1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
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
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
|
// SPDX-License-Identifier: GPL-2.0
/*
* Copyright (C) 2015 - 2016 Cavium, Inc.
*/
#include <linux/bitfield.h>
#include <linux/kernel.h>
#include <linux/init.h>
#include <linux/pci.h>
#include <linux/of_address.h>
#include <linux/of_pci.h>
#include <linux/pci-acpi.h>
#include <linux/pci-ecam.h>
#include <linux/platform_device.h>
#include <linux/io-64-nonatomic-lo-hi.h>
#include "../pci.h"
#if defined(CONFIG_PCI_HOST_THUNDER_PEM) || (defined(CONFIG_ACPI) && defined(CONFIG_PCI_QUIRKS))
#define PEM_CFG_WR 0x28
#define PEM_CFG_RD 0x30
/*
* Enhanced Configuration Access Mechanism (ECAM)
*
* N.B. This is a non-standard platform-specific ECAM bus shift value. For
* standard values defined in the PCI Express Base Specification see
* include/linux/pci-ecam.h.
*/
#define THUNDER_PCIE_ECAM_BUS_SHIFT 24
struct thunder_pem_pci {
u32 ea_entry[3];
void __iomem *pem_reg_base;
};
static int thunder_pem_bridge_read(struct pci_bus *bus, unsigned int devfn,
int where, int size, u32 *val)
{
u64 read_val, tmp_val;
struct pci_config_window *cfg = bus->sysdata;
struct thunder_pem_pci *pem_pci = (struct thunder_pem_pci *)cfg->priv;
if (devfn != 0 || where >= 2048)
return PCIBIOS_DEVICE_NOT_FOUND;
/*
* 32-bit accesses only. Write the address to the low order
* bits of PEM_CFG_RD, then trigger the read by reading back.
* The config data lands in the upper 32-bits of PEM_CFG_RD.
*/
read_val = where & ~3ull;
writeq(read_val, pem_pci->pem_reg_base + PEM_CFG_RD);
read_val = readq(pem_pci->pem_reg_base + PEM_CFG_RD);
read_val >>= 32;
/*
* The config space contains some garbage, fix it up. Also
* synthesize an EA capability for the BAR used by MSI-X.
*/
switch (where & ~3) {
case 0x40:
read_val &= 0xffff00ff;
read_val |= 0x00007000; /* Skip MSI CAP */
break;
case 0x70: /* Express Cap */
/*
* Change PME interrupt to vector 2 on T88 where it
* reads as 0, else leave it alone.
*/
if (!(read_val & (0x1f << 25)))
read_val |= (2u << 25);
break;
case 0xb0: /* MSI-X Cap */
/* TableSize=2 or 4, Next Cap is EA */
read_val &= 0xc00000ff;
/*
* If Express Cap(0x70) raw PME vector reads as 0 we are on
* T88 and TableSize is reported as 4, else TableSize
* is 2.
*/
writeq(0x70, pem_pci->pem_reg_base + PEM_CFG_RD);
tmp_val = readq(pem_pci->pem_reg_base + PEM_CFG_RD);
tmp_val >>= 32;
if (!(tmp_val & (0x1f << 25)))
read_val |= 0x0003bc00;
else
read_val |= 0x0001bc00;
break;
case 0xb4:
/* Table offset=0, BIR=0 */
read_val = 0x00000000;
break;
case 0xb8:
/* BPA offset=0xf0000, BIR=0 */
read_val = 0x000f0000;
break;
case 0xbc:
/* EA, 1 entry, no next Cap */
read_val = 0x00010014;
break;
case 0xc0:
/* DW2 for type-1 */
read_val = 0x00000000;
break;
case 0xc4:
/* Entry BEI=0, PP=0x00, SP=0xff, ES=3 */
read_val = 0x80ff0003;
break;
case 0xc8:
read_val = pem_pci->ea_entry[0];
break;
case 0xcc:
read_val = pem_pci->ea_entry[1];
break;
case 0xd0:
read_val = pem_pci->ea_entry[2];
break;
default:
break;
}
read_val >>= (8 * (where & 3));
switch (size) {
case 1:
read_val &= 0xff;
break;
case 2:
read_val &= 0xffff;
break;
default:
break;
}
*val = read_val;
return PCIBIOS_SUCCESSFUL;
}
static int thunder_pem_config_read(struct pci_bus *bus, unsigned int devfn,
int where, int size, u32 *val)
{
struct pci_config_window *cfg = bus->sysdata;
if (bus->number < cfg->busr.start ||
bus->number > cfg->busr.end)
return PCIBIOS_DEVICE_NOT_FOUND;
/*
* The first device on the bus is the PEM PCIe bridge.
* Special case its config access.
*/
if (bus->number == cfg->busr.start)
return thunder_pem_bridge_read(bus, devfn, where, size, val);
return pci_generic_config_read(bus, devfn, where, size, val);
}
/*
* Some of the w1c_bits below also include read-only or non-writable
* reserved bits, this makes the code simpler and is OK as the bits
* are not affected by writing zeros to them.
*/
static u32 thunder_pem_bridge_w1c_bits(u64 where_aligned)
{
u32 w1c_bits = 0;
switch (where_aligned) {
case 0x04: /* Command/Status */
case 0x1c: /* Base and I/O Limit/Secondary Status */
w1c_bits = 0xff000000;
break;
case 0x44: /* Power Management Control and Status */
w1c_bits = 0xfffffe00;
break;
case 0x78: /* Device Control/Device Status */
case 0x80: /* Link Control/Link Status */
case 0x88: /* Slot Control/Slot Status */
case 0x90: /* Root Status */
case 0xa0: /* Link Control 2 Registers/Link Status 2 */
w1c_bits = 0xffff0000;
break;
case 0x104: /* Uncorrectable Error Status */
case 0x110: /* Correctable Error Status */
case 0x130: /* Error Status */
case 0x160: /* Link Control 4 */
w1c_bits = 0xffffffff;
break;
default:
break;
}
return w1c_bits;
}
/* Some bits must be written to one so they appear to be read-only. */
static u32 thunder_pem_bridge_w1_bits(u64 where_aligned)
{
u32 w1_bits;
switch (where_aligned) {
case 0x1c: /* I/O Base / I/O Limit, Secondary Status */
/* Force 32-bit I/O addressing. */
w1_bits = 0x0101;
break;
case 0x24: /* Prefetchable Memory Base / Prefetchable Memory Limit */
/* Force 64-bit addressing */
w1_bits = 0x00010001;
break;
default:
w1_bits = 0;
break;
}
return w1_bits;
}
static int thunder_pem_bridge_write(struct pci_bus *bus, unsigned int devfn,
int where, int size, u32 val)
{
struct pci_config_window *cfg = bus->sysdata;
struct thunder_pem_pci *pem_pci = (struct thunder_pem_pci *)cfg->priv;
u64 write_val, read_val;
u64 where_aligned = where & ~3ull;
u32 mask = 0;
if (devfn != 0 || where >= 2048)
return PCIBIOS_DEVICE_NOT_FOUND;
/*
* 32-bit accesses only. If the write is for a size smaller
* than 32-bits, we must first read the 32-bit value and merge
* in the desired bits and then write the whole 32-bits back
* out.
*/
switch (size) {
case 1:
writeq(where_aligned, pem_pci->pem_reg_base + PEM_CFG_RD);
read_val = readq(pem_pci->pem_reg_base + PEM_CFG_RD);
read_val >>= 32;
mask = ~(0xff << (8 * (where & 3)));
read_val &= mask;
val = (val & 0xff) << (8 * (where & 3));
val |= (u32)read_val;
break;
case 2:
writeq(where_aligned, pem_pci->pem_reg_base + PEM_CFG_RD);
read_val = readq(pem_pci->pem_reg_base + PEM_CFG_RD);
read_val >>= 32;
mask = ~(0xffff << (8 * (where & 3)));
read_val &= mask;
val = (val & 0xffff) << (8 * (where & 3));
val |= (u32)read_val;
break;
default:
break;
}
/*
* By expanding the write width to 32 bits, we may
* inadvertently hit some W1C bits that were not intended to
* be written. Calculate the mask that must be applied to the
* data to be written to avoid these cases.
*/
if (mask) {
u32 w1c_bits = thunder_pem_bridge_w1c_bits(where);
if (w1c_bits) {
mask &= w1c_bits;
val &= ~mask;
}
}
/*
* Some bits must be read-only with value of one. Since the
* access method allows these to be cleared if a zero is
* written, force them to one before writing.
*/
val |= thunder_pem_bridge_w1_bits(where_aligned);
/*
* Low order bits are the config address, the high order 32
* bits are the data to be written.
*/
write_val = (((u64)val) << 32) | where_aligned;
writeq(write_val, pem_pci->pem_reg_base + PEM_CFG_WR);
return PCIBIOS_SUCCESSFUL;
}
static int thunder_pem_config_write(struct pci_bus *bus, unsigned int devfn,
int where, int size, u32 val)
{
struct pci_config_window *cfg = bus->sysdata;
if (bus->number < cfg->busr.start ||
bus->number > cfg->busr.end)
return PCIBIOS_DEVICE_NOT_FOUND;
/*
* The first device on the bus is the PEM PCIe bridge.
* Special case its config access.
*/
if (bus->number == cfg->busr.start)
return thunder_pem_bridge_write(bus, devfn, where, size, val);
return pci_generic_config_write(bus, devfn, where, size, val);
}
static int thunder_pem_init(struct device *dev, struct pci_config_window *cfg,
struct resource *res_pem)
{
struct thunder_pem_pci *pem_pci;
resource_size_t bar4_start;
pem_pci = devm_kzalloc(dev, sizeof(*pem_pci), GFP_KERNEL);
if (!pem_pci)
return -ENOMEM;
pem_pci->pem_reg_base = devm_ioremap(dev, res_pem->start, 0x10000);
if (!pem_pci->pem_reg_base)
return -ENOMEM;
/*
* The MSI-X BAR for the PEM and AER interrupts is located at
* a fixed offset from the PEM register base. Generate a
* fragment of the synthesized Enhanced Allocation capability
* structure here for the BAR.
*/
bar4_start = res_pem->start + 0xf00000;
pem_pci->ea_entry[0] = lower_32_bits(bar4_start) | 2;
pem_pci->ea_entry[1] = lower_32_bits(res_pem->end - bar4_start) & ~3u;
pem_pci->ea_entry[2] = upper_32_bits(bar4_start);
cfg->priv = pem_pci;
return 0;
}
#if defined(CONFIG_ACPI) && defined(CONFIG_PCI_QUIRKS)
#define PEM_RES_BASE 0x87e0c0000000ULL
#define PEM_NODE_MASK GENMASK_ULL(45, 44)
#define PEM_INDX_MASK GENMASK_ULL(26, 24)
#define PEM_MIN_DOM_IN_NODE 4
#define PEM_MAX_DOM_IN_NODE 10
static void thunder_pem_reserve_range(struct device *dev, int seg,
struct resource *r)
{
resource_size_t start = r->start, end = r->end;
struct resource *res;
const char *regionid;
regionid = kasprintf(GFP_KERNEL, "PEM RC:%d", seg);
if (!regionid)
return;
res = request_mem_region(start, end - start + 1, regionid);
if (res)
res->flags &= ~IORESOURCE_BUSY;
else
kfree(regionid);
dev_info(dev, "%pR %s reserved\n", r,
res ? "has been" : "could not be");
}
static void thunder_pem_legacy_fw(struct acpi_pci_root *root,
struct resource *res_pem)
{
int node = acpi_get_node(root->device->handle);
int index;
if (node == NUMA_NO_NODE)
node = 0;
index = root->segment - PEM_MIN_DOM_IN_NODE;
index -= node * PEM_MAX_DOM_IN_NODE;
res_pem->start = PEM_RES_BASE | FIELD_PREP(PEM_NODE_MASK, node) |
FIELD_PREP(PEM_INDX_MASK, index);
res_pem->flags = IORESOURCE_MEM;
}
static int thunder_pem_acpi_init(struct pci_config_window *cfg)
{
struct device *dev = cfg->parent;
struct acpi_device *adev = to_acpi_device(dev);
struct acpi_pci_root *root = acpi_driver_data(adev);
struct resource *res_pem;
int ret;
res_pem = devm_kzalloc(&adev->dev, sizeof(*res_pem), GFP_KERNEL);
if (!res_pem)
return -ENOMEM;
ret = acpi_get_rc_resources(dev, "CAVA02B", root->segment, res_pem);
/*
* If we fail to gather resources it means that we run with old
* FW where we need to calculate PEM-specific resources manually.
*/
if (ret) {
thunder_pem_legacy_fw(root, res_pem);
/*
* Reserve 64K size PEM specific resources. The full 16M range
* size is required for thunder_pem_init() call.
*/
res_pem->end = res_pem->start + SZ_64K - 1;
thunder_pem_reserve_range(dev, root->segment, res_pem);
res_pem->end = res_pem->start + SZ_16M - 1;
/* Reserve PCI configuration space as well. */
thunder_pem_reserve_range(dev, root->segment, &cfg->res);
}
return thunder_pem_init(dev, cfg, res_pem);
}
const struct pci_ecam_ops thunder_pem_ecam_ops = {
.bus_shift = THUNDER_PCIE_ECAM_BUS_SHIFT,
.init = thunder_pem_acpi_init,
.pci_ops = {
.map_bus = pci_ecam_map_bus,
.read = thunder_pem_config_read,
.write = thunder_pem_config_write,
}
};
#endif
#ifdef CONFIG_PCI_HOST_THUNDER_PEM
static int thunder_pem_platform_init(struct pci_config_window *cfg)
{
struct device *dev = cfg->parent;
struct platform_device *pdev = to_platform_device(dev);
struct resource *res_pem;
if (!dev->of_node)
return -EINVAL;
/*
* The second register range is the PEM bridge to the PCIe
* bus. It has a different config access method than those
* devices behind the bridge.
*/
res_pem = platform_get_resource(pdev, IORESOURCE_MEM, 1);
if (!res_pem) {
dev_err(dev, "missing \"reg[1]\"property\n");
return -EINVAL;
}
return thunder_pem_init(dev, cfg, res_pem);
}
static const struct pci_ecam_ops pci_thunder_pem_ops = {
.bus_shift = THUNDER_PCIE_ECAM_BUS_SHIFT,
.init = thunder_pem_platform_init,
.pci_ops = {
.map_bus = pci_ecam_map_bus,
.read = thunder_pem_config_read,
.write = thunder_pem_config_write,
}
};
static const struct of_device_id thunder_pem_of_match[] = {
{
.compatible = "cavium,pci-host-thunder-pem",
.data = &pci_thunder_pem_ops,
},
{ },
};
static struct platform_driver thunder_pem_driver = {
.driver = {
.name = KBUILD_MODNAME,
.of_match_table = thunder_pem_of_match,
.suppress_bind_attrs = true,
},
.probe = pci_host_common_probe,
};
builtin_platform_driver(thunder_pem_driver);
#endif
#endif
|