_CO_MASK);
		if ((crs == CXL_PSL9_DSISR_An_PF_SLR) ||
		    (crs == CXL_PSL9_DSISR_An_PF_RGC) ||
		    (crs == CXL_PSL9_DSISR_An_PF_RGP) ||
		    (crs == CXL_PSL9_DSISR_An_PF_HRH) ||
		    (crs == CXL_PSL9_DSISR_An_PF_STEG) ||
		    (crs == CXL_PSL9_DSISR_An_URTCH)) {
			return true;
		}
	}

C
Christophe Lombard 已提交
215 216
	return false;
}
217

218 219 220 221 222 223
void cxl_handle_fault(struct work_struct *fault_work)
{
	struct cxl_context *ctx =
		container_of(fault_work, struct cxl_context, fault_work);
	u64 dsisr = ctx->dsisr;
	u64 dar = ctx->dar;
224
	struct mm_struct *mm = NULL;
225

226 227 228 229 230 231 232 233 234 235 236
	if (cpu_has_feature(CPU_FTR_HVMODE)) {
		if (cxl_p2n_read(ctx->afu, CXL_PSL_DSISR_An) != dsisr ||
		    cxl_p2n_read(ctx->afu, CXL_PSL_DAR_An) != dar ||
		    cxl_p2n_read(ctx->afu, CXL_PSL_PEHandle_An) != ctx->pe) {
			/* Most likely explanation is harmless - a dedicated
			 * process has detached and these were cleared by the
			 * PSL purge, but warn about it just in case
			 */
			dev_notice(&ctx->afu->dev, "cxl_handle_fault: Translation fault regs changed\n");
			return;
		}
237 238
	}

239 240 241 242 243 244
	/* Early return if the context is being / has been detached */
	if (ctx->status == CLOSED) {
		cxl_ack_ae(ctx);
		return;
	}

245 246 247
	pr_devel("CXL BOTTOM HALF handling fault for afu pe: %i. "
		"DSISR: %#llx DAR: %#llx\n", ctx->pe, dsisr, dar);

248
	if (!ctx->kernel) {
249 250 251 252 253

		mm = get_mem_context(ctx);
		if (mm == NULL) {
			pr_devel("%s: unable to get mm for pe=%d pid=%i\n",
				 __func__, ctx->pe, pid_nr(ctx->pid));
254 255
			cxl_ack_ae(ctx);
			return;
256 257 258
		} else {
			pr_devel("Handling page fault for pe=%d pid=%i\n",
				 ctx->pe, pid_nr(ctx->pid));
259
		}
260 261
	}

C
Christophe Lombard 已提交
262
	if (cxl_is_segment_miss(ctx, dsisr))
263
		cxl_handle_segment_miss(ctx, mm, dar);
C
Christophe Lombard 已提交
264
	else if (cxl_is_page_fault(ctx, dsisr))
265 266 267 268
		cxl_handle_page_fault(ctx, mm, dsisr, dar);
	else
		WARN(1, "cxl_handle_fault has nothing to handle\n");

269 270
	if (mm)
		mmput(mm);
271 272 273 274 275 276
}

static void cxl_prefault_one(struct cxl_context *ctx, u64 ea)
{
	struct mm_struct *mm;

277 278
	mm = get_mem_context(ctx);
	if (mm == NULL) {
279 280 281 282 283
		pr_devel("cxl_prefault_one unable to get mm %i\n",
			 pid_nr(ctx->pid));
		return;
	}

284
	cxl_fault_segment(ctx, mm, ea);
285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306

	mmput(mm);
}

static u64 next_segment(u64 ea, u64 vsid)
{
	if (vsid & SLB_VSID_B_1T)
		ea |= (1ULL << 40) - 1;
	else
		ea |= (1ULL << 28) - 1;

	return ea + 1;
}

static void cxl_prefault_vma(struct cxl_context *ctx)
{
	u64 ea, last_esid = 0;
	struct copro_slb slb;
	struct vm_area_struct *vma;
	int rc;
	struct mm_struct *mm;

307 308
	mm = get_mem_context(ctx);
	if (mm == NULL) {
309 310
		pr_devel("cxl_prefault_vm unable to get mm %i\n",
			 pid_nr(ctx->pid));
311
		return;
312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346
	}

	down_read(&mm->mmap_sem);
	for (vma = mm->mmap; vma; vma = vma->vm_next) {
		for (ea = vma->vm_start; ea < vma->vm_end;
				ea = next_segment(ea, slb.vsid)) {
			rc = copro_calculate_slb(mm, ea, &slb);
			if (rc)
				continue;

			if (last_esid == slb.esid)
				continue;

			cxl_load_segment(ctx, &slb);
			last_esid = slb.esid;
		}
	}
	up_read(&mm->mmap_sem);

	mmput(mm);
}

void cxl_prefault(struct cxl_context *ctx, u64 wed)
{
	switch (ctx->afu->prefault_mode) {
	case CXL_PREFAULT_WED:
		cxl_prefault_one(ctx, wed);
		break;
	case CXL_PREFAULT_ALL:
		cxl_prefault_vma(ctx);
		break;
	default:
		break;
	}
}