Message-Id: <1220110858.14894.46.camel@lappy.programming.kicks-ass.net>
Date:	Sat, 30 Aug 2008 17:40:58 +0200
From:	Peter Zijlstra <a.p.zijlstra@...llo.nl>
To:	paulmck@...ux.vnet.ibm.com
Cc:	linux-kernel@...r.kernel.org, cl@...ux-foundation.org,
	mingo@...e.hu, akpm@...ux-foundation.org, manfred@...orfullife.com,
	dipankar@...ibm.com, josht@...ux.vnet.ibm.com, schamp@....com,
	niv@...ibm.com, dvhltc@...ibm.com, ego@...ibm.com,
	laijs@...fujitsu.com, rostedt@...dmis.org,
	Mathieu Desnoyers <mathieu.desnoyers@...ymtl.ca>
Subject: Re: [PATCH, RFC, tip/core/rcu] v3 scalable classic RCU
	implementation

On Sat, 2008-08-30 at 07:10 -0700, Paul E. McKenney wrote:
> On Sat, Aug 30, 2008 at 11:33:00AM +0200, Peter Zijlstra wrote:
> > On Fri, 2008-08-29 at 17:49 -0700, Paul E. McKenney wrote:
> > 
> > > Some shortcomings:
> > > 
> > > o	Entering and leaving dynticks idle mode is a quiescent state,
> > > 	but the current patch doesn't take advantage of this (noted
> > > 	by Manfred).  It appears that it should be possible to make
> > > 	nmi_enter() and nmi_exit() provide an in_nmi(), which would make
> > > 	it possible for rcu_irq_enter() and rcu_irq_exit() to figure
> > > 	out whether it is safe to tell RCU about the quiescent state --
> > > 	and also greatly simplify the code.
> > 
> > Already done and available in the -tip tree, curtesy of Mathieu.
> 
> Very cool!!!  I see one of his patches at http://lkml.org/lkml/2008/4/17/342,
> but how do I find out which branch of -tip this is on?  (I am learning
> git, but it is a slow process...)
> 
> This would also simplify preemptable RCU's dyntick interface, removing
> the need for proofs.

Not sure - my git-foo isn't good enough either :-(

All I can offer is that its available in tip/master (the collective
merge of all of tip's branches) as commit:
0d84b78a606f1562532cd576ee8733caf5a4aed3, which I found using
git-annotate include/linux/hardirq.h

How to find from which particular topic branch it came from, I too am
clueless.

---
commit 0d84b78a606f1562532cd576ee8733caf5a4aed3
Author: Mathieu Desnoyers <mathieu.desnoyers@...ymtl.ca>
Date:   Mon May 12 21:21:07 2008 +0200

    x86 NMI-safe INT3 and Page Fault
    
    Implements an alternative iret with popf and return so trap and exception
    handlers can return to the NMI handler without issuing iret. iret would cause
    NMIs to be reenabled prematurely. x86_32 uses popf and far return. x86_64 has to
    copy the return instruction pointer to the top of the previous stack, issue a
    popf, loads the previous esp and issue a near return (ret).
    
    It allows placing immediate values (and therefore optimized trace_marks) in NMI
    code since returning from a breakpoint would be valid. Accessing vmalloc'd
    memory, which allows executing module code or accessing vmapped or vmalloc'd
    areas from NMI context, would also be valid. This is very useful to tracers like
    LTTng.
    
    This patch makes all faults, traps and exception safe to be called from NMI
    context *except* single-stepping, which requires iret to restore the TF (trap
    flag) and jump to the return address in a single instruction. Sorry, no kprobes
    support in NMI handlers because of this limitation.  We cannot single-step an
    NMI handler, because iret must set the TF flag and return back to the
    instruction to single-step in a single instruction. This cannot be emulated with
    popf/lret, because lret would be single-stepped. It does not apply to immediate
    values because they do not use single-stepping. This code detects if the TF
    flag is set and uses the iret path for single-stepping, even if it reactivates
    NMIs prematurely.
    
    Test to detect if nested under a NMI handler is only done upon the return from
    trap/exception to kernel, which is not frequent. Other return paths (return from
    trap/exception to userspace, return from interrupt) keep the exact same behavior
    (no slowdown).
    
    Depends on :
    change-alpha-active-count-bit.patch
    change-avr32-active-count-bit.patch
    
    TODO : test with lguest, xen, kvm.
    
    ** This patch depends on the "Stringify support commas" patchset **
    ** Also depends on fix-x86_64-page-fault-scheduler-race patch **
    
    tested on x86_32 (tests implemented in a separate patch) :
    - instrumented the return path to export the EIP, CS and EFLAGS values when
      taken so we know the return path code has been executed.
    - trace_mark, using immediate values, with 10ms delay with the breakpoint
      activated. Runs well through the return path.
    - tested vmalloc faults in NMI handler by placing a non-optimized marker in the
      NMI handler (so no breakpoint is executed) and connecting a probe which
      touches every pages of a 20MB vmalloc'd buffer. It executes trough the return
      path without problem.
    - Tested with and without preemption
    
    tested on x86_64
    - instrumented the return path to export the EIP, CS and EFLAGS values when
      taken so we know the return path code has been executed.
    - trace_mark, using immediate values, with 10ms delay with the breakpoint
      activated. Runs well through the return path.
    
    To test on x86_64 :
    - Test without preemption
    - Test vmalloc faults
    - Test on Intel 64 bits CPUs. (AMD64 was fine)
    
    Changelog since v1 :
    - x86_64 fixes.
    Changelog since v2 :
    - fix paravirt build
    Changelog since v3 :
    - Include modifications suggested by Jeremy
    Changelog since v4 :
    - including hardirq.h in entry_32/64.S is a bad idea (non ifndef'd C code),
      define HARDNMI_MASK in the .S files directly.
    Changelog since v5 :
    - Add HARDNMI_MASK to irq_count() and make die() more verbose for NMIs.
    Changelog since v7 :
    - Implement paravirtualized nmi_return.
    Changelog since v8 :
    - refreshed the patch for asm-offsets. Those were left out of v8.
    - now depends on "Stringify support commas" patch.
    Changelog since v9 :
    - Only test the nmi nested preempt count flag upon return from exceptions, not
      on return from interrupts. Only the kernel return path has this test.
    - Add Xen, VMI, lguest support. Use their iret pavavirt ops in lieu of
      nmi_return.
    
    -- Ported to sched-devel.git
    
    Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@...ymtl.ca>
    CC: akpm@...l.org
    CC: mingo@...e.hu
    CC: "H. Peter Anvin" <hpa@...or.com>
    CC: Jeremy Fitzhardinge <jeremy@...p.org>
    CC: Steven Rostedt <rostedt@...dmis.org>
    CC: "Frank Ch. Eigler" <fche@...hat.com>
    Signed-off-by: Ingo Molnar <mingo@...e.hu>
    Signed-off-by: Thomas Gleixner <tglx@...utronix.de>

diff --git a/arch/x86/kernel/asm-offsets_32.c b/arch/x86/kernel/asm-offsets_32.c
index 9258808..73474e0 100644
--- a/arch/x86/kernel/asm-offsets_32.c
+++ b/arch/x86/kernel/asm-offsets_32.c
@@ -111,6 +111,7 @@ void foo(void)
 	OFFSET(PV_IRQ_irq_disable, pv_irq_ops, irq_disable);
 	OFFSET(PV_IRQ_irq_enable, pv_irq_ops, irq_enable);
 	OFFSET(PV_CPU_iret, pv_cpu_ops, iret);
+	OFFSET(PV_CPU_nmi_return, pv_cpu_ops, nmi_return);
 	OFFSET(PV_CPU_irq_enable_syscall_ret, pv_cpu_ops, irq_enable_syscall_ret);
 	OFFSET(PV_CPU_read_cr0, pv_cpu_ops, read_cr0);
 #endif
diff --git a/arch/x86/kernel/asm-offsets_64.c b/arch/x86/kernel/asm-offsets_64.c
index f126c05..a5bbec3 100644
--- a/arch/x86/kernel/asm-offsets_64.c
+++ b/arch/x86/kernel/asm-offsets_64.c
@@ -62,6 +62,7 @@ int main(void)
 	OFFSET(PV_IRQ_irq_disable, pv_irq_ops, irq_disable);
 	OFFSET(PV_IRQ_irq_enable, pv_irq_ops, irq_enable);
 	OFFSET(PV_CPU_iret, pv_cpu_ops, iret);
+	OFFSET(PV_CPU_nmi_return, pv_cpu_ops, nmi_return);
 	OFFSET(PV_CPU_irq_enable_syscall_ret, pv_cpu_ops, irq_enable_syscall_ret);
 	OFFSET(PV_CPU_swapgs, pv_cpu_ops, swapgs);
 	OFFSET(PV_MMU_read_cr2, pv_mmu_ops, read_cr2);
diff --git a/arch/x86/kernel/entry_32.S b/arch/x86/kernel/entry_32.S
index e6517ce..2d88211 100644
--- a/arch/x86/kernel/entry_32.S
+++ b/arch/x86/kernel/entry_32.S
@@ -68,6 +68,8 @@
 
 #define nr_syscalls ((syscall_table_size)/4)
 
+#define HARDNMI_MASK 0x40000000
+
 #ifdef CONFIG_PREEMPT
 #define preempt_stop(clobbers)	DISABLE_INTERRUPTS(clobbers); TRACE_IRQS_OFF
 #else
@@ -232,8 +234,32 @@ END(ret_from_fork)
 	# userspace resumption stub bypassing syscall exit tracing
 	ALIGN
 	RING0_PTREGS_FRAME
+
 ret_from_exception:
 	preempt_stop(CLBR_ANY)
+	GET_THREAD_INFO(%ebp)
+	movl PT_EFLAGS(%esp), %eax	# mix EFLAGS and CS
+	movb PT_CS(%esp), %al
+	andl $(X86_EFLAGS_VM | SEGMENT_RPL_MASK), %eax
+	cmpl $USER_RPL, %eax
+	jae resume_userspace	# returning to v8086 or userspace
+	testl $HARDNMI_MASK,TI_preempt_count(%ebp)
+	jz resume_kernel		/* Not nested over NMI ? */
+	testw $X86_EFLAGS_TF, PT_EFLAGS(%esp)
+	jnz resume_kernel		/*
+					 * If single-stepping an NMI handler,
+					 * use the normal iret path instead of
+					 * the popf/lret because lret would be
+					 * single-stepped. It should not
+					 * happen : it will reactivate NMIs
+					 * prematurely.
+					 */
+	TRACE_IRQS_IRET
+	RESTORE_REGS
+	addl $4, %esp			# skip orig_eax/error_code
+	CFI_ADJUST_CFA_OFFSET -4
+	INTERRUPT_RETURN_NMI_SAFE
+
 ret_from_intr:
 	GET_THREAD_INFO(%ebp)
 check_userspace:
@@ -873,6 +899,10 @@ ENTRY(native_iret)
 .previous
 END(native_iret)
 
+ENTRY(native_nmi_return)
+	NATIVE_INTERRUPT_RETURN_NMI_SAFE # Should we deal with popf exception ?
+END(native_nmi_return)
+
 ENTRY(native_irq_enable_syscall_ret)
 	sti
 	sysexit
diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S
index fe25e5f..5f8edc7 100644
--- a/arch/x86/kernel/entry_64.S
+++ b/arch/x86/kernel/entry_64.S
@@ -156,6 +156,8 @@ END(mcount)
 #endif /* CONFIG_DYNAMIC_FTRACE */
 #endif /* CONFIG_FTRACE */
 
+#define HARDNMI_MASK 0x40000000
+
 #ifndef CONFIG_PREEMPT
 #define retint_kernel retint_restore_args
 #endif	
@@ -698,6 +700,9 @@ ENTRY(native_iret)
 	.section __ex_table,"a"
 	.quad native_iret, bad_iret
 	.previous
+
+ENTRY(native_nmi_return)
+	NATIVE_INTERRUPT_RETURN_NMI_SAFE
 #endif
 
 	.section .fixup,"ax"
@@ -753,6 +758,23 @@ retint_signal:
 	GET_THREAD_INFO(%rcx)
 	jmp retint_check
 
+	/* Returning to kernel space from exception. */
+	/* rcx:	 threadinfo. interrupts off. */
+ENTRY(retexc_kernel)
+	testl $HARDNMI_MASK,threadinfo_preempt_count(%rcx)
+	jz retint_kernel		/* Not nested over NMI ? */
+	testw $X86_EFLAGS_TF,EFLAGS-ARGOFFSET(%rsp)	/* trap flag? */
+	jnz retint_kernel		/*
+					 * If single-stepping an NMI handler,
+					 * use the normal iret path instead of
+					 * the popf/lret because lret would be
+					 * single-stepped. It should not
+					 * happen : it will reactivate NMIs
+					 * prematurely.
+					 */
+	RESTORE_ARGS 0,8,0
+	INTERRUPT_RETURN_NMI_SAFE
+
 #ifdef CONFIG_PREEMPT
 	/* Returning to kernel space. Check if we need preemption */
 	/* rcx:	 threadinfo. interrupts off. */
@@ -911,9 +933,17 @@ paranoid_swapgs\trace:
 	TRACE_IRQS_IRETQ 0
 	.endif
 	SWAPGS_UNSAFE_STACK
-paranoid_restore\trace:
+paranoid_restore_no_nmi\trace:
 	RESTORE_ALL 8
 	jmp irq_return
+paranoid_restore\trace:
+	GET_THREAD_INFO(%rcx)
+	testl $HARDNMI_MASK,threadinfo_preempt_count(%rcx)
+	jz paranoid_restore_no_nmi\trace	/* Nested over NMI ? */
+	testw $X86_EFLAGS_TF,EFLAGS-0(%rsp)	/* trap flag? */
+	jnz paranoid_restore_no_nmi\trace
+	RESTORE_ALL 8
+	INTERRUPT_RETURN_NMI_SAFE
 paranoid_userspace\trace:
 	GET_THREAD_INFO(%rcx)
 	movl threadinfo_flags(%rcx),%ebx
@@ -1012,7 +1042,7 @@ error_exit:
 	TRACE_IRQS_OFF
 	GET_THREAD_INFO(%rcx)	
 	testl %eax,%eax
-	jne  retint_kernel
+	jne  retexc_kernel
 	LOCKDEP_SYS_EXIT_IRQ
 	movl  threadinfo_flags(%rcx),%edx
 	movl  $_TIF_WORK_MASK,%edi
diff --git a/arch/x86/kernel/paravirt.c b/arch/x86/kernel/paravirt.c
index 74f0c5e..bb174a8 100644
--- a/arch/x86/kernel/paravirt.c
+++ b/arch/x86/kernel/paravirt.c
@@ -139,6 +139,7 @@ unsigned paravirt_patch_default(u8 type, u16 clobbers, void *insnbuf,
 		/* If the operation is a nop, then nop the callsite */
 		ret = paravirt_patch_nop();
 	else if (type == PARAVIRT_PATCH(pv_cpu_ops.iret) ||
+		 type == PARAVIRT_PATCH(pv_cpu_ops.nmi_return) ||
 		 type == PARAVIRT_PATCH(pv_cpu_ops.irq_enable_syscall_ret))
 		/* If operation requires a jmp, then jmp */
 		ret = paravirt_patch_jmp(insnbuf, opfunc, addr, len);
@@ -190,6 +191,7 @@ static void native_flush_tlb_single(unsigned long addr)
 
 /* These are in entry.S */
 extern void native_iret(void);
+extern void native_nmi_return(void);
 extern void native_irq_enable_syscall_ret(void);
 
 static int __init print_banner(void)
@@ -328,6 +330,7 @@ struct pv_cpu_ops pv_cpu_ops = {
 
 	.irq_enable_syscall_ret = native_irq_enable_syscall_ret,
 	.iret = native_iret,
+	.nmi_return = native_nmi_return,
 	.swapgs = native_swapgs,
 
 	.set_iopl_mask = native_set_iopl_mask,
diff --git a/arch/x86/kernel/paravirt_patch_32.c b/arch/x86/kernel/paravirt_patch_32.c
index 82fc5fc..8ed31c7 100644
--- a/arch/x86/kernel/paravirt_patch_32.c
+++ b/arch/x86/kernel/paravirt_patch_32.c
@@ -1,10 +1,13 @@
-#include <asm/paravirt.h>
+#include <linux/stringify.h>
+#include <linux/irqflags.h>
 
 DEF_NATIVE(pv_irq_ops, irq_disable, "cli");
 DEF_NATIVE(pv_irq_ops, irq_enable, "sti");
 DEF_NATIVE(pv_irq_ops, restore_fl, "push %eax; popf");
 DEF_NATIVE(pv_irq_ops, save_fl, "pushf; pop %eax");
 DEF_NATIVE(pv_cpu_ops, iret, "iret");
+DEF_NATIVE(pv_cpu_ops, nmi_return,
+	__stringify(NATIVE_INTERRUPT_RETURN_NMI_SAFE));
 DEF_NATIVE(pv_cpu_ops, irq_enable_syscall_ret, "sti; sysexit");
 DEF_NATIVE(pv_mmu_ops, read_cr2, "mov %cr2, %eax");
 DEF_NATIVE(pv_mmu_ops, write_cr3, "mov %eax, %cr3");
@@ -29,6 +32,7 @@ unsigned native_patch(u8 type, u16 clobbers, void *ibuf,
 		PATCH_SITE(pv_irq_ops, restore_fl);
 		PATCH_SITE(pv_irq_ops, save_fl);
 		PATCH_SITE(pv_cpu_ops, iret);
+		PATCH_SITE(pv_cpu_ops, nmi_return);
 		PATCH_SITE(pv_cpu_ops, irq_enable_syscall_ret);
 		PATCH_SITE(pv_mmu_ops, read_cr2);
 		PATCH_SITE(pv_mmu_ops, read_cr3);
diff --git a/arch/x86/kernel/paravirt_patch_64.c b/arch/x86/kernel/paravirt_patch_64.c
index 7d904e1..56eccea 100644
--- a/arch/x86/kernel/paravirt_patch_64.c
+++ b/arch/x86/kernel/paravirt_patch_64.c
@@ -1,12 +1,15 @@
+#include <linux/irqflags.h>
+#include <linux/stringify.h>
 #include <asm/paravirt.h>
 #include <asm/asm-offsets.h>
-#include <linux/stringify.h>
 
 DEF_NATIVE(pv_irq_ops, irq_disable, "cli");
 DEF_NATIVE(pv_irq_ops, irq_enable, "sti");
 DEF_NATIVE(pv_irq_ops, restore_fl, "pushq %rdi; popfq");
 DEF_NATIVE(pv_irq_ops, save_fl, "pushfq; popq %rax");
 DEF_NATIVE(pv_cpu_ops, iret, "iretq");
+DEF_NATIVE(pv_cpu_ops, nmi_return,
+	__stringify(NATIVE_INTERRUPT_RETURN_NMI_SAFE));
 DEF_NATIVE(pv_mmu_ops, read_cr2, "movq %cr2, %rax");
 DEF_NATIVE(pv_mmu_ops, read_cr3, "movq %cr3, %rax");
 DEF_NATIVE(pv_mmu_ops, write_cr3, "movq %rdi, %cr3");
@@ -35,6 +38,7 @@ unsigned native_patch(u8 type, u16 clobbers, void *ibuf,
 		PATCH_SITE(pv_irq_ops, irq_enable);
 		PATCH_SITE(pv_irq_ops, irq_disable);
 		PATCH_SITE(pv_cpu_ops, iret);
+		PATCH_SITE(pv_cpu_ops, nmi_return);
 		PATCH_SITE(pv_cpu_ops, irq_enable_syscall_ret);
 		PATCH_SITE(pv_cpu_ops, swapgs);
 		PATCH_SITE(pv_mmu_ops, read_cr2);
diff --git a/arch/x86/kernel/traps_32.c b/arch/x86/kernel/traps_32.c
index bde6f63..f3a59cd 100644
--- a/arch/x86/kernel/traps_32.c
+++ b/arch/x86/kernel/traps_32.c
@@ -475,6 +475,9 @@ void die(const char *str, struct pt_regs *regs, long err)
 	if (kexec_should_crash(current))
 		crash_kexec(regs);
 
+	if (in_nmi())
+		panic("Fatal exception in non-maskable interrupt");
+
 	if (in_interrupt())
 		panic("Fatal exception in interrupt");
 
diff --git a/arch/x86/kernel/traps_64.c b/arch/x86/kernel/traps_64.c
index adff76e..3dacb75 100644
--- a/arch/x86/kernel/traps_64.c
+++ b/arch/x86/kernel/traps_64.c
@@ -555,6 +555,10 @@ void __kprobes oops_end(unsigned long flags, struct pt_regs *regs, int signr)
 		oops_exit();
 		return;
 	}
+	if (in_nmi())
+		panic("Fatal exception in non-maskable interrupt");
+	if (in_interrupt())
+		panic("Fatal exception in interrupt");
 	if (panic_on_oops)
 		panic("Fatal exception");
 	oops_exit();
diff --git a/arch/x86/kernel/vmi_32.c b/arch/x86/kernel/vmi_32.c
index 956f389..01d687d 100644
--- a/arch/x86/kernel/vmi_32.c
+++ b/arch/x86/kernel/vmi_32.c
@@ -151,6 +151,8 @@ static unsigned vmi_patch(u8 type, u16 clobbers, void *insns,
 					      insns, ip);
 		case PARAVIRT_PATCH(pv_cpu_ops.iret):
 			return patch_internal(VMI_CALL_IRET, len, insns, ip);
+		case PARAVIRT_PATCH(pv_cpu_ops.nmi_return):
+			return patch_internal(VMI_CALL_IRET, len, insns, ip);
 		case PARAVIRT_PATCH(pv_cpu_ops.irq_enable_syscall_ret):
 			return patch_internal(VMI_CALL_SYSEXIT, len, insns, ip);
 		default:
diff --git a/arch/x86/lguest/boot.c b/arch/x86/lguest/boot.c
index af65b2d..f5cbb74 100644
--- a/arch/x86/lguest/boot.c
+++ b/arch/x86/lguest/boot.c
@@ -958,6 +958,7 @@ __init void lguest_init(void)
 	pv_cpu_ops.cpuid = lguest_cpuid;
 	pv_cpu_ops.load_idt = lguest_load_idt;
 	pv_cpu_ops.iret = lguest_iret;
+	pv_cpu_ops.nmi_return = lguest_iret;
 	pv_cpu_ops.load_sp0 = lguest_load_sp0;
 	pv_cpu_ops.load_tr_desc = lguest_load_tr_desc;
 	pv_cpu_ops.set_ldt = lguest_set_ldt;
diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c
index c8a56e4..33272ce 100644
--- a/arch/x86/xen/enlighten.c
+++ b/arch/x86/xen/enlighten.c
@@ -1008,6 +1008,7 @@ static const struct pv_cpu_ops xen_cpu_ops __initdata = {
 	.read_pmc = native_read_pmc,
 
 	.iret = xen_iret,
+	.nmi_return = xen_iret,
 	.irq_enable_syscall_ret = xen_sysexit,
 
 	.load_tr_desc = paravirt_nop,
diff --git a/include/asm-x86/irqflags.h b/include/asm-x86/irqflags.h
index 24d71b1..c3009fd 100644
--- a/include/asm-x86/irqflags.h
+++ b/include/asm-x86/irqflags.h
@@ -51,6 +51,61 @@ static inline void native_halt(void)
 
 #endif
 
+#ifdef CONFIG_X86_64
+/*
+ * Only returns from a trap or exception to a NMI context (intra-privilege
+ * level near return) to the same SS and CS segments. Should be used
+ * upon trap or exception return when nested over a NMI context so no iret is
+ * issued. It takes care of modifying the eflags, rsp and returning to the
+ * previous function.
+ *
+ * The stack, at that point, looks like :
+ *
+ * 0(rsp)  RIP
+ * 8(rsp)  CS
+ * 16(rsp) EFLAGS
+ * 24(rsp) RSP
+ * 32(rsp) SS
+ *
+ * Upon execution :
+ * Copy EIP to the top of the return stack
+ * Update top of return stack address
+ * Pop eflags into the eflags register
+ * Make the return stack current
+ * Near return (popping the return address from the return stack)
+ */
+#define NATIVE_INTERRUPT_RETURN_NMI_SAFE	pushq %rax;		\
+						movq %rsp, %rax;	\
+						movq 24+8(%rax), %rsp;	\
+						pushq 0+8(%rax);	\
+						pushq 16+8(%rax);	\
+						movq (%rax), %rax;	\
+						popfq;			\
+						ret
+#else
+/*
+ * Protected mode only, no V8086. Implies that protected mode must
+ * be entered before NMIs or MCEs are enabled. Only returns from a trap or
+ * exception to a NMI context (intra-privilege level far return). Should be used
+ * upon trap or exception return when nested over a NMI context so no iret is
+ * issued.
+ *
+ * The stack, at that point, looks like :
+ *
+ * 0(esp) EIP
+ * 4(esp) CS
+ * 8(esp) EFLAGS
+ *
+ * Upon execution :
+ * Copy the stack eflags to top of stack
+ * Pop eflags into the eflags register
+ * Far return: pop EIP and CS into their register, and additionally pop EFLAGS.
+ */
+#define NATIVE_INTERRUPT_RETURN_NMI_SAFE	pushl 8(%esp);	\
+						popfl;		\
+						lret $4
+#endif
+
 #ifdef CONFIG_PARAVIRT
 #include <asm/paravirt.h>
 #else
@@ -109,6 +164,7 @@ static inline unsigned long __raw_local_irq_save(void)
 
 #define ENABLE_INTERRUPTS(x)	sti
 #define DISABLE_INTERRUPTS(x)	cli
+#define INTERRUPT_RETURN_NMI_SAFE	NATIVE_INTERRUPT_RETURN_NMI_SAFE
 
 #ifdef CONFIG_X86_64
 #define INTERRUPT_RETURN	iretq
diff --git a/include/asm-x86/paravirt.h b/include/asm-x86/paravirt.h
index 0f13b94..d5087e0 100644
--- a/include/asm-x86/paravirt.h
+++ b/include/asm-x86/paravirt.h
@@ -141,9 +141,10 @@ struct pv_cpu_ops {
 	u64 (*read_pmc)(int counter);
 	unsigned long long (*read_tscp)(unsigned int *aux);
 
-	/* These two are jmp to, not actually called. */
+	/* These three are jmp to, not actually called. */
 	void (*irq_enable_syscall_ret)(void);
 	void (*iret)(void);
+	void (*nmi_return)(void);
 
 	void (*swapgs)(void);
 
@@ -1385,6 +1386,10 @@ static inline unsigned long __raw_local_irq_save(void)
 	PARA_SITE(PARA_PATCH(pv_cpu_ops, PV_CPU_iret), CLBR_NONE,	\
 		  jmp *%cs:pv_cpu_ops+PV_CPU_iret)
 
+#define INTERRUPT_RETURN_NMI_SAFE					\
+	PARA_SITE(PARA_PATCH(pv_cpu_ops, PV_CPU_nmi_return), CLBR_NONE,	\
+		  jmp *%cs:pv_cpu_ops+PV_CPU_nmi_return)
+
 #define DISABLE_INTERRUPTS(clobbers)					\
 	PARA_SITE(PARA_PATCH(pv_irq_ops, PV_IRQ_irq_disable), clobbers, \
 		  PV_SAVE_REGS;			\
diff --git a/include/linux/hardirq.h b/include/linux/hardirq.h
index 181006c..b39f49d 100644
--- a/include/linux/hardirq.h
+++ b/include/linux/hardirq.h
@@ -22,10 +22,13 @@
  * PREEMPT_MASK: 0x000000ff
  * SOFTIRQ_MASK: 0x0000ff00
  * HARDIRQ_MASK: 0x0fff0000
+ * HARDNMI_MASK: 0x40000000
  */
 #define PREEMPT_BITS	8
 #define SOFTIRQ_BITS	8
 
+#define HARDNMI_BITS	1
+
 #ifndef HARDIRQ_BITS
 #define HARDIRQ_BITS	12
 
@@ -45,16 +48,19 @@
 #define PREEMPT_SHIFT	0
 #define SOFTIRQ_SHIFT	(PREEMPT_SHIFT + PREEMPT_BITS)
 #define HARDIRQ_SHIFT	(SOFTIRQ_SHIFT + SOFTIRQ_BITS)
+#define HARDNMI_SHIFT	(30)
 
 #define __IRQ_MASK(x)	((1UL << (x))-1)
 
 #define PREEMPT_MASK	(__IRQ_MASK(PREEMPT_BITS) << PREEMPT_SHIFT)
 #define SOFTIRQ_MASK	(__IRQ_MASK(SOFTIRQ_BITS) << SOFTIRQ_SHIFT)
 #define HARDIRQ_MASK	(__IRQ_MASK(HARDIRQ_BITS) << HARDIRQ_SHIFT)
+#define HARDNMI_MASK	(__IRQ_MASK(HARDNMI_BITS) << HARDNMI_SHIFT)
 
 #define PREEMPT_OFFSET	(1UL << PREEMPT_SHIFT)
 #define SOFTIRQ_OFFSET	(1UL << SOFTIRQ_SHIFT)
 #define HARDIRQ_OFFSET	(1UL << HARDIRQ_SHIFT)
+#define HARDNMI_OFFSET	(1UL << HARDNMI_SHIFT)
 
 #if PREEMPT_ACTIVE < (1 << (HARDIRQ_SHIFT + HARDIRQ_BITS))
 #error PREEMPT_ACTIVE is too low!
@@ -62,7 +68,9 @@
 
 #define hardirq_count()	(preempt_count() & HARDIRQ_MASK)
 #define softirq_count()	(preempt_count() & SOFTIRQ_MASK)
-#define irq_count()	(preempt_count() & (HARDIRQ_MASK | SOFTIRQ_MASK))
+#define irq_count() \
+	(preempt_count() & (HARDNMI_MASK | HARDIRQ_MASK | SOFTIRQ_MASK))
+#define hardnmi_count()	(preempt_count() & HARDNMI_MASK)
 
 /*
  * Are we doing bottom half or hardware interrupt processing?
@@ -71,6 +79,7 @@
 #define in_irq()		(hardirq_count())
 #define in_softirq()		(softirq_count())
 #define in_interrupt()		(irq_count())
+#define in_nmi()		(hardnmi_count())
 
 #if defined(CONFIG_PREEMPT)
 # define PREEMPT_INATOMIC_BASE kernel_locked()
@@ -161,7 +170,19 @@ extern void irq_enter(void);
  */
 extern void irq_exit(void);
 
-#define nmi_enter()		do { lockdep_off(); __irq_enter(); } while (0)
-#define nmi_exit()		do { __irq_exit(); lockdep_on(); } while (0)
+#define nmi_enter()					\
+	do {						\
+		lockdep_off();				\
+		BUG_ON(hardnmi_count());		\
+		add_preempt_count(HARDNMI_OFFSET);	\
+		__irq_enter();				\
+	} while (0)
+
+#define nmi_exit()					\
+	do {						\
+		__irq_exit();				\
+		sub_preempt_count(HARDNMI_OFFSET);	\
+		lockdep_on();				\
+	} while (0)
 
 #endif /* LINUX_HARDIRQ_H */


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Powered by blists - more mailing lists