Commit 650e5455 authored by Linus Torvalds's avatar Linus Torvalds
Browse files

Merge branch 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 

Pull x86 fixes from Ingo Molnar:
 "A handful of x86 fixes:

   - a syscall ABI fix, fixing an Android breakage
   - a Xen PV guest fix relating to the RTC device, causing a
     non-working console
   - a Xen guest syscall stack frame fix
   - an MCE hotplug CPU crash fix"

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/numachip: Fix NumaConnect2 MMCFG PCI access
  x86/entry: Restore traditional SYSENTER calling convention
  x86/entry: Fix some comments
  x86/paravirt: Prevent rtc_cmos platform device init on PV guests
  x86/xen: Avoid fast syscall path for Xen PV guests
  x86/mce: Ensure offline CPUs don't participate in rendezvous process
parents de030179 dd7a5ab4

arch/x86/entry/common.c

+3 −3
Original line number Diff line number Diff line
@@ -421,7 +421,7 @@ __visible long do_fast_syscall_32(struct pt_regs *regs) 
	regs->ip = landing_pad;



	/*

	 * Fetch ECX from where the vDSO stashed it.

	 * Fetch EBP from where the vDSO stashed it.

	 *

	 * WARNING: We are in CONTEXT_USER and RCU isn't paying attention!

	 */
@@ -432,10 +432,10 @@ __visible long do_fast_syscall_32(struct pt_regs *regs) 
		 * Micro-optimization: the pointer we're following is explicitly

		 * 32 bits, so it can't be out of range.

		 */

		__get_user(*(u32 *)&regs->cx,

		__get_user(*(u32 *)&regs->bp,

			    (u32 __user __force *)(unsigned long)(u32)regs->sp)

#else

		get_user(*(u32 *)&regs->cx,

		get_user(*(u32 *)&regs->bp,

			 (u32 __user __force *)(unsigned long)(u32)regs->sp)

#endif

		) {

arch/x86/entry/entry_32.S

+4 −3
Original line number Diff line number Diff line
@@ -292,7 +292,7 @@ ENTRY(entry_SYSENTER_32) 
	movl	TSS_sysenter_sp0(%esp), %esp

sysenter_past_esp:

	pushl	$__USER_DS		/* pt_regs->ss */

	pushl	%ecx			/* pt_regs->cx */

	pushl	%ebp			/* pt_regs->sp (stashed in bp) */

	pushfl				/* pt_regs->flags (except IF = 0) */

	orl	$X86_EFLAGS_IF, (%esp)	/* Fix IF */

	pushl	$__USER_CS		/* pt_regs->cs */
@@ -308,8 +308,9 @@ sysenter_past_esp: 


	movl	%esp, %eax

	call	do_fast_syscall_32

	testl	%eax, %eax

	jz	.Lsyscall_32_done

	/* XEN PV guests always use IRET path */

	ALTERNATIVE "testl %eax, %eax; jz .Lsyscall_32_done", \

		    "jmp .Lsyscall_32_done", X86_FEATURE_XENPV



/* Opportunistic SYSEXIT */

	TRACE_IRQS_ON			/* User mode traces as IRQs on. */

arch/x86/entry/entry_64_compat.S

+11 −9
Original line number Diff line number Diff line
@@ -63,7 +63,7 @@ ENTRY(entry_SYSENTER_compat) 


	/* Construct struct pt_regs on stack */

	pushq	$__USER32_DS		/* pt_regs->ss */

	pushq	%rcx			/* pt_regs->sp */

	pushq	%rbp			/* pt_regs->sp (stashed in bp) */



	/*

	 * Push flags.  This is nasty.  First, interrupts are currently
@@ -82,14 +82,14 @@ ENTRY(entry_SYSENTER_compat) 
	pushq	%rdi			/* pt_regs->di */

	pushq	%rsi			/* pt_regs->si */

	pushq	%rdx			/* pt_regs->dx */

	pushq	%rcx			/* pt_regs->cx (will be overwritten) */

	pushq	%rcx			/* pt_regs->cx */

	pushq	$-ENOSYS		/* pt_regs->ax */

	pushq   %r8                     /* pt_regs->r8  = 0 */

	pushq   %r8                     /* pt_regs->r9  = 0 */

	pushq   %r8                     /* pt_regs->r10 = 0 */

	pushq   %r8                     /* pt_regs->r11 = 0 */

	pushq   %rbx                    /* pt_regs->rbx */

	pushq   %rbp                    /* pt_regs->rbp */

	pushq   %rbp                    /* pt_regs->rbp (will be overwritten) */

	pushq   %r8                     /* pt_regs->r12 = 0 */

	pushq   %r8                     /* pt_regs->r13 = 0 */

	pushq   %r8                     /* pt_regs->r14 = 0 */
@@ -121,8 +121,9 @@ sysenter_flags_fixed: 


	movq	%rsp, %rdi

	call	do_fast_syscall_32

	testl	%eax, %eax

	jz	.Lsyscall_32_done

	/* XEN PV guests always use IRET path */

	ALTERNATIVE "testl %eax, %eax; jz .Lsyscall_32_done", \

		    "jmp .Lsyscall_32_done", X86_FEATURE_XENPV

	jmp	sysret32_from_system_call



sysenter_fix_flags:
@@ -178,7 +179,7 @@ ENTRY(entry_SYSCALL_compat) 
	pushq	%rdi			/* pt_regs->di */

	pushq	%rsi			/* pt_regs->si */

	pushq	%rdx			/* pt_regs->dx */

	pushq	%rcx			/* pt_regs->cx (will be overwritten) */

	pushq	%rbp			/* pt_regs->cx (stashed in bp) */

	pushq	$-ENOSYS		/* pt_regs->ax */

	xorq    %r8,%r8

	pushq   %r8                     /* pt_regs->r8  = 0 */
@@ -186,7 +187,7 @@ ENTRY(entry_SYSCALL_compat) 
	pushq   %r8                     /* pt_regs->r10 = 0 */

	pushq   %r8                     /* pt_regs->r11 = 0 */

	pushq   %rbx                    /* pt_regs->rbx */

	pushq   %rbp                    /* pt_regs->rbp */

	pushq   %rbp                    /* pt_regs->rbp (will be overwritten) */

	pushq   %r8                     /* pt_regs->r12 = 0 */

	pushq   %r8                     /* pt_regs->r13 = 0 */

	pushq   %r8                     /* pt_regs->r14 = 0 */
@@ -200,8 +201,9 @@ ENTRY(entry_SYSCALL_compat) 


	movq	%rsp, %rdi

	call	do_fast_syscall_32

	testl	%eax, %eax

	jz	.Lsyscall_32_done

	/* XEN PV guests always use IRET path */

	ALTERNATIVE "testl %eax, %eax; jz .Lsyscall_32_done", \

		    "jmp .Lsyscall_32_done", X86_FEATURE_XENPV



	/* Opportunistic SYSRET */

sysret32_from_system_call:

arch/x86/entry/vdso/vdso32/system_call.S

+43 −11
Original line number Diff line number Diff line 
/*

 * Code for the vDSO.  This version uses the old int $0x80 method.

 * AT_SYSINFO entry point

*/



#include <asm/dwarf2.h>
@@ -21,35 +21,67 @@ __kernel_vsyscall: 
	/*

	 * Reshuffle regs so that all of any of the entry instructions

	 * will preserve enough state.

	 *

	 * A really nice entry sequence would be:

	 *  pushl %edx

	 *  pushl %ecx

	 *  movl  %esp, %ecx

	 *

	 * Unfortunately, naughty Android versions between July and December

	 * 2015 actually hardcode the traditional Linux SYSENTER entry

	 * sequence.  That is severely broken for a number of reasons (ask

	 * anyone with an AMD CPU, for example).  Nonetheless, we try to keep

	 * it working approximately as well as it ever worked.

	 *

	 * This link may eludicate some of the history:

	 *   https://android-review.googlesource.com/#/q/Iac3295376d61ef83e713ac9b528f3b50aa780cd7

	 * personally, I find it hard to understand what's going on there.

	 *

	 * Note to future user developers: DO NOT USE SYSENTER IN YOUR CODE.

	 * Execute an indirect call to the address in the AT_SYSINFO auxv

	 * entry.  That is the ONLY correct way to make a fast 32-bit system

	 * call on Linux.  (Open-coding int $0x80 is also fine, but it's

	 * slow.)

	 */

	pushl	%ecx

	CFI_ADJUST_CFA_OFFSET	4

	CFI_REL_OFFSET		ecx, 0

	pushl	%edx

	CFI_ADJUST_CFA_OFFSET	4

	CFI_REL_OFFSET		edx, 0

	pushl	%ecx

	pushl	%ebp

	CFI_ADJUST_CFA_OFFSET	4

	CFI_REL_OFFSET		ecx, 0

	movl	%esp, %ecx

	CFI_REL_OFFSET		ebp, 0



	#define SYSENTER_SEQUENCE	"movl %esp, %ebp; sysenter"

	#define SYSCALL_SEQUENCE	"movl %ecx, %ebp; syscall"



#ifdef CONFIG_X86_64

	/* If SYSENTER (Intel) or SYSCALL32 (AMD) is available, use it. */

	ALTERNATIVE_2 "", "sysenter", X86_FEATURE_SYSENTER32, \

	                  "syscall",  X86_FEATURE_SYSCALL32

	ALTERNATIVE_2 "", SYSENTER_SEQUENCE, X86_FEATURE_SYSENTER32, \

	                  SYSCALL_SEQUENCE,  X86_FEATURE_SYSCALL32

#else

	ALTERNATIVE "", "sysenter", X86_FEATURE_SEP

	ALTERNATIVE "", SYSENTER_SEQUENCE, X86_FEATURE_SEP

#endif



	/* Enter using int $0x80 */

	movl	(%esp), %ecx

	int	$0x80

GLOBAL(int80_landing_pad)



	/* Restore ECX and EDX in case they were clobbered. */

	popl	%ecx

	CFI_RESTORE		ecx

	/*

	 * Restore EDX and ECX in case they were clobbered.  EBP is not

	 * clobbered (the kernel restores it), but it's cleaner and

	 * probably faster to pop it than to adjust ESP using addl.

	 */

	popl	%ebp

	CFI_RESTORE		ebp

	CFI_ADJUST_CFA_OFFSET	-4

	popl	%edx

	CFI_RESTORE		edx

	CFI_ADJUST_CFA_OFFSET	-4

	popl	%ecx

	CFI_RESTORE		ecx

	CFI_ADJUST_CFA_OFFSET	-4

	ret

	CFI_ENDPROC

arch/x86/include/asm/cpufeature.h

+1 −0
Original line number Diff line number Diff line
@@ -216,6 +216,7 @@ 
#define X86_FEATURE_PAUSEFILTER ( 8*32+13) /* AMD filtered pause intercept */

#define X86_FEATURE_PFTHRESHOLD ( 8*32+14) /* AMD pause filter threshold */

#define X86_FEATURE_VMMCALL     ( 8*32+15) /* Prefer vmmcall to vmcall */

#define X86_FEATURE_XENPV       ( 8*32+16) /* "" Xen paravirtual guest */





/* Intel-defined CPU features, CPUID level 0x00000007:0 (ebx), word 9 */

CRA Git | Maintained and supported by SUSTech CRA and CCSE