aboutsummaryrefslogtreecommitdiffstats
path: root/arch/powerpc/kernel
diff options
context:
space:
mode:
Diffstat (limited to 'arch/powerpc/kernel')
-rw-r--r--arch/powerpc/kernel/Makefile3
-rw-r--r--arch/powerpc/kernel/asm-offsets.c21
-rw-r--r--arch/powerpc/kernel/crash.c2
-rw-r--r--arch/powerpc/kernel/dma-swiotlb.c1
-rw-r--r--arch/powerpc/kernel/exceptions-64s.S8
-rw-r--r--arch/powerpc/kernel/head_64.S7
-rw-r--r--arch/powerpc/kernel/head_8xx.S313
-rw-r--r--arch/powerpc/kernel/head_fsl_booke.S22
-rw-r--r--arch/powerpc/kernel/io.c4
-rw-r--r--arch/powerpc/kernel/irq.c137
-rw-r--r--arch/powerpc/kernel/lparcfg.c4
-rw-r--r--arch/powerpc/kernel/misc_32.S18
-rw-r--r--arch/powerpc/kernel/nvram_64.c56
-rw-r--r--arch/powerpc/kernel/perf_callchain.c20
-rw-r--r--arch/powerpc/kernel/ppc_ksyms.c3
-rw-r--r--arch/powerpc/kernel/proc_powerpc.c (renamed from arch/powerpc/kernel/proc_ppc64.c)102
-rw-r--r--arch/powerpc/kernel/rtas_flash.c10
-rw-r--r--arch/powerpc/kernel/rtasd.c539
-rw-r--r--arch/powerpc/kernel/setup_64.c5
-rw-r--r--arch/powerpc/kernel/smp.c3
-rw-r--r--arch/powerpc/kernel/sysfs.c19
-rw-r--r--arch/powerpc/kernel/time.c1
-rw-r--r--arch/powerpc/kernel/traps.c22
-rw-r--r--arch/powerpc/kernel/vector.S2
24 files changed, 968 insertions, 354 deletions
diff --git a/arch/powerpc/kernel/Makefile b/arch/powerpc/kernel/Makefile
index b23664a0b86c..c002b0410219 100644
--- a/arch/powerpc/kernel/Makefile
+++ b/arch/powerpc/kernel/Makefile
@@ -42,10 +42,11 @@ obj-$(CONFIG_ALTIVEC) += vecemu.o
42obj-$(CONFIG_PPC_970_NAP) += idle_power4.o 42obj-$(CONFIG_PPC_970_NAP) += idle_power4.o
43obj-$(CONFIG_PPC_OF) += of_device.o of_platform.o prom_parse.o 43obj-$(CONFIG_PPC_OF) += of_device.o of_platform.o prom_parse.o
44obj-$(CONFIG_PPC_CLOCK) += clock.o 44obj-$(CONFIG_PPC_CLOCK) += clock.o
45procfs-$(CONFIG_PPC64) := proc_ppc64.o 45procfs-y := proc_powerpc.o
46obj-$(CONFIG_PROC_FS) += $(procfs-y) 46obj-$(CONFIG_PROC_FS) += $(procfs-y)
47rtaspci-$(CONFIG_PPC64)-$(CONFIG_PCI) := rtas_pci.o 47rtaspci-$(CONFIG_PPC64)-$(CONFIG_PCI) := rtas_pci.o
48obj-$(CONFIG_PPC_RTAS) += rtas.o rtas-rtc.o $(rtaspci-y-y) 48obj-$(CONFIG_PPC_RTAS) += rtas.o rtas-rtc.o $(rtaspci-y-y)
49obj-$(CONFIG_PPC_RTAS_DAEMON) += rtasd.o
49obj-$(CONFIG_RTAS_FLASH) += rtas_flash.o 50obj-$(CONFIG_RTAS_FLASH) += rtas_flash.o
50obj-$(CONFIG_RTAS_PROC) += rtas-proc.o 51obj-$(CONFIG_RTAS_PROC) += rtas-proc.o
51obj-$(CONFIG_LPARCFG) += lparcfg.o 52obj-$(CONFIG_LPARCFG) += lparcfg.o
diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c
index 0812b0f414bb..a6c2b63227b3 100644
--- a/arch/powerpc/kernel/asm-offsets.c
+++ b/arch/powerpc/kernel/asm-offsets.c
@@ -190,6 +190,11 @@ int main(void)
190 DEFINE(PACA_SYSTEM_TIME, offsetof(struct paca_struct, system_time)); 190 DEFINE(PACA_SYSTEM_TIME, offsetof(struct paca_struct, system_time));
191 DEFINE(PACA_DATA_OFFSET, offsetof(struct paca_struct, data_offset)); 191 DEFINE(PACA_DATA_OFFSET, offsetof(struct paca_struct, data_offset));
192 DEFINE(PACA_TRAP_SAVE, offsetof(struct paca_struct, trap_save)); 192 DEFINE(PACA_TRAP_SAVE, offsetof(struct paca_struct, trap_save));
193#ifdef CONFIG_KVM_BOOK3S_64_HANDLER
194 DEFINE(PACA_KVM_IN_GUEST, offsetof(struct paca_struct, kvm_in_guest));
195 DEFINE(PACA_KVM_SLB, offsetof(struct paca_struct, kvm_slb));
196 DEFINE(PACA_KVM_SLB_MAX, offsetof(struct paca_struct, kvm_slb_max));
197#endif
193#endif /* CONFIG_PPC64 */ 198#endif /* CONFIG_PPC64 */
194 199
195 /* RTAS */ 200 /* RTAS */
@@ -398,14 +403,24 @@ int main(void)
398 DEFINE(VCPU_LAST_INST, offsetof(struct kvm_vcpu, arch.last_inst)); 403 DEFINE(VCPU_LAST_INST, offsetof(struct kvm_vcpu, arch.last_inst));
399 DEFINE(VCPU_FAULT_DEAR, offsetof(struct kvm_vcpu, arch.fault_dear)); 404 DEFINE(VCPU_FAULT_DEAR, offsetof(struct kvm_vcpu, arch.fault_dear));
400 DEFINE(VCPU_FAULT_ESR, offsetof(struct kvm_vcpu, arch.fault_esr)); 405 DEFINE(VCPU_FAULT_ESR, offsetof(struct kvm_vcpu, arch.fault_esr));
406
407 /* book3s_64 */
408#ifdef CONFIG_PPC64
409 DEFINE(VCPU_FAULT_DSISR, offsetof(struct kvm_vcpu, arch.fault_dsisr));
410 DEFINE(VCPU_HOST_RETIP, offsetof(struct kvm_vcpu, arch.host_retip));
411 DEFINE(VCPU_HOST_R2, offsetof(struct kvm_vcpu, arch.host_r2));
412 DEFINE(VCPU_HOST_MSR, offsetof(struct kvm_vcpu, arch.host_msr));
413 DEFINE(VCPU_SHADOW_MSR, offsetof(struct kvm_vcpu, arch.shadow_msr));
414 DEFINE(VCPU_TRAMPOLINE_LOWMEM, offsetof(struct kvm_vcpu, arch.trampoline_lowmem));
415 DEFINE(VCPU_TRAMPOLINE_ENTER, offsetof(struct kvm_vcpu, arch.trampoline_enter));
416 DEFINE(VCPU_HIGHMEM_HANDLER, offsetof(struct kvm_vcpu, arch.highmem_handler));
417 DEFINE(VCPU_HFLAGS, offsetof(struct kvm_vcpu, arch.hflags));
418#endif
401#endif 419#endif
402#ifdef CONFIG_44x 420#ifdef CONFIG_44x
403 DEFINE(PGD_T_LOG2, PGD_T_LOG2); 421 DEFINE(PGD_T_LOG2, PGD_T_LOG2);
404 DEFINE(PTE_T_LOG2, PTE_T_LOG2); 422 DEFINE(PTE_T_LOG2, PTE_T_LOG2);
405#endif 423#endif
406#ifdef CONFIG_FSL_BOOKE
407 DEFINE(TLBCAM_SIZE, sizeof(struct tlbcam));
408#endif
409 424
410#ifdef CONFIG_KVM_EXIT_TIMING 425#ifdef CONFIG_KVM_EXIT_TIMING
411 DEFINE(VCPU_TIMING_EXIT_TBU, offsetof(struct kvm_vcpu, 426 DEFINE(VCPU_TIMING_EXIT_TBU, offsetof(struct kvm_vcpu,
diff --git a/arch/powerpc/kernel/crash.c b/arch/powerpc/kernel/crash.c
index 0a8439aafdd1..6f4613dd05ef 100644
--- a/arch/powerpc/kernel/crash.c
+++ b/arch/powerpc/kernel/crash.c
@@ -373,7 +373,7 @@ void default_machine_crash_shutdown(struct pt_regs *regs)
373 hard_irq_disable(); 373 hard_irq_disable();
374 374
375 for_each_irq(i) { 375 for_each_irq(i) {
376 struct irq_desc *desc = irq_desc + i; 376 struct irq_desc *desc = irq_to_desc(i);
377 377
378 if (desc->status & IRQ_INPROGRESS) 378 if (desc->status & IRQ_INPROGRESS)
379 desc->chip->eoi(i); 379 desc->chip->eoi(i);
diff --git a/arch/powerpc/kernel/dma-swiotlb.c b/arch/powerpc/kernel/dma-swiotlb.c
index e96cbbd9b449..59c928564a03 100644
--- a/arch/powerpc/kernel/dma-swiotlb.c
+++ b/arch/powerpc/kernel/dma-swiotlb.c
@@ -21,7 +21,6 @@
21#include <asm/dma.h> 21#include <asm/dma.h>
22#include <asm/abs_addr.h> 22#include <asm/abs_addr.h>
23 23
24int swiotlb __read_mostly;
25unsigned int ppc_swiotlb_enable; 24unsigned int ppc_swiotlb_enable;
26 25
27/* 26/*
diff --git a/arch/powerpc/kernel/exceptions-64s.S b/arch/powerpc/kernel/exceptions-64s.S
index c7eb4e0eb86c..e3be98ffe2a7 100644
--- a/arch/powerpc/kernel/exceptions-64s.S
+++ b/arch/powerpc/kernel/exceptions-64s.S
@@ -41,6 +41,7 @@ __start_interrupts:
41 . = 0x200 41 . = 0x200
42_machine_check_pSeries: 42_machine_check_pSeries:
43 HMT_MEDIUM 43 HMT_MEDIUM
44 DO_KVM 0x200
44 mtspr SPRN_SPRG_SCRATCH0,r13 /* save r13 */ 45 mtspr SPRN_SPRG_SCRATCH0,r13 /* save r13 */
45 EXCEPTION_PROLOG_PSERIES(PACA_EXMC, machine_check_common) 46 EXCEPTION_PROLOG_PSERIES(PACA_EXMC, machine_check_common)
46 47
@@ -48,6 +49,7 @@ _machine_check_pSeries:
48 .globl data_access_pSeries 49 .globl data_access_pSeries
49data_access_pSeries: 50data_access_pSeries:
50 HMT_MEDIUM 51 HMT_MEDIUM
52 DO_KVM 0x300
51 mtspr SPRN_SPRG_SCRATCH0,r13 53 mtspr SPRN_SPRG_SCRATCH0,r13
52BEGIN_FTR_SECTION 54BEGIN_FTR_SECTION
53 mfspr r13,SPRN_SPRG_PACA 55 mfspr r13,SPRN_SPRG_PACA
@@ -77,6 +79,7 @@ ALT_FTR_SECTION_END_IFCLR(CPU_FTR_SLB)
77 .globl data_access_slb_pSeries 79 .globl data_access_slb_pSeries
78data_access_slb_pSeries: 80data_access_slb_pSeries:
79 HMT_MEDIUM 81 HMT_MEDIUM
82 DO_KVM 0x380
80 mtspr SPRN_SPRG_SCRATCH0,r13 83 mtspr SPRN_SPRG_SCRATCH0,r13
81 mfspr r13,SPRN_SPRG_PACA /* get paca address into r13 */ 84 mfspr r13,SPRN_SPRG_PACA /* get paca address into r13 */
82 std r3,PACA_EXSLB+EX_R3(r13) 85 std r3,PACA_EXSLB+EX_R3(r13)
@@ -115,6 +118,7 @@ data_access_slb_pSeries:
115 .globl instruction_access_slb_pSeries 118 .globl instruction_access_slb_pSeries
116instruction_access_slb_pSeries: 119instruction_access_slb_pSeries:
117 HMT_MEDIUM 120 HMT_MEDIUM
121 DO_KVM 0x480
118 mtspr SPRN_SPRG_SCRATCH0,r13 122 mtspr SPRN_SPRG_SCRATCH0,r13
119 mfspr r13,SPRN_SPRG_PACA /* get paca address into r13 */ 123 mfspr r13,SPRN_SPRG_PACA /* get paca address into r13 */
120 std r3,PACA_EXSLB+EX_R3(r13) 124 std r3,PACA_EXSLB+EX_R3(r13)
@@ -154,6 +158,7 @@ instruction_access_slb_pSeries:
154 .globl system_call_pSeries 158 .globl system_call_pSeries
155system_call_pSeries: 159system_call_pSeries:
156 HMT_MEDIUM 160 HMT_MEDIUM
161 DO_KVM 0xc00
157BEGIN_FTR_SECTION 162BEGIN_FTR_SECTION
158 cmpdi r0,0x1ebe 163 cmpdi r0,0x1ebe
159 beq- 1f 164 beq- 1f
@@ -187,14 +192,17 @@ END_FTR_SECTION_IFSET(CPU_FTR_REAL_LE)
187 */ 192 */
188performance_monitor_pSeries_1: 193performance_monitor_pSeries_1:
189 . = 0xf00 194 . = 0xf00
195 DO_KVM 0xf00
190 b performance_monitor_pSeries 196 b performance_monitor_pSeries
191 197
192altivec_unavailable_pSeries_1: 198altivec_unavailable_pSeries_1:
193 . = 0xf20 199 . = 0xf20
200 DO_KVM 0xf20
194 b altivec_unavailable_pSeries 201 b altivec_unavailable_pSeries
195 202
196vsx_unavailable_pSeries_1: 203vsx_unavailable_pSeries_1:
197 . = 0xf40 204 . = 0xf40
205 DO_KVM 0xf40
198 b vsx_unavailable_pSeries 206 b vsx_unavailable_pSeries
199 207
200#ifdef CONFIG_CBE_RAS 208#ifdef CONFIG_CBE_RAS
diff --git a/arch/powerpc/kernel/head_64.S b/arch/powerpc/kernel/head_64.S
index c38afdb45d7b..925807488022 100644
--- a/arch/powerpc/kernel/head_64.S
+++ b/arch/powerpc/kernel/head_64.S
@@ -37,6 +37,7 @@
37#include <asm/firmware.h> 37#include <asm/firmware.h>
38#include <asm/page_64.h> 38#include <asm/page_64.h>
39#include <asm/irqflags.h> 39#include <asm/irqflags.h>
40#include <asm/kvm_book3s_64_asm.h>
40 41
41/* The physical memory is layed out such that the secondary processor 42/* The physical memory is layed out such that the secondary processor
42 * spin code sits at 0x0000...0x00ff. On server, the vectors follow 43 * spin code sits at 0x0000...0x00ff. On server, the vectors follow
@@ -165,6 +166,12 @@ exception_marker:
165#include "exceptions-64s.S" 166#include "exceptions-64s.S"
166#endif 167#endif
167 168
169/* KVM trampoline code needs to be close to the interrupt handlers */
170
171#ifdef CONFIG_KVM_BOOK3S_64_HANDLER
172#include "../kvm/book3s_64_rmhandlers.S"
173#endif
174
168_GLOBAL(generic_secondary_thread_init) 175_GLOBAL(generic_secondary_thread_init)
169 mr r24,r3 176 mr r24,r3
170 177
diff --git a/arch/powerpc/kernel/head_8xx.S b/arch/powerpc/kernel/head_8xx.S
index 6ded19d01891..678f98cd5e64 100644
--- a/arch/powerpc/kernel/head_8xx.S
+++ b/arch/powerpc/kernel/head_8xx.S
@@ -206,6 +206,8 @@ MachineCheck:
206 EXCEPTION_PROLOG 206 EXCEPTION_PROLOG
207 mfspr r4,SPRN_DAR 207 mfspr r4,SPRN_DAR
208 stw r4,_DAR(r11) 208 stw r4,_DAR(r11)
209 li r5,0x00f0
210 mtspr SPRN_DAR,r5 /* Tag DAR, to be used in DTLB Error */
209 mfspr r5,SPRN_DSISR 211 mfspr r5,SPRN_DSISR
210 stw r5,_DSISR(r11) 212 stw r5,_DSISR(r11)
211 addi r3,r1,STACK_FRAME_OVERHEAD 213 addi r3,r1,STACK_FRAME_OVERHEAD
@@ -222,6 +224,8 @@ DataAccess:
222 stw r10,_DSISR(r11) 224 stw r10,_DSISR(r11)
223 mr r5,r10 225 mr r5,r10
224 mfspr r4,SPRN_DAR 226 mfspr r4,SPRN_DAR
227 li r10,0x00f0
228 mtspr SPRN_DAR,r10 /* Tag DAR, to be used in DTLB Error */
225 EXC_XFER_EE_LITE(0x300, handle_page_fault) 229 EXC_XFER_EE_LITE(0x300, handle_page_fault)
226 230
227/* Instruction access exception. 231/* Instruction access exception.
@@ -244,6 +248,8 @@ Alignment:
244 EXCEPTION_PROLOG 248 EXCEPTION_PROLOG
245 mfspr r4,SPRN_DAR 249 mfspr r4,SPRN_DAR
246 stw r4,_DAR(r11) 250 stw r4,_DAR(r11)
251 li r5,0x00f0
252 mtspr SPRN_DAR,r5 /* Tag DAR, to be used in DTLB Error */
247 mfspr r5,SPRN_DSISR 253 mfspr r5,SPRN_DSISR
248 stw r5,_DSISR(r11) 254 stw r5,_DSISR(r11)
249 addi r3,r1,STACK_FRAME_OVERHEAD 255 addi r3,r1,STACK_FRAME_OVERHEAD
@@ -333,26 +339,20 @@ InstructionTLBMiss:
333 mfspr r11, SPRN_MD_TWC /* ....and get the pte address */ 339 mfspr r11, SPRN_MD_TWC /* ....and get the pte address */
334 lwz r10, 0(r11) /* Get the pte */ 340 lwz r10, 0(r11) /* Get the pte */
335 341
336#ifdef CONFIG_SWAP 342 andi. r11, r10, _PAGE_ACCESSED | _PAGE_PRESENT
337 /* do not set the _PAGE_ACCESSED bit of a non-present page */ 343 cmpwi cr0, r11, _PAGE_ACCESSED | _PAGE_PRESENT
338 andi. r11, r10, _PAGE_PRESENT 344 bne- cr0, 2f
339 beq 4f 345
340 ori r10, r10, _PAGE_ACCESSED 346 /* Clear PP lsb, 0x400 */
341 mfspr r11, SPRN_MD_TWC /* get the pte address again */ 347 rlwinm r10, r10, 0, 22, 20
342 stw r10, 0(r11)
3434:
344#else
345 ori r10, r10, _PAGE_ACCESSED
346 stw r10, 0(r11)
347#endif
348 348
349 /* The Linux PTE won't go exactly into the MMU TLB. 349 /* The Linux PTE won't go exactly into the MMU TLB.
350 * Software indicator bits 21, 22 and 28 must be clear. 350 * Software indicator bits 22 and 28 must be clear.
351 * Software indicator bits 24, 25, 26, and 27 must be 351 * Software indicator bits 24, 25, 26, and 27 must be
352 * set. All other Linux PTE bits control the behavior 352 * set. All other Linux PTE bits control the behavior
353 * of the MMU. 353 * of the MMU.
354 */ 354 */
3552: li r11, 0x00f0 355 li r11, 0x00f0
356 rlwimi r10, r11, 0, 24, 28 /* Set 24-27, clear 28 */ 356 rlwimi r10, r11, 0, 24, 28 /* Set 24-27, clear 28 */
357 DO_8xx_CPU6(0x2d80, r3) 357 DO_8xx_CPU6(0x2d80, r3)
358 mtspr SPRN_MI_RPN, r10 /* Update TLB entry */ 358 mtspr SPRN_MI_RPN, r10 /* Update TLB entry */
@@ -365,6 +365,22 @@ InstructionTLBMiss:
365 lwz r3, 8(r0) 365 lwz r3, 8(r0)
366#endif 366#endif
367 rfi 367 rfi
3682:
369 mfspr r11, SPRN_SRR1
370 /* clear all error bits as TLB Miss
371 * sets a few unconditionally
372 */
373 rlwinm r11, r11, 0, 0xffff
374 mtspr SPRN_SRR1, r11
375
376 mfspr r10, SPRN_M_TW /* Restore registers */
377 lwz r11, 0(r0)
378 mtcr r11
379 lwz r11, 4(r0)
380#ifdef CONFIG_8xx_CPU6
381 lwz r3, 8(r0)
382#endif
383 b InstructionAccess
368 384
369 . = 0x1200 385 . = 0x1200
370DataStoreTLBMiss: 386DataStoreTLBMiss:
@@ -406,29 +422,45 @@ DataStoreTLBMiss:
406 * above. 422 * above.
407 */ 423 */
408 rlwimi r11, r10, 0, 27, 27 424 rlwimi r11, r10, 0, 27, 27
425 /* Insert the WriteThru flag into the TWC from the Linux PTE.
426 * It is bit 25 in the Linux PTE and bit 30 in the TWC
427 */
428 rlwimi r11, r10, 32-5, 30, 30
409 DO_8xx_CPU6(0x3b80, r3) 429 DO_8xx_CPU6(0x3b80, r3)
410 mtspr SPRN_MD_TWC, r11 430 mtspr SPRN_MD_TWC, r11
411 431
412#ifdef CONFIG_SWAP 432 /* Both _PAGE_ACCESSED and _PAGE_PRESENT has to be set.
413 /* do not set the _PAGE_ACCESSED bit of a non-present page */ 433 * We also need to know if the insn is a load/store, so:
414 andi. r11, r10, _PAGE_PRESENT 434 * Clear _PAGE_PRESENT and load that which will
415 beq 4f 435 * trap into DTLB Error with store bit set accordinly.
416 ori r10, r10, _PAGE_ACCESSED 436 */
4174: 437 /* PRESENT=0x1, ACCESSED=0x20
418 /* and update pte in table */ 438 * r11 = ((r10 & PRESENT) & ((r10 & ACCESSED) >> 5));
419#else 439 * r10 = (r10 & ~PRESENT) | r11;
420 ori r10, r10, _PAGE_ACCESSED 440 */
421#endif 441 rlwinm r11, r10, 32-5, _PAGE_PRESENT
422 mfspr r11, SPRN_MD_TWC /* get the pte address again */ 442 and r11, r11, r10
423 stw r10, 0(r11) 443 rlwimi r10, r11, 0, _PAGE_PRESENT
444
445 /* Honour kernel RO, User NA */
446 /* 0x200 == Extended encoding, bit 22 */
447 /* r11 = (r10 & _PAGE_USER) >> 2 */
448 rlwinm r11, r10, 32-2, 0x200
449 or r10, r11, r10
450 /* r11 = (r10 & _PAGE_RW) >> 1 */
451 rlwinm r11, r10, 32-1, 0x200
452 or r10, r11, r10
453 /* invert RW and 0x200 bits */
454 xori r10, r10, _PAGE_RW | 0x200
424 455
425 /* The Linux PTE won't go exactly into the MMU TLB. 456 /* The Linux PTE won't go exactly into the MMU TLB.
426 * Software indicator bits 21, 22 and 28 must be clear. 457 * Software indicator bits 22 and 28 must be clear.
427 * Software indicator bits 24, 25, 26, and 27 must be 458 * Software indicator bits 24, 25, 26, and 27 must be
428 * set. All other Linux PTE bits control the behavior 459 * set. All other Linux PTE bits control the behavior
429 * of the MMU. 460 * of the MMU.
430 */ 461 */
4312: li r11, 0x00f0 4622: li r11, 0x00f0
463 mtspr SPRN_DAR,r11 /* Tag DAR */
432 rlwimi r10, r11, 0, 24, 28 /* Set 24-27, clear 28 */ 464 rlwimi r10, r11, 0, 24, 28 /* Set 24-27, clear 28 */
433 DO_8xx_CPU6(0x3d80, r3) 465 DO_8xx_CPU6(0x3d80, r3)
434 mtspr SPRN_MD_RPN, r10 /* Update TLB entry */ 466 mtspr SPRN_MD_RPN, r10 /* Update TLB entry */
@@ -469,97 +501,10 @@ DataTLBError:
469 stw r10, 0(r0) 501 stw r10, 0(r0)
470 stw r11, 4(r0) 502 stw r11, 4(r0)
471 503
472 /* First, make sure this was a store operation.
473 */
474 mfspr r10, SPRN_DSISR
475 andis. r11, r10, 0x0200 /* If set, indicates store op */
476 beq 2f
477
478 /* The EA of a data TLB miss is automatically stored in the MD_EPN
479 * register. The EA of a data TLB error is automatically stored in
480 * the DAR, but not the MD_EPN register. We must copy the 20 most
481 * significant bits of the EA from the DAR to MD_EPN before we
482 * start walking the page tables. We also need to copy the CASID
483 * value from the M_CASID register.
484 * Addendum: The EA of a data TLB error is _supposed_ to be stored
485 * in DAR, but it seems that this doesn't happen in some cases, such
486 * as when the error is due to a dcbi instruction to a page with a
487 * TLB that doesn't have the changed bit set. In such cases, there
488 * does not appear to be any way to recover the EA of the error
489 * since it is neither in DAR nor MD_EPN. As a workaround, the
490 * _PAGE_HWWRITE bit is set for all kernel data pages when the PTEs
491 * are initialized in mapin_ram(). This will avoid the problem,
492 * assuming we only use the dcbi instruction on kernel addresses.
493 */
494 mfspr r10, SPRN_DAR 504 mfspr r10, SPRN_DAR
495 rlwinm r11, r10, 0, 0, 19 505 cmpwi cr0, r10, 0x00f0
496 ori r11, r11, MD_EVALID 506 beq- FixupDAR /* must be a buggy dcbX, icbi insn. */
497 mfspr r10, SPRN_M_CASID 507DARFixed:/* Return from dcbx instruction bug workaround, r10 holds value of DAR */
498 rlwimi r11, r10, 0, 28, 31
499 DO_8xx_CPU6(0x3780, r3)
500 mtspr SPRN_MD_EPN, r11
501
502 mfspr r10, SPRN_M_TWB /* Get level 1 table entry address */
503
504 /* If we are faulting a kernel address, we have to use the
505 * kernel page tables.
506 */
507 andi. r11, r10, 0x0800
508 beq 3f
509 lis r11, swapper_pg_dir@h
510 ori r11, r11, swapper_pg_dir@l
511 rlwimi r10, r11, 0, 2, 19
5123:
513 lwz r11, 0(r10) /* Get the level 1 entry */
514 rlwinm. r10, r11,0,0,19 /* Extract page descriptor page address */
515 beq 2f /* If zero, bail */
516
517 /* We have a pte table, so fetch the pte from the table.
518 */
519 ori r11, r11, 1 /* Set valid bit in physical L2 page */
520 DO_8xx_CPU6(0x3b80, r3)
521 mtspr SPRN_MD_TWC, r11 /* Load pte table base address */
522 mfspr r11, SPRN_MD_TWC /* ....and get the pte address */
523 lwz r10, 0(r11) /* Get the pte */
524
525 andi. r11, r10, _PAGE_RW /* Is it writeable? */
526 beq 2f /* Bail out if not */
527
528 /* Update 'changed', among others.
529 */
530#ifdef CONFIG_SWAP
531 ori r10, r10, _PAGE_DIRTY|_PAGE_HWWRITE
532 /* do not set the _PAGE_ACCESSED bit of a non-present page */
533 andi. r11, r10, _PAGE_PRESENT
534 beq 4f
535 ori r10, r10, _PAGE_ACCESSED
5364:
537#else
538 ori r10, r10, _PAGE_DIRTY|_PAGE_ACCESSED|_PAGE_HWWRITE
539#endif
540 mfspr r11, SPRN_MD_TWC /* Get pte address again */
541 stw r10, 0(r11) /* and update pte in table */
542
543 /* The Linux PTE won't go exactly into the MMU TLB.
544 * Software indicator bits 21, 22 and 28 must be clear.
545 * Software indicator bits 24, 25, 26, and 27 must be
546 * set. All other Linux PTE bits control the behavior
547 * of the MMU.
548 */
549 li r11, 0x00f0
550 rlwimi r10, r11, 0, 24, 28 /* Set 24-27, clear 28 */
551 DO_8xx_CPU6(0x3d80, r3)
552 mtspr SPRN_MD_RPN, r10 /* Update TLB entry */
553
554 mfspr r10, SPRN_M_TW /* Restore registers */
555 lwz r11, 0(r0)
556 mtcr r11
557 lwz r11, 4(r0)
558#ifdef CONFIG_8xx_CPU6
559 lwz r3, 8(r0)
560#endif
561 rfi
5622:
563 mfspr r10, SPRN_M_TW /* Restore registers */ 508 mfspr r10, SPRN_M_TW /* Restore registers */
564 lwz r11, 0(r0) 509 lwz r11, 0(r0)
565 mtcr r11 510 mtcr r11
@@ -588,6 +533,140 @@ DataTLBError:
588 533
589 . = 0x2000 534 . = 0x2000
590 535
536/* This is the procedure to calculate the data EA for buggy dcbx,dcbi instructions
537 * by decoding the registers used by the dcbx instruction and adding them.
538 * DAR is set to the calculated address and r10 also holds the EA on exit.
539 */
540 /* define if you don't want to use self modifying code */
541#define NO_SELF_MODIFYING_CODE
542FixupDAR:/* Entry point for dcbx workaround. */
543 /* fetch instruction from memory. */
544 mfspr r10, SPRN_SRR0
545 DO_8xx_CPU6(0x3780, r3)
546 mtspr SPRN_MD_EPN, r10
547 mfspr r11, SPRN_M_TWB /* Get level 1 table entry address */
548 cmplwi cr0, r11, 0x0800
549 blt- 3f /* Branch if user space */
550 lis r11, (swapper_pg_dir-PAGE_OFFSET)@h
551 ori r11, r11, (swapper_pg_dir-PAGE_OFFSET)@l
552 rlwimi r11, r10, 32-20, 0xffc /* r11 = r11&~0xffc|(r10>>20)&0xffc */
5533: lwz r11, 0(r11) /* Get the level 1 entry */
554 DO_8xx_CPU6(0x3b80, r3)
555 mtspr SPRN_MD_TWC, r11 /* Load pte table base address */
556 mfspr r11, SPRN_MD_TWC /* ....and get the pte address */
557 lwz r11, 0(r11) /* Get the pte */
558 /* concat physical page address(r11) and page offset(r10) */
559 rlwimi r11, r10, 0, 20, 31
560 lwz r11,0(r11)
561/* Check if it really is a dcbx instruction. */
562/* dcbt and dcbtst does not generate DTLB Misses/Errors,
563 * no need to include them here */
564 srwi r10, r11, 26 /* check if major OP code is 31 */
565 cmpwi cr0, r10, 31
566 bne- 141f
567 rlwinm r10, r11, 0, 21, 30
568 cmpwi cr0, r10, 2028 /* Is dcbz? */
569 beq+ 142f
570 cmpwi cr0, r10, 940 /* Is dcbi? */
571 beq+ 142f
572 cmpwi cr0, r10, 108 /* Is dcbst? */
573 beq+ 144f /* Fix up store bit! */
574 cmpwi cr0, r10, 172 /* Is dcbf? */
575 beq+ 142f
576 cmpwi cr0, r10, 1964 /* Is icbi? */
577 beq+ 142f
578141: mfspr r10, SPRN_DAR /* r10 must hold DAR at exit */
579 b DARFixed /* Nope, go back to normal TLB processing */
580
581144: mfspr r10, SPRN_DSISR
582 rlwinm r10, r10,0,7,5 /* Clear store bit for buggy dcbst insn */
583 mtspr SPRN_DSISR, r10
584142: /* continue, it was a dcbx, dcbi instruction. */
585#ifdef CONFIG_8xx_CPU6
586 lwz r3, 8(r0) /* restore r3 from memory */
587#endif
588#ifndef NO_SELF_MODIFYING_CODE
589 andis. r10,r11,0x1f /* test if reg RA is r0 */
590 li r10,modified_instr@l
591 dcbtst r0,r10 /* touch for store */
592 rlwinm r11,r11,0,0,20 /* Zero lower 10 bits */
593 oris r11,r11,640 /* Transform instr. to a "add r10,RA,RB" */
594 ori r11,r11,532
595 stw r11,0(r10) /* store add/and instruction */
596 dcbf 0,r10 /* flush new instr. to memory. */
597 icbi 0,r10 /* invalidate instr. cache line */
598 lwz r11, 4(r0) /* restore r11 from memory */
599 mfspr r10, SPRN_M_TW /* restore r10 from M_TW */
600 isync /* Wait until new instr is loaded from memory */
601modified_instr:
602 .space 4 /* this is where the add instr. is stored */
603 bne+ 143f
604 subf r10,r0,r10 /* r10=r10-r0, only if reg RA is r0 */
605143: mtdar r10 /* store faulting EA in DAR */
606 b DARFixed /* Go back to normal TLB handling */
607#else
608 mfctr r10
609 mtdar r10 /* save ctr reg in DAR */
610 rlwinm r10, r11, 24, 24, 28 /* offset into jump table for reg RB */
611 addi r10, r10, 150f@l /* add start of table */
612 mtctr r10 /* load ctr with jump address */
613 xor r10, r10, r10 /* sum starts at zero */
614 bctr /* jump into table */
615150:
616 add r10, r10, r0 ;b 151f
617 add r10, r10, r1 ;b 151f
618 add r10, r10, r2 ;b 151f
619 add r10, r10, r3 ;b 151f
620 add r10, r10, r4 ;b 151f
621 add r10, r10, r5 ;b 151f
622 add r10, r10, r6 ;b 151f
623 add r10, r10, r7 ;b 151f
624 add r10, r10, r8 ;b 151f
625 add r10, r10, r9 ;b 151f
626 mtctr r11 ;b 154f /* r10 needs special handling */
627 mtctr r11 ;b 153f /* r11 needs special handling */
628 add r10, r10, r12 ;b 151f
629 add r10, r10, r13 ;b 151f
630 add r10, r10, r14 ;b 151f
631 add r10, r10, r15 ;b 151f
632 add r10, r10, r16 ;b 151f
633 add r10, r10, r17 ;b 151f
634 add r10, r10, r18 ;b 151f
635 add r10, r10, r19 ;b 151f
636 add r10, r10, r20 ;b 151f
637 add r10, r10, r21 ;b 151f
638 add r10, r10, r22 ;b 151f
639 add r10, r10, r23 ;b 151f
640 add r10, r10, r24 ;b 151f
641 add r10, r10, r25 ;b 151f
642 add r10, r10, r26 ;b 151f
643 add r10, r10, r27 ;b 151f
644 add r10, r10, r28 ;b 151f
645 add r10, r10, r29 ;b 151f
646 add r10, r10, r30 ;b 151f
647 add r10, r10, r31
648151:
649 rlwinm. r11,r11,19,24,28 /* offset into jump table for reg RA */
650 beq 152f /* if reg RA is zero, don't add it */
651 addi r11, r11, 150b@l /* add start of table */
652 mtctr r11 /* load ctr with jump address */
653 rlwinm r11,r11,0,16,10 /* make sure we don't execute this more than once */
654 bctr /* jump into table */
655152:
656 mfdar r11
657 mtctr r11 /* restore ctr reg from DAR */
658 mtdar r10 /* save fault EA to DAR */
659 b DARFixed /* Go back to normal TLB handling */
660
661 /* special handling for r10,r11 since these are modified already */
662153: lwz r11, 4(r0) /* load r11 from memory */
663 b 155f
664154: mfspr r11, SPRN_M_TW /* load r10 from M_TW */
665155: add r10, r10, r11 /* add it */
666 mfctr r11 /* restore r11 */
667 b 151b
668#endif
669
591 .globl giveup_fpu 670 .globl giveup_fpu
592giveup_fpu: 671giveup_fpu:
593 blr 672 blr
diff --git a/arch/powerpc/kernel/head_fsl_booke.S b/arch/powerpc/kernel/head_fsl_booke.S
index 975788ca05d2..7f4bd7f3b6af 100644
--- a/arch/powerpc/kernel/head_fsl_booke.S
+++ b/arch/powerpc/kernel/head_fsl_booke.S
@@ -944,28 +944,6 @@ _GLOBAL(__setup_e500mc_ivors)
944 blr 944 blr
945 945
946/* 946/*
947 * extern void loadcam_entry(unsigned int index)
948 *
949 * Load TLBCAM[index] entry in to the L2 CAM MMU
950 */
951_GLOBAL(loadcam_entry)
952 lis r4,TLBCAM@ha
953 addi r4,r4,TLBCAM@l
954 mulli r5,r3,TLBCAM_SIZE
955 add r3,r5,r4
956 lwz r4,0(r3)
957 mtspr SPRN_MAS0,r4
958 lwz r4,4(r3)
959 mtspr SPRN_MAS1,r4
960 lwz r4,8(r3)
961 mtspr SPRN_MAS2,r4
962 lwz r4,12(r3)
963 mtspr SPRN_MAS3,r4
964 tlbwe
965 isync
966 blr
967
968/*
969 * extern void giveup_altivec(struct task_struct *prev) 947 * extern void giveup_altivec(struct task_struct *prev)
970 * 948 *
971 * The e500 core does not have an AltiVec unit. 949 * The e500 core does not have an AltiVec unit.
diff --git a/arch/powerpc/kernel/io.c b/arch/powerpc/kernel/io.c
index 1882bf419fa6..8dc7547c2377 100644
--- a/arch/powerpc/kernel/io.c
+++ b/arch/powerpc/kernel/io.c
@@ -161,7 +161,7 @@ void _memcpy_fromio(void *dest, const volatile void __iomem *src,
161 dest++; 161 dest++;
162 n--; 162 n--;
163 } 163 }
164 while(n > 4) { 164 while(n >= 4) {
165 *((u32 *)dest) = *((volatile u32 *)vsrc); 165 *((u32 *)dest) = *((volatile u32 *)vsrc);
166 eieio(); 166 eieio();
167 vsrc += 4; 167 vsrc += 4;
@@ -190,7 +190,7 @@ void _memcpy_toio(volatile void __iomem *dest, const void *src, unsigned long n)
190 vdest++; 190 vdest++;
191 n--; 191 n--;
192 } 192 }
193 while(n > 4) { 193 while(n >= 4) {
194 *((volatile u32 *)vdest) = *((volatile u32 *)src); 194 *((volatile u32 *)vdest) = *((volatile u32 *)src);
195 src += 4; 195 src += 4;
196 vdest += 4; 196 vdest += 4;
diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c
index 02a334662cc0..f6dca4f4b295 100644
--- a/arch/powerpc/kernel/irq.c
+++ b/arch/powerpc/kernel/irq.c
@@ -87,7 +87,10 @@ extern int tau_interrupts(int);
87#endif /* CONFIG_PPC32 */ 87#endif /* CONFIG_PPC32 */
88 88
89#ifdef CONFIG_PPC64 89#ifdef CONFIG_PPC64
90
91#ifndef CONFIG_SPARSE_IRQ
90EXPORT_SYMBOL(irq_desc); 92EXPORT_SYMBOL(irq_desc);
93#endif
91 94
92int distribute_irqs = 1; 95int distribute_irqs = 1;
93 96
@@ -189,33 +192,7 @@ int show_interrupts(struct seq_file *p, void *v)
189 for_each_online_cpu(j) 192 for_each_online_cpu(j)
190 seq_printf(p, "CPU%d ", j); 193 seq_printf(p, "CPU%d ", j);
191 seq_putc(p, '\n'); 194 seq_putc(p, '\n');
192 } 195 } else if (i == nr_irqs) {
193
194 if (i < NR_IRQS) {
195 desc = get_irq_desc(i);
196 spin_lock_irqsave(&desc->lock, flags);
197 action = desc->action;
198 if (!action || !action->handler)
199 goto skip;
200 seq_printf(p, "%3d: ", i);
201#ifdef CONFIG_SMP
202 for_each_online_cpu(j)
203 seq_printf(p, "%10u ", kstat_irqs_cpu(i, j));
204#else
205 seq_printf(p, "%10u ", kstat_irqs(i));
206#endif /* CONFIG_SMP */
207 if (desc->chip)
208 seq_printf(p, " %s ", desc->chip->typename);
209 else
210 seq_puts(p, " None ");
211 seq_printf(p, "%s", (desc->status & IRQ_LEVEL) ? "Level " : "Edge ");
212 seq_printf(p, " %s", action->name);
213 for (action = action->next; action; action = action->next)
214 seq_printf(p, ", %s", action->name);
215 seq_putc(p, '\n');
216skip:
217 spin_unlock_irqrestore(&desc->lock, flags);
218 } else if (i == NR_IRQS) {
219#if defined(CONFIG_PPC32) && defined(CONFIG_TAU_INT) 196#if defined(CONFIG_PPC32) && defined(CONFIG_TAU_INT)
220 if (tau_initialized){ 197 if (tau_initialized){
221 seq_puts(p, "TAU: "); 198 seq_puts(p, "TAU: ");
@@ -225,30 +202,68 @@ skip:
225 } 202 }
226#endif /* CONFIG_PPC32 && CONFIG_TAU_INT*/ 203#endif /* CONFIG_PPC32 && CONFIG_TAU_INT*/
227 seq_printf(p, "BAD: %10u\n", ppc_spurious_interrupts); 204 seq_printf(p, "BAD: %10u\n", ppc_spurious_interrupts);
205
206 return 0;
228 } 207 }
208
209 desc = irq_to_desc(i);
210 if (!desc)
211 return 0;
212
213 spin_lock_irqsave(&desc->lock, flags);
214
215 action = desc->action;
216 if (!action || !action->handler)
217 goto skip;
218
219 seq_printf(p, "%3d: ", i);
220#ifdef CONFIG_SMP
221 for_each_online_cpu(j)
222 seq_printf(p, "%10u ", kstat_irqs_cpu(i, j));
223#else
224 seq_printf(p, "%10u ", kstat_irqs(i));
225#endif /* CONFIG_SMP */
226
227 if (desc->chip)
228 seq_printf(p, " %s ", desc->chip->name);
229 else
230 seq_puts(p, " None ");
231
232 seq_printf(p, "%s", (desc->status & IRQ_LEVEL) ? "Level " : "Edge ");
233 seq_printf(p, " %s", action->name);
234
235 for (action = action->next; action; action = action->next)
236 seq_printf(p, ", %s", action->name);
237 seq_putc(p, '\n');
238
239skip:
240 spin_unlock_irqrestore(&desc->lock, flags);
241
229 return 0; 242 return 0;
230} 243}
231 244
232#ifdef CONFIG_HOTPLUG_CPU 245#ifdef CONFIG_HOTPLUG_CPU
233void fixup_irqs(cpumask_t map) 246void fixup_irqs(cpumask_t map)
234{ 247{
248 struct irq_desc *desc;
235 unsigned int irq; 249 unsigned int irq;
236 static int warned; 250 static int warned;
237 251
238 for_each_irq(irq) { 252 for_each_irq(irq) {
239 cpumask_t mask; 253 cpumask_t mask;
240 254
241 if (irq_desc[irq].status & IRQ_PER_CPU) 255 desc = irq_to_desc(irq);
256 if (desc && desc->status & IRQ_PER_CPU)
242 continue; 257 continue;
243 258
244 cpumask_and(&mask, irq_desc[irq].affinity, &map); 259 cpumask_and(&mask, desc->affinity, &map);
245 if (any_online_cpu(mask) == NR_CPUS) { 260 if (any_online_cpu(mask) == NR_CPUS) {
246 printk("Breaking affinity for irq %i\n", irq); 261 printk("Breaking affinity for irq %i\n", irq);
247 mask = map; 262 mask = map;
248 } 263 }
249 if (irq_desc[irq].chip->set_affinity) 264 if (desc->chip->set_affinity)
250 irq_desc[irq].chip->set_affinity(irq, &mask); 265 desc->chip->set_affinity(irq, &mask);
251 else if (irq_desc[irq].action && !(warned++)) 266 else if (desc->action && !(warned++))
252 printk("Cannot set affinity for irq %i\n", irq); 267 printk("Cannot set affinity for irq %i\n", irq);
253 } 268 }
254 269
@@ -275,7 +290,7 @@ static inline void handle_one_irq(unsigned int irq)
275 return; 290 return;
276 } 291 }
277 292
278 desc = irq_desc + irq; 293 desc = irq_to_desc(irq);
279 saved_sp_limit = current->thread.ksp_limit; 294 saved_sp_limit = current->thread.ksp_limit;
280 295
281 irqtp->task = curtp->task; 296 irqtp->task = curtp->task;
@@ -541,7 +556,7 @@ struct irq_host *irq_alloc_host(struct device_node *of_node,
541 smp_wmb(); 556 smp_wmb();
542 557
543 /* Clear norequest flags */ 558 /* Clear norequest flags */
544 get_irq_desc(i)->status &= ~IRQ_NOREQUEST; 559 irq_to_desc(i)->status &= ~IRQ_NOREQUEST;
545 560
546 /* Legacy flags are left to default at this point, 561 /* Legacy flags are left to default at this point,
547 * one can then use irq_create_mapping() to 562 * one can then use irq_create_mapping() to
@@ -607,8 +622,16 @@ void irq_set_virq_count(unsigned int count)
607static int irq_setup_virq(struct irq_host *host, unsigned int virq, 622static int irq_setup_virq(struct irq_host *host, unsigned int virq,
608 irq_hw_number_t hwirq) 623 irq_hw_number_t hwirq)
609{ 624{
625 struct irq_desc *desc;
626
627 desc = irq_to_desc_alloc_node(virq, 0);
628 if (!desc) {
629 pr_debug("irq: -> allocating desc failed\n");
630 goto error;
631 }
632
610 /* Clear IRQ_NOREQUEST flag */ 633 /* Clear IRQ_NOREQUEST flag */
611 get_irq_desc(virq)->status &= ~IRQ_NOREQUEST; 634 desc->status &= ~IRQ_NOREQUEST;
612 635
613 /* map it */ 636 /* map it */
614 smp_wmb(); 637 smp_wmb();
@@ -617,11 +640,14 @@ static int irq_setup_virq(struct irq_host *host, unsigned int virq,
617 640
618 if (host->ops->map(host, virq, hwirq)) { 641 if (host->ops->map(host, virq, hwirq)) {
619 pr_debug("irq: -> mapping failed, freeing\n"); 642 pr_debug("irq: -> mapping failed, freeing\n");
620 irq_free_virt(virq, 1); 643 goto error;
621 return -1;
622 } 644 }
623 645
624 return 0; 646 return 0;
647
648error:
649 irq_free_virt(virq, 1);
650 return -1;
625} 651}
626 652
627unsigned int irq_create_direct_mapping(struct irq_host *host) 653unsigned int irq_create_direct_mapping(struct irq_host *host)
@@ -705,7 +731,7 @@ unsigned int irq_create_mapping(struct irq_host *host,
705EXPORT_SYMBOL_GPL(irq_create_mapping); 731EXPORT_SYMBOL_GPL(irq_create_mapping);
706 732
707unsigned int irq_create_of_mapping(struct device_node *controller, 733unsigned int irq_create_of_mapping(struct device_node *controller,
708 u32 *intspec, unsigned int intsize) 734 const u32 *intspec, unsigned int intsize)
709{ 735{
710 struct irq_host *host; 736 struct irq_host *host;
711 irq_hw_number_t hwirq; 737 irq_hw_number_t hwirq;
@@ -738,7 +764,7 @@ unsigned int irq_create_of_mapping(struct device_node *controller,
738 764
739 /* Set type if specified and different than the current one */ 765 /* Set type if specified and different than the current one */
740 if (type != IRQ_TYPE_NONE && 766 if (type != IRQ_TYPE_NONE &&
741 type != (get_irq_desc(virq)->status & IRQF_TRIGGER_MASK)) 767 type != (irq_to_desc(virq)->status & IRQF_TRIGGER_MASK))
742 set_irq_type(virq, type); 768 set_irq_type(virq, type);
743 return virq; 769 return virq;
744} 770}
@@ -810,7 +836,7 @@ void irq_dispose_mapping(unsigned int virq)
810 irq_map[virq].hwirq = host->inval_irq; 836 irq_map[virq].hwirq = host->inval_irq;
811 837
812 /* Set some flags */ 838 /* Set some flags */
813 get_irq_desc(virq)->status |= IRQ_NOREQUEST; 839 irq_to_desc(virq)->status |= IRQ_NOREQUEST;
814 840
815 /* Free it */ 841 /* Free it */
816 irq_free_virt(virq, 1); 842 irq_free_virt(virq, 1);
@@ -1002,12 +1028,24 @@ void irq_free_virt(unsigned int virq, unsigned int count)
1002 spin_unlock_irqrestore(&irq_big_lock, flags); 1028 spin_unlock_irqrestore(&irq_big_lock, flags);
1003} 1029}
1004 1030
1005void irq_early_init(void) 1031int arch_early_irq_init(void)
1006{ 1032{
1007 unsigned int i; 1033 struct irq_desc *desc;
1034 int i;
1035
1036 for (i = 0; i < NR_IRQS; i++) {
1037 desc = irq_to_desc(i);
1038 if (desc)
1039 desc->status |= IRQ_NOREQUEST;
1040 }
1008 1041
1009 for (i = 0; i < NR_IRQS; i++) 1042 return 0;
1010 get_irq_desc(i)->status |= IRQ_NOREQUEST; 1043}
1044
1045int arch_init_chip_data(struct irq_desc *desc, int node)
1046{
1047 desc->status |= IRQ_NOREQUEST;
1048 return 0;
1011} 1049}
1012 1050
1013/* We need to create the radix trees late */ 1051/* We need to create the radix trees late */
@@ -1069,16 +1107,19 @@ static int virq_debug_show(struct seq_file *m, void *private)
1069 seq_printf(m, "%-5s %-7s %-15s %s\n", "virq", "hwirq", 1107 seq_printf(m, "%-5s %-7s %-15s %s\n", "virq", "hwirq",
1070 "chip name", "host name"); 1108 "chip name", "host name");
1071 1109
1072 for (i = 1; i < NR_IRQS; i++) { 1110 for (i = 1; i < nr_irqs; i++) {
1073 desc = get_irq_desc(i); 1111 desc = irq_to_desc(i);
1112 if (!desc)
1113 continue;
1114
1074 spin_lock_irqsave(&desc->lock, flags); 1115 spin_lock_irqsave(&desc->lock, flags);
1075 1116
1076 if (desc->action && desc->action->handler) { 1117 if (desc->action && desc->action->handler) {
1077 seq_printf(m, "%5d ", i); 1118 seq_printf(m, "%5d ", i);
1078 seq_printf(m, "0x%05lx ", virq_to_hw(i)); 1119 seq_printf(m, "0x%05lx ", virq_to_hw(i));
1079 1120
1080 if (desc->chip && desc->chip->typename) 1121 if (desc->chip && desc->chip->name)
1081 p = desc->chip->typename; 1122 p = desc->chip->name;
1082 else 1123 else
1083 p = none; 1124 p = none;
1084 seq_printf(m, "%-15s ", p); 1125 seq_printf(m, "%-15s ", p);
diff --git a/arch/powerpc/kernel/lparcfg.c b/arch/powerpc/kernel/lparcfg.c
index ed0ac4e4b8d8..79a00bb9c64c 100644
--- a/arch/powerpc/kernel/lparcfg.c
+++ b/arch/powerpc/kernel/lparcfg.c
@@ -781,9 +781,9 @@ static int __init lparcfg_init(void)
781 !firmware_has_feature(FW_FEATURE_ISERIES)) 781 !firmware_has_feature(FW_FEATURE_ISERIES))
782 mode |= S_IWUSR; 782 mode |= S_IWUSR;
783 783
784 ent = proc_create("ppc64/lparcfg", mode, NULL, &lparcfg_fops); 784 ent = proc_create("powerpc/lparcfg", mode, NULL, &lparcfg_fops);
785 if (!ent) { 785 if (!ent) {
786 printk(KERN_ERR "Failed to create ppc64/lparcfg\n"); 786 printk(KERN_ERR "Failed to create powerpc/lparcfg\n");
787 return -EIO; 787 return -EIO;
788 } 788 }
789 789
diff --git a/arch/powerpc/kernel/misc_32.S b/arch/powerpc/kernel/misc_32.S
index da9c0c4c10f3..8649f536f8df 100644
--- a/arch/powerpc/kernel/misc_32.S
+++ b/arch/powerpc/kernel/misc_32.S
@@ -502,15 +502,7 @@ _GLOBAL(clear_pages)
502 li r0,PAGE_SIZE/L1_CACHE_BYTES 502 li r0,PAGE_SIZE/L1_CACHE_BYTES
503 slw r0,r0,r4 503 slw r0,r0,r4
504 mtctr r0 504 mtctr r0
505#ifdef CONFIG_8xx
506 li r4, 0
5071: stw r4, 0(r3)
508 stw r4, 4(r3)
509 stw r4, 8(r3)
510 stw r4, 12(r3)
511#else
5121: dcbz 0,r3 5051: dcbz 0,r3
513#endif
514 addi r3,r3,L1_CACHE_BYTES 506 addi r3,r3,L1_CACHE_BYTES
515 bdnz 1b 507 bdnz 1b
516 blr 508 blr
@@ -535,15 +527,6 @@ _GLOBAL(copy_page)
535 addi r3,r3,-4 527 addi r3,r3,-4
536 addi r4,r4,-4 528 addi r4,r4,-4
537 529
538#ifdef CONFIG_8xx
539 /* don't use prefetch on 8xx */
540 li r0,4096/L1_CACHE_BYTES
541 mtctr r0
5421: COPY_16_BYTES
543 bdnz 1b
544 blr
545
546#else /* not 8xx, we can prefetch */
547 li r5,4 530 li r5,4
548 531
549#if MAX_COPY_PREFETCH > 1 532#if MAX_COPY_PREFETCH > 1
@@ -584,7 +567,6 @@ _GLOBAL(copy_page)
584 li r0,MAX_COPY_PREFETCH 567 li r0,MAX_COPY_PREFETCH
585 li r11,4 568 li r11,4
586 b 2b 569 b 2b
587#endif /* CONFIG_8xx */
588 570
589/* 571/*
590 * void atomic_clear_mask(atomic_t mask, atomic_t *addr) 572 * void atomic_clear_mask(atomic_t mask, atomic_t *addr)
diff --git a/arch/powerpc/kernel/nvram_64.c b/arch/powerpc/kernel/nvram_64.c
index 0ed31f220482..ad461e735aec 100644
--- a/arch/powerpc/kernel/nvram_64.c
+++ b/arch/powerpc/kernel/nvram_64.c
@@ -139,8 +139,8 @@ out:
139 139
140} 140}
141 141
142static int dev_nvram_ioctl(struct inode *inode, struct file *file, 142static long dev_nvram_ioctl(struct file *file, unsigned int cmd,
143 unsigned int cmd, unsigned long arg) 143 unsigned long arg)
144{ 144{
145 switch(cmd) { 145 switch(cmd) {
146#ifdef CONFIG_PPC_PMAC 146#ifdef CONFIG_PPC_PMAC
@@ -169,11 +169,11 @@ static int dev_nvram_ioctl(struct inode *inode, struct file *file,
169} 169}
170 170
171const struct file_operations nvram_fops = { 171const struct file_operations nvram_fops = {
172 .owner = THIS_MODULE, 172 .owner = THIS_MODULE,
173 .llseek = dev_nvram_llseek, 173 .llseek = dev_nvram_llseek,
174 .read = dev_nvram_read, 174 .read = dev_nvram_read,
175 .write = dev_nvram_write, 175 .write = dev_nvram_write,
176 .ioctl = dev_nvram_ioctl, 176 .unlocked_ioctl = dev_nvram_ioctl,
177}; 177};
178 178
179static struct miscdevice nvram_dev = { 179static struct miscdevice nvram_dev = {
@@ -184,7 +184,7 @@ static struct miscdevice nvram_dev = {
184 184
185 185
186#ifdef DEBUG_NVRAM 186#ifdef DEBUG_NVRAM
187static void nvram_print_partitions(char * label) 187static void __init nvram_print_partitions(char * label)
188{ 188{
189 struct list_head * p; 189 struct list_head * p;
190 struct nvram_partition * tmp_part; 190 struct nvram_partition * tmp_part;
@@ -202,7 +202,7 @@ static void nvram_print_partitions(char * label)
202#endif 202#endif
203 203
204 204
205static int nvram_write_header(struct nvram_partition * part) 205static int __init nvram_write_header(struct nvram_partition * part)
206{ 206{
207 loff_t tmp_index; 207 loff_t tmp_index;
208 int rc; 208 int rc;
@@ -214,7 +214,7 @@ static int nvram_write_header(struct nvram_partition * part)
214} 214}
215 215
216 216
217static unsigned char nvram_checksum(struct nvram_header *p) 217static unsigned char __init nvram_checksum(struct nvram_header *p)
218{ 218{
219 unsigned int c_sum, c_sum2; 219 unsigned int c_sum, c_sum2;
220 unsigned short *sp = (unsigned short *)p->name; /* assume 6 shorts */ 220 unsigned short *sp = (unsigned short *)p->name; /* assume 6 shorts */
@@ -228,32 +228,7 @@ static unsigned char nvram_checksum(struct nvram_header *p)
228 return c_sum; 228 return c_sum;
229} 229}
230 230
231 231static int __init nvram_remove_os_partition(void)
232/*
233 * Find an nvram partition, sig can be 0 for any
234 * partition or name can be NULL for any name, else
235 * tries to match both
236 */
237struct nvram_partition *nvram_find_partition(int sig, const char *name)
238{
239 struct nvram_partition * part;
240 struct list_head * p;
241
242 list_for_each(p, &nvram_part->partition) {
243 part = list_entry(p, struct nvram_partition, partition);
244
245 if (sig && part->header.signature != sig)
246 continue;
247 if (name && 0 != strncmp(name, part->header.name, 12))
248 continue;
249 return part;
250 }
251 return NULL;
252}
253EXPORT_SYMBOL(nvram_find_partition);
254
255
256static int nvram_remove_os_partition(void)
257{ 232{
258 struct list_head *i; 233 struct list_head *i;
259 struct list_head *j; 234 struct list_head *j;
@@ -319,7 +294,7 @@ static int nvram_remove_os_partition(void)
319 * Will create a partition starting at the first free 294 * Will create a partition starting at the first free
320 * space found if space has enough room. 295 * space found if space has enough room.
321 */ 296 */
322static int nvram_create_os_partition(void) 297static int __init nvram_create_os_partition(void)
323{ 298{
324 struct nvram_partition *part; 299 struct nvram_partition *part;
325 struct nvram_partition *new_part; 300 struct nvram_partition *new_part;
@@ -422,7 +397,7 @@ static int nvram_create_os_partition(void)
422 * 5.) If the max chunk cannot be allocated then try finding a chunk 397 * 5.) If the max chunk cannot be allocated then try finding a chunk
423 * that will satisfy the minum needed (NVRAM_MIN_REQ). 398 * that will satisfy the minum needed (NVRAM_MIN_REQ).
424 */ 399 */
425static int nvram_setup_partition(void) 400static int __init nvram_setup_partition(void)
426{ 401{
427 struct list_head * p; 402 struct list_head * p;
428 struct nvram_partition * part; 403 struct nvram_partition * part;
@@ -480,7 +455,7 @@ static int nvram_setup_partition(void)
480} 455}
481 456
482 457
483static int nvram_scan_partitions(void) 458static int __init nvram_scan_partitions(void)
484{ 459{
485 loff_t cur_index = 0; 460 loff_t cur_index = 0;
486 struct nvram_header phead; 461 struct nvram_header phead;
@@ -706,6 +681,9 @@ int nvram_clear_error_log(void)
706 int clear_word = ERR_FLAG_ALREADY_LOGGED; 681 int clear_word = ERR_FLAG_ALREADY_LOGGED;
707 int rc; 682 int rc;
708 683
684 if (nvram_error_log_index == -1)
685 return -1;
686
709 tmp_index = nvram_error_log_index; 687 tmp_index = nvram_error_log_index;
710 688
711 rc = ppc_md.nvram_write((char *)&clear_word, sizeof(int), &tmp_index); 689 rc = ppc_md.nvram_write((char *)&clear_word, sizeof(int), &tmp_index);
diff --git a/arch/powerpc/kernel/perf_callchain.c b/arch/powerpc/kernel/perf_callchain.c
index 0a03cf70d247..936f04dbfc6f 100644
--- a/arch/powerpc/kernel/perf_callchain.c
+++ b/arch/powerpc/kernel/perf_callchain.c
@@ -119,13 +119,6 @@ static void perf_callchain_kernel(struct pt_regs *regs,
119} 119}
120 120
121#ifdef CONFIG_PPC64 121#ifdef CONFIG_PPC64
122
123#ifdef CONFIG_HUGETLB_PAGE
124#define is_huge_psize(pagesize) (HPAGE_SHIFT && mmu_huge_psizes[pagesize])
125#else
126#define is_huge_psize(pagesize) 0
127#endif
128
129/* 122/*
130 * On 64-bit we don't want to invoke hash_page on user addresses from 123 * On 64-bit we don't want to invoke hash_page on user addresses from
131 * interrupt context, so if the access faults, we read the page tables 124 * interrupt context, so if the access faults, we read the page tables
@@ -135,7 +128,7 @@ static int read_user_stack_slow(void __user *ptr, void *ret, int nb)
135{ 128{
136 pgd_t *pgdir; 129 pgd_t *pgdir;
137 pte_t *ptep, pte; 130 pte_t *ptep, pte;
138 int pagesize; 131 unsigned shift;
139 unsigned long addr = (unsigned long) ptr; 132 unsigned long addr = (unsigned long) ptr;
140 unsigned long offset; 133 unsigned long offset;
141 unsigned long pfn; 134 unsigned long pfn;
@@ -145,17 +138,14 @@ static int read_user_stack_slow(void __user *ptr, void *ret, int nb)
145 if (!pgdir) 138 if (!pgdir)
146 return -EFAULT; 139 return -EFAULT;
147 140
148 pagesize = get_slice_psize(current->mm, addr); 141 ptep = find_linux_pte_or_hugepte(pgdir, addr, &shift);
142 if (!shift)
143 shift = PAGE_SHIFT;
149 144
150 /* align address to page boundary */ 145 /* align address to page boundary */
151 offset = addr & ((1ul << mmu_psize_defs[pagesize].shift) - 1); 146 offset = addr & ((1UL << shift) - 1);
152 addr -= offset; 147 addr -= offset;
153 148
154 if (is_huge_psize(pagesize))
155 ptep = huge_pte_offset(current->mm, addr);
156 else
157 ptep = find_linux_pte(pgdir, addr);
158
159 if (ptep == NULL) 149 if (ptep == NULL)
160 return -EFAULT; 150 return -EFAULT;
161 pte = *ptep; 151 pte = *ptep;
diff --git a/arch/powerpc/kernel/ppc_ksyms.c b/arch/powerpc/kernel/ppc_ksyms.c
index c8b27bb4dbde..425451453e96 100644
--- a/arch/powerpc/kernel/ppc_ksyms.c
+++ b/arch/powerpc/kernel/ppc_ksyms.c
@@ -96,8 +96,6 @@ EXPORT_SYMBOL(copy_4K_page);
96EXPORT_SYMBOL(isa_io_base); 96EXPORT_SYMBOL(isa_io_base);
97EXPORT_SYMBOL(isa_mem_base); 97EXPORT_SYMBOL(isa_mem_base);
98EXPORT_SYMBOL(pci_dram_offset); 98EXPORT_SYMBOL(pci_dram_offset);
99EXPORT_SYMBOL(pci_alloc_consistent);
100EXPORT_SYMBOL(pci_free_consistent);
101#endif /* CONFIG_PCI */ 99#endif /* CONFIG_PCI */
102 100
103EXPORT_SYMBOL(start_thread); 101EXPORT_SYMBOL(start_thread);
@@ -162,7 +160,6 @@ EXPORT_SYMBOL(screen_info);
162 160
163#ifdef CONFIG_PPC32 161#ifdef CONFIG_PPC32
164EXPORT_SYMBOL(timer_interrupt); 162EXPORT_SYMBOL(timer_interrupt);
165EXPORT_SYMBOL(irq_desc);
166EXPORT_SYMBOL(tb_ticks_per_jiffy); 163EXPORT_SYMBOL(tb_ticks_per_jiffy);
167EXPORT_SYMBOL(cacheable_memcpy); 164EXPORT_SYMBOL(cacheable_memcpy);
168EXPORT_SYMBOL(cacheable_memzero); 165EXPORT_SYMBOL(cacheable_memzero);
diff --git a/arch/powerpc/kernel/proc_ppc64.c b/arch/powerpc/kernel/proc_powerpc.c
index c647ddef40dc..1ed3b8d7981e 100644
--- a/arch/powerpc/kernel/proc_ppc64.c
+++ b/arch/powerpc/kernel/proc_powerpc.c
@@ -28,55 +28,7 @@
28#include <asm/uaccess.h> 28#include <asm/uaccess.h>
29#include <asm/prom.h> 29#include <asm/prom.h>
30 30
31static loff_t page_map_seek( struct file *file, loff_t off, int whence); 31#ifdef CONFIG_PPC64
32static ssize_t page_map_read( struct file *file, char __user *buf, size_t nbytes,
33 loff_t *ppos);
34static int page_map_mmap( struct file *file, struct vm_area_struct *vma );
35
36static const struct file_operations page_map_fops = {
37 .llseek = page_map_seek,
38 .read = page_map_read,
39 .mmap = page_map_mmap
40};
41
42/*
43 * Create the ppc64 and ppc64/rtas directories early. This allows us to
44 * assume that they have been previously created in drivers.
45 */
46static int __init proc_ppc64_create(void)
47{
48 struct proc_dir_entry *root;
49
50 root = proc_mkdir("ppc64", NULL);
51 if (!root)
52 return 1;
53
54 if (!of_find_node_by_path("/rtas"))
55 return 0;
56
57 if (!proc_mkdir("rtas", root))
58 return 1;
59
60 if (!proc_symlink("rtas", NULL, "ppc64/rtas"))
61 return 1;
62
63 return 0;
64}
65core_initcall(proc_ppc64_create);
66
67static int __init proc_ppc64_init(void)
68{
69 struct proc_dir_entry *pde;
70
71 pde = proc_create_data("ppc64/systemcfg", S_IFREG|S_IRUGO, NULL,
72 &page_map_fops, vdso_data);
73 if (!pde)
74 return 1;
75 pde->size = PAGE_SIZE;
76
77 return 0;
78}
79__initcall(proc_ppc64_init);
80 32
81static loff_t page_map_seek( struct file *file, loff_t off, int whence) 33static loff_t page_map_seek( struct file *file, loff_t off, int whence)
82{ 34{
@@ -120,3 +72,55 @@ static int page_map_mmap( struct file *file, struct vm_area_struct *vma )
120 return 0; 72 return 0;
121} 73}
122 74
75static const struct file_operations page_map_fops = {
76 .llseek = page_map_seek,
77 .read = page_map_read,
78 .mmap = page_map_mmap
79};
80
81
82static int __init proc_ppc64_init(void)
83{
84 struct proc_dir_entry *pde;
85
86 pde = proc_create_data("powerpc/systemcfg", S_IFREG|S_IRUGO, NULL,
87 &page_map_fops, vdso_data);
88 if (!pde)
89 return 1;
90 pde->size = PAGE_SIZE;
91
92 return 0;
93}
94__initcall(proc_ppc64_init);
95
96#endif /* CONFIG_PPC64 */
97
98/*
99 * Create the ppc64 and ppc64/rtas directories early. This allows us to
100 * assume that they have been previously created in drivers.
101 */
102static int __init proc_ppc64_create(void)
103{
104 struct proc_dir_entry *root;
105
106 root = proc_mkdir("powerpc", NULL);
107 if (!root)
108 return 1;
109
110#ifdef CONFIG_PPC64
111 if (!proc_symlink("ppc64", NULL, "powerpc"))
112 pr_err("Failed to create link /proc/ppc64 -> /proc/powerpc\n");
113#endif
114
115 if (!of_find_node_by_path("/rtas"))
116 return 0;
117
118 if (!proc_mkdir("rtas", root))
119 return 1;
120
121 if (!proc_symlink("rtas", NULL, "powerpc/rtas"))
122 return 1;
123
124 return 0;
125}
126core_initcall(proc_ppc64_create);
diff --git a/arch/powerpc/kernel/rtas_flash.c b/arch/powerpc/kernel/rtas_flash.c
index 13011a96a977..a85117d5c9a4 100644
--- a/arch/powerpc/kernel/rtas_flash.c
+++ b/arch/powerpc/kernel/rtas_flash.c
@@ -6,7 +6,7 @@
6 * as published by the Free Software Foundation; either version 6 * as published by the Free Software Foundation; either version
7 * 2 of the License, or (at your option) any later version. 7 * 2 of the License, or (at your option) any later version.
8 * 8 *
9 * /proc/ppc64/rtas/firmware_flash interface 9 * /proc/powerpc/rtas/firmware_flash interface
10 * 10 *
11 * This file implements a firmware_flash interface to pump a firmware 11 * This file implements a firmware_flash interface to pump a firmware
12 * image into the kernel. At reboot time rtas_restart() will see the 12 * image into the kernel. At reboot time rtas_restart() will see the
@@ -740,7 +740,7 @@ static int __init rtas_flash_init(void)
740 return 1; 740 return 1;
741 } 741 }
742 742
743 firmware_flash_pde = create_flash_pde("ppc64/rtas/" 743 firmware_flash_pde = create_flash_pde("powerpc/rtas/"
744 FIRMWARE_FLASH_NAME, 744 FIRMWARE_FLASH_NAME,
745 &rtas_flash_operations); 745 &rtas_flash_operations);
746 if (firmware_flash_pde == NULL) { 746 if (firmware_flash_pde == NULL) {
@@ -754,7 +754,7 @@ static int __init rtas_flash_init(void)
754 if (rc != 0) 754 if (rc != 0)
755 goto cleanup; 755 goto cleanup;
756 756
757 firmware_update_pde = create_flash_pde("ppc64/rtas/" 757 firmware_update_pde = create_flash_pde("powerpc/rtas/"
758 FIRMWARE_UPDATE_NAME, 758 FIRMWARE_UPDATE_NAME,
759 &rtas_flash_operations); 759 &rtas_flash_operations);
760 if (firmware_update_pde == NULL) { 760 if (firmware_update_pde == NULL) {
@@ -768,7 +768,7 @@ static int __init rtas_flash_init(void)
768 if (rc != 0) 768 if (rc != 0)
769 goto cleanup; 769 goto cleanup;
770 770
771 validate_pde = create_flash_pde("ppc64/rtas/" VALIDATE_FLASH_NAME, 771 validate_pde = create_flash_pde("powerpc/rtas/" VALIDATE_FLASH_NAME,
772 &validate_flash_operations); 772 &validate_flash_operations);
773 if (validate_pde == NULL) { 773 if (validate_pde == NULL) {
774 rc = -ENOMEM; 774 rc = -ENOMEM;
@@ -781,7 +781,7 @@ static int __init rtas_flash_init(void)
781 if (rc != 0) 781 if (rc != 0)
782 goto cleanup; 782 goto cleanup;
783 783
784 manage_pde = create_flash_pde("ppc64/rtas/" MANAGE_FLASH_NAME, 784 manage_pde = create_flash_pde("powerpc/rtas/" MANAGE_FLASH_NAME,
785 &manage_flash_operations); 785 &manage_flash_operations);
786 if (manage_pde == NULL) { 786 if (manage_pde == NULL) {
787 rc = -ENOMEM; 787 rc = -ENOMEM;
diff --git a/arch/powerpc/kernel/rtasd.c b/arch/powerpc/kernel/rtasd.c
new file mode 100644
index 000000000000..2e4832ab2108
--- /dev/null
+++ b/arch/powerpc/kernel/rtasd.c
@@ -0,0 +1,539 @@
1/*
2 * Copyright (C) 2001 Anton Blanchard <anton@au.ibm.com>, IBM
3 *
4 * This program is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU General Public License
6 * as published by the Free Software Foundation; either version
7 * 2 of the License, or (at your option) any later version.
8 *
9 * Communication to userspace based on kernel/printk.c
10 */
11
12#include <linux/types.h>
13#include <linux/errno.h>
14#include <linux/sched.h>
15#include <linux/kernel.h>
16#include <linux/poll.h>
17#include <linux/proc_fs.h>
18#include <linux/init.h>
19#include <linux/vmalloc.h>
20#include <linux/spinlock.h>
21#include <linux/cpu.h>
22#include <linux/workqueue.h>
23
24#include <asm/uaccess.h>
25#include <asm/io.h>
26#include <asm/rtas.h>
27#include <asm/prom.h>
28#include <asm/nvram.h>
29#include <asm/atomic.h>
30#include <asm/machdep.h>
31
32
33static DEFINE_SPINLOCK(rtasd_log_lock);
34
35static DECLARE_WAIT_QUEUE_HEAD(rtas_log_wait);
36
37static char *rtas_log_buf;
38static unsigned long rtas_log_start;
39static unsigned long rtas_log_size;
40
41static int surveillance_timeout = -1;
42
43static unsigned int rtas_error_log_max;
44static unsigned int rtas_error_log_buffer_max;
45
46/* RTAS service tokens */
47static unsigned int event_scan;
48static unsigned int rtas_event_scan_rate;
49
50static int full_rtas_msgs = 0;
51
52/* Stop logging to nvram after first fatal error */
53static int logging_enabled; /* Until we initialize everything,
54 * make sure we don't try logging
55 * anything */
56static int error_log_cnt;
57
58/*
59 * Since we use 32 bit RTAS, the physical address of this must be below
60 * 4G or else bad things happen. Allocate this in the kernel data and
61 * make it big enough.
62 */
63static unsigned char logdata[RTAS_ERROR_LOG_MAX];
64
65static char *rtas_type[] = {
66 "Unknown", "Retry", "TCE Error", "Internal Device Failure",
67 "Timeout", "Data Parity", "Address Parity", "Cache Parity",
68 "Address Invalid", "ECC Uncorrected", "ECC Corrupted",
69};
70
71static char *rtas_event_type(int type)
72{
73 if ((type > 0) && (type < 11))
74 return rtas_type[type];
75
76 switch (type) {
77 case RTAS_TYPE_EPOW:
78 return "EPOW";
79 case RTAS_TYPE_PLATFORM:
80 return "Platform Error";
81 case RTAS_TYPE_IO:
82 return "I/O Event";
83 case RTAS_TYPE_INFO:
84 return "Platform Information Event";
85 case RTAS_TYPE_DEALLOC:
86 return "Resource Deallocation Event";
87 case RTAS_TYPE_DUMP:
88 return "Dump Notification Event";
89 }
90
91 return rtas_type[0];
92}
93
94/* To see this info, grep RTAS /var/log/messages and each entry
95 * will be collected together with obvious begin/end.
96 * There will be a unique identifier on the begin and end lines.
97 * This will persist across reboots.
98 *
99 * format of error logs returned from RTAS:
100 * bytes (size) : contents
101 * --------------------------------------------------------
102 * 0-7 (8) : rtas_error_log
103 * 8-47 (40) : extended info
104 * 48-51 (4) : vendor id
105 * 52-1023 (vendor specific) : location code and debug data
106 */
107static void printk_log_rtas(char *buf, int len)
108{
109
110 int i,j,n = 0;
111 int perline = 16;
112 char buffer[64];
113 char * str = "RTAS event";
114
115 if (full_rtas_msgs) {
116 printk(RTAS_DEBUG "%d -------- %s begin --------\n",
117 error_log_cnt, str);
118
119 /*
120 * Print perline bytes on each line, each line will start
121 * with RTAS and a changing number, so syslogd will
122 * print lines that are otherwise the same. Separate every
123 * 4 bytes with a space.
124 */
125 for (i = 0; i < len; i++) {
126 j = i % perline;
127 if (j == 0) {
128 memset(buffer, 0, sizeof(buffer));
129 n = sprintf(buffer, "RTAS %d:", i/perline);
130 }
131
132 if ((i % 4) == 0)
133 n += sprintf(buffer+n, " ");
134
135 n += sprintf(buffer+n, "%02x", (unsigned char)buf[i]);
136
137 if (j == (perline-1))
138 printk(KERN_DEBUG "%s\n", buffer);
139 }
140 if ((i % perline) != 0)
141 printk(KERN_DEBUG "%s\n", buffer);
142
143 printk(RTAS_DEBUG "%d -------- %s end ----------\n",
144 error_log_cnt, str);
145 } else {
146 struct rtas_error_log *errlog = (struct rtas_error_log *)buf;
147
148 printk(RTAS_DEBUG "event: %d, Type: %s, Severity: %d\n",
149 error_log_cnt, rtas_event_type(errlog->type),
150 errlog->severity);
151 }
152}
153
154static int log_rtas_len(char * buf)
155{
156 int len;
157 struct rtas_error_log *err;
158
159 /* rtas fixed header */
160 len = 8;
161 err = (struct rtas_error_log *)buf;
162 if (err->extended_log_length) {
163
164 /* extended header */
165 len += err->extended_log_length;
166 }
167
168 if (rtas_error_log_max == 0)
169 rtas_error_log_max = rtas_get_error_log_max();
170
171 if (len > rtas_error_log_max)
172 len = rtas_error_log_max;
173
174 return len;
175}
176
177/*
178 * First write to nvram, if fatal error, that is the only
179 * place we log the info. The error will be picked up
180 * on the next reboot by rtasd. If not fatal, run the
181 * method for the type of error. Currently, only RTAS
182 * errors have methods implemented, but in the future
183 * there might be a need to store data in nvram before a
184 * call to panic().
185 *
186 * XXX We write to nvram periodically, to indicate error has
187 * been written and sync'd, but there is a possibility
188 * that if we don't shutdown correctly, a duplicate error
189 * record will be created on next reboot.
190 */
191void pSeries_log_error(char *buf, unsigned int err_type, int fatal)
192{
193 unsigned long offset;
194 unsigned long s;
195 int len = 0;
196
197 pr_debug("rtasd: logging event\n");
198 if (buf == NULL)
199 return;
200
201 spin_lock_irqsave(&rtasd_log_lock, s);
202
203 /* get length and increase count */
204 switch (err_type & ERR_TYPE_MASK) {
205 case ERR_TYPE_RTAS_LOG:
206 len = log_rtas_len(buf);
207 if (!(err_type & ERR_FLAG_BOOT))
208 error_log_cnt++;
209 break;
210 case ERR_TYPE_KERNEL_PANIC:
211 default:
212 WARN_ON_ONCE(!irqs_disabled()); /* @@@ DEBUG @@@ */
213 spin_unlock_irqrestore(&rtasd_log_lock, s);
214 return;
215 }
216
217#ifdef CONFIG_PPC64
218 /* Write error to NVRAM */
219 if (logging_enabled && !(err_type & ERR_FLAG_BOOT))
220 nvram_write_error_log(buf, len, err_type, error_log_cnt);
221#endif /* CONFIG_PPC64 */
222
223 /*
224 * rtas errors can occur during boot, and we do want to capture
225 * those somewhere, even if nvram isn't ready (why not?), and even
226 * if rtasd isn't ready. Put them into the boot log, at least.
227 */
228 if ((err_type & ERR_TYPE_MASK) == ERR_TYPE_RTAS_LOG)
229 printk_log_rtas(buf, len);
230
231 /* Check to see if we need to or have stopped logging */
232 if (fatal || !logging_enabled) {
233 logging_enabled = 0;
234 WARN_ON_ONCE(!irqs_disabled()); /* @@@ DEBUG @@@ */
235 spin_unlock_irqrestore(&rtasd_log_lock, s);
236 return;
237 }
238
239 /* call type specific method for error */
240 switch (err_type & ERR_TYPE_MASK) {
241 case ERR_TYPE_RTAS_LOG:
242 offset = rtas_error_log_buffer_max *
243 ((rtas_log_start+rtas_log_size) & LOG_NUMBER_MASK);
244
245 /* First copy over sequence number */
246 memcpy(&rtas_log_buf[offset], (void *) &error_log_cnt, sizeof(int));
247
248 /* Second copy over error log data */
249 offset += sizeof(int);
250 memcpy(&rtas_log_buf[offset], buf, len);
251
252 if (rtas_log_size < LOG_NUMBER)
253 rtas_log_size += 1;
254 else
255 rtas_log_start += 1;
256
257 WARN_ON_ONCE(!irqs_disabled()); /* @@@ DEBUG @@@ */
258 spin_unlock_irqrestore(&rtasd_log_lock, s);
259 wake_up_interruptible(&rtas_log_wait);
260 break;
261 case ERR_TYPE_KERNEL_PANIC:
262 default:
263 WARN_ON_ONCE(!irqs_disabled()); /* @@@ DEBUG @@@ */
264 spin_unlock_irqrestore(&rtasd_log_lock, s);
265 return;
266 }
267
268}
269
270static int rtas_log_open(struct inode * inode, struct file * file)
271{
272 return 0;
273}
274
275static int rtas_log_release(struct inode * inode, struct file * file)
276{
277 return 0;
278}
279
280/* This will check if all events are logged, if they are then, we
281 * know that we can safely clear the events in NVRAM.
282 * Next we'll sit and wait for something else to log.
283 */
284static ssize_t rtas_log_read(struct file * file, char __user * buf,
285 size_t count, loff_t *ppos)
286{
287 int error;
288 char *tmp;
289 unsigned long s;
290 unsigned long offset;
291
292 if (!buf || count < rtas_error_log_buffer_max)
293 return -EINVAL;
294
295 count = rtas_error_log_buffer_max;
296
297 if (!access_ok(VERIFY_WRITE, buf, count))
298 return -EFAULT;
299
300 tmp = kmalloc(count, GFP_KERNEL);
301 if (!tmp)
302 return -ENOMEM;
303
304 spin_lock_irqsave(&rtasd_log_lock, s);
305
306 /* if it's 0, then we know we got the last one (the one in NVRAM) */
307 while (rtas_log_size == 0) {
308 if (file->f_flags & O_NONBLOCK) {
309 spin_unlock_irqrestore(&rtasd_log_lock, s);
310 error = -EAGAIN;
311 goto out;
312 }
313
314 if (!logging_enabled) {
315 spin_unlock_irqrestore(&rtasd_log_lock, s);
316 error = -ENODATA;
317 goto out;
318 }
319#ifdef CONFIG_PPC64
320 nvram_clear_error_log();
321#endif /* CONFIG_PPC64 */
322
323 spin_unlock_irqrestore(&rtasd_log_lock, s);
324 error = wait_event_interruptible(rtas_log_wait, rtas_log_size);
325 if (error)
326 goto out;
327 spin_lock_irqsave(&rtasd_log_lock, s);
328 }
329
330 offset = rtas_error_log_buffer_max * (rtas_log_start & LOG_NUMBER_MASK);
331 memcpy(tmp, &rtas_log_buf[offset], count);
332
333 rtas_log_start += 1;
334 rtas_log_size -= 1;
335 spin_unlock_irqrestore(&rtasd_log_lock, s);
336
337 error = copy_to_user(buf, tmp, count) ? -EFAULT : count;
338out:
339 kfree(tmp);
340 return error;
341}
342
343static unsigned int rtas_log_poll(struct file *file, poll_table * wait)
344{
345 poll_wait(file, &rtas_log_wait, wait);
346 if (rtas_log_size)
347 return POLLIN | POLLRDNORM;
348 return 0;
349}
350
351static const struct file_operations proc_rtas_log_operations = {
352 .read = rtas_log_read,
353 .poll = rtas_log_poll,
354 .open = rtas_log_open,
355 .release = rtas_log_release,
356};
357
358static int enable_surveillance(int timeout)
359{
360 int error;
361
362 error = rtas_set_indicator(SURVEILLANCE_TOKEN, 0, timeout);
363
364 if (error == 0)
365 return 0;
366
367 if (error == -EINVAL) {
368 printk(KERN_DEBUG "rtasd: surveillance not supported\n");
369 return 0;
370 }
371
372 printk(KERN_ERR "rtasd: could not update surveillance\n");
373 return -1;
374}
375
376static void do_event_scan(void)
377{
378 int error;
379 do {
380 memset(logdata, 0, rtas_error_log_max);
381 error = rtas_call(event_scan, 4, 1, NULL,
382 RTAS_EVENT_SCAN_ALL_EVENTS, 0,
383 __pa(logdata), rtas_error_log_max);
384 if (error == -1) {
385 printk(KERN_ERR "event-scan failed\n");
386 break;
387 }
388
389 if (error == 0)
390 pSeries_log_error(logdata, ERR_TYPE_RTAS_LOG, 0);
391
392 } while(error == 0);
393}
394
395static void rtas_event_scan(struct work_struct *w);
396DECLARE_DELAYED_WORK(event_scan_work, rtas_event_scan);
397
398/*
399 * Delay should be at least one second since some machines have problems if
400 * we call event-scan too quickly.
401 */
402static unsigned long event_scan_delay = 1*HZ;
403static int first_pass = 1;
404
405static void rtas_event_scan(struct work_struct *w)
406{
407 unsigned int cpu;
408
409 do_event_scan();
410
411 get_online_cpus();
412
413 cpu = next_cpu(smp_processor_id(), cpu_online_map);
414 if (cpu == NR_CPUS) {
415 cpu = first_cpu(cpu_online_map);
416
417 if (first_pass) {
418 first_pass = 0;
419 event_scan_delay = 30*HZ/rtas_event_scan_rate;
420
421 if (surveillance_timeout != -1) {
422 pr_debug("rtasd: enabling surveillance\n");
423 enable_surveillance(surveillance_timeout);
424 pr_debug("rtasd: surveillance enabled\n");
425 }
426 }
427 }
428
429 schedule_delayed_work_on(cpu, &event_scan_work,
430 __round_jiffies_relative(event_scan_delay, cpu));
431
432 put_online_cpus();
433}
434
435#ifdef CONFIG_PPC64
436static void retreive_nvram_error_log(void)
437{
438 unsigned int err_type ;
439 int rc ;
440
441 /* See if we have any error stored in NVRAM */
442 memset(logdata, 0, rtas_error_log_max);
443 rc = nvram_read_error_log(logdata, rtas_error_log_max,
444 &err_type, &error_log_cnt);
445 /* We can use rtas_log_buf now */
446 logging_enabled = 1;
447 if (!rc) {
448 if (err_type != ERR_FLAG_ALREADY_LOGGED) {
449 pSeries_log_error(logdata, err_type | ERR_FLAG_BOOT, 0);
450 }
451 }
452}
453#else /* CONFIG_PPC64 */
454static void retreive_nvram_error_log(void)
455{
456}
457#endif /* CONFIG_PPC64 */
458
459static void start_event_scan(void)
460{
461 printk(KERN_DEBUG "RTAS daemon started\n");
462 pr_debug("rtasd: will sleep for %d milliseconds\n",
463 (30000 / rtas_event_scan_rate));
464
465 /* Retreive errors from nvram if any */
466 retreive_nvram_error_log();
467
468 schedule_delayed_work_on(first_cpu(cpu_online_map), &event_scan_work,
469 event_scan_delay);
470}
471
472static int __init rtas_init(void)
473{
474 struct proc_dir_entry *entry;
475
476 if (!machine_is(pseries) && !machine_is(chrp))
477 return 0;
478
479 /* No RTAS */
480 event_scan = rtas_token("event-scan");
481 if (event_scan == RTAS_UNKNOWN_SERVICE) {
482 printk(KERN_INFO "rtasd: No event-scan on system\n");
483 return -ENODEV;
484 }
485
486 rtas_event_scan_rate = rtas_token("rtas-event-scan-rate");
487 if (rtas_event_scan_rate == RTAS_UNKNOWN_SERVICE) {
488 printk(KERN_ERR "rtasd: no rtas-event-scan-rate on system\n");
489 return -ENODEV;
490 }
491
492 /* Make room for the sequence number */
493 rtas_error_log_max = rtas_get_error_log_max();
494 rtas_error_log_buffer_max = rtas_error_log_max + sizeof(int);
495
496 rtas_log_buf = vmalloc(rtas_error_log_buffer_max*LOG_NUMBER);
497 if (!rtas_log_buf) {
498 printk(KERN_ERR "rtasd: no memory\n");
499 return -ENOMEM;
500 }
501
502 entry = proc_create("powerpc/rtas/error_log", S_IRUSR, NULL,
503 &proc_rtas_log_operations);
504 if (!entry)
505 printk(KERN_ERR "Failed to create error_log proc entry\n");
506
507 start_event_scan();
508
509 return 0;
510}
511__initcall(rtas_init);
512
513static int __init surveillance_setup(char *str)
514{
515 int i;
516
517 /* We only do surveillance on pseries */
518 if (!machine_is(pseries))
519 return 0;
520
521 if (get_option(&str,&i)) {
522 if (i >= 0 && i <= 255)
523 surveillance_timeout = i;
524 }
525
526 return 1;
527}
528__setup("surveillance=", surveillance_setup);
529
530static int __init rtasmsgs_setup(char *str)
531{
532 if (strcmp(str, "on") == 0)
533 full_rtas_msgs = 1;
534 else if (strcmp(str, "off") == 0)
535 full_rtas_msgs = 0;
536
537 return 1;
538}
539__setup("rtasmsgs=", rtasmsgs_setup);
diff --git a/arch/powerpc/kernel/setup_64.c b/arch/powerpc/kernel/setup_64.c
index df2c9e932b37..6568406b2a30 100644
--- a/arch/powerpc/kernel/setup_64.c
+++ b/arch/powerpc/kernel/setup_64.c
@@ -356,11 +356,6 @@ void __init setup_system(void)
356 */ 356 */
357 initialize_cache_info(); 357 initialize_cache_info();
358 358
359 /*
360 * Initialize irq remapping subsystem
361 */
362 irq_early_init();
363
364#ifdef CONFIG_PPC_RTAS 359#ifdef CONFIG_PPC_RTAS
365 /* 360 /*
366 * Initialize RTAS if available 361 * Initialize RTAS if available
diff --git a/arch/powerpc/kernel/smp.c b/arch/powerpc/kernel/smp.c
index 9b86a74d2815..97196eefef3e 100644
--- a/arch/powerpc/kernel/smp.c
+++ b/arch/powerpc/kernel/smp.c
@@ -218,6 +218,9 @@ void crash_send_ipi(void (*crash_ipi_callback)(struct pt_regs *))
218 218
219static void stop_this_cpu(void *dummy) 219static void stop_this_cpu(void *dummy)
220{ 220{
221 /* Remove this CPU */
222 set_cpu_online(smp_processor_id(), false);
223
221 local_irq_disable(); 224 local_irq_disable();
222 while (1) 225 while (1)
223 ; 226 ;
diff --git a/arch/powerpc/kernel/sysfs.c b/arch/powerpc/kernel/sysfs.c
index 956ab33fd73f..e235e52dc4fe 100644
--- a/arch/powerpc/kernel/sysfs.c
+++ b/arch/powerpc/kernel/sysfs.c
@@ -461,6 +461,25 @@ static void unregister_cpu_online(unsigned int cpu)
461 461
462 cacheinfo_cpu_offline(cpu); 462 cacheinfo_cpu_offline(cpu);
463} 463}
464
465#ifdef CONFIG_ARCH_CPU_PROBE_RELEASE
466ssize_t arch_cpu_probe(const char *buf, size_t count)
467{
468 if (ppc_md.cpu_probe)
469 return ppc_md.cpu_probe(buf, count);
470
471 return -EINVAL;
472}
473
474ssize_t arch_cpu_release(const char *buf, size_t count)
475{
476 if (ppc_md.cpu_release)
477 return ppc_md.cpu_release(buf, count);
478
479 return -EINVAL;
480}
481#endif /* CONFIG_ARCH_CPU_PROBE_RELEASE */
482
464#endif /* CONFIG_HOTPLUG_CPU */ 483#endif /* CONFIG_HOTPLUG_CPU */
465 484
466static int __cpuinit sysfs_cpu_notify(struct notifier_block *self, 485static int __cpuinit sysfs_cpu_notify(struct notifier_block *self,
diff --git a/arch/powerpc/kernel/time.c b/arch/powerpc/kernel/time.c
index 674800b242d6..9ba2cc88591d 100644
--- a/arch/powerpc/kernel/time.c
+++ b/arch/powerpc/kernel/time.c
@@ -269,6 +269,7 @@ void account_system_vtime(struct task_struct *tsk)
269 per_cpu(cputime_scaled_last_delta, smp_processor_id()) = deltascaled; 269 per_cpu(cputime_scaled_last_delta, smp_processor_id()) = deltascaled;
270 local_irq_restore(flags); 270 local_irq_restore(flags);
271} 271}
272EXPORT_SYMBOL_GPL(account_system_vtime);
272 273
273/* 274/*
274 * Transfer the user and system times accumulated in the paca 275 * Transfer the user and system times accumulated in the paca
diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c
index 9d1f9354d6ca..804f0f30f227 100644
--- a/arch/powerpc/kernel/traps.c
+++ b/arch/powerpc/kernel/traps.c
@@ -198,28 +198,6 @@ void _exception(int signr, struct pt_regs *regs, int code, unsigned long addr)
198 info.si_code = code; 198 info.si_code = code;
199 info.si_addr = (void __user *) addr; 199 info.si_addr = (void __user *) addr;
200 force_sig_info(signr, &info, current); 200 force_sig_info(signr, &info, current);
201
202 /*
203 * Init gets no signals that it doesn't have a handler for.
204 * That's all very well, but if it has caused a synchronous
205 * exception and we ignore the resulting signal, it will just
206 * generate the same exception over and over again and we get
207 * nowhere. Better to kill it and let the kernel panic.
208 */
209 if (is_global_init(current)) {
210 __sighandler_t handler;
211
212 spin_lock_irq(&current->sighand->siglock);
213 handler = current->sighand->action[signr-1].sa.sa_handler;
214 spin_unlock_irq(&current->sighand->siglock);
215 if (handler == SIG_DFL) {
216 /* init has generated a synchronous exception
217 and it doesn't have a handler for the signal */
218 printk(KERN_CRIT "init has generated signal %d "
219 "but has no handler for it\n", signr);
220 do_exit(signr);
221 }
222 }
223} 201}
224 202
225#ifdef CONFIG_PPC64 203#ifdef CONFIG_PPC64
diff --git a/arch/powerpc/kernel/vector.S b/arch/powerpc/kernel/vector.S
index 67b6916f0e94..fe460482fa68 100644
--- a/arch/powerpc/kernel/vector.S
+++ b/arch/powerpc/kernel/vector.S
@@ -58,7 +58,7 @@ _GLOBAL(load_up_altivec)
58 * all 1's 58 * all 1's
59 */ 59 */
60 mfspr r4,SPRN_VRSAVE 60 mfspr r4,SPRN_VRSAVE
61 cmpdi 0,r4,0 61 cmpwi 0,r4,0
62 bne+ 1f 62 bne+ 1f
63 li r4,-1 63 li r4,-1
64 mtspr SPRN_VRSAVE,r4 64 mtspr SPRN_VRSAVE,r4
generated by cgit v1.2.2 (git 2.25.0) at 2025-02-01 07:55:40 -0500