4 files changed, 161 insertions, 54 deletions

diff --git a/arch/x86_64/Kconfig.debug b/arch/x86_64/Kconfig.debug
index 1d92ab56c0f9..775d211a5cf9 100644
--- a/arch/x86_64/Kconfig.debug
+++ b/arch/x86_64/Kconfig.debug
@@ -1,5 +1,9 @@
 menu "Kernel hacking"
 
+config TRACE_IRQFLAGS_SUPPORT
+        bool
+        default y
+
 source "lib/Kconfig.debug"
 
 config DEBUG_RODATA
diff --git a/arch/x86_64/ia32/ia32entry.S b/arch/x86_64/ia32/ia32entry.S
index c536fa98ea37..9b5bb413a6e9 100644
--- a/arch/x86_64/ia32/ia32entry.S
+++ b/arch/x86_64/ia32/ia32entry.S
@@ -13,6 +13,7 @@
 #include <asm/thread_info.h>    
 #include <asm/segment.h>
 #include <asm/vsyscall32.h>
+#include <asm/irqflags.h>
 #include <linux/linkage.h>
 
 #define IA32_NR_syscalls ((ia32_syscall_end - ia32_sys_call_table)/8)
@@ -75,6 +76,10 @@ ENTRY(ia32_sysenter_target)
         swapgs
         movq    %gs:pda_kernelstack, %rsp
         addq    $(PDA_STACKOFFSET),%rsp 
+        /*
+         * No need to follow this irqs on/off section: the syscall
+         * disabled irqs, here we enable it straight after entry:
+         */
         sti     
         movl    %ebp,%ebp               /* zero extension */
         pushq   $__USER32_DS
@@ -118,6 +123,7 @@ sysenter_do_call:
         movq    %rax,RAX-ARGOFFSET(%rsp)
         GET_THREAD_INFO(%r10)
         cli
+        TRACE_IRQS_OFF
         testl   $_TIF_ALLWORK_MASK,threadinfo_flags(%r10)
         jnz     int_ret_from_sys_call
         andl    $~TS_COMPAT,threadinfo_status(%r10)
@@ -132,6 +138,7 @@ sysenter_do_call:
         CFI_REGISTER rsp,rcx
         movl    $VSYSCALL32_SYSEXIT,%edx        /* User %eip */
         CFI_REGISTER rip,rdx
+        TRACE_IRQS_ON
         swapgs
         sti             /* sti only takes effect after the next instruction */
         /* sysexit */
@@ -186,6 +193,10 @@ ENTRY(ia32_cstar_target)
         movl    %esp,%r8d
         CFI_REGISTER    rsp,r8
         movq    %gs:pda_kernelstack,%rsp
+        /*
+         * No need to follow this irqs on/off section: the syscall
+         * disabled irqs and here we enable it straight after entry:
+         */
         sti
         SAVE_ARGS 8,1,1
         movl    %eax,%eax       /* zero extension */
@@ -220,6 +231,7 @@ cstar_do_call:
         movq %rax,RAX-ARGOFFSET(%rsp)
         GET_THREAD_INFO(%r10)
         cli
+        TRACE_IRQS_OFF
         testl $_TIF_ALLWORK_MASK,threadinfo_flags(%r10)
         jnz  int_ret_from_sys_call
         andl $~TS_COMPAT,threadinfo_status(%r10)
@@ -228,6 +240,7 @@ cstar_do_call:
         CFI_REGISTER rip,rcx
         movl EFLAGS-ARGOFFSET(%rsp),%r11d       
         /*CFI_REGISTER rflags,r11*/
+        TRACE_IRQS_ON
         movl RSP-ARGOFFSET(%rsp),%esp
         CFI_RESTORE rsp
         swapgs
@@ -286,7 +299,11 @@ ENTRY(ia32_syscall)
         /*CFI_REL_OFFSET        rflags,EFLAGS-RIP*/
         /*CFI_REL_OFFSET        cs,CS-RIP*/
         CFI_REL_OFFSET  rip,RIP-RIP
-        swapgs  
+        swapgs
+        /*
+         * No need to follow this irqs on/off section: the syscall
+         * disabled irqs and here we enable it straight after entry:
+         */
         sti
         movl %eax,%eax
         pushq %rax
diff --git a/arch/x86_64/kernel/entry.S b/arch/x86_64/kernel/entry.S
index ed92c2983606..d464dded68c0 100644
--- a/arch/x86_64/kernel/entry.S
+++ b/arch/x86_64/kernel/entry.S
@@ -41,13 +41,24 @@
 #include <asm/thread_info.h>
 #include <asm/hw_irq.h>
 #include <asm/page.h>
+#include <asm/irqflags.h>
 
         .code64
 
 #ifndef CONFIG_PREEMPT
 #define retint_kernel retint_restore_args
 #endif  
-        
+
+
+.macro TRACE_IRQS_IRETQ offset=ARGOFFSET
+#ifdef CONFIG_TRACE_IRQFLAGS
+        bt   $9,EFLAGS-\offset(%rsp)    /* interrupts off? */
+        jnc  1f
+        TRACE_IRQS_ON
+1:
+#endif
+.endm
+
 /*
  * C code is not supposed to know about undefined top of stack. Every time 
  * a C function with an pt_regs argument is called from the SYSCALL based 
@@ -194,6 +205,10 @@ ENTRY(system_call)
         swapgs
         movq    %rsp,%gs:pda_oldrsp 
         movq    %gs:pda_kernelstack,%rsp
+        /*
+         * No need to follow this irqs off/on section - it's straight
+         * and short:
+         */
         sti                                     
         SAVE_ARGS 8,1
         movq  %rax,ORIG_RAX-ARGOFFSET(%rsp) 
@@ -219,10 +234,15 @@ ret_from_sys_call:
 sysret_check:           
         GET_THREAD_INFO(%rcx)
         cli
+        TRACE_IRQS_OFF
         movl threadinfo_flags(%rcx),%edx
         andl %edi,%edx
         CFI_REMEMBER_STATE
         jnz  sysret_careful 
+        /*
+         * sysretq will re-enable interrupts:
+         */
+        TRACE_IRQS_ON
         movq RIP-ARGOFFSET(%rsp),%rcx
         CFI_REGISTER    rip,rcx
         RESTORE_ARGS 0,-ARG_SKIP,1
@@ -237,6 +257,7 @@ sysret_careful:
         CFI_RESTORE_STATE
         bt $TIF_NEED_RESCHED,%edx
         jnc sysret_signal
+        TRACE_IRQS_ON
         sti
         pushq %rdi
         CFI_ADJUST_CFA_OFFSET 8
@@ -247,6 +268,7 @@ sysret_careful:
 
         /* Handle a signal */ 
 sysret_signal:
+        TRACE_IRQS_ON
         sti
         testl $(_TIF_SIGPENDING|_TIF_NOTIFY_RESUME|_TIF_SINGLESTEP),%edx
         jz    1f
@@ -261,6 +283,7 @@ sysret_signal:
         /* Use IRET because user could have changed frame. This
            works because ptregscall_common has called FIXUP_TOP_OF_STACK. */
         cli
+        TRACE_IRQS_OFF
         jmp int_with_check
         
 badsys:
@@ -309,6 +332,7 @@ ENTRY(int_ret_from_sys_call)
         CFI_REL_OFFSET  r10,R10-ARGOFFSET
         CFI_REL_OFFSET  r11,R11-ARGOFFSET
         cli
+        TRACE_IRQS_OFF
         testl $3,CS-ARGOFFSET(%rsp)
         je retint_restore_args
         movl $_TIF_ALLWORK_MASK,%edi
@@ -327,6 +351,7 @@ int_with_check:
 int_careful:
         bt $TIF_NEED_RESCHED,%edx
         jnc  int_very_careful
+        TRACE_IRQS_ON
         sti
         pushq %rdi
         CFI_ADJUST_CFA_OFFSET 8
@@ -334,10 +359,12 @@ int_careful:
         popq %rdi
         CFI_ADJUST_CFA_OFFSET -8
         cli
+        TRACE_IRQS_OFF
         jmp int_with_check
 
         /* handle signals and tracing -- both require a full stack frame */
 int_very_careful:
+        TRACE_IRQS_ON
         sti
         SAVE_REST
         /* Check for syscall exit trace */      
@@ -351,6 +378,7 @@ int_very_careful:
         CFI_ADJUST_CFA_OFFSET -8
         andl $~(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT|_TIF_SINGLESTEP),%edi
         cli
+        TRACE_IRQS_OFF
         jmp int_restore_rest
         
 int_signal:
@@ -363,6 +391,7 @@ int_signal:
 int_restore_rest:
         RESTORE_REST
         cli
+        TRACE_IRQS_OFF
         jmp int_with_check
         CFI_ENDPROC
 END(int_ret_from_sys_call)
@@ -484,6 +513,10 @@ END(stub_rt_sigreturn)
         swapgs  
 1:      incl    %gs:pda_irqcount        # RED-PEN should check preempt count
         cmoveq %gs:pda_irqstackptr,%rsp
+        /*
+         * We entered an interrupt context - irqs are off:
+         */
+        TRACE_IRQS_OFF
         call \func
         .endm
 
@@ -493,6 +526,7 @@ ENTRY(common_interrupt)
         /* 0(%rsp): oldrsp-ARGOFFSET */
 ret_from_intr:
         cli     
+        TRACE_IRQS_OFF
         decl %gs:pda_irqcount
         leaveq
         CFI_DEF_CFA_REGISTER    rsp
@@ -515,9 +549,21 @@ retint_check:
         CFI_REMEMBER_STATE
         jnz  retint_careful
 retint_swapgs:          
+        /*
+         * The iretq could re-enable interrupts:
+         */
+        cli
+        TRACE_IRQS_IRETQ
         swapgs 
+        jmp restore_args
+
 retint_restore_args:                            
         cli
+        /*
+         * The iretq could re-enable interrupts:
+         */
+        TRACE_IRQS_IRETQ
+restore_args:
         RESTORE_ARGS 0,8,0                                              
 iret_label:     
         iretq
@@ -530,6 +576,7 @@ iret_label:
         /* running with kernel gs */
 bad_iret:
         movq $11,%rdi   /* SIGSEGV */
+        TRACE_IRQS_ON
         sti
         jmp do_exit                     
         .previous       
@@ -539,6 +586,7 @@ retint_careful:
         CFI_RESTORE_STATE
         bt    $TIF_NEED_RESCHED,%edx
         jnc   retint_signal
+        TRACE_IRQS_ON
         sti
         pushq %rdi
         CFI_ADJUST_CFA_OFFSET   8
@@ -547,11 +595,13 @@ retint_careful:
         CFI_ADJUST_CFA_OFFSET   -8
         GET_THREAD_INFO(%rcx)
         cli
+        TRACE_IRQS_OFF
         jmp retint_check
         
 retint_signal:
         testl $(_TIF_SIGPENDING|_TIF_NOTIFY_RESUME|_TIF_SINGLESTEP),%edx
         jz    retint_swapgs
+        TRACE_IRQS_ON
         sti
         SAVE_REST
         movq $-1,ORIG_RAX(%rsp)                         
@@ -560,6 +610,7 @@ retint_signal:
         call do_notify_resume
         RESTORE_REST
         cli
+        TRACE_IRQS_OFF
         movl $_TIF_NEED_RESCHED,%edi
         GET_THREAD_INFO(%rcx)
         jmp retint_check
@@ -666,7 +717,7 @@ END(spurious_interrupt)
 
         /* error code is on the stack already */
         /* handle NMI like exceptions that can happen everywhere */
-        .macro paranoidentry sym, ist=0
+        .macro paranoidentry sym, ist=0, irqtrace=1
         SAVE_ALL
         cld
         movl $1,%ebx
@@ -691,8 +742,73 @@ END(spurious_interrupt)
         addq    $EXCEPTION_STKSZ, per_cpu__init_tss + TSS_ist + (\ist - 1) * 8(%rbp)
         .endif
         cli
+        .if \irqtrace
+        TRACE_IRQS_OFF
+        .endif
         .endm
-        
+
+        /*
+         * "Paranoid" exit path from exception stack.
+         * Paranoid because this is used by NMIs and cannot take
+         * any kernel state for granted.
+         * We don't do kernel preemption checks here, because only
+         * NMI should be common and it does not enable IRQs and
+         * cannot get reschedule ticks.
+         *
+         * "trace" is 0 for the NMI handler only, because irq-tracing
+         * is fundamentally NMI-unsafe. (we cannot change the soft and
+         * hard flags at once, atomically)
+         */
+        .macro paranoidexit trace=1
+        /* ebx: no swapgs flag */
+paranoid_exit\trace:
+        testl %ebx,%ebx                         /* swapgs needed? */
+        jnz paranoid_restore\trace
+        testl $3,CS(%rsp)
+        jnz   paranoid_userspace\trace
+paranoid_swapgs\trace:
+        TRACE_IRQS_IRETQ 0
+        swapgs
+paranoid_restore\trace:
+        RESTORE_ALL 8
+        iretq
+paranoid_userspace\trace:
+        GET_THREAD_INFO(%rcx)
+        movl threadinfo_flags(%rcx),%ebx
+        andl $_TIF_WORK_MASK,%ebx
+        jz paranoid_swapgs\trace
+        movq %rsp,%rdi                  /* &pt_regs */
+        call sync_regs
+        movq %rax,%rsp                  /* switch stack for scheduling */
+        testl $_TIF_NEED_RESCHED,%ebx
+        jnz paranoid_schedule\trace
+        movl %ebx,%edx                  /* arg3: thread flags */
+        .if \trace
+        TRACE_IRQS_ON
+        .endif
+        sti
+        xorl %esi,%esi                  /* arg2: oldset */
+        movq %rsp,%rdi                  /* arg1: &pt_regs */
+        call do_notify_resume
+        cli
+        .if \trace
+        TRACE_IRQS_OFF
+        .endif
+        jmp paranoid_userspace\trace
+paranoid_schedule\trace:
+        .if \trace
+        TRACE_IRQS_ON
+        .endif
+        sti
+        call schedule
+        cli
+        .if \trace
+        TRACE_IRQS_OFF
+        .endif
+        jmp paranoid_userspace\trace
+        CFI_ENDPROC
+        .endm
+
 /*
  * Exception entry point. This expects an error code/orig_rax on the stack
  * and the exception handler in %rax.   
@@ -748,6 +864,7 @@ error_exit:
         movl %ebx,%eax          
         RESTORE_REST
         cli
+        TRACE_IRQS_OFF
         GET_THREAD_INFO(%rcx)   
         testl %eax,%eax
         jne  retint_kernel
@@ -755,6 +872,10 @@ error_exit:
         movl  $_TIF_WORK_MASK,%edi
         andl  %edi,%edx
         jnz  retint_careful
+        /*
+         * The iret might restore flags:
+         */
+        TRACE_IRQS_IRETQ
         swapgs 
         RESTORE_ARGS 0,8,0                                              
         jmp iret_label
@@ -916,8 +1037,7 @@ KPROBE_ENTRY(debug)
         pushq $0
         CFI_ADJUST_CFA_OFFSET 8         
         paranoidentry do_debug, DEBUG_STACK
-        jmp paranoid_exit
-        CFI_ENDPROC
+        paranoidexit
 END(debug)
         .previous .text
 
@@ -926,49 +1046,13 @@ KPROBE_ENTRY(nmi)
         INTR_FRAME
         pushq $-1
         CFI_ADJUST_CFA_OFFSET 8
-        paranoidentry do_nmi
-        /*
-         * "Paranoid" exit path from exception stack.
-         * Paranoid because this is used by NMIs and cannot take
-         * any kernel state for granted.
-         * We don't do kernel preemption checks here, because only
-         * NMI should be common and it does not enable IRQs and
-         * cannot get reschedule ticks.
-         */
-        /* ebx: no swapgs flag */
-paranoid_exit:
-        testl %ebx,%ebx                         /* swapgs needed? */
-        jnz paranoid_restore
-        testl $3,CS(%rsp)
-        jnz   paranoid_userspace
-paranoid_swapgs:        
-        swapgs
-paranoid_restore:       
-        RESTORE_ALL 8
-        iretq
-paranoid_userspace:     
-        GET_THREAD_INFO(%rcx)
-        movl threadinfo_flags(%rcx),%ebx
-        andl $_TIF_WORK_MASK,%ebx
-        jz paranoid_swapgs
-        movq %rsp,%rdi                  /* &pt_regs */
-        call sync_regs
-        movq %rax,%rsp                  /* switch stack for scheduling */
-        testl $_TIF_NEED_RESCHED,%ebx
-        jnz paranoid_schedule
-        movl %ebx,%edx                  /* arg3: thread flags */
-        sti
-        xorl %esi,%esi                  /* arg2: oldset */
-        movq %rsp,%rdi                  /* arg1: &pt_regs */
-        call do_notify_resume
-        cli
-        jmp paranoid_userspace
-paranoid_schedule:
-        sti
-        call schedule
-        cli
-        jmp paranoid_userspace
-        CFI_ENDPROC
+        paranoidentry do_nmi, 0, 0
+#ifdef CONFIG_TRACE_IRQFLAGS
+        paranoidexit 0
+#else
+        jmp paranoid_exit1
+        CFI_ENDPROC
+#endif
 END(nmi)
         .previous .text
 
@@ -977,7 +1061,7 @@ KPROBE_ENTRY(int3)
         pushq $0
         CFI_ADJUST_CFA_OFFSET 8
         paranoidentry do_int3, DEBUG_STACK
-        jmp paranoid_exit
+        jmp paranoid_exit1
         CFI_ENDPROC
 END(int3)
         .previous .text
@@ -1006,7 +1090,7 @@ END(reserved)
 ENTRY(double_fault)
         XCPT_FRAME
         paranoidentry do_double_fault
-        jmp paranoid_exit
+        jmp paranoid_exit1
         CFI_ENDPROC
 END(double_fault)
 
@@ -1022,7 +1106,7 @@ END(segment_not_present)
 ENTRY(stack_segment)
         XCPT_FRAME
         paranoidentry do_stack_segment
-        jmp paranoid_exit
+        jmp paranoid_exit1
         CFI_ENDPROC
 END(stack_segment)
 
@@ -1050,7 +1134,7 @@ ENTRY(machine_check)
         pushq $0
         CFI_ADJUST_CFA_OFFSET 8 
         paranoidentry do_machine_check
-        jmp paranoid_exit
+        jmp paranoid_exit1
         CFI_ENDPROC
 END(machine_check)
 #endif
diff --git a/arch/x86_64/kernel/irq.c b/arch/x86_64/kernel/irq.c
index a1f1df5f7bfc..5221a53e90c1 100644
--- a/arch/x86_64/kernel/irq.c
+++ b/arch/x86_64/kernel/irq.c
@@ -177,8 +177,10 @@ asmlinkage void do_softirq(void)
         local_irq_save(flags);
         pending = local_softirq_pending();
         /* Switch to interrupt stack */
-        if (pending)
+        if (pending) {
                 call_softirq();
+                WARN_ON_ONCE(softirq_count());
+        }
         local_irq_restore(flags);
 }
 EXPORT_SYMBOL(do_softirq);