aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorLinus Torvalds <torvalds@linux-foundation.org>2016-12-12 16:49:57 -0500
committerLinus Torvalds <torvalds@linux-foundation.org>2016-12-12 16:49:57 -0500
commit5645688f9d0d5a32f030f9c5429e1a58bedca23b (patch)
tree0b576ba953fb26d521c6b8c3364848acb00ceef3
parent4ade5b2268b9ff05e48a9cb99689c4fd15fbe9c3 (diff)
parent53938ee427bf27525a63721b7e25d86b8f31f161 (diff)
Merge branch 'x86-asm-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 asm updates from Ingo Molnar: "The main changes in this development cycle were: - a large number of call stack dumping/printing improvements: higher robustness, better cross-context dumping, improved output, etc. (Josh Poimboeuf) - vDSO getcpu() performance improvement for future Intel CPUs with the RDPID instruction (Andy Lutomirski) - add two new Intel AVX512 features and the CPUID support infrastructure for it: AVX512IFMA and AVX512VBMI. (Gayatri Kammela, He Chen) - more copy-user unification (Borislav Petkov) - entry code assembly macro simplifications (Alexander Kuleshov) - vDSO C/R support improvements (Dmitry Safonov) - misc fixes and cleanups (Borislav Petkov, Paul Bolle)" * 'x86-asm-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (40 commits) scripts/decode_stacktrace.sh: Fix address line detection on x86 x86/boot/64: Use defines for page size x86/dumpstack: Make stack name tags more comprehensible selftests/x86: Add test_vdso to test getcpu() x86/vdso: Use RDPID in preference to LSL when available x86/dumpstack: Handle NULL stack pointer in show_trace_log_lvl() x86/cpufeatures: Enable new AVX512 cpu features x86/cpuid: Provide get_scattered_cpuid_leaf() x86/cpuid: Cleanup cpuid_regs definitions x86/copy_user: Unify the code by removing the 64-bit asm _copy_*_user() variants x86/unwind: Ensure stack grows down x86/vdso: Set vDSO pointer only after success x86/prctl/uapi: Remove #ifdef for CHECKPOINT_RESTORE x86/unwind: Detect bad stack return address x86/dumpstack: Warn on stack recursion x86/unwind: Warn on bad frame pointer x86/decoder: Use stderr if insn sanity test fails x86/decoder: Use stdout if insn decoder test is successful mm/page_alloc: Remove kernel address exposure in free_reserved_area() x86/dumpstack: Remove raw stack dump ...
Diffstat
-rw-r--r--Documentation/kernel-parameters.txt3
-rw-r--r--Documentation/sysctl/kernel.txt8
-rw-r--r--Documentation/x86/x86_64/boot-options.txt4
-rw-r--r--arch/x86/entry/calling.h33
-rw-r--r--arch/x86/entry/entry_32.S141
-rw-r--r--arch/x86/entry/entry_64.S16
-rw-r--r--arch/x86/entry/vdso/vma.c10
-rw-r--r--arch/x86/events/intel/pt.c45
-rw-r--r--arch/x86/include/asm/cpufeatures.h3
-rw-r--r--arch/x86/include/asm/kdebug.h1
-rw-r--r--arch/x86/include/asm/processor.h14
-rw-r--r--arch/x86/include/asm/stacktrace.h8
-rw-r--r--arch/x86/include/asm/unwind.h16
-rw-r--r--arch/x86/include/asm/vgtod.h7
-rw-r--r--arch/x86/include/uapi/asm/prctl.h8
-rw-r--r--arch/x86/kernel/cpu/scattered.c57
-rw-r--r--arch/x86/kernel/cpuid.c4
-rw-r--r--arch/x86/kernel/dumpstack.c68
-rw-r--r--arch/x86/kernel/dumpstack_32.c56
-rw-r--r--arch/x86/kernel/dumpstack_64.c79
-rw-r--r--arch/x86/kernel/fpu/xstate.c2
-rw-r--r--arch/x86/kernel/head_32.S49
-rw-r--r--arch/x86/kernel/head_64.S52
-rw-r--r--arch/x86/kernel/process_32.c7
-rw-r--r--arch/x86/kernel/process_64.c13
-rw-r--r--arch/x86/kernel/smpboot.c4
-rw-r--r--arch/x86/kernel/unwind_frame.c161
-rw-r--r--arch/x86/kernel/vmlinux.lds.S2
-rw-r--r--arch/x86/lib/copy_user_64.S47
-rw-r--r--arch/x86/lib/usercopy.c49
-rw-r--r--arch/x86/lib/usercopy_32.c49
-rw-r--r--arch/x86/mm/fault.c3
-rw-r--r--arch/x86/platform/uv/uv_nmi.c4
-rw-r--r--arch/x86/tools/insn_sanity.c3
-rw-r--r--arch/x86/tools/test_get_len.c2
-rw-r--r--kernel/sysctl.c7
-rw-r--r--mm/page_alloc.c4
-rwxr-xr-xscripts/decode_stacktrace.sh3
-rwxr-xr-xscripts/faddr2line33
-rw-r--r--tools/testing/selftests/x86/Makefile2
-rw-r--r--tools/testing/selftests/x86/test_vdso.c123
41 files changed, 703 insertions, 497 deletions
diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt
index 86a31dfc036e..7c817dc45d45 100644
--- a/Documentation/kernel-parameters.txt
+++ b/Documentation/kernel-parameters.txt
@@ -1963,9 +1963,6 @@ bytes respectively. Such letter suffixes can also be entirely omitted.
1963 kmemcheck=2 (one-shot mode) 1963 kmemcheck=2 (one-shot mode)
1964 Default: 2 (one-shot mode) 1964 Default: 2 (one-shot mode)
1965 1965
1966 kstack=N [X86] Print N words from the kernel stack
1967 in oops dumps.
1968
1969 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs. 1966 kvm.ignore_msrs=[KVM] Ignore guest accesses to unhandled MSRs.
1970 Default is 0 (don't ignore, but inject #GP) 1967 Default is 0 (don't ignore, but inject #GP)
1971 1968
diff --git a/Documentation/sysctl/kernel.txt b/Documentation/sysctl/kernel.txt
index ffab8b5caa60..065f18478c1c 100644
--- a/Documentation/sysctl/kernel.txt
+++ b/Documentation/sysctl/kernel.txt
@@ -40,7 +40,6 @@ show up in /proc/sys/kernel:
40- hung_task_warnings 40- hung_task_warnings
41- kexec_load_disabled 41- kexec_load_disabled
42- kptr_restrict 42- kptr_restrict
43- kstack_depth_to_print [ X86 only ]
44- l2cr [ PPC only ] 43- l2cr [ PPC only ]
45- modprobe ==> Documentation/debugging-modules.txt 44- modprobe ==> Documentation/debugging-modules.txt
46- modules_disabled 45- modules_disabled
@@ -395,13 +394,6 @@ When kptr_restrict is set to (2), kernel pointers printed using
395 394
396============================================================== 395==============================================================
397 396
398kstack_depth_to_print: (X86 only)
399
400Controls the number of words to print when dumping the raw
401kernel stack.
402
403==============================================================
404
405l2cr: (PPC only) 397l2cr: (PPC only)
406 398
407This flag controls the L2 cache of G3 processor boards. If 399This flag controls the L2 cache of G3 processor boards. If
diff --git a/Documentation/x86/x86_64/boot-options.txt b/Documentation/x86/x86_64/boot-options.txt
index 0965a71f9942..61b611e9eeaf 100644
--- a/Documentation/x86/x86_64/boot-options.txt
+++ b/Documentation/x86/x86_64/boot-options.txt
@@ -277,10 +277,6 @@ IOMMU (input/output memory management unit)
277 space might stop working. Use this option if you have devices that 277 space might stop working. Use this option if you have devices that
278 are accessed from userspace directly on some PCI host bridge. 278 are accessed from userspace directly on some PCI host bridge.
279 279
280Debugging
281
282 kstack=N Print N words from the kernel stack in oops dumps.
283
284Miscellaneous 280Miscellaneous
285 281
286 nogbpages 282 nogbpages
diff --git a/arch/x86/entry/calling.h b/arch/x86/entry/calling.h
index 9a9e5884066c..05ed3d393da7 100644
--- a/arch/x86/entry/calling.h
+++ b/arch/x86/entry/calling.h
@@ -90,8 +90,8 @@ For 32-bit we have the following conventions - kernel is built with
90 90
91#define SIZEOF_PTREGS 21*8 91#define SIZEOF_PTREGS 21*8
92 92
93 .macro ALLOC_PT_GPREGS_ON_STACK addskip=0 93 .macro ALLOC_PT_GPREGS_ON_STACK
94 addq $-(15*8+\addskip), %rsp 94 addq $-(15*8), %rsp
95 .endm 95 .endm
96 96
97 .macro SAVE_C_REGS_HELPER offset=0 rax=1 rcx=1 r8910=1 r11=1 97 .macro SAVE_C_REGS_HELPER offset=0 rax=1 rcx=1 r8910=1 r11=1
@@ -147,15 +147,6 @@ For 32-bit we have the following conventions - kernel is built with
147 movq 5*8+\offset(%rsp), %rbx 147 movq 5*8+\offset(%rsp), %rbx
148 .endm 148 .endm
149 149
150 .macro ZERO_EXTRA_REGS
151 xorl %r15d, %r15d
152 xorl %r14d, %r14d
153 xorl %r13d, %r13d
154 xorl %r12d, %r12d
155 xorl %ebp, %ebp
156 xorl %ebx, %ebx
157 .endm
158
159 .macro RESTORE_C_REGS_HELPER rstor_rax=1, rstor_rcx=1, rstor_r11=1, rstor_r8910=1, rstor_rdx=1 150 .macro RESTORE_C_REGS_HELPER rstor_rax=1, rstor_rcx=1, rstor_r11=1, rstor_r8910=1, rstor_rdx=1
160 .if \rstor_r11 151 .if \rstor_r11
161 movq 6*8(%rsp), %r11 152 movq 6*8(%rsp), %r11
@@ -201,6 +192,26 @@ For 32-bit we have the following conventions - kernel is built with
201 .byte 0xf1 192 .byte 0xf1
202 .endm 193 .endm
203 194
195/*
196 * This is a sneaky trick to help the unwinder find pt_regs on the stack. The
197 * frame pointer is replaced with an encoded pointer to pt_regs. The encoding
198 * is just setting the LSB, which makes it an invalid stack address and is also
199 * a signal to the unwinder that it's a pt_regs pointer in disguise.
200 *
201 * NOTE: This macro must be used *after* SAVE_EXTRA_REGS because it corrupts
202 * the original rbp.
203 */
204.macro ENCODE_FRAME_POINTER ptregs_offset=0
205#ifdef CONFIG_FRAME_POINTER
206 .if \ptregs_offset
207 leaq \ptregs_offset(%rsp), %rbp
208 .else
209 mov %rsp, %rbp
210 .endif
211 orq $0x1, %rbp
212#endif
213.endm
214
204#endif /* CONFIG_X86_64 */ 215#endif /* CONFIG_X86_64 */
205 216
206/* 217/*
diff --git a/arch/x86/entry/entry_32.S b/arch/x86/entry/entry_32.S
index 21b352a11b49..acc0c6f36f3f 100644
--- a/arch/x86/entry/entry_32.S
+++ b/arch/x86/entry/entry_32.S
@@ -45,6 +45,7 @@
45#include <asm/asm.h> 45#include <asm/asm.h>
46#include <asm/smap.h> 46#include <asm/smap.h>
47#include <asm/export.h> 47#include <asm/export.h>
48#include <asm/frame.h>
48 49
49 .section .entry.text, "ax" 50 .section .entry.text, "ax"
50 51
@@ -175,6 +176,22 @@
175 SET_KERNEL_GS %edx 176 SET_KERNEL_GS %edx
176.endm 177.endm
177 178
179/*
180 * This is a sneaky trick to help the unwinder find pt_regs on the stack. The
181 * frame pointer is replaced with an encoded pointer to pt_regs. The encoding
182 * is just setting the LSB, which makes it an invalid stack address and is also
183 * a signal to the unwinder that it's a pt_regs pointer in disguise.
184 *
185 * NOTE: This macro must be used *after* SAVE_ALL because it corrupts the
186 * original rbp.
187 */
188.macro ENCODE_FRAME_POINTER
189#ifdef CONFIG_FRAME_POINTER
190 mov %esp, %ebp
191 orl $0x1, %ebp
192#endif
193.endm
194
178.macro RESTORE_INT_REGS 195.macro RESTORE_INT_REGS
179 popl %ebx 196 popl %ebx
180 popl %ecx 197 popl %ecx
@@ -238,6 +255,23 @@ ENTRY(__switch_to_asm)
238END(__switch_to_asm) 255END(__switch_to_asm)
239 256
240/* 257/*
258 * The unwinder expects the last frame on the stack to always be at the same
259 * offset from the end of the page, which allows it to validate the stack.
260 * Calling schedule_tail() directly would break that convention because its an
261 * asmlinkage function so its argument has to be pushed on the stack. This
262 * wrapper creates a proper "end of stack" frame header before the call.
263 */
264ENTRY(schedule_tail_wrapper)
265 FRAME_BEGIN
266
267 pushl %eax
268 call schedule_tail
269 popl %eax
270
271 FRAME_END
272 ret
273ENDPROC(schedule_tail_wrapper)
274/*
241 * A newly forked process directly context switches into this address. 275 * A newly forked process directly context switches into this address.
242 * 276 *
243 * eax: prev task we switched from 277 * eax: prev task we switched from
@@ -245,9 +279,7 @@ END(__switch_to_asm)
245 * edi: kernel thread arg 279 * edi: kernel thread arg
246 */ 280 */
247ENTRY(ret_from_fork) 281ENTRY(ret_from_fork)
248 pushl %eax 282 call schedule_tail_wrapper
249 call schedule_tail
250 popl %eax
251 283
252 testl %ebx, %ebx 284 testl %ebx, %ebx
253 jnz 1f /* kernel threads are uncommon */ 285 jnz 1f /* kernel threads are uncommon */
@@ -307,13 +339,13 @@ END(ret_from_exception)
307#ifdef CONFIG_PREEMPT 339#ifdef CONFIG_PREEMPT
308ENTRY(resume_kernel) 340ENTRY(resume_kernel)
309 DISABLE_INTERRUPTS(CLBR_ANY) 341 DISABLE_INTERRUPTS(CLBR_ANY)
310need_resched: 342.Lneed_resched:
311 cmpl $0, PER_CPU_VAR(__preempt_count) 343 cmpl $0, PER_CPU_VAR(__preempt_count)
312 jnz restore_all 344 jnz restore_all
313 testl $X86_EFLAGS_IF, PT_EFLAGS(%esp) # interrupts off (exception path) ? 345 testl $X86_EFLAGS_IF, PT_EFLAGS(%esp) # interrupts off (exception path) ?
314 jz restore_all 346 jz restore_all
315 call preempt_schedule_irq 347 call preempt_schedule_irq
316 jmp need_resched 348 jmp .Lneed_resched
317END(resume_kernel) 349END(resume_kernel)
318#endif 350#endif
319 351
@@ -334,7 +366,7 @@ GLOBAL(__begin_SYSENTER_singlestep_region)
334 */ 366 */
335ENTRY(xen_sysenter_target) 367ENTRY(xen_sysenter_target)
336 addl $5*4, %esp /* remove xen-provided frame */ 368 addl $5*4, %esp /* remove xen-provided frame */
337 jmp sysenter_past_esp 369 jmp .Lsysenter_past_esp
338#endif 370#endif
339 371
340/* 372/*
@@ -371,7 +403,7 @@ ENTRY(xen_sysenter_target)
371 */ 403 */
372ENTRY(entry_SYSENTER_32) 404ENTRY(entry_SYSENTER_32)
373 movl TSS_sysenter_sp0(%esp), %esp 405 movl TSS_sysenter_sp0(%esp), %esp
374sysenter_past_esp: 406.Lsysenter_past_esp:
375 pushl $__USER_DS /* pt_regs->ss */ 407 pushl $__USER_DS /* pt_regs->ss */
376 pushl %ebp /* pt_regs->sp (stashed in bp) */ 408 pushl %ebp /* pt_regs->sp (stashed in bp) */
377 pushfl /* pt_regs->flags (except IF = 0) */ 409 pushfl /* pt_regs->flags (except IF = 0) */
@@ -504,9 +536,9 @@ ENTRY(entry_INT80_32)
504 536
505restore_all: 537restore_all:
506 TRACE_IRQS_IRET 538 TRACE_IRQS_IRET
507restore_all_notrace: 539.Lrestore_all_notrace:
508#ifdef CONFIG_X86_ESPFIX32 540#ifdef CONFIG_X86_ESPFIX32
509 ALTERNATIVE "jmp restore_nocheck", "", X86_BUG_ESPFIX 541 ALTERNATIVE "jmp .Lrestore_nocheck", "", X86_BUG_ESPFIX
510 542
511 movl PT_EFLAGS(%esp), %eax # mix EFLAGS, SS and CS 543 movl PT_EFLAGS(%esp), %eax # mix EFLAGS, SS and CS
512 /* 544 /*
@@ -518,22 +550,23 @@ restore_all_notrace:
518 movb PT_CS(%esp), %al 550 movb PT_CS(%esp), %al
519 andl $(X86_EFLAGS_VM | (SEGMENT_TI_MASK << 8) | SEGMENT_RPL_MASK), %eax 551 andl $(X86_EFLAGS_VM | (SEGMENT_TI_MASK << 8) | SEGMENT_RPL_MASK), %eax
520 cmpl $((SEGMENT_LDT << 8) | USER_RPL), %eax 552 cmpl $((SEGMENT_LDT << 8) | USER_RPL), %eax
521 je ldt_ss # returning to user-space with LDT SS 553 je .Lldt_ss # returning to user-space with LDT SS
522#endif 554#endif
523restore_nocheck: 555.Lrestore_nocheck:
524 RESTORE_REGS 4 # skip orig_eax/error_code 556 RESTORE_REGS 4 # skip orig_eax/error_code
525irq_return: 557.Lirq_return:
526 INTERRUPT_RETURN 558 INTERRUPT_RETURN
559
527.section .fixup, "ax" 560.section .fixup, "ax"
528ENTRY(iret_exc ) 561ENTRY(iret_exc )
529 pushl $0 # no error code 562 pushl $0 # no error code
530 pushl $do_iret_error 563 pushl $do_iret_error
531 jmp error_code 564 jmp common_exception
532.previous 565.previous
533 _ASM_EXTABLE(irq_return, iret_exc) 566 _ASM_EXTABLE(.Lirq_return, iret_exc)
534 567
535#ifdef CONFIG_X86_ESPFIX32 568#ifdef CONFIG_X86_ESPFIX32
536ldt_ss: 569.Lldt_ss:
537/* 570/*
538 * Setup and switch to ESPFIX stack 571 * Setup and switch to ESPFIX stack
539 * 572 *
@@ -562,7 +595,7 @@ ldt_ss:
562 */ 595 */
563 DISABLE_INTERRUPTS(CLBR_EAX) 596 DISABLE_INTERRUPTS(CLBR_EAX)
564 lss (%esp), %esp /* switch to espfix segment */ 597 lss (%esp), %esp /* switch to espfix segment */
565 jmp restore_nocheck 598 jmp .Lrestore_nocheck
566#endif 599#endif
567ENDPROC(entry_INT80_32) 600ENDPROC(entry_INT80_32)
568 601
@@ -624,6 +657,7 @@ common_interrupt:
624 ASM_CLAC 657 ASM_CLAC
625 addl $-0x80, (%esp) /* Adjust vector into the [-256, -1] range */ 658 addl $-0x80, (%esp) /* Adjust vector into the [-256, -1] range */
626 SAVE_ALL 659 SAVE_ALL
660 ENCODE_FRAME_POINTER
627 TRACE_IRQS_OFF 661 TRACE_IRQS_OFF
628 movl %esp, %eax 662 movl %esp, %eax
629 call do_IRQ 663 call do_IRQ
@@ -635,6 +669,7 @@ ENTRY(name) \
635 ASM_CLAC; \ 669 ASM_CLAC; \
636 pushl $~(nr); \ 670 pushl $~(nr); \
637 SAVE_ALL; \ 671 SAVE_ALL; \
672 ENCODE_FRAME_POINTER; \
638 TRACE_IRQS_OFF \ 673 TRACE_IRQS_OFF \
639 movl %esp, %eax; \ 674 movl %esp, %eax; \
640 call fn; \ 675 call fn; \
@@ -659,7 +694,7 @@ ENTRY(coprocessor_error)
659 ASM_CLAC 694 ASM_CLAC
660 pushl $0 695 pushl $0
661 pushl $do_coprocessor_error 696 pushl $do_coprocessor_error
662 jmp error_code 697 jmp common_exception
663END(coprocessor_error) 698END(coprocessor_error)
664 699
665ENTRY(simd_coprocessor_error) 700ENTRY(simd_coprocessor_error)
@@ -673,14 +708,14 @@ ENTRY(simd_coprocessor_error)
673#else 708#else
674 pushl $do_simd_coprocessor_error 709 pushl $do_simd_coprocessor_error
675#endif 710#endif
676 jmp error_code 711 jmp common_exception
677END(simd_coprocessor_error) 712END(simd_coprocessor_error)
678 713
679ENTRY(device_not_available) 714ENTRY(device_not_available)
680 ASM_CLAC 715 ASM_CLAC
681 pushl $-1 # mark this as an int 716 pushl $-1 # mark this as an int
682 pushl $do_device_not_available 717 pushl $do_device_not_available
683 jmp error_code 718 jmp common_exception
684END(device_not_available) 719END(device_not_available)
685 720
686#ifdef CONFIG_PARAVIRT 721#ifdef CONFIG_PARAVIRT
@@ -694,59 +729,59 @@ ENTRY(overflow)
694 ASM_CLAC 729 ASM_CLAC
695 pushl $0 730 pushl $0
696 pushl $do_overflow 731 pushl $do_overflow
697 jmp error_code 732 jmp common_exception
698END(overflow) 733END(overflow)
699 734
700ENTRY(bounds) 735ENTRY(bounds)
701 ASM_CLAC 736 ASM_CLAC
702 pushl $0 737 pushl $0
703 pushl $do_bounds 738 pushl $do_bounds
704 jmp error_code 739 jmp common_exception
705END(bounds) 740END(bounds)
706 741
707ENTRY(invalid_op) 742ENTRY(invalid_op)
708 ASM_CLAC 743 ASM_CLAC
709 pushl $0 744 pushl $0
710 pushl $do_invalid_op 745 pushl $do_invalid_op
711 jmp error_code 746 jmp common_exception
712END(invalid_op) 747END(invalid_op)
713 748
714ENTRY(coprocessor_segment_overrun) 749ENTRY(coprocessor_segment_overrun)
715 ASM_CLAC 750 ASM_CLAC
716 pushl $0 751 pushl $0
717 pushl $do_coprocessor_segment_overrun 752 pushl $do_coprocessor_segment_overrun
718 jmp error_code 753 jmp common_exception
719END(coprocessor_segment_overrun) 754END(coprocessor_segment_overrun)
720 755
721ENTRY(invalid_TSS) 756ENTRY(invalid_TSS)
722 ASM_CLAC 757 ASM_CLAC
723 pushl $do_invalid_TSS 758 pushl $do_invalid_TSS
724 jmp error_code 759 jmp common_exception
725END(invalid_TSS) 760END(invalid_TSS)
726 761
727ENTRY(segment_not_present) 762ENTRY(segment_not_present)
728 ASM_CLAC 763 ASM_CLAC
729 pushl $do_segment_not_present 764 pushl $do_segment_not_present
730 jmp error_code 765 jmp common_exception
731END(segment_not_present) 766END(segment_not_present)
732 767
733ENTRY(stack_segment) 768ENTRY(stack_segment)
734 ASM_CLAC 769 ASM_CLAC
735 pushl $do_stack_segment 770 pushl $do_stack_segment
736 jmp error_code 771 jmp common_exception
737END(stack_segment) 772END(stack_segment)
738 773
739ENTRY(alignment_check) 774ENTRY(alignment_check)
740 ASM_CLAC 775 ASM_CLAC
741 pushl $do_alignment_check 776 pushl $do_alignment_check
742 jmp error_code 777 jmp common_exception
743END(alignment_check) 778END(alignment_check)
744 779
745ENTRY(divide_error) 780ENTRY(divide_error)
746 ASM_CLAC 781 ASM_CLAC
747 pushl $0 # no error code 782 pushl $0 # no error code
748 pushl $do_divide_error 783 pushl $do_divide_error
749 jmp error_code 784 jmp common_exception
750END(divide_error) 785END(divide_error)
751 786
752#ifdef CONFIG_X86_MCE 787#ifdef CONFIG_X86_MCE
@@ -754,7 +789,7 @@ ENTRY(machine_check)
754 ASM_CLAC 789 ASM_CLAC
755 pushl $0 790 pushl $0
756 pushl machine_check_vector 791 pushl machine_check_vector
757 jmp error_code 792 jmp common_exception
758END(machine_check) 793END(machine_check)
759#endif 794#endif
760 795
@@ -762,13 +797,14 @@ ENTRY(spurious_interrupt_bug)
762 ASM_CLAC 797 ASM_CLAC
763 pushl $0 798 pushl $0
764 pushl $do_spurious_interrupt_bug 799 pushl $do_spurious_interrupt_bug
765 jmp error_code 800 jmp common_exception
766END(spurious_interrupt_bug) 801END(spurious_interrupt_bug)
767 802
768#ifdef CONFIG_XEN 803#ifdef CONFIG_XEN
769ENTRY(xen_hypervisor_callback) 804ENTRY(xen_hypervisor_callback)
770 pushl $-1 /* orig_ax = -1 => not a system call */ 805 pushl $-1 /* orig_ax = -1 => not a system call */
771 SAVE_ALL 806 SAVE_ALL
807 ENCODE_FRAME_POINTER
772 TRACE_IRQS_OFF 808 TRACE_IRQS_OFF
773 809
774 /* 810 /*
@@ -823,6 +859,7 @@ ENTRY(xen_failsafe_callback)
823 jmp iret_exc 859 jmp iret_exc
8245: pushl $-1 /* orig_ax = -1 => not a system call */ 8605: pushl $-1 /* orig_ax = -1 => not a system call */
825 SAVE_ALL 861 SAVE_ALL
862 ENCODE_FRAME_POINTER
826 jmp ret_from_exception 863 jmp ret_from_exception
827 864
828.section .fixup, "ax" 865.section .fixup, "ax"
@@ -882,7 +919,7 @@ ftrace_call:
882 popl %edx 919 popl %edx
883 popl %ecx 920 popl %ecx
884 popl %eax 921 popl %eax
885ftrace_ret: 922.Lftrace_ret:
886#ifdef CONFIG_FUNCTION_GRAPH_TRACER 923#ifdef CONFIG_FUNCTION_GRAPH_TRACER
887.globl ftrace_graph_call 924.globl ftrace_graph_call
888ftrace_graph_call: 925ftrace_graph_call:
@@ -952,7 +989,7 @@ GLOBAL(ftrace_regs_call)
952 popl %gs 989 popl %gs
953 addl $8, %esp /* Skip orig_ax and ip */ 990 addl $8, %esp /* Skip orig_ax and ip */
954 popf /* Pop flags at end (no addl to corrupt flags) */ 991 popf /* Pop flags at end (no addl to corrupt flags) */
955 jmp ftrace_ret 992 jmp .Lftrace_ret
956 993
957 popf 994 popf
958 jmp ftrace_stub 995 jmp ftrace_stub
@@ -963,7 +1000,7 @@ ENTRY(mcount)
963 jb ftrace_stub /* Paging not enabled yet? */ 1000 jb ftrace_stub /* Paging not enabled yet? */
964 1001
965 cmpl $ftrace_stub, ftrace_trace_function 1002 cmpl $ftrace_stub, ftrace_trace_function
966 jnz trace 1003 jnz .Ltrace
967#ifdef CONFIG_FUNCTION_GRAPH_TRACER 1004#ifdef CONFIG_FUNCTION_GRAPH_TRACER
968 cmpl $ftrace_stub, ftrace_graph_return 1005 cmpl $ftrace_stub, ftrace_graph_return
969 jnz ftrace_graph_caller 1006 jnz ftrace_graph_caller
@@ -976,7 +1013,7 @@ ftrace_stub:
976 ret 1013 ret
977 1014
978 /* taken from glibc */ 1015 /* taken from glibc */
979trace: 1016.Ltrace:
980 pushl %eax 1017 pushl %eax
981 pushl %ecx 1018 pushl %ecx
982 pushl %edx 1019 pushl %edx
@@ -1027,7 +1064,7 @@ return_to_handler:
1027ENTRY(trace_page_fault) 1064ENTRY(trace_page_fault)
1028 ASM_CLAC 1065 ASM_CLAC
1029 pushl $trace_do_page_fault 1066 pushl $trace_do_page_fault
1030 jmp error_code 1067 jmp common_exception
1031END(trace_page_fault) 1068END(trace_page_fault)
1032#endif 1069#endif
1033 1070
@@ -1035,7 +1072,10 @@ ENTRY(page_fault)
1035 ASM_CLAC 1072 ASM_CLAC
1036 pushl $do_page_fault 1073 pushl $do_page_fault
1037 ALIGN 1074 ALIGN
1038error_code: 1075 jmp common_exception
1076END(page_fault)
1077
1078common_exception:
1039 /* the function address is in %gs's slot on the stack */ 1079 /* the function address is in %gs's slot on the stack */
1040 pushl %fs 1080 pushl %fs
1041 pushl %es 1081 pushl %es
@@ -1047,6 +1087,7 @@ error_code:
1047 pushl %edx 1087 pushl %edx
1048 pushl %ecx 1088 pushl %ecx
1049 pushl %ebx 1089 pushl %ebx
1090 ENCODE_FRAME_POINTER
1050 cld 1091 cld
1051 movl $(__KERNEL_PERCPU), %ecx 1092 movl $(__KERNEL_PERCPU), %ecx
1052 movl %ecx, %fs 1093 movl %ecx, %fs
@@ -1064,7 +1105,7 @@ error_code:
1064 movl %esp, %eax # pt_regs pointer 1105 movl %esp, %eax # pt_regs pointer
1065 call *%edi 1106 call *%edi
1066 jmp ret_from_exception 1107 jmp ret_from_exception
1067END(page_fault) 1108END(common_exception)
1068 1109
1069ENTRY(debug) 1110ENTRY(debug)
1070 /* 1111 /*
@@ -1079,6 +1120,7 @@ ENTRY(debug)
1079 ASM_CLAC 1120 ASM_CLAC
1080 pushl $-1 # mark this as an int 1121 pushl $-1 # mark this as an int
1081 SAVE_ALL 1122 SAVE_ALL
1123 ENCODE_FRAME_POINTER
1082 xorl %edx, %edx # error code 0 1124 xorl %edx, %edx # error code 0
1083 movl %esp, %eax # pt_regs pointer 1125 movl %esp, %eax # pt_regs pointer
1084 1126
@@ -1094,11 +1136,11 @@ ENTRY(debug)
1094 1136
1095.Ldebug_from_sysenter_stack: 1137.Ldebug_from_sysenter_stack:
1096 /* We're on the SYSENTER stack. Switch off. */ 1138 /* We're on the SYSENTER stack. Switch off. */
1097 movl %esp, %ebp 1139 movl %esp, %ebx
1098 movl PER_CPU_VAR(cpu_current_top_of_stack), %esp 1140 movl PER_CPU_VAR(cpu_current_top_of_stack), %esp
1099 TRACE_IRQS_OFF 1141 TRACE_IRQS_OFF
1100 call do_debug 1142 call do_debug
1101 movl %ebp, %esp 1143 movl %ebx, %esp
1102 jmp ret_from_exception 1144 jmp ret_from_exception
1103END(debug) 1145END(debug)
1104 1146
@@ -1116,11 +1158,12 @@ ENTRY(nmi)
1116 movl %ss, %eax 1158 movl %ss, %eax
1117 cmpw $__ESPFIX_SS, %ax 1159 cmpw $__ESPFIX_SS, %ax
1118 popl %eax 1160 popl %eax
1119 je nmi_espfix_stack 1161 je .Lnmi_espfix_stack
1120#endif 1162#endif
1121 1163
1122 pushl %eax # pt_regs->orig_ax 1164 pushl %eax # pt_regs->orig_ax
1123 SAVE_ALL 1165 SAVE_ALL
1166 ENCODE_FRAME_POINTER
1124 xorl %edx, %edx # zero error code 1167 xorl %edx, %edx # zero error code
1125 movl %esp, %eax # pt_regs pointer 1168 movl %esp, %eax # pt_regs pointer
1126 1169
@@ -1132,21 +1175,21 @@ ENTRY(nmi)
1132 1175
1133 /* Not on SYSENTER stack. */ 1176 /* Not on SYSENTER stack. */
1134 call do_nmi 1177 call do_nmi
1135 jmp restore_all_notrace 1178 jmp .Lrestore_all_notrace
1136 1179
1137.Lnmi_from_sysenter_stack: 1180.Lnmi_from_sysenter_stack:
1138 /* 1181 /*
1139 * We're on the SYSENTER stack. Switch off. No one (not even debug) 1182 * We're on the SYSENTER stack. Switch off. No one (not even debug)
1140 * is using the thread stack right now, so it's safe for us to use it. 1183 * is using the thread stack right now, so it's safe for us to use it.
1141 */ 1184 */
1142 movl %esp, %ebp 1185 movl %esp, %ebx
1143 movl PER_CPU_VAR(cpu_current_top_of_stack), %esp 1186 movl PER_CPU_VAR(cpu_current_top_of_stack), %esp
1144 call do_nmi 1187 call do_nmi
1145 movl %ebp, %esp 1188 movl %ebx, %esp
1146 jmp restore_all_notrace 1189 jmp .Lrestore_all_notrace
1147 1190
1148#ifdef CONFIG_X86_ESPFIX32 1191#ifdef CONFIG_X86_ESPFIX32
1149nmi_espfix_stack: 1192.Lnmi_espfix_stack:
1150 /* 1193 /*
1151 * create the pointer to lss back 1194 * create the pointer to lss back
1152 */ 1195 */
@@ -1159,12 +1202,13 @@ nmi_espfix_stack:
1159 .endr 1202 .endr
1160 pushl %eax 1203 pushl %eax
1161 SAVE_ALL 1204 SAVE_ALL
1205 ENCODE_FRAME_POINTER
1162 FIXUP_ESPFIX_STACK # %eax == %esp 1206 FIXUP_ESPFIX_STACK # %eax == %esp
1163 xorl %edx, %edx # zero error code 1207 xorl %edx, %edx # zero error code
1164 call do_nmi 1208 call do_nmi
1165 RESTORE_REGS 1209 RESTORE_REGS
1166 lss 12+4(%esp), %esp # back to espfix stack 1210 lss 12+4(%esp), %esp # back to espfix stack
1167 jmp irq_return 1211 jmp .Lirq_return
1168#endif 1212#endif
1169END(nmi) 1213END(nmi)
1170 1214
@@ -1172,6 +1216,7 @@ ENTRY(int3)
1172 ASM_CLAC 1216 ASM_CLAC
1173 pushl $-1 # mark this as an int 1217 pushl $-1 # mark this as an int
1174 SAVE_ALL 1218 SAVE_ALL
1219 ENCODE_FRAME_POINTER
1175 TRACE_IRQS_OFF 1220 TRACE_IRQS_OFF
1176 xorl %edx, %edx # zero error code 1221 xorl %edx, %edx # zero error code
1177 movl %esp, %eax # pt_regs pointer 1222 movl %esp, %eax # pt_regs pointer
@@ -1181,14 +1226,14 @@ END(int3)
1181 1226
1182ENTRY(general_protection) 1227ENTRY(general_protection)
1183 pushl $do_general_protection 1228 pushl $do_general_protection
1184 jmp error_code 1229 jmp common_exception
1185END(general_protection) 1230END(general_protection)
1186 1231
1187#ifdef CONFIG_KVM_GUEST 1232#ifdef CONFIG_KVM_GUEST
1188ENTRY(async_page_fault) 1233ENTRY(async_page_fault)
1189 ASM_CLAC 1234 ASM_CLAC
1190 pushl $do_async_page_fault 1235 pushl $do_async_page_fault
1191 jmp error_code 1236 jmp common_exception
1192END(async_page_fault) 1237END(async_page_fault)
1193#endif 1238#endif
1194 1239
diff --git a/arch/x86/entry/entry_64.S b/arch/x86/entry/entry_64.S
index ef766a358b37..5b219707c2f2 100644
--- a/arch/x86/entry/entry_64.S
+++ b/arch/x86/entry/entry_64.S
@@ -38,12 +38,6 @@
38#include <asm/export.h> 38#include <asm/export.h>
39#include <linux/err.h> 39#include <linux/err.h>
40 40
41/* Avoid __ASSEMBLER__'ifying <linux/audit.h> just for this. */
42#include <linux/elf-em.h>
43#define AUDIT_ARCH_X86_64 (EM_X86_64|__AUDIT_ARCH_64BIT|__AUDIT_ARCH_LE)
44#define __AUDIT_ARCH_64BIT 0x80000000
45#define __AUDIT_ARCH_LE 0x40000000
46
47.code64 41.code64
48.section .entry.text, "ax" 42.section .entry.text, "ax"
49 43
@@ -469,6 +463,7 @@ END(irq_entries_start)
469 ALLOC_PT_GPREGS_ON_STACK 463 ALLOC_PT_GPREGS_ON_STACK
470 SAVE_C_REGS 464 SAVE_C_REGS
471 SAVE_EXTRA_REGS 465 SAVE_EXTRA_REGS
466 ENCODE_FRAME_POINTER
472 467
473 testb $3, CS(%rsp) 468 testb $3, CS(%rsp)
474 jz 1f 469 jz 1f
@@ -985,6 +980,7 @@ ENTRY(xen_failsafe_callback)
985 ALLOC_PT_GPREGS_ON_STACK 980 ALLOC_PT_GPREGS_ON_STACK
986 SAVE_C_REGS 981 SAVE_C_REGS
987 SAVE_EXTRA_REGS 982 SAVE_EXTRA_REGS
983 ENCODE_FRAME_POINTER
988 jmp error_exit 984 jmp error_exit
989END(xen_failsafe_callback) 985END(xen_failsafe_callback)
990 986
@@ -1028,6 +1024,7 @@ ENTRY(paranoid_entry)
1028 cld 1024 cld
1029 SAVE_C_REGS 8 1025 SAVE_C_REGS 8
1030 SAVE_EXTRA_REGS 8 1026 SAVE_EXTRA_REGS 8
1027 ENCODE_FRAME_POINTER 8
1031 movl $1, %ebx 1028 movl $1, %ebx
1032 movl $MSR_GS_BASE, %ecx 1029 movl $MSR_GS_BASE, %ecx
1033 rdmsr 1030 rdmsr
@@ -1075,6 +1072,7 @@ ENTRY(error_entry)
1075 cld 1072 cld
1076 SAVE_C_REGS 8 1073 SAVE_C_REGS 8
1077 SAVE_EXTRA_REGS 8 1074 SAVE_EXTRA_REGS 8
1075 ENCODE_FRAME_POINTER 8
1078 xorl %ebx, %ebx 1076 xorl %ebx, %ebx
1079 testb $3, CS+8(%rsp) 1077 testb $3, CS+8(%rsp)
1080 jz .Lerror_kernelspace 1078 jz .Lerror_kernelspace
@@ -1257,6 +1255,7 @@ ENTRY(nmi)
1257 pushq %r13 /* pt_regs->r13 */ 1255 pushq %r13 /* pt_regs->r13 */
1258 pushq %r14 /* pt_regs->r14 */ 1256 pushq %r14 /* pt_regs->r14 */
1259 pushq %r15 /* pt_regs->r15 */ 1257 pushq %r15 /* pt_regs->r15 */
1258 ENCODE_FRAME_POINTER
1260 1259
1261 /* 1260 /*
1262 * At this point we no longer need to worry about stack damage 1261 * At this point we no longer need to worry about stack damage
@@ -1270,11 +1269,10 @@ ENTRY(nmi)
1270 1269
1271 /* 1270 /*
1272 * Return back to user mode. We must *not* do the normal exit 1271 * Return back to user mode. We must *not* do the normal exit
1273 * work, because we don't want to enable interrupts. Fortunately, 1272 * work, because we don't want to enable interrupts.
1274 * do_nmi doesn't modify pt_regs.
1275 */ 1273 */
1276 SWAPGS 1274 SWAPGS
1277 jmp restore_c_regs_and_iret 1275 jmp restore_regs_and_iret
1278 1276
1279.Lnmi_from_kernel: 1277.Lnmi_from_kernel:
1280 /* 1278 /*
diff --git a/arch/x86/entry/vdso/vma.c b/arch/x86/entry/vdso/vma.c
index 23c881caabd1..e739002427ed 100644
--- a/arch/x86/entry/vdso/vma.c
+++ b/arch/x86/entry/vdso/vma.c
@@ -161,8 +161,6 @@ static int map_vdso(const struct vdso_image *image, unsigned long addr)
161 } 161 }
162 162
163 text_start = addr - image->sym_vvar_start; 163 text_start = addr - image->sym_vvar_start;
164 current->mm->context.vdso = (void __user *)text_start;
165 current->mm->context.vdso_image = image;
166 164
167 /* 165 /*
168 * MAYWRITE to allow gdb to COW and set breakpoints 166 * MAYWRITE to allow gdb to COW and set breakpoints
@@ -189,14 +187,12 @@ static int map_vdso(const struct vdso_image *image, unsigned long addr)
189 if (IS_ERR(vma)) { 187 if (IS_ERR(vma)) {
190 ret = PTR_ERR(vma); 188 ret = PTR_ERR(vma);
191 do_munmap(mm, text_start, image->size); 189 do_munmap(mm, text_start, image->size);
190 } else {
191 current->mm->context.vdso = (void __user *)text_start;
192 current->mm->context.vdso_image = image;
192 } 193 }
193 194
194up_fail: 195up_fail:
195 if (ret) {
196 current->mm->context.vdso = NULL;
197 current->mm->context.vdso_image = NULL;
198 }
199
200 up_write(&mm->mmap_sem); 196 up_write(&mm->mmap_sem);
201 return ret; 197 return ret;
202} 198}
diff --git a/arch/x86/events/intel/pt.c b/arch/x86/events/intel/pt.c
index c5047b8f777b..1c1b9fe705c8 100644
--- a/arch/x86/events/intel/pt.c
+++ b/arch/x86/events/intel/pt.c
@@ -36,13 +36,6 @@ static DEFINE_PER_CPU(struct pt, pt_ctx);
36 36
37static struct pt_pmu pt_pmu; 37static struct pt_pmu pt_pmu;
38 38
39enum cpuid_regs {
40 CR_EAX = 0,
41 CR_ECX,
42 CR_EDX,
43 CR_EBX
44};
45
46/* 39/*
47 * Capabilities of Intel PT hardware, such as number of address bits or 40 * Capabilities of Intel PT hardware, such as number of address bits or
48 * supported output schemes, are cached and exported to userspace as "caps" 41 * supported output schemes, are cached and exported to userspace as "caps"
@@ -64,21 +57,21 @@ static struct pt_cap_desc {
64 u8 reg; 57 u8 reg;
65 u32 mask; 58 u32 mask;
66} pt_caps[] = { 59} pt_caps[] = {
67 PT_CAP(max_subleaf, 0, CR_EAX, 0xffffffff), 60 PT_CAP(max_subleaf, 0, CPUID_EAX, 0xffffffff),
68 PT_CAP(cr3_filtering, 0, CR_EBX, BIT(0)), 61 PT_CAP(cr3_filtering, 0, CPUID_EBX, BIT(0)),
69 PT_CAP(psb_cyc, 0, CR_EBX, BIT(1)), 62 PT_CAP(psb_cyc, 0, CPUID_EBX, BIT(1)),
70 PT_CAP(ip_filtering, 0, CR_EBX, BIT(2)), 63 PT_CAP(ip_filtering, 0, CPUID_EBX, BIT(2)),
71 PT_CAP(mtc, 0, CR_EBX, BIT(3)), 64 PT_CAP(mtc, 0, CPUID_EBX, BIT(3)),
72 PT_CAP(ptwrite, 0, CR_EBX, BIT(4)), 65 PT_CAP(ptwrite, 0, CPUID_EBX, BIT(4)),
73 PT_CAP(power_event_trace, 0, CR_EBX, BIT(5)), 66 PT_CAP(power_event_trace, 0, CPUID_EBX, BIT(5)),
74 PT_CAP(topa_output, 0, CR_ECX, BIT(0)), 67 PT_CAP(topa_output, 0, CPUID_ECX, BIT(0)),
75 PT_CAP(topa_multiple_entries, 0, CR_ECX, BIT(1)), 68 PT_CAP(topa_multiple_entries, 0, CPUID_ECX, BIT(1)),
76 PT_CAP(single_range_output, 0, CR_ECX, BIT(2)), 69 PT_CAP(single_range_output, 0, CPUID_ECX, BIT(2)),
77 PT_CAP(payloads_lip, 0, CR_ECX, BIT(31)), 70 PT_CAP(payloads_lip, 0, CPUID_ECX, BIT(31)),
78 PT_CAP(num_address_ranges, 1, CR_EAX, 0x3), 71 PT_CAP(num_address_ranges, 1, CPUID_EAX, 0x3),
79 PT_CAP(mtc_periods, 1, CR_EAX, 0xffff0000), 72 PT_CAP(mtc_periods, 1, CPUID_EAX, 0xffff0000),
80 PT_CAP(cycle_thresholds, 1, CR_EBX, 0xffff), 73 PT_CAP(cycle_thresholds, 1, CPUID_EBX, 0xffff),
81 PT_CAP(psb_periods, 1, CR_EBX, 0xffff0000), 74 PT_CAP(psb_periods, 1, CPUID_EBX, 0xffff0000),
82}; 75};
83 76
84static u32 pt_cap_get(enum pt_capabilities cap) 77static u32 pt_cap_get(enum pt_capabilities cap)
@@ -213,10 +206,10 @@ static int __init pt_pmu_hw_init(void)
213 206
214 for (i = 0; i < PT_CPUID_LEAVES; i++) { 207 for (i = 0; i < PT_CPUID_LEAVES; i++) {
215 cpuid_count(20, i, 208 cpuid_count(20, i,
216 &pt_pmu.caps[CR_EAX + i*PT_CPUID_REGS_NUM], 209 &pt_pmu.caps[CPUID_EAX + i*PT_CPUID_REGS_NUM],
217 &pt_pmu.caps[CR_EBX + i*PT_CPUID_REGS_NUM], 210 &pt_pmu.caps[CPUID_EBX + i*PT_CPUID_REGS_NUM],
218 &pt_pmu.caps[CR_ECX + i*PT_CPUID_REGS_NUM], 211 &pt_pmu.caps[CPUID_ECX + i*PT_CPUID_REGS_NUM],
219 &pt_pmu.caps[CR_EDX + i*PT_CPUID_REGS_NUM]); 212 &pt_pmu.caps[CPUID_EDX + i*PT_CPUID_REGS_NUM]);
220 } 213 }
221 214
222 ret = -ENOMEM; 215 ret = -ENOMEM;
diff --git a/arch/x86/include/asm/cpufeatures.h b/arch/x86/include/asm/cpufeatures.h
index d625b651e526..4dba597c5807 100644
--- a/arch/x86/include/asm/cpufeatures.h
+++ b/arch/x86/include/asm/cpufeatures.h
@@ -227,6 +227,7 @@
227#define X86_FEATURE_RDSEED ( 9*32+18) /* The RDSEED instruction */ 227#define X86_FEATURE_RDSEED ( 9*32+18) /* The RDSEED instruction */
228#define X86_FEATURE_ADX ( 9*32+19) /* The ADCX and ADOX instructions */ 228#define X86_FEATURE_ADX ( 9*32+19) /* The ADCX and ADOX instructions */
229#define X86_FEATURE_SMAP ( 9*32+20) /* Supervisor Mode Access Prevention */ 229#define X86_FEATURE_SMAP ( 9*32+20) /* Supervisor Mode Access Prevention */
230#define X86_FEATURE_AVX512IFMA ( 9*32+21) /* AVX-512 Integer Fused Multiply-Add instructions */
230#define X86_FEATURE_CLFLUSHOPT ( 9*32+23) /* CLFLUSHOPT instruction */ 231#define X86_FEATURE_CLFLUSHOPT ( 9*32+23) /* CLFLUSHOPT instruction */
231#define X86_FEATURE_CLWB ( 9*32+24) /* CLWB instruction */ 232#define X86_FEATURE_CLWB ( 9*32+24) /* CLWB instruction */
232#define X86_FEATURE_AVX512PF ( 9*32+26) /* AVX-512 Prefetch */ 233#define X86_FEATURE_AVX512PF ( 9*32+26) /* AVX-512 Prefetch */
@@ -280,8 +281,10 @@
280#define X86_FEATURE_AVIC (15*32+13) /* Virtual Interrupt Controller */ 281#define X86_FEATURE_AVIC (15*32+13) /* Virtual Interrupt Controller */
281 282
282/* Intel-defined CPU features, CPUID level 0x00000007:0 (ecx), word 16 */ 283/* Intel-defined CPU features, CPUID level 0x00000007:0 (ecx), word 16 */
284#define X86_FEATURE_AVX512VBMI (16*32+ 1) /* AVX512 Vector Bit Manipulation instructions*/
283#define X86_FEATURE_PKU (16*32+ 3) /* Protection Keys for Userspace */ 285#define X86_FEATURE_PKU (16*32+ 3) /* Protection Keys for Userspace */
284#define X86_FEATURE_OSPKE (16*32+ 4) /* OS Protection Keys Enable */ 286#define X86_FEATURE_OSPKE (16*32+ 4) /* OS Protection Keys Enable */
287#define X86_FEATURE_RDPID (16*32+ 22) /* RDPID instruction */
285 288
286/* AMD-defined CPU features, CPUID level 0x80000007 (ebx), word 17 */ 289/* AMD-defined CPU features, CPUID level 0x80000007 (ebx), word 17 */
287#define X86_FEATURE_OVERFLOW_RECOV (17*32+0) /* MCA overflow recovery support */ 290#define X86_FEATURE_OVERFLOW_RECOV (17*32+0) /* MCA overflow recovery support */
diff --git a/arch/x86/include/asm/kdebug.h b/arch/x86/include/asm/kdebug.h
index d31881188431..29a594a3b82a 100644
--- a/arch/x86/include/asm/kdebug.h
+++ b/arch/x86/include/asm/kdebug.h
@@ -21,7 +21,6 @@ enum die_val {
21 DIE_NMIUNKNOWN, 21 DIE_NMIUNKNOWN,
22}; 22};
23 23
24extern void printk_address(unsigned long address);
25extern void die(const char *, struct pt_regs *,long); 24extern void die(const char *, struct pt_regs *,long);
26extern int __must_check __die(const char *, struct pt_regs *, long); 25extern int __must_check __die(const char *, struct pt_regs *, long);
27extern void show_stack_regs(struct pt_regs *regs); 26extern void show_stack_regs(struct pt_regs *regs);
diff --git a/arch/x86/include/asm/processor.h b/arch/x86/include/asm/processor.h
index c84605bb2a15..1f6a92903b09 100644
--- a/arch/x86/include/asm/processor.h
+++ b/arch/x86/include/asm/processor.h
@@ -137,6 +137,17 @@ struct cpuinfo_x86 {
137 u32 microcode; 137 u32 microcode;
138}; 138};
139 139
140struct cpuid_regs {
141 u32 eax, ebx, ecx, edx;
142};
143
144enum cpuid_regs_idx {
145 CPUID_EAX = 0,
146 CPUID_EBX,
147 CPUID_ECX,
148 CPUID_EDX,
149};
150
140#define X86_VENDOR_INTEL 0 151#define X86_VENDOR_INTEL 0
141#define X86_VENDOR_CYRIX 1 152#define X86_VENDOR_CYRIX 1
142#define X86_VENDOR_AMD 2 153#define X86_VENDOR_AMD 2
@@ -178,6 +189,9 @@ extern void identify_secondary_cpu(struct cpuinfo_x86 *);
178extern void print_cpu_info(struct cpuinfo_x86 *); 189extern void print_cpu_info(struct cpuinfo_x86 *);
179void print_cpu_msr(struct cpuinfo_x86 *); 190void print_cpu_msr(struct cpuinfo_x86 *);
180extern void init_scattered_cpuid_features(struct cpuinfo_x86 *c); 191extern void init_scattered_cpuid_features(struct cpuinfo_x86 *c);
192extern u32 get_scattered_cpuid_leaf(unsigned int level,
193 unsigned int sub_leaf,
194 enum cpuid_regs_idx reg);
181extern unsigned int init_intel_cacheinfo(struct cpuinfo_x86 *c); 195extern unsigned int init_intel_cacheinfo(struct cpuinfo_x86 *c);
182extern void init_amd_cacheinfo(struct cpuinfo_x86 *c); 196extern void init_amd_cacheinfo(struct cpuinfo_x86 *c);
183 197
diff --git a/arch/x86/include/asm/stacktrace.h b/arch/x86/include/asm/stacktrace.h
index 37f2e0b377ad..a3269c897ec5 100644
--- a/arch/x86/include/asm/stacktrace.h
+++ b/arch/x86/include/asm/stacktrace.h
@@ -30,8 +30,7 @@ bool in_task_stack(unsigned long *stack, struct task_struct *task,
30int get_stack_info(unsigned long *stack, struct task_struct *task, 30int get_stack_info(unsigned long *stack, struct task_struct *task,
31 struct stack_info *info, unsigned long *visit_mask); 31 struct stack_info *info, unsigned long *visit_mask);
32 32
33void stack_type_str(enum stack_type type, const char **begin, 33const char *stack_type_name(enum stack_type type);
34 const char **end);
35 34
36static inline bool on_stack(struct stack_info *info, void *addr, size_t len) 35static inline bool on_stack(struct stack_info *info, void *addr, size_t len)
37{ 36{
@@ -43,8 +42,6 @@ static inline bool on_stack(struct stack_info *info, void *addr, size_t len)
43 addr + len > begin && addr + len <= end); 42 addr + len > begin && addr + len <= end);
44} 43}
45 44
46extern int kstack_depth_to_print;
47
48#ifdef CONFIG_X86_32 45#ifdef CONFIG_X86_32
49#define STACKSLOTS_PER_LINE 8 46#define STACKSLOTS_PER_LINE 8
50#else 47#else
@@ -86,9 +83,6 @@ get_stack_pointer(struct task_struct *task, struct pt_regs *regs)
86void show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs, 83void show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs,
87 unsigned long *stack, char *log_lvl); 84 unsigned long *stack, char *log_lvl);
88 85
89void show_stack_log_lvl(struct task_struct *task, struct pt_regs *regs,
90 unsigned long *sp, char *log_lvl);
91
92extern unsigned int code_bytes; 86extern unsigned int code_bytes;
93 87
94/* The form of the top of the frame on the stack */ 88/* The form of the top of the frame on the stack */
diff --git a/arch/x86/include/asm/unwind.h b/arch/x86/include/asm/unwind.h
index 46de9ac4b990..c5a7f3a930dd 100644
--- a/arch/x86/include/asm/unwind.h
+++ b/arch/x86/include/asm/unwind.h
@@ -13,6 +13,7 @@ struct unwind_state {
13 int graph_idx; 13 int graph_idx;
14#ifdef CONFIG_FRAME_POINTER 14#ifdef CONFIG_FRAME_POINTER
15 unsigned long *bp; 15 unsigned long *bp;
16 struct pt_regs *regs;
16#else 17#else
17 unsigned long *sp; 18 unsigned long *sp;
18#endif 19#endif
@@ -47,7 +48,15 @@ unsigned long *unwind_get_return_address_ptr(struct unwind_state *state)
47 if (unwind_done(state)) 48 if (unwind_done(state))
48 return NULL; 49 return NULL;
49 50
50 return state->bp + 1; 51 return state->regs ? &state->regs->ip : state->bp + 1;
52}
53
54static inline struct pt_regs *unwind_get_entry_regs(struct unwind_state *state)
55{
56 if (unwind_done(state))
57 return NULL;
58
59 return state->regs;
51} 60}
52 61
53#else /* !CONFIG_FRAME_POINTER */ 62#else /* !CONFIG_FRAME_POINTER */
@@ -58,6 +67,11 @@ unsigned long *unwind_get_return_address_ptr(struct unwind_state *state)
58 return NULL; 67 return NULL;
59} 68}
60 69
70static inline struct pt_regs *unwind_get_entry_regs(struct unwind_state *state)
71{
72 return NULL;
73}
74
61#endif /* CONFIG_FRAME_POINTER */ 75#endif /* CONFIG_FRAME_POINTER */
62 76
63#endif /* _ASM_X86_UNWIND_H */ 77#endif /* _ASM_X86_UNWIND_H */
diff --git a/arch/x86/include/asm/vgtod.h b/arch/x86/include/asm/vgtod.h
index e728699db774..3a01996db58f 100644
--- a/arch/x86/include/asm/vgtod.h
+++ b/arch/x86/include/asm/vgtod.h
@@ -89,8 +89,13 @@ static inline unsigned int __getcpu(void)
89 * works on all CPUs. This is volatile so that it orders 89 * works on all CPUs. This is volatile so that it orders
90 * correctly wrt barrier() and to keep gcc from cleverly 90 * correctly wrt barrier() and to keep gcc from cleverly
91 * hoisting it out of the calling function. 91 * hoisting it out of the calling function.
92 *
93 * If RDPID is available, use it.
92 */ 94 */
93 asm volatile ("lsl %1,%0" : "=r" (p) : "r" (__PER_CPU_SEG)); 95 alternative_io ("lsl %[p],%[seg]",
96 ".byte 0xf3,0x0f,0xc7,0xf8", /* RDPID %eax/rax */
97 X86_FEATURE_RDPID,
98 [p] "=a" (p), [seg] "r" (__PER_CPU_SEG));
94 99
95 return p; 100 return p;
96} 101}
diff --git a/arch/x86/include/uapi/asm/prctl.h b/arch/x86/include/uapi/asm/prctl.h
index ae135de547f5..835aa51c7f6e 100644
--- a/arch/x86/include/uapi/asm/prctl.h
+++ b/arch/x86/include/uapi/asm/prctl.h
@@ -6,10 +6,8 @@
6#define ARCH_GET_FS 0x1003 6#define ARCH_GET_FS 0x1003
7#define ARCH_GET_GS 0x1004 7#define ARCH_GET_GS 0x1004
8 8
9#ifdef CONFIG_CHECKPOINT_RESTORE 9#define ARCH_MAP_VDSO_X32 0x2001
10# define ARCH_MAP_VDSO_X32 0x2001 10#define ARCH_MAP_VDSO_32 0x2002
11# define ARCH_MAP_VDSO_32 0x2002 11#define ARCH_MAP_VDSO_64 0x2003
12# define ARCH_MAP_VDSO_64 0x2003
13#endif
14 12
15#endif /* _ASM_X86_PRCTL_H */ 13#endif /* _ASM_X86_PRCTL_H */
diff --git a/arch/x86/kernel/cpu/scattered.c b/arch/x86/kernel/cpu/scattered.c
index 1db8dc490b66..d1316f9c8329 100644
--- a/arch/x86/kernel/cpu/scattered.c
+++ b/arch/x86/kernel/cpu/scattered.c
@@ -17,11 +17,17 @@ struct cpuid_bit {
17 u32 sub_leaf; 17 u32 sub_leaf;
18}; 18};
19 19
20enum cpuid_regs { 20/* Please keep the leaf sorted by cpuid_bit.level for faster search. */
21 CR_EAX = 0, 21static const struct cpuid_bit cpuid_bits[] = {
22 CR_ECX, 22 { X86_FEATURE_APERFMPERF, CPUID_ECX, 0, 0x00000006, 0 },
23 CR_EDX, 23 { X86_FEATURE_EPB, CPUID_ECX, 3, 0x00000006, 0 },
24 CR_EBX 24 { X86_FEATURE_INTEL_PT, CPUID_EBX, 25, 0x00000007, 0 },
25 { X86_FEATURE_AVX512_4VNNIW, CPUID_EDX, 2, 0x00000007, 0 },
26 { X86_FEATURE_AVX512_4FMAPS, CPUID_EDX, 3, 0x00000007, 0 },
27 { X86_FEATURE_HW_PSTATE, CPUID_EDX, 7, 0x80000007, 0 },
28 { X86_FEATURE_CPB, CPUID_EDX, 9, 0x80000007, 0 },
29 { X86_FEATURE_PROC_FEEDBACK, CPUID_EDX, 11, 0x80000007, 0 },
30 { 0, 0, 0, 0, 0 }
25}; 31};
26 32
27void init_scattered_cpuid_features(struct cpuinfo_x86 *c) 33void init_scattered_cpuid_features(struct cpuinfo_x86 *c)
@@ -30,18 +36,6 @@ void init_scattered_cpuid_features(struct cpuinfo_x86 *c)
30 u32 regs[4]; 36 u32 regs[4];
31 const struct cpuid_bit *cb; 37 const struct cpuid_bit *cb;
32 38
33 static const struct cpuid_bit cpuid_bits[] = {
34 { X86_FEATURE_INTEL_PT, CR_EBX,25, 0x00000007, 0 },
35 { X86_FEATURE_AVX512_4VNNIW, CR_EDX, 2, 0x00000007, 0 },
36 { X86_FEATURE_AVX512_4FMAPS, CR_EDX, 3, 0x00000007, 0 },
37 { X86_FEATURE_APERFMPERF, CR_ECX, 0, 0x00000006, 0 },
38 { X86_FEATURE_EPB, CR_ECX, 3, 0x00000006, 0 },
39 { X86_FEATURE_HW_PSTATE, CR_EDX, 7, 0x80000007, 0 },
40 { X86_FEATURE_CPB, CR_EDX, 9, 0x80000007, 0 },
41 { X86_FEATURE_PROC_FEEDBACK, CR_EDX,11, 0x80000007, 0 },
42 { 0, 0, 0, 0, 0 }
43 };
44
45 for (cb = cpuid_bits; cb->feature; cb++) { 39 for (cb = cpuid_bits; cb->feature; cb++) {
46 40
47 /* Verify that the level is valid */ 41 /* Verify that the level is valid */
@@ -50,10 +44,35 @@ void init_scattered_cpuid_features(struct cpuinfo_x86 *c)
50 max_level > (cb->level | 0xffff)) 44 max_level > (cb->level | 0xffff))
51 continue; 45 continue;
52 46
53 cpuid_count(cb->level, cb->sub_leaf, &regs[CR_EAX], 47 cpuid_count(cb->level, cb->sub_leaf, &regs[CPUID_EAX],
54 &regs[CR_EBX], &regs[CR_ECX], &regs[CR_EDX]); 48 &regs[CPUID_EBX], &regs[CPUID_ECX],
49 &regs[CPUID_EDX]);
55 50
56 if (regs[cb->reg] & (1 << cb->bit)) 51 if (regs[cb->reg] & (1 << cb->bit))
57 set_cpu_cap(c, cb->feature); 52 set_cpu_cap(c, cb->feature);
58 } 53 }
59} 54}
55
56u32 get_scattered_cpuid_leaf(unsigned int level, unsigned int sub_leaf,
57 enum cpuid_regs_idx reg)
58{
59 const struct cpuid_bit *cb;
60 u32 cpuid_val = 0;
61
62 for (cb = cpuid_bits; cb->feature; cb++) {
63
64 if (level > cb->level)
65 continue;
66
67 if (level < cb->level)
68 break;
69
70 if (reg == cb->reg && sub_leaf == cb->sub_leaf) {
71 if (cpu_has(&boot_cpu_data, cb->feature))
72 cpuid_val |= BIT(cb->bit);
73 }
74 }
75
76 return cpuid_val;
77}
78EXPORT_SYMBOL_GPL(get_scattered_cpuid_leaf);
diff --git a/arch/x86/kernel/cpuid.c b/arch/x86/kernel/cpuid.c
index 2836de390f95..9095c80723d6 100644
--- a/arch/x86/kernel/cpuid.c
+++ b/arch/x86/kernel/cpuid.c
@@ -46,10 +46,6 @@
46 46
47static struct class *cpuid_class; 47static struct class *cpuid_class;
48 48
49struct cpuid_regs {
50 u32 eax, ebx, ecx, edx;
51};
52
53static void cpuid_smp_cpuid(void *cmd_block) 49static void cpuid_smp_cpuid(void *cmd_block)
54{ 50{
55 struct cpuid_regs *cmd = (struct cpuid_regs *)cmd_block; 51 struct cpuid_regs *cmd = (struct cpuid_regs *)cmd_block;
diff --git a/arch/x86/kernel/dumpstack.c b/arch/x86/kernel/dumpstack.c
index 85f854b98a9d..0cfd01d2754c 100644
--- a/arch/x86/kernel/dumpstack.c
+++ b/arch/x86/kernel/dumpstack.c
@@ -22,7 +22,6 @@
22int panic_on_unrecovered_nmi; 22int panic_on_unrecovered_nmi;
23int panic_on_io_nmi; 23int panic_on_io_nmi;
24unsigned int code_bytes = 64; 24unsigned int code_bytes = 64;
25int kstack_depth_to_print = 3 * STACKSLOTS_PER_LINE;
26static int die_counter; 25static int die_counter;
27 26
28bool in_task_stack(unsigned long *stack, struct task_struct *task, 27bool in_task_stack(unsigned long *stack, struct task_struct *task,
@@ -46,14 +45,7 @@ static void printk_stack_address(unsigned long address, int reliable,
46 char *log_lvl) 45 char *log_lvl)
47{ 46{
48 touch_nmi_watchdog(); 47 touch_nmi_watchdog();
49 printk("%s [<%p>] %s%pB\n", 48 printk("%s %s%pB\n", log_lvl, reliable ? "" : "? ", (void *)address);
50 log_lvl, (void *)address, reliable ? "" : "? ",
51 (void *)address);
52}
53
54void printk_address(unsigned long address)
55{
56 pr_cont(" [<%p>] %pS\n", (void *)address, (void *)address);
57} 49}
58 50
59void show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs, 51void show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs,
@@ -67,6 +59,7 @@ void show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs,
67 printk("%sCall Trace:\n", log_lvl); 59 printk("%sCall Trace:\n", log_lvl);
68 60
69 unwind_start(&state, task, regs, stack); 61 unwind_start(&state, task, regs, stack);
62 stack = stack ? : get_stack_pointer(task, regs);
70 63
71 /* 64 /*
72 * Iterate through the stacks, starting with the current stack pointer. 65 * Iterate through the stacks, starting with the current stack pointer.
@@ -82,8 +75,8 @@ void show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs,
82 * - softirq stack 75 * - softirq stack
83 * - hardirq stack 76 * - hardirq stack
84 */ 77 */
85 for (; stack; stack = stack_info.next_sp) { 78 for (regs = NULL; stack; stack = stack_info.next_sp) {
86 const char *str_begin, *str_end; 79 const char *stack_name;
87 80
88 /* 81 /*
89 * If we overflowed the task stack into a guard page, jump back 82 * If we overflowed the task stack into a guard page, jump back
@@ -95,9 +88,9 @@ void show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs,
95 if (get_stack_info(stack, task, &stack_info, &visit_mask)) 88 if (get_stack_info(stack, task, &stack_info, &visit_mask))
96 break; 89 break;
97 90
98 stack_type_str(stack_info.type, &str_begin, &str_end); 91 stack_name = stack_type_name(stack_info.type);
99 if (str_begin) 92 if (stack_name)
100 printk("%s <%s> ", log_lvl, str_begin); 93 printk("%s <%s>\n", log_lvl, stack_name);
101 94
102 /* 95 /*
103 * Scan the stack, printing any text addresses we find. At the 96 * Scan the stack, printing any text addresses we find. At the
@@ -119,6 +112,15 @@ void show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs,
119 if (!__kernel_text_address(addr)) 112 if (!__kernel_text_address(addr))
120 continue; 113 continue;
121 114
115 /*
116 * Don't print regs->ip again if it was already printed
117 * by __show_regs() below.
118 */
119 if (regs && stack == &regs->ip) {
120 unwind_next_frame(&state);
121 continue;
122 }
123
122 if (stack == ret_addr_p) 124 if (stack == ret_addr_p)
123 reliable = 1; 125 reliable = 1;
124 126
@@ -146,10 +148,15 @@ void show_trace_log_lvl(struct task_struct *task, struct pt_regs *regs,
146 * of the addresses will just be printed as unreliable. 148 * of the addresses will just be printed as unreliable.
147 */ 149 */
148 unwind_next_frame(&state); 150 unwind_next_frame(&state);
151
152 /* if the frame has entry regs, print them */
153 regs = unwind_get_entry_regs(&state);
154 if (regs)
155 __show_regs(regs, 0);
149 } 156 }
150 157
151 if (str_end) 158 if (stack_name)
152 printk("%s <%s> ", log_lvl, str_end); 159 printk("%s </%s>\n", log_lvl, stack_name);
153 } 160 }
154} 161}
155 162
@@ -164,12 +171,12 @@ void show_stack(struct task_struct *task, unsigned long *sp)
164 if (!sp && task == current) 171 if (!sp && task == current)
165 sp = get_stack_pointer(current, NULL); 172 sp = get_stack_pointer(current, NULL);
166 173
167 show_stack_log_lvl(task, NULL, sp, ""); 174 show_trace_log_lvl(task, NULL, sp, KERN_DEFAULT);
168} 175}
169 176
170void show_stack_regs(struct pt_regs *regs) 177void show_stack_regs(struct pt_regs *regs)
171{ 178{
172 show_stack_log_lvl(current, regs, NULL, ""); 179 show_trace_log_lvl(current, regs, NULL, KERN_DEFAULT);
173} 180}
174 181
175static arch_spinlock_t die_lock = __ARCH_SPIN_LOCK_UNLOCKED; 182static arch_spinlock_t die_lock = __ARCH_SPIN_LOCK_UNLOCKED;
@@ -261,14 +268,11 @@ int __die(const char *str, struct pt_regs *regs, long err)
261 sp = kernel_stack_pointer(regs); 268 sp = kernel_stack_pointer(regs);
262 savesegment(ss, ss); 269 savesegment(ss, ss);
263 } 270 }
264 printk(KERN_EMERG "EIP: [<%08lx>] ", regs->ip); 271 printk(KERN_EMERG "EIP: %pS SS:ESP: %04x:%08lx\n",
265 print_symbol("%s", regs->ip); 272 (void *)regs->ip, ss, sp);
266 printk(" SS:ESP %04x:%08lx\n", ss, sp);
267#else 273#else
268 /* Executive summary in case the oops scrolled away */ 274 /* Executive summary in case the oops scrolled away */
269 printk(KERN_ALERT "RIP "); 275 printk(KERN_ALERT "RIP: %pS RSP: %016lx\n", (void *)regs->ip, regs->sp);
270 printk_address(regs->ip);
271 printk(" RSP <%016lx>\n", regs->sp);
272#endif 276#endif
273 return 0; 277 return 0;
274} 278}
@@ -291,22 +295,6 @@ void die(const char *str, struct pt_regs *regs, long err)
291 oops_end(flags, regs, sig); 295 oops_end(flags, regs, sig);
292} 296}
293 297
294static int __init kstack_setup(char *s)
295{
296 ssize_t ret;
297 unsigned long val;
298
299 if (!s)
300 return -EINVAL;
301
302 ret = kstrtoul(s, 0, &val);
303 if (ret)
304 return ret;
305 kstack_depth_to_print = val;
306 return 0;
307}
308early_param("kstack", kstack_setup);
309
310static int __init code_bytes_setup(char *s) 298static int __init code_bytes_setup(char *s)
311{ 299{
312 ssize_t ret; 300 ssize_t ret;
diff --git a/arch/x86/kernel/dumpstack_32.c b/arch/x86/kernel/dumpstack_32.c
index 06eb322b5f9f..bb3b5b9a6899 100644
--- a/arch/x86/kernel/dumpstack_32.c
+++ b/arch/x86/kernel/dumpstack_32.c
@@ -16,18 +16,15 @@
16 16
17#include <asm/stacktrace.h> 17#include <asm/stacktrace.h>
18 18
19void stack_type_str(enum stack_type type, const char **begin, const char **end) 19const char *stack_type_name(enum stack_type type)
20{ 20{
21 switch (type) { 21 if (type == STACK_TYPE_IRQ)
22 case STACK_TYPE_IRQ: 22 return "IRQ";
23 case STACK_TYPE_SOFTIRQ: 23
24 *begin = "IRQ"; 24 if (type == STACK_TYPE_SOFTIRQ)
25 *end = "EOI"; 25 return "SOFTIRQ";
26 break; 26
27 default: 27 return NULL;
28 *begin = NULL;
29 *end = NULL;
30 }
31} 28}
32 29
33static bool in_hardirq_stack(unsigned long *stack, struct stack_info *info) 30static bool in_hardirq_stack(unsigned long *stack, struct stack_info *info)
@@ -109,8 +106,10 @@ recursion_check:
109 * just break out and report an unknown stack type. 106 * just break out and report an unknown stack type.
110 */ 107 */
111 if (visit_mask) { 108 if (visit_mask) {
112 if (*visit_mask & (1UL << info->type)) 109 if (*visit_mask & (1UL << info->type)) {
110 printk_deferred_once(KERN_WARNING "WARNING: stack recursion on stack type %d\n", info->type);
113 goto unknown; 111 goto unknown;
112 }
114 *visit_mask |= 1UL << info->type; 113 *visit_mask |= 1UL << info->type;
115 } 114 }
116 115
@@ -121,36 +120,6 @@ unknown:
121 return -EINVAL; 120 return -EINVAL;
122} 121}
123 122
124void show_stack_log_lvl(struct task_struct *task, struct pt_regs *regs,
125 unsigned long *sp, char *log_lvl)
126{
127 unsigned long *stack;
128 int i;
129
130 if (!try_get_task_stack(task))
131 return;
132
133 sp = sp ? : get_stack_pointer(task, regs);
134
135 stack = sp;
136 for (i = 0; i < kstack_depth_to_print; i++) {
137 if (kstack_end(stack))
138 break;
139 if ((i % STACKSLOTS_PER_LINE) == 0) {
140 if (i != 0)
141 pr_cont("\n");
142 printk("%s %08lx", log_lvl, *stack++);
143 } else
144 pr_cont(" %08lx", *stack++);
145 touch_nmi_watchdog();
146 }
147 pr_cont("\n");
148 show_trace_log_lvl(task, regs, sp, log_lvl);
149
150 put_task_stack(task);
151}
152
153
154void show_regs(struct pt_regs *regs) 123void show_regs(struct pt_regs *regs)
155{ 124{
156 int i; 125 int i;
@@ -168,8 +137,7 @@ void show_regs(struct pt_regs *regs)
168 unsigned char c; 137 unsigned char c;
169 u8 *ip; 138 u8 *ip;
170 139
171 pr_emerg("Stack:\n"); 140 show_trace_log_lvl(current, regs, NULL, KERN_EMERG);
172 show_stack_log_lvl(current, regs, NULL, KERN_EMERG);
173 141
174 pr_emerg("Code:"); 142 pr_emerg("Code:");
175 143
diff --git a/arch/x86/kernel/dumpstack_64.c b/arch/x86/kernel/dumpstack_64.c
index 36cf1a498227..fac189efcc34 100644
--- a/arch/x86/kernel/dumpstack_64.c
+++ b/arch/x86/kernel/dumpstack_64.c
@@ -28,23 +28,17 @@ static unsigned long exception_stack_sizes[N_EXCEPTION_STACKS] = {
28 [DEBUG_STACK - 1] = DEBUG_STKSZ 28 [DEBUG_STACK - 1] = DEBUG_STKSZ
29}; 29};
30 30
31void stack_type_str(enum stack_type type, const char **begin, const char **end) 31const char *stack_type_name(enum stack_type type)
32{ 32{
33 BUILD_BUG_ON(N_EXCEPTION_STACKS != 4); 33 BUILD_BUG_ON(N_EXCEPTION_STACKS != 4);
34 34
35 switch (type) { 35 if (type == STACK_TYPE_IRQ)
36 case STACK_TYPE_IRQ: 36 return "IRQ";
37 *begin = "IRQ"; 37
38 *end = "EOI"; 38 if (type >= STACK_TYPE_EXCEPTION && type <= STACK_TYPE_EXCEPTION_LAST)
39 break; 39 return exception_stack_names[type - STACK_TYPE_EXCEPTION];
40 case STACK_TYPE_EXCEPTION ... STACK_TYPE_EXCEPTION_LAST: 40
41 *begin = exception_stack_names[type - STACK_TYPE_EXCEPTION]; 41 return NULL;
42 *end = "EOE";
43 break;
44 default:
45 *begin = NULL;
46 *end = NULL;
47 }
48} 42}
49 43
50static bool in_exception_stack(unsigned long *stack, struct stack_info *info) 44static bool in_exception_stack(unsigned long *stack, struct stack_info *info)
@@ -128,8 +122,10 @@ recursion_check:
128 * just break out and report an unknown stack type. 122 * just break out and report an unknown stack type.
129 */ 123 */
130 if (visit_mask) { 124 if (visit_mask) {
131 if (*visit_mask & (1UL << info->type)) 125 if (*visit_mask & (1UL << info->type)) {
126 printk_deferred_once(KERN_WARNING "WARNING: stack recursion on stack type %d\n", info->type);
132 goto unknown; 127 goto unknown;
128 }
133 *visit_mask |= 1UL << info->type; 129 *visit_mask |= 1UL << info->type;
134 } 130 }
135 131
@@ -140,56 +136,6 @@ unknown:
140 return -EINVAL; 136 return -EINVAL;
141} 137}
142 138
143void show_stack_log_lvl(struct task_struct *task, struct pt_regs *regs,
144 unsigned long *sp, char *log_lvl)
145{
146 unsigned long *irq_stack_end;
147 unsigned long *irq_stack;
148 unsigned long *stack;
149 int i;
150
151 if (!try_get_task_stack(task))
152 return;
153
154 irq_stack_end = (unsigned long *)this_cpu_read(irq_stack_ptr);
155 irq_stack = irq_stack_end - (IRQ_STACK_SIZE / sizeof(long));
156
157 sp = sp ? : get_stack_pointer(task, regs);
158
159 stack = sp;
160 for (i = 0; i < kstack_depth_to_print; i++) {
161 unsigned long word;
162
163 if (stack >= irq_stack && stack <= irq_stack_end) {
164 if (stack == irq_stack_end) {
165 stack = (unsigned long *) (irq_stack_end[-1]);
166 pr_cont(" <EOI> ");
167 }
168 } else {
169 if (kstack_end(stack))
170 break;
171 }
172
173 if (probe_kernel_address(stack, word))
174 break;
175
176 if ((i % STACKSLOTS_PER_LINE) == 0) {
177 if (i != 0)
178 pr_cont("\n");
179 printk("%s %016lx", log_lvl, word);
180 } else
181 pr_cont(" %016lx", word);
182
183 stack++;
184 touch_nmi_watchdog();
185 }
186
187 pr_cont("\n");
188 show_trace_log_lvl(task, regs, sp, log_lvl);
189
190 put_task_stack(task);
191}
192
193void show_regs(struct pt_regs *regs) 139void show_regs(struct pt_regs *regs)
194{ 140{
195 int i; 141 int i;
@@ -207,8 +153,7 @@ void show_regs(struct pt_regs *regs)
207 unsigned char c; 153 unsigned char c;
208 u8 *ip; 154 u8 *ip;
209 155
210 printk(KERN_DEFAULT "Stack:\n"); 156 show_trace_log_lvl(current, regs, NULL, KERN_DEFAULT);
211 show_stack_log_lvl(current, regs, NULL, KERN_DEFAULT);
212 157
213 printk(KERN_DEFAULT "Code: "); 158 printk(KERN_DEFAULT "Code: ");
214 159
diff --git a/arch/x86/kernel/fpu/xstate.c b/arch/x86/kernel/fpu/xstate.c
index 095ef7ddd6ae..ce47452879fd 100644
--- a/arch/x86/kernel/fpu/xstate.c
+++ b/arch/x86/kernel/fpu/xstate.c
@@ -65,6 +65,7 @@ void fpu__xstate_clear_all_cpu_caps(void)
65 setup_clear_cpu_cap(X86_FEATURE_AVX); 65 setup_clear_cpu_cap(X86_FEATURE_AVX);
66 setup_clear_cpu_cap(X86_FEATURE_AVX2); 66 setup_clear_cpu_cap(X86_FEATURE_AVX2);
67 setup_clear_cpu_cap(X86_FEATURE_AVX512F); 67 setup_clear_cpu_cap(X86_FEATURE_AVX512F);
68 setup_clear_cpu_cap(X86_FEATURE_AVX512IFMA);
68 setup_clear_cpu_cap(X86_FEATURE_AVX512PF); 69 setup_clear_cpu_cap(X86_FEATURE_AVX512PF);
69 setup_clear_cpu_cap(X86_FEATURE_AVX512ER); 70 setup_clear_cpu_cap(X86_FEATURE_AVX512ER);
70 setup_clear_cpu_cap(X86_FEATURE_AVX512CD); 71 setup_clear_cpu_cap(X86_FEATURE_AVX512CD);
@@ -73,6 +74,7 @@ void fpu__xstate_clear_all_cpu_caps(void)
73 setup_clear_cpu_cap(X86_FEATURE_AVX512VL); 74 setup_clear_cpu_cap(X86_FEATURE_AVX512VL);
74 setup_clear_cpu_cap(X86_FEATURE_MPX); 75 setup_clear_cpu_cap(X86_FEATURE_MPX);
75 setup_clear_cpu_cap(X86_FEATURE_XGETBV1); 76 setup_clear_cpu_cap(X86_FEATURE_XGETBV1);
77 setup_clear_cpu_cap(X86_FEATURE_AVX512VBMI);
76 setup_clear_cpu_cap(X86_FEATURE_PKU); 78 setup_clear_cpu_cap(X86_FEATURE_PKU);
77 setup_clear_cpu_cap(X86_FEATURE_AVX512_4VNNIW); 79 setup_clear_cpu_cap(X86_FEATURE_AVX512_4VNNIW);
78 setup_clear_cpu_cap(X86_FEATURE_AVX512_4FMAPS); 80 setup_clear_cpu_cap(X86_FEATURE_AVX512_4FMAPS);
diff --git a/arch/x86/kernel/head_32.S b/arch/x86/kernel/head_32.S
index 2dabea46f039..4e8577d03372 100644
--- a/arch/x86/kernel/head_32.S
+++ b/arch/x86/kernel/head_32.S
@@ -63,6 +63,8 @@
63#define PAGE_TABLE_SIZE(pages) ((pages) / PTRS_PER_PGD) 63#define PAGE_TABLE_SIZE(pages) ((pages) / PTRS_PER_PGD)
64#endif 64#endif
65 65
66#define SIZEOF_PTREGS 17*4
67
66/* 68/*
67 * Number of possible pages in the lowmem region. 69 * Number of possible pages in the lowmem region.
68 * 70 *
@@ -248,19 +250,19 @@ page_pde_offset = (__PAGE_OFFSET >> 20);
248#ifdef CONFIG_PARAVIRT 250#ifdef CONFIG_PARAVIRT
249 /* This is can only trip for a broken bootloader... */ 251 /* This is can only trip for a broken bootloader... */
250 cmpw $0x207, pa(boot_params + BP_version) 252 cmpw $0x207, pa(boot_params + BP_version)
251 jb default_entry 253 jb .Ldefault_entry
252 254
253 /* Paravirt-compatible boot parameters. Look to see what architecture 255 /* Paravirt-compatible boot parameters. Look to see what architecture
254 we're booting under. */ 256 we're booting under. */
255 movl pa(boot_params + BP_hardware_subarch), %eax 257 movl pa(boot_params + BP_hardware_subarch), %eax
256 cmpl $num_subarch_entries, %eax 258 cmpl $num_subarch_entries, %eax
257 jae bad_subarch 259 jae .Lbad_subarch
258 260
259 movl pa(subarch_entries)(,%eax,4), %eax 261 movl pa(subarch_entries)(,%eax,4), %eax
260 subl $__PAGE_OFFSET, %eax 262 subl $__PAGE_OFFSET, %eax
261 jmp *%eax 263 jmp *%eax
262 264
263bad_subarch: 265.Lbad_subarch:
264WEAK(lguest_entry) 266WEAK(lguest_entry)
265WEAK(xen_entry) 267WEAK(xen_entry)
266 /* Unknown implementation; there's really 268 /* Unknown implementation; there's really
@@ -270,14 +272,14 @@ WEAK(xen_entry)
270 __INITDATA 272 __INITDATA
271 273
272subarch_entries: 274subarch_entries:
273 .long default_entry /* normal x86/PC */ 275 .long .Ldefault_entry /* normal x86/PC */
274 .long lguest_entry /* lguest hypervisor */ 276 .long lguest_entry /* lguest hypervisor */
275 .long xen_entry /* Xen hypervisor */ 277 .long xen_entry /* Xen hypervisor */
276 .long default_entry /* Moorestown MID */ 278 .long .Ldefault_entry /* Moorestown MID */
277num_subarch_entries = (. - subarch_entries) / 4 279num_subarch_entries = (. - subarch_entries) / 4
278.previous 280.previous
279#else 281#else
280 jmp default_entry 282 jmp .Ldefault_entry
281#endif /* CONFIG_PARAVIRT */ 283#endif /* CONFIG_PARAVIRT */
282 284
283#ifdef CONFIG_HOTPLUG_CPU 285#ifdef CONFIG_HOTPLUG_CPU
@@ -289,7 +291,8 @@ num_subarch_entries = (. - subarch_entries) / 4
289ENTRY(start_cpu0) 291ENTRY(start_cpu0)
290 movl initial_stack, %ecx 292 movl initial_stack, %ecx
291 movl %ecx, %esp 293 movl %ecx, %esp
292 jmp *(initial_code) 294 call *(initial_code)
2951: jmp 1b
293ENDPROC(start_cpu0) 296ENDPROC(start_cpu0)
294#endif 297#endif
295 298
@@ -317,7 +320,7 @@ ENTRY(startup_32_smp)
317 call load_ucode_ap 320 call load_ucode_ap
318#endif 321#endif
319 322
320default_entry: 323.Ldefault_entry:
321#define CR0_STATE (X86_CR0_PE | X86_CR0_MP | X86_CR0_ET | \ 324#define CR0_STATE (X86_CR0_PE | X86_CR0_MP | X86_CR0_ET | \
322 X86_CR0_NE | X86_CR0_WP | X86_CR0_AM | \ 325 X86_CR0_NE | X86_CR0_WP | X86_CR0_AM | \
323 X86_CR0_PG) 326 X86_CR0_PG)
@@ -347,7 +350,7 @@ default_entry:
347 pushfl 350 pushfl
348 popl %eax # get EFLAGS 351 popl %eax # get EFLAGS
349 testl $X86_EFLAGS_ID,%eax # did EFLAGS.ID remained set? 352 testl $X86_EFLAGS_ID,%eax # did EFLAGS.ID remained set?
350 jz enable_paging # hw disallowed setting of ID bit 353 jz .Lenable_paging # hw disallowed setting of ID bit
351 # which means no CPUID and no CR4 354 # which means no CPUID and no CR4
352 355
353 xorl %eax,%eax 356 xorl %eax,%eax
@@ -357,13 +360,13 @@ default_entry:
357 movl $1,%eax 360 movl $1,%eax
358 cpuid 361 cpuid
359 andl $~1,%edx # Ignore CPUID.FPU 362 andl $~1,%edx # Ignore CPUID.FPU
360 jz enable_paging # No flags or only CPUID.FPU = no CR4 363 jz .Lenable_paging # No flags or only CPUID.FPU = no CR4
361 364
362 movl pa(mmu_cr4_features),%eax 365 movl pa(mmu_cr4_features),%eax
363 movl %eax,%cr4 366 movl %eax,%cr4
364 367
365 testb $X86_CR4_PAE, %al # check if PAE is enabled 368 testb $X86_CR4_PAE, %al # check if PAE is enabled
366 jz enable_paging 369 jz .Lenable_paging
367 370
368 /* Check if extended functions are implemented */ 371 /* Check if extended functions are implemented */
369 movl $0x80000000, %eax 372 movl $0x80000000, %eax
@@ -371,7 +374,7 @@ default_entry:
371 /* Value must be in the range 0x80000001 to 0x8000ffff */ 374 /* Value must be in the range 0x80000001 to 0x8000ffff */
372 subl $0x80000001, %eax 375 subl $0x80000001, %eax
373 cmpl $(0x8000ffff-0x80000001), %eax 376 cmpl $(0x8000ffff-0x80000001), %eax
374 ja enable_paging 377 ja .Lenable_paging
375 378
376 /* Clear bogus XD_DISABLE bits */ 379 /* Clear bogus XD_DISABLE bits */
377 call verify_cpu 380 call verify_cpu
@@ -380,7 +383,7 @@ default_entry:
380 cpuid 383 cpuid
381 /* Execute Disable bit supported? */ 384 /* Execute Disable bit supported? */
382 btl $(X86_FEATURE_NX & 31), %edx 385 btl $(X86_FEATURE_NX & 31), %edx
383 jnc enable_paging 386 jnc .Lenable_paging
384 387
385 /* Setup EFER (Extended Feature Enable Register) */ 388 /* Setup EFER (Extended Feature Enable Register) */
386 movl $MSR_EFER, %ecx 389 movl $MSR_EFER, %ecx
@@ -390,7 +393,7 @@ default_entry:
390 /* Make changes effective */ 393 /* Make changes effective */
391 wrmsr 394 wrmsr
392 395
393enable_paging: 396.Lenable_paging:
394 397
395/* 398/*
396 * Enable paging 399 * Enable paging
@@ -419,7 +422,7 @@ enable_paging:
419 */ 422 */
420 movb $4,X86 # at least 486 423 movb $4,X86 # at least 486
421 cmpl $-1,X86_CPUID 424 cmpl $-1,X86_CPUID
422 je is486 425 je .Lis486
423 426
424 /* get vendor info */ 427 /* get vendor info */
425 xorl %eax,%eax # call CPUID with 0 -> return vendor ID 428 xorl %eax,%eax # call CPUID with 0 -> return vendor ID
@@ -430,7 +433,7 @@ enable_paging:
430 movl %ecx,X86_VENDOR_ID+8 # last 4 chars 433 movl %ecx,X86_VENDOR_ID+8 # last 4 chars
431 434
432 orl %eax,%eax # do we have processor info as well? 435 orl %eax,%eax # do we have processor info as well?
433 je is486 436 je .Lis486
434 437
435 movl $1,%eax # Use the CPUID instruction to get CPU type 438 movl $1,%eax # Use the CPUID instruction to get CPU type
436 cpuid 439 cpuid
@@ -444,7 +447,7 @@ enable_paging:
444 movb %cl,X86_MASK 447 movb %cl,X86_MASK
445 movl %edx,X86_CAPABILITY 448 movl %edx,X86_CAPABILITY
446 449
447is486: 450.Lis486:
448 movl $0x50022,%ecx # set AM, WP, NE and MP 451 movl $0x50022,%ecx # set AM, WP, NE and MP
449 movl %cr0,%eax 452 movl %cr0,%eax
450 andl $0x80000011,%eax # Save PG,PE,ET 453 andl $0x80000011,%eax # Save PG,PE,ET
@@ -470,8 +473,9 @@ is486:
470 xorl %eax,%eax # Clear LDT 473 xorl %eax,%eax # Clear LDT
471 lldt %ax 474 lldt %ax
472 475
473 pushl $0 # fake return address for unwinder 476 call *(initial_code)
474 jmp *(initial_code) 4771: jmp 1b
478ENDPROC(startup_32_smp)
475 479
476#include "verify_cpu.S" 480#include "verify_cpu.S"
477 481
@@ -709,7 +713,12 @@ ENTRY(initial_page_table)
709.data 713.data
710.balign 4 714.balign 4
711ENTRY(initial_stack) 715ENTRY(initial_stack)
712 .long init_thread_union+THREAD_SIZE 716 /*
717 * The SIZEOF_PTREGS gap is a convention which helps the in-kernel
718 * unwinder reliably detect the end of the stack.
719 */
720 .long init_thread_union + THREAD_SIZE - SIZEOF_PTREGS - \
721 TOP_OF_KERNEL_STACK_PADDING;
713 722
714__INITRODATA 723__INITRODATA
715int_msg: 724int_msg:
diff --git a/arch/x86/kernel/head_64.S b/arch/x86/kernel/head_64.S
index b4421cc191b0..a15d381e6020 100644
--- a/arch/x86/kernel/head_64.S
+++ b/arch/x86/kernel/head_64.S
@@ -66,13 +66,8 @@ startup_64:
66 * tables and then reload them. 66 * tables and then reload them.
67 */ 67 */
68 68
69 /* 69 /* Set up the stack for verify_cpu(), similar to initial_stack below */
70 * Setup stack for verify_cpu(). "-8" because initial_stack is defined 70 leaq (__end_init_task - SIZEOF_PTREGS)(%rip), %rsp
71 * this way, see below. Our best guess is a NULL ptr for stack
72 * termination heuristics and we don't want to break anything which
73 * might depend on it (kgdb, ...).
74 */
75 leaq (__end_init_task - 8)(%rip), %rsp
76 71
77 /* Sanitize CPU configuration */ 72 /* Sanitize CPU configuration */
78 call verify_cpu 73 call verify_cpu
@@ -117,20 +112,20 @@ startup_64:
117 movq %rdi, %rax 112 movq %rdi, %rax
118 shrq $PGDIR_SHIFT, %rax 113 shrq $PGDIR_SHIFT, %rax
119 114
120 leaq (4096 + _KERNPG_TABLE)(%rbx), %rdx 115 leaq (PAGE_SIZE + _KERNPG_TABLE)(%rbx), %rdx
121 movq %rdx, 0(%rbx,%rax,8) 116 movq %rdx, 0(%rbx,%rax,8)
122 movq %rdx, 8(%rbx,%rax,8) 117 movq %rdx, 8(%rbx,%rax,8)
123 118
124 addq $4096, %rdx 119 addq $PAGE_SIZE, %rdx
125 movq %rdi, %rax 120 movq %rdi, %rax
126 shrq $PUD_SHIFT, %rax 121 shrq $PUD_SHIFT, %rax
127 andl $(PTRS_PER_PUD-1), %eax 122 andl $(PTRS_PER_PUD-1), %eax
128 movq %rdx, 4096(%rbx,%rax,8) 123 movq %rdx, PAGE_SIZE(%rbx,%rax,8)
129 incl %eax 124 incl %eax
130 andl $(PTRS_PER_PUD-1), %eax 125 andl $(PTRS_PER_PUD-1), %eax
131 movq %rdx, 4096(%rbx,%rax,8) 126 movq %rdx, PAGE_SIZE(%rbx,%rax,8)
132 127
133 addq $8192, %rbx 128 addq $PAGE_SIZE * 2, %rbx
134 movq %rdi, %rax 129 movq %rdi, %rax
135 shrq $PMD_SHIFT, %rdi 130 shrq $PMD_SHIFT, %rdi
136 addq $(__PAGE_KERNEL_LARGE_EXEC & ~_PAGE_GLOBAL), %rax 131 addq $(__PAGE_KERNEL_LARGE_EXEC & ~_PAGE_GLOBAL), %rax
@@ -265,13 +260,17 @@ ENTRY(secondary_startup_64)
265 movl $MSR_GS_BASE,%ecx 260 movl $MSR_GS_BASE,%ecx
266 movl initial_gs(%rip),%eax 261 movl initial_gs(%rip),%eax
267 movl initial_gs+4(%rip),%edx 262 movl initial_gs+4(%rip),%edx
268 wrmsr 263 wrmsr
269 264
270 /* rsi is pointer to real mode structure with interesting info. 265 /* rsi is pointer to real mode structure with interesting info.
271 pass it to C */ 266 pass it to C */
272 movq %rsi, %rdi 267 movq %rsi, %rdi
273 268 jmp start_cpu
274 /* Finally jump to run C code and to be on real kernel address 269ENDPROC(secondary_startup_64)
270
271ENTRY(start_cpu)
272 /*
273 * Jump to run C code and to be on a real kernel address.
275 * Since we are running on identity-mapped space we have to jump 274 * Since we are running on identity-mapped space we have to jump
276 * to the full 64bit address, this is only possible as indirect 275 * to the full 64bit address, this is only possible as indirect
277 * jump. In addition we need to ensure %cs is set so we make this 276 * jump. In addition we need to ensure %cs is set so we make this
@@ -295,12 +294,13 @@ ENTRY(secondary_startup_64)
295 * REX.W + FF /5 JMP m16:64 Jump far, absolute indirect, 294 * REX.W + FF /5 JMP m16:64 Jump far, absolute indirect,
296 * address given in m16:64. 295 * address given in m16:64.
297 */ 296 */
298 movq initial_code(%rip),%rax 297 call 1f # put return address on stack for unwinder
299 pushq $0 # fake return address to stop unwinder 2981: xorq %rbp, %rbp # clear frame pointer
299 movq initial_code(%rip), %rax
300 pushq $__KERNEL_CS # set correct cs 300 pushq $__KERNEL_CS # set correct cs
301 pushq %rax # target address in negative space 301 pushq %rax # target address in negative space
302 lretq 302 lretq
303ENDPROC(secondary_startup_64) 303ENDPROC(start_cpu)
304 304
305#include "verify_cpu.S" 305#include "verify_cpu.S"
306 306
@@ -308,15 +308,11 @@ ENDPROC(secondary_startup_64)
308/* 308/*
309 * Boot CPU0 entry point. It's called from play_dead(). Everything has been set 309 * Boot CPU0 entry point. It's called from play_dead(). Everything has been set
310 * up already except stack. We just set up stack here. Then call 310 * up already except stack. We just set up stack here. Then call
311 * start_secondary(). 311 * start_secondary() via start_cpu().
312 */ 312 */
313ENTRY(start_cpu0) 313ENTRY(start_cpu0)
314 movq initial_stack(%rip),%rsp 314 movq initial_stack(%rip), %rsp
315 movq initial_code(%rip),%rax 315 jmp start_cpu
316 pushq $0 # fake return address to stop unwinder
317 pushq $__KERNEL_CS # set correct cs
318 pushq %rax # target address in negative space
319 lretq
320ENDPROC(start_cpu0) 316ENDPROC(start_cpu0)
321#endif 317#endif
322 318
@@ -328,7 +324,11 @@ ENDPROC(start_cpu0)
328 GLOBAL(initial_gs) 324 GLOBAL(initial_gs)
329 .quad INIT_PER_CPU_VAR(irq_stack_union) 325 .quad INIT_PER_CPU_VAR(irq_stack_union)
330 GLOBAL(initial_stack) 326 GLOBAL(initial_stack)
331 .quad init_thread_union+THREAD_SIZE-8 327 /*
328 * The SIZEOF_PTREGS gap is a convention which helps the in-kernel
329 * unwinder reliably detect the end of the stack.
330 */
331 .quad init_thread_union + THREAD_SIZE - SIZEOF_PTREGS
332 __FINITDATA 332 __FINITDATA
333 333
334bad_address: 334bad_address:
diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c
index bd7be8efdc4c..e3223bc78cb6 100644
--- a/arch/x86/kernel/process_32.c
+++ b/arch/x86/kernel/process_32.c
@@ -72,10 +72,9 @@ void __show_regs(struct pt_regs *regs, int all)
72 savesegment(gs, gs); 72 savesegment(gs, gs);
73 } 73 }
74 74
75 printk(KERN_DEFAULT "EIP: %04x:[<%08lx>] EFLAGS: %08lx CPU: %d\n", 75 printk(KERN_DEFAULT "EIP: %pS\n", (void *)regs->ip);
76 (u16)regs->cs, regs->ip, regs->flags, 76 printk(KERN_DEFAULT "EFLAGS: %08lx CPU: %d\n", regs->flags,
77 smp_processor_id()); 77 smp_processor_id());
78 print_symbol("EIP is at %s\n", regs->ip);
79 78
80 printk(KERN_DEFAULT "EAX: %08lx EBX: %08lx ECX: %08lx EDX: %08lx\n", 79 printk(KERN_DEFAULT "EAX: %08lx EBX: %08lx ECX: %08lx EDX: %08lx\n",
81 regs->ax, regs->bx, regs->cx, regs->dx); 80 regs->ax, regs->bx, regs->cx, regs->dx);
diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c
index b3760b3c1ca0..c99f1ca35eb5 100644
--- a/arch/x86/kernel/process_64.c
+++ b/arch/x86/kernel/process_64.c
@@ -61,10 +61,15 @@ void __show_regs(struct pt_regs *regs, int all)
61 unsigned int fsindex, gsindex; 61 unsigned int fsindex, gsindex;
62 unsigned int ds, cs, es; 62 unsigned int ds, cs, es;
63 63
64 printk(KERN_DEFAULT "RIP: %04lx:[<%016lx>] ", regs->cs & 0xffff, regs->ip); 64 printk(KERN_DEFAULT "RIP: %04lx:%pS\n", regs->cs & 0xffff,
65 printk_address(regs->ip); 65 (void *)regs->ip);
66 printk(KERN_DEFAULT "RSP: %04lx:%016lx EFLAGS: %08lx\n", regs->ss, 66 printk(KERN_DEFAULT "RSP: %04lx:%016lx EFLAGS: %08lx", regs->ss,
67 regs->sp, regs->flags); 67 regs->sp, regs->flags);
68 if (regs->orig_ax != -1)
69 pr_cont(" ORIG_RAX: %016lx\n", regs->orig_ax);
70 else
71 pr_cont("\n");
72
68 printk(KERN_DEFAULT "RAX: %016lx RBX: %016lx RCX: %016lx\n", 73 printk(KERN_DEFAULT "RAX: %016lx RBX: %016lx RCX: %016lx\n",
69 regs->ax, regs->bx, regs->cx); 74 regs->ax, regs->bx, regs->cx);
70 printk(KERN_DEFAULT "RDX: %016lx RSI: %016lx RDI: %016lx\n", 75 printk(KERN_DEFAULT "RDX: %016lx RSI: %016lx RDI: %016lx\n",
diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c
index 118e792a7be6..8add3a5b8635 100644
--- a/arch/x86/kernel/smpboot.c
+++ b/arch/x86/kernel/smpboot.c
@@ -987,9 +987,7 @@ static int do_boot_cpu(int apicid, int cpu, struct task_struct *idle)
987 int cpu0_nmi_registered = 0; 987 int cpu0_nmi_registered = 0;
988 unsigned long timeout; 988 unsigned long timeout;
989 989
990 idle->thread.sp = (unsigned long) (((struct pt_regs *) 990 idle->thread.sp = (unsigned long)task_pt_regs(idle);
991 (THREAD_SIZE + task_stack_page(idle))) - 1);
992
993 early_gdt_descr.address = (unsigned long)get_cpu_gdt_table(cpu); 991 early_gdt_descr.address = (unsigned long)get_cpu_gdt_table(cpu);
994 initial_code = (unsigned long)start_secondary; 992 initial_code = (unsigned long)start_secondary;
995 initial_stack = idle->thread.sp; 993 initial_stack = idle->thread.sp;
diff --git a/arch/x86/kernel/unwind_frame.c b/arch/x86/kernel/unwind_frame.c
index a2456d4d286a..ea7b7f9a3b9e 100644
--- a/arch/x86/kernel/unwind_frame.c
+++ b/arch/x86/kernel/unwind_frame.c
@@ -14,13 +14,55 @@ unsigned long unwind_get_return_address(struct unwind_state *state)
14 if (unwind_done(state)) 14 if (unwind_done(state))
15 return 0; 15 return 0;
16 16
17 if (state->regs && user_mode(state->regs))
18 return 0;
19
17 addr = ftrace_graph_ret_addr(state->task, &state->graph_idx, *addr_p, 20 addr = ftrace_graph_ret_addr(state->task, &state->graph_idx, *addr_p,
18 addr_p); 21 addr_p);
19 22
20 return __kernel_text_address(addr) ? addr : 0; 23 if (!__kernel_text_address(addr)) {
24 printk_deferred_once(KERN_WARNING
25 "WARNING: unrecognized kernel stack return address %p at %p in %s:%d\n",
26 (void *)addr, addr_p, state->task->comm,
27 state->task->pid);
28 return 0;
29 }
30
31 return addr;
21} 32}
22EXPORT_SYMBOL_GPL(unwind_get_return_address); 33EXPORT_SYMBOL_GPL(unwind_get_return_address);
23 34
35static size_t regs_size(struct pt_regs *regs)
36{
37 /* x86_32 regs from kernel mode are two words shorter: */
38 if (IS_ENABLED(CONFIG_X86_32) && !user_mode(regs))
39 return sizeof(*regs) - 2*sizeof(long);
40
41 return sizeof(*regs);
42}
43
44static bool is_last_task_frame(struct unwind_state *state)
45{
46 unsigned long bp = (unsigned long)state->bp;
47 unsigned long regs = (unsigned long)task_pt_regs(state->task);
48
49 return bp == regs - FRAME_HEADER_SIZE;
50}
51
52/*
53 * This determines if the frame pointer actually contains an encoded pointer to
54 * pt_regs on the stack. See ENCODE_FRAME_POINTER.
55 */
56static struct pt_regs *decode_frame_pointer(unsigned long *bp)
57{
58 unsigned long regs = (unsigned long)bp;
59
60 if (!(regs & 0x1))
61 return NULL;
62
63 return (struct pt_regs *)(regs & ~0x1);
64}
65
24static bool update_stack_state(struct unwind_state *state, void *addr, 66static bool update_stack_state(struct unwind_state *state, void *addr,
25 size_t len) 67 size_t len)
26{ 68{
@@ -43,26 +85,117 @@ static bool update_stack_state(struct unwind_state *state, void *addr,
43 85
44bool unwind_next_frame(struct unwind_state *state) 86bool unwind_next_frame(struct unwind_state *state)
45{ 87{
46 unsigned long *next_bp; 88 struct pt_regs *regs;
89 unsigned long *next_bp, *next_frame;
90 size_t next_len;
91 enum stack_type prev_type = state->stack_info.type;
47 92
48 if (unwind_done(state)) 93 if (unwind_done(state))
49 return false; 94 return false;
50 95
51 next_bp = (unsigned long *)*state->bp; 96 /* have we reached the end? */
97 if (state->regs && user_mode(state->regs))
98 goto the_end;
99
100 if (is_last_task_frame(state)) {
101 regs = task_pt_regs(state->task);
102
103 /*
104 * kthreads (other than the boot CPU's idle thread) have some
105 * partial regs at the end of their stack which were placed
106 * there by copy_thread_tls(). But the regs don't have any
107 * useful information, so we can skip them.
108 *
109 * This user_mode() check is slightly broader than a PF_KTHREAD
110 * check because it also catches the awkward situation where a
111 * newly forked kthread transitions into a user task by calling
112 * do_execve(), which eventually clears PF_KTHREAD.
113 */
114 if (!user_mode(regs))
115 goto the_end;
116
117 /*
118 * We're almost at the end, but not quite: there's still the
119 * syscall regs frame. Entry code doesn't encode the regs
120 * pointer for syscalls, so we have to set it manually.
121 */
122 state->regs = regs;
123 state->bp = NULL;
124 return true;
125 }
126
127 /* get the next frame pointer */
128 if (state->regs)
129 next_bp = (unsigned long *)state->regs->bp;
130 else
131 next_bp = (unsigned long *)*state->bp;
132
133 /* is the next frame pointer an encoded pointer to pt_regs? */
134 regs = decode_frame_pointer(next_bp);
135 if (regs) {
136 next_frame = (unsigned long *)regs;
137 next_len = sizeof(*regs);
138 } else {
139 next_frame = next_bp;
140 next_len = FRAME_HEADER_SIZE;
141 }
52 142
53 /* make sure the next frame's data is accessible */ 143 /* make sure the next frame's data is accessible */
54 if (!update_stack_state(state, next_bp, FRAME_HEADER_SIZE)) 144 if (!update_stack_state(state, next_frame, next_len)) {
55 return false; 145 /*
146 * Don't warn on bad regs->bp. An interrupt in entry code
147 * might cause a false positive warning.
148 */
149 if (state->regs)
150 goto the_end;
151
152 goto bad_address;
153 }
154
155 /* Make sure it only unwinds up and doesn't overlap the last frame: */
156 if (state->stack_info.type == prev_type) {
157 if (state->regs && (void *)next_frame < (void *)state->regs + regs_size(state->regs))
158 goto bad_address;
159
160 if (state->bp && (void *)next_frame < (void *)state->bp + FRAME_HEADER_SIZE)
161 goto bad_address;
162 }
56 163
57 /* move to the next frame */ 164 /* move to the next frame */
58 state->bp = next_bp; 165 if (regs) {
166 state->regs = regs;
167 state->bp = NULL;
168 } else {
169 state->bp = next_bp;
170 state->regs = NULL;
171 }
172
59 return true; 173 return true;
174
175bad_address:
176 if (state->regs) {
177 printk_deferred_once(KERN_WARNING
178 "WARNING: kernel stack regs at %p in %s:%d has bad 'bp' value %p\n",
179 state->regs, state->task->comm,
180 state->task->pid, next_frame);
181 } else {
182 printk_deferred_once(KERN_WARNING
183 "WARNING: kernel stack frame pointer at %p in %s:%d has bad value %p\n",
184 state->bp, state->task->comm,
185 state->task->pid, next_frame);
186 }
187the_end:
188 state->stack_info.type = STACK_TYPE_UNKNOWN;
189 return false;
60} 190}
61EXPORT_SYMBOL_GPL(unwind_next_frame); 191EXPORT_SYMBOL_GPL(unwind_next_frame);
62 192
63void __unwind_start(struct unwind_state *state, struct task_struct *task, 193void __unwind_start(struct unwind_state *state, struct task_struct *task,
64 struct pt_regs *regs, unsigned long *first_frame) 194 struct pt_regs *regs, unsigned long *first_frame)
65{ 195{
196 unsigned long *bp, *frame;
197 size_t len;
198
66 memset(state, 0, sizeof(*state)); 199 memset(state, 0, sizeof(*state));
67 state->task = task; 200 state->task = task;
68 201
@@ -73,12 +206,22 @@ void __unwind_start(struct unwind_state *state, struct task_struct *task,
73 } 206 }
74 207
75 /* set up the starting stack frame */ 208 /* set up the starting stack frame */
76 state->bp = get_frame_pointer(task, regs); 209 bp = get_frame_pointer(task, regs);
210 regs = decode_frame_pointer(bp);
211 if (regs) {
212 state->regs = regs;
213 frame = (unsigned long *)regs;
214 len = sizeof(*regs);
215 } else {
216 state->bp = bp;
217 frame = bp;
218 len = FRAME_HEADER_SIZE;
219 }
77 220
78 /* initialize stack info and make sure the frame data is accessible */ 221 /* initialize stack info and make sure the frame data is accessible */
79 get_stack_info(state->bp, state->task, &state->stack_info, 222 get_stack_info(frame, state->task, &state->stack_info,
80 &state->stack_mask); 223 &state->stack_mask);
81 update_stack_state(state, state->bp, FRAME_HEADER_SIZE); 224 update_stack_state(state, frame, len);
82 225
83 /* 226 /*
84 * The caller can provide the address of the first frame directly 227 * The caller can provide the address of the first frame directly
diff --git a/arch/x86/kernel/vmlinux.lds.S b/arch/x86/kernel/vmlinux.lds.S
index dbf67f64d5ec..e79f15f108a8 100644
--- a/arch/x86/kernel/vmlinux.lds.S
+++ b/arch/x86/kernel/vmlinux.lds.S
@@ -91,10 +91,10 @@ SECTIONS
91 /* Text and read-only data */ 91 /* Text and read-only data */
92 .text : AT(ADDR(.text) - LOAD_OFFSET) { 92 .text : AT(ADDR(.text) - LOAD_OFFSET) {
93 _text = .; 93 _text = .;
94 _stext = .;
94 /* bootstrapping code */ 95 /* bootstrapping code */
95 HEAD_TEXT 96 HEAD_TEXT
96 . = ALIGN(8); 97 . = ALIGN(8);
97 _stext = .;
98 TEXT_TEXT 98 TEXT_TEXT
99 SCHED_TEXT 99 SCHED_TEXT
100 CPUIDLE_TEXT 100 CPUIDLE_TEXT
diff --git a/arch/x86/lib/copy_user_64.S b/arch/x86/lib/copy_user_64.S
index d376e4b48f88..c5959576c315 100644
--- a/arch/x86/lib/copy_user_64.S
+++ b/arch/x86/lib/copy_user_64.S
@@ -16,53 +16,6 @@
16#include <asm/smap.h> 16#include <asm/smap.h>
17#include <asm/export.h> 17#include <asm/export.h>
18 18
19/* Standard copy_to_user with segment limit checking */
20ENTRY(_copy_to_user)
21 mov PER_CPU_VAR(current_task), %rax
22 movq %rdi,%rcx
23 addq %rdx,%rcx
24 jc bad_to_user
25 cmpq TASK_addr_limit(%rax),%rcx
26 ja bad_to_user
27 ALTERNATIVE_2 "jmp copy_user_generic_unrolled", \
28 "jmp copy_user_generic_string", \
29 X86_FEATURE_REP_GOOD, \
30 "jmp copy_user_enhanced_fast_string", \
31 X86_FEATURE_ERMS
32ENDPROC(_copy_to_user)
33EXPORT_SYMBOL(_copy_to_user)
34
35/* Standard copy_from_user with segment limit checking */
36ENTRY(_copy_from_user)
37 mov PER_CPU_VAR(current_task), %rax
38 movq %rsi,%rcx
39 addq %rdx,%rcx
40 jc bad_from_user
41 cmpq TASK_addr_limit(%rax),%rcx
42 ja bad_from_user
43 ALTERNATIVE_2 "jmp copy_user_generic_unrolled", \
44 "jmp copy_user_generic_string", \
45 X86_FEATURE_REP_GOOD, \
46 "jmp copy_user_enhanced_fast_string", \
47 X86_FEATURE_ERMS
48ENDPROC(_copy_from_user)
49EXPORT_SYMBOL(_copy_from_user)
50
51
52 .section .fixup,"ax"
53 /* must zero dest */
54ENTRY(bad_from_user)
55bad_from_user:
56 movl %edx,%ecx
57 xorl %eax,%eax
58 rep
59 stosb
60bad_to_user:
61 movl %edx,%eax
62 ret
63ENDPROC(bad_from_user)
64 .previous
65
66/* 19/*
67 * copy_user_generic_unrolled - memory copy with exception handling. 20 * copy_user_generic_unrolled - memory copy with exception handling.
68 * This version is for CPUs like P4 that don't have efficient micro 21 * This version is for CPUs like P4 that don't have efficient micro
diff --git a/arch/x86/lib/usercopy.c b/arch/x86/lib/usercopy.c
index b4908789484e..c074799bddae 100644
--- a/arch/x86/lib/usercopy.c
+++ b/arch/x86/lib/usercopy.c
@@ -34,3 +34,52 @@ copy_from_user_nmi(void *to, const void __user *from, unsigned long n)
34 return ret; 34 return ret;
35} 35}
36EXPORT_SYMBOL_GPL(copy_from_user_nmi); 36EXPORT_SYMBOL_GPL(copy_from_user_nmi);
37
38/**
39 * copy_to_user: - Copy a block of data into user space.
40 * @to: Destination address, in user space.
41 * @from: Source address, in kernel space.
42 * @n: Number of bytes to copy.
43 *
44 * Context: User context only. This function may sleep if pagefaults are
45 * enabled.
46 *
47 * Copy data from kernel space to user space.
48 *
49 * Returns number of bytes that could not be copied.
50 * On success, this will be zero.
51 */
52unsigned long _copy_to_user(void __user *to, const void *from, unsigned n)
53{
54 if (access_ok(VERIFY_WRITE, to, n))
55 n = __copy_to_user(to, from, n);
56 return n;
57}
58EXPORT_SYMBOL(_copy_to_user);
59
60/**
61 * copy_from_user: - Copy a block of data from user space.
62 * @to: Destination address, in kernel space.
63 * @from: Source address, in user space.
64 * @n: Number of bytes to copy.
65 *
66 * Context: User context only. This function may sleep if pagefaults are
67 * enabled.
68 *
69 * Copy data from user space to kernel space.
70 *
71 * Returns number of bytes that could not be copied.
72 * On success, this will be zero.
73 *
74 * If some data could not be copied, this function will pad the copied
75 * data to the requested size using zero bytes.
76 */
77unsigned long _copy_from_user(void *to, const void __user *from, unsigned n)
78{
79 if (access_ok(VERIFY_READ, from, n))
80 n = __copy_from_user(to, from, n);
81 else
82 memset(to, 0, n);
83 return n;
84}
85EXPORT_SYMBOL(_copy_from_user);
diff --git a/arch/x86/lib/usercopy_32.c b/arch/x86/lib/usercopy_32.c
index 3bc7baf2a711..0b281217c890 100644
--- a/arch/x86/lib/usercopy_32.c
+++ b/arch/x86/lib/usercopy_32.c
@@ -640,52 +640,3 @@ unsigned long __copy_from_user_ll_nocache_nozero(void *to, const void __user *fr
640 return n; 640 return n;
641} 641}
642EXPORT_SYMBOL(__copy_from_user_ll_nocache_nozero); 642EXPORT_SYMBOL(__copy_from_user_ll_nocache_nozero);
643
644/**
645 * copy_to_user: - Copy a block of data into user space.
646 * @to: Destination address, in user space.
647 * @from: Source address, in kernel space.
648 * @n: Number of bytes to copy.
649 *
650 * Context: User context only. This function may sleep if pagefaults are
651 * enabled.
652 *
653 * Copy data from kernel space to user space.
654 *
655 * Returns number of bytes that could not be copied.
656 * On success, this will be zero.
657 */
658unsigned long _copy_to_user(void __user *to, const void *from, unsigned n)
659{
660 if (access_ok(VERIFY_WRITE, to, n))
661 n = __copy_to_user(to, from, n);
662 return n;
663}
664EXPORT_SYMBOL(_copy_to_user);
665
666/**
667 * copy_from_user: - Copy a block of data from user space.
668 * @to: Destination address, in kernel space.
669 * @from: Source address, in user space.
670 * @n: Number of bytes to copy.
671 *
672 * Context: User context only. This function may sleep if pagefaults are
673 * enabled.
674 *
675 * Copy data from user space to kernel space.
676 *
677 * Returns number of bytes that could not be copied.
678 * On success, this will be zero.
679 *
680 * If some data could not be copied, this function will pad the copied
681 * data to the requested size using zero bytes.
682 */
683unsigned long _copy_from_user(void *to, const void __user *from, unsigned n)
684{
685 if (access_ok(VERIFY_READ, from, n))
686 n = __copy_from_user(to, from, n);
687 else
688 memset(to, 0, n);
689 return n;
690}
691EXPORT_SYMBOL(_copy_from_user);
diff --git a/arch/x86/mm/fault.c b/arch/x86/mm/fault.c
index 9f72ca3b2669..17c55a536fdd 100644
--- a/arch/x86/mm/fault.c
+++ b/arch/x86/mm/fault.c
@@ -679,8 +679,7 @@ show_fault_oops(struct pt_regs *regs, unsigned long error_code,
679 printk(KERN_CONT "paging request"); 679 printk(KERN_CONT "paging request");
680 680
681 printk(KERN_CONT " at %p\n", (void *) address); 681 printk(KERN_CONT " at %p\n", (void *) address);
682 printk(KERN_ALERT "IP:"); 682 printk(KERN_ALERT "IP: %pS\n", (void *)regs->ip);
683 printk_address(regs->ip);
684 683
685 dump_pagetable(address); 684 dump_pagetable(address);
686} 685}
diff --git a/arch/x86/platform/uv/uv_nmi.c b/arch/x86/platform/uv/uv_nmi.c
index cd5173a2733f..8410e7d0a5b5 100644
--- a/arch/x86/platform/uv/uv_nmi.c
+++ b/arch/x86/platform/uv/uv_nmi.c
@@ -387,8 +387,8 @@ static void uv_nmi_dump_cpu_ip_hdr(void)
387/* Dump Instruction Pointer info */ 387/* Dump Instruction Pointer info */
388static void uv_nmi_dump_cpu_ip(int cpu, struct pt_regs *regs) 388static void uv_nmi_dump_cpu_ip(int cpu, struct pt_regs *regs)
389{ 389{
390 pr_info("UV: %4d %6d %-32.32s ", cpu, current->pid, current->comm); 390 pr_info("UV: %4d %6d %-32.32s %pS",
391 printk_address(regs->ip); 391 cpu, current->pid, current->comm, (void *)regs->ip);
392} 392}
393 393
394/* 394/*
diff --git a/arch/x86/tools/insn_sanity.c b/arch/x86/tools/insn_sanity.c
index ba70ff232917..1972565ab106 100644
--- a/arch/x86/tools/insn_sanity.c
+++ b/arch/x86/tools/insn_sanity.c
@@ -269,7 +269,8 @@ int main(int argc, char **argv)
269 insns++; 269 insns++;
270 } 270 }
271 271
272 fprintf(stdout, "%s: %s: decoded and checked %d %s instructions with %d errors (seed:0x%x)\n", 272 fprintf((errors) ? stderr : stdout,
273 "%s: %s: decoded and checked %d %s instructions with %d errors (seed:0x%x)\n",
273 prog, 274 prog,
274 (errors) ? "Failure" : "Success", 275 (errors) ? "Failure" : "Success",
275 insns, 276 insns,
diff --git a/arch/x86/tools/test_get_len.c b/arch/x86/tools/test_get_len.c
index 56f04db0c9c0..ecf31e0358c8 100644
--- a/arch/x86/tools/test_get_len.c
+++ b/arch/x86/tools/test_get_len.c
@@ -167,7 +167,7 @@ int main(int argc, char **argv)
167 fprintf(stderr, "Warning: decoded and checked %d" 167 fprintf(stderr, "Warning: decoded and checked %d"
168 " instructions with %d warnings\n", insns, warnings); 168 " instructions with %d warnings\n", insns, warnings);
169 else 169 else
170 fprintf(stderr, "Succeed: decoded and checked %d" 170 fprintf(stdout, "Success: decoded and checked %d"
171 " instructions\n", insns); 171 " instructions\n", insns);
172 return 0; 172 return 0;
173} 173}
diff --git a/kernel/sysctl.c b/kernel/sysctl.c
index 739fb17371af..39b3368f6de6 100644
--- a/kernel/sysctl.c
+++ b/kernel/sysctl.c
@@ -983,13 +983,6 @@ static struct ctl_table kern_table[] = {
983 .proc_handler = proc_dointvec, 983 .proc_handler = proc_dointvec,
984 }, 984 },
985 { 985 {
986 .procname = "kstack_depth_to_print",
987 .data = &kstack_depth_to_print,
988 .maxlen = sizeof(int),
989 .mode = 0644,
990 .proc_handler = proc_dointvec,
991 },
992 {
993 .procname = "io_delay_type", 986 .procname = "io_delay_type",
994 .data = &io_delay_type, 987 .data = &io_delay_type,
995 .maxlen = sizeof(int), 988 .maxlen = sizeof(int),
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 6de9440e3ae2..61b0988bba8c 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -6399,8 +6399,8 @@ unsigned long free_reserved_area(void *start, void *end, int poison, char *s)
6399 } 6399 }
6400 6400
6401 if (pages && s) 6401 if (pages && s)
6402 pr_info("Freeing %s memory: %ldK (%p - %p)\n", 6402 pr_info("Freeing %s memory: %ldK\n",
6403 s, pages << (PAGE_SHIFT - 10), start, end); 6403 s, pages << (PAGE_SHIFT - 10));
6404 6404
6405 return pages; 6405 return pages;
6406} 6406}
diff --git a/scripts/decode_stacktrace.sh b/scripts/decode_stacktrace.sh
index c332684e1b5a..5206d99ddeb8 100755
--- a/scripts/decode_stacktrace.sh
+++ b/scripts/decode_stacktrace.sh
@@ -139,7 +139,8 @@ handle_line() {
139 139
140while read line; do 140while read line; do
141 # Let's see if we have an address in the line 141 # Let's see if we have an address in the line
142 if [[ $line =~ \[\<([^]]+)\>\] ]]; then 142 if [[ $line =~ \[\<([^]]+)\>\] ]] ||
143 [[ $line =~ [^+\ ]+\+0x[0-9a-f]+/0x[0-9a-f]+ ]]; then
143 # Translate address to line numbers 144 # Translate address to line numbers
144 handle_line "$line" 145 handle_line "$line"
145 # Is it a code line? 146 # Is it a code line?
diff --git a/scripts/faddr2line b/scripts/faddr2line
index 450b33257339..29df825d375c 100755
--- a/scripts/faddr2line
+++ b/scripts/faddr2line
@@ -105,9 +105,18 @@ __faddr2line() {
105 # In rare cases there might be duplicates. 105 # In rare cases there might be duplicates.
106 while read symbol; do 106 while read symbol; do
107 local fields=($symbol) 107 local fields=($symbol)
108 local sym_base=0x${fields[1]} 108 local sym_base=0x${fields[0]}
109 local sym_size=${fields[2]} 109 local sym_type=${fields[1]}
110 local sym_type=${fields[3]} 110 local sym_end=0x${fields[3]}
111
112 # calculate the size
113 local sym_size=$(($sym_end - $sym_base))
114 if [[ -z $sym_size ]] || [[ $sym_size -le 0 ]]; then
115 warn "bad symbol size: base: $sym_base end: $sym_end"
116 DONE=1
117 return
118 fi
119 sym_size=0x$(printf %x $sym_size)
111 120
112 # calculate the address 121 # calculate the address
113 local addr=$(($sym_base + $offset)) 122 local addr=$(($sym_base + $offset))
@@ -116,26 +125,26 @@ __faddr2line() {
116 DONE=1 125 DONE=1
117 return 126 return
118 fi 127 fi
119 local hexaddr=0x$(printf %x $addr) 128 addr=0x$(printf %x $addr)
120 129
121 # weed out non-function symbols 130 # weed out non-function symbols
122 if [[ $sym_type != "FUNC" ]]; then 131 if [[ $sym_type != t ]] && [[ $sym_type != T ]]; then
123 [[ $print_warnings = 1 ]] && 132 [[ $print_warnings = 1 ]] &&
124 echo "skipping $func address at $hexaddr due to non-function symbol" 133 echo "skipping $func address at $addr due to non-function symbol of type '$sym_type'"
125 continue 134 continue
126 fi 135 fi
127 136
128 # if the user provided a size, make sure it matches the symbol's size 137 # if the user provided a size, make sure it matches the symbol's size
129 if [[ -n $size ]] && [[ $size -ne $sym_size ]]; then 138 if [[ -n $size ]] && [[ $size -ne $sym_size ]]; then
130 [[ $print_warnings = 1 ]] && 139 [[ $print_warnings = 1 ]] &&
131 echo "skipping $func address at $hexaddr due to size mismatch ($size != $sym_size)" 140 echo "skipping $func address at $addr due to size mismatch ($size != $sym_size)"
132 continue; 141 continue;
133 fi 142 fi
134 143
135 # make sure the provided offset is within the symbol's range 144 # make sure the provided offset is within the symbol's range
136 if [[ $offset -gt $sym_size ]]; then 145 if [[ $offset -gt $sym_size ]]; then
137 [[ $print_warnings = 1 ]] && 146 [[ $print_warnings = 1 ]] &&
138 echo "skipping $func address at $hexaddr due to size mismatch ($offset > $sym_size)" 147 echo "skipping $func address at $addr due to size mismatch ($offset > $sym_size)"
139 continue 148 continue
140 fi 149 fi
141 150
@@ -143,12 +152,12 @@ __faddr2line() {
143 [[ $FIRST = 0 ]] && echo 152 [[ $FIRST = 0 ]] && echo
144 FIRST=0 153 FIRST=0
145 154
146 local hexsize=0x$(printf %x $sym_size) 155 # pass real address to addr2line
147 echo "$func+$offset/$hexsize:" 156 echo "$func+$offset/$sym_size:"
148 addr2line -fpie $objfile $hexaddr | sed "s; $dir_prefix\(\./\)*; ;" 157 addr2line -fpie $objfile $addr | sed "s; $dir_prefix\(\./\)*; ;"
149 DONE=1 158 DONE=1
150 159
151 done < <(readelf -sW $objfile | awk -v f=$func '$8 == f {print}') 160 done < <(nm -n $objfile | awk -v fn=$func '$3 == fn { found=1; line=$0; start=$1; next } found == 1 { found=0; print line, $1 }')
152} 161}
153 162
154[[ $# -lt 2 ]] && usage 163[[ $# -lt 2 ]] && usage
diff --git a/tools/testing/selftests/x86/Makefile b/tools/testing/selftests/x86/Makefile
index a89f80a5b711..8c1cb423cfe6 100644
--- a/tools/testing/selftests/x86/Makefile
+++ b/tools/testing/selftests/x86/Makefile
@@ -6,7 +6,7 @@ include ../lib.mk
6 6
7TARGETS_C_BOTHBITS := single_step_syscall sysret_ss_attrs syscall_nt ptrace_syscall test_mremap_vdso \ 7TARGETS_C_BOTHBITS := single_step_syscall sysret_ss_attrs syscall_nt ptrace_syscall test_mremap_vdso \
8 check_initial_reg_state sigreturn ldt_gdt iopl \ 8 check_initial_reg_state sigreturn ldt_gdt iopl \
9 protection_keys 9 protection_keys test_vdso
10TARGETS_C_32BIT_ONLY := entry_from_vm86 syscall_arg_fault test_syscall_vdso unwind_vdso \ 10TARGETS_C_32BIT_ONLY := entry_from_vm86 syscall_arg_fault test_syscall_vdso unwind_vdso \
11 test_FCMOV test_FCOMI test_FISTTP \ 11 test_FCMOV test_FCOMI test_FISTTP \
12 vdso_restorer 12 vdso_restorer
diff --git a/tools/testing/selftests/x86/test_vdso.c b/tools/testing/selftests/x86/test_vdso.c
new file mode 100644
index 000000000000..65d7a2bf7e14
--- /dev/null
+++ b/tools/testing/selftests/x86/test_vdso.c
@@ -0,0 +1,123 @@
1/*
2 * ldt_gdt.c - Test cases for LDT and GDT access
3 * Copyright (c) 2011-2015 Andrew Lutomirski
4 */
5
6#define _GNU_SOURCE
7
8#include <stdio.h>
9#include <sys/time.h>
10#include <time.h>
11#include <stdlib.h>
12#include <unistd.h>
13#include <sys/syscall.h>
14#include <dlfcn.h>
15#include <string.h>
16#include <errno.h>
17#include <sched.h>
18#include <stdbool.h>
19
20#ifndef SYS_getcpu
21# ifdef __x86_64__
22# define SYS_getcpu 309
23# else
24# define SYS_getcpu 318
25# endif
26#endif
27
28int nerrs = 0;
29
30#ifdef __x86_64__
31# define VSYS(x) (x)
32#else
33# define VSYS(x) 0
34#endif
35
36typedef long (*getcpu_t)(unsigned *, unsigned *, void *);
37
38const getcpu_t vgetcpu = (getcpu_t)VSYS(0xffffffffff600800);
39getcpu_t vdso_getcpu;
40
41void fill_function_pointers()
42{
43 void *vdso = dlopen("linux-vdso.so.1",
44 RTLD_LAZY | RTLD_LOCAL | RTLD_NOLOAD);
45 if (!vdso)
46 vdso = dlopen("linux-gate.so.1",
47 RTLD_LAZY | RTLD_LOCAL | RTLD_NOLOAD);
48 if (!vdso) {
49 printf("[WARN]\tfailed to find vDSO\n");
50 return;
51 }
52
53 vdso_getcpu = (getcpu_t)dlsym(vdso, "__vdso_getcpu");
54 if (!vdso_getcpu)
55 printf("Warning: failed to find getcpu in vDSO\n");
56}
57
58static long sys_getcpu(unsigned * cpu, unsigned * node,
59 void* cache)
60{
61 return syscall(__NR_getcpu, cpu, node, cache);
62}
63
64static void test_getcpu(void)
65{
66 printf("[RUN]\tTesting getcpu...\n");
67
68 for (int cpu = 0; ; cpu++) {
69 cpu_set_t cpuset;
70 CPU_ZERO(&cpuset);
71 CPU_SET(cpu, &cpuset);
72 if (sched_setaffinity(0, sizeof(cpuset), &cpuset) != 0)
73 return;
74
75 unsigned cpu_sys, cpu_vdso, cpu_vsys,
76 node_sys, node_vdso, node_vsys;
77 long ret_sys, ret_vdso = 1, ret_vsys = 1;
78 unsigned node;
79
80 ret_sys = sys_getcpu(&cpu_sys, &node_sys, 0);
81 if (vdso_getcpu)
82 ret_vdso = vdso_getcpu(&cpu_vdso, &node_vdso, 0);
83 if (vgetcpu)
84 ret_vsys = vgetcpu(&cpu_vsys, &node_vsys, 0);
85
86 if (!ret_sys)
87 node = node_sys;
88 else if (!ret_vdso)
89 node = node_vdso;
90 else if (!ret_vsys)
91 node = node_vsys;
92
93 bool ok = true;
94 if (!ret_sys && (cpu_sys != cpu || node_sys != node))
95 ok = false;
96 if (!ret_vdso && (cpu_vdso != cpu || node_vdso != node))
97 ok = false;
98 if (!ret_vsys && (cpu_vsys != cpu || node_vsys != node))
99 ok = false;
100
101 printf("[%s]\tCPU %u:", ok ? "OK" : "FAIL", cpu);
102 if (!ret_sys)
103 printf(" syscall: cpu %u, node %u", cpu_sys, node_sys);
104 if (!ret_vdso)
105 printf(" vdso: cpu %u, node %u", cpu_vdso, node_vdso);
106 if (!ret_vsys)
107 printf(" vsyscall: cpu %u, node %u", cpu_vsys,
108 node_vsys);
109 printf("\n");
110
111 if (!ok)
112 nerrs++;
113 }
114}
115
116int main(int argc, char **argv)
117{
118 fill_function_pointers();
119
120 test_getcpu();
121
122 return nerrs ? 1 : 0;
123}
generated by cgit v1.2.2 (git 2.25.0) at 2025-08-22 02:33:18 -0400