kernel_optimize_test/arch/sparc/kernel/rtrap_64.S
David S. Miller 28a1f533ae sparc64: Fix hardirq tracing in trap return path.
We can overflow the hardirq stack if we set the %pil here
so early, just let the normal control flow do it.

This is fine as we are allowed to do the actual IRQ enable
at any point after we call trace_hardirqs_on.

Signed-off-by: David S. Miller <davem@davemloft.net>
2010-04-20 00:48:37 -07:00

415 lines
10 KiB
ArmAsm

/*
* rtrap.S: Preparing for return from trap on Sparc V9.
*
* Copyright (C) 1997,1998 Jakub Jelinek (jj@sunsite.mff.cuni.cz)
* Copyright (C) 1997 David S. Miller (davem@caip.rutgers.edu)
*/
#include <asm/asi.h>
#include <asm/pstate.h>
#include <asm/ptrace.h>
#include <asm/spitfire.h>
#include <asm/head.h>
#include <asm/visasm.h>
#include <asm/processor.h>
#define RTRAP_PSTATE (PSTATE_TSO|PSTATE_PEF|PSTATE_PRIV|PSTATE_IE)
#define RTRAP_PSTATE_IRQOFF (PSTATE_TSO|PSTATE_PEF|PSTATE_PRIV)
#define RTRAP_PSTATE_AG_IRQOFF (PSTATE_TSO|PSTATE_PEF|PSTATE_PRIV|PSTATE_AG)
.text
.align 32
__handle_softirq:
call do_softirq
nop
ba,a,pt %xcc, __handle_softirq_continue
nop
__handle_preemption:
call schedule
wrpr %g0, RTRAP_PSTATE, %pstate
ba,pt %xcc, __handle_preemption_continue
wrpr %g0, RTRAP_PSTATE_IRQOFF, %pstate
__handle_user_windows:
call fault_in_user_windows
wrpr %g0, RTRAP_PSTATE, %pstate
wrpr %g0, RTRAP_PSTATE_IRQOFF, %pstate
/* Redo sched+sig checks */
ldx [%g6 + TI_FLAGS], %l0
andcc %l0, _TIF_NEED_RESCHED, %g0
be,pt %xcc, 1f
nop
call schedule
wrpr %g0, RTRAP_PSTATE, %pstate
wrpr %g0, RTRAP_PSTATE_IRQOFF, %pstate
ldx [%g6 + TI_FLAGS], %l0
1: andcc %l0, _TIF_DO_NOTIFY_RESUME_MASK, %g0
be,pt %xcc, __handle_user_windows_continue
nop
mov %l5, %o1
add %sp, PTREGS_OFF, %o0
mov %l0, %o2
call do_notify_resume
wrpr %g0, RTRAP_PSTATE, %pstate
wrpr %g0, RTRAP_PSTATE_IRQOFF, %pstate
/* Signal delivery can modify pt_regs tstate, so we must
* reload it.
*/
ldx [%sp + PTREGS_OFF + PT_V9_TSTATE], %l1
sethi %hi(0xf << 20), %l4
and %l1, %l4, %l4
ba,pt %xcc, __handle_user_windows_continue
andn %l1, %l4, %l1
__handle_userfpu:
rd %fprs, %l5
andcc %l5, FPRS_FEF, %g0
sethi %hi(TSTATE_PEF), %o0
be,a,pn %icc, __handle_userfpu_continue
andn %l1, %o0, %l1
ba,a,pt %xcc, __handle_userfpu_continue
__handle_signal:
mov %l5, %o1
add %sp, PTREGS_OFF, %o0
mov %l0, %o2
call do_notify_resume
wrpr %g0, RTRAP_PSTATE, %pstate
wrpr %g0, RTRAP_PSTATE_IRQOFF, %pstate
/* Signal delivery can modify pt_regs tstate, so we must
* reload it.
*/
ldx [%sp + PTREGS_OFF + PT_V9_TSTATE], %l1
sethi %hi(0xf << 20), %l4
and %l1, %l4, %l4
ba,pt %xcc, __handle_signal_continue
andn %l1, %l4, %l1
/* When returning from a NMI (%pil==15) interrupt we want to
* avoid running softirqs, doing IRQ tracing, preempting, etc.
*/
.globl rtrap_nmi
rtrap_nmi: ldx [%sp + PTREGS_OFF + PT_V9_TSTATE], %l1
sethi %hi(0xf << 20), %l4
and %l1, %l4, %l4
andn %l1, %l4, %l1
srl %l4, 20, %l4
ba,pt %xcc, rtrap_no_irq_enable
wrpr %l4, %pil
.align 64
.globl rtrap_irq, rtrap, irqsz_patchme, rtrap_xcall
rtrap_irq:
rtrap:
#ifndef CONFIG_SMP
sethi %hi(__cpu_data), %l0
lduw [%l0 + %lo(__cpu_data)], %l1
#else
sethi %hi(__cpu_data), %l0
or %l0, %lo(__cpu_data), %l0
lduw [%l0 + %g5], %l1
#endif
cmp %l1, 0
/* mm/ultra.S:xcall_report_regs KNOWS about this load. */
bne,pn %icc, __handle_softirq
ldx [%sp + PTREGS_OFF + PT_V9_TSTATE], %l1
__handle_softirq_continue:
rtrap_xcall:
sethi %hi(0xf << 20), %l4
and %l1, %l4, %l4
andn %l1, %l4, %l1
srl %l4, 20, %l4
#ifdef CONFIG_TRACE_IRQFLAGS
brnz,pn %l4, rtrap_no_irq_enable
nop
call trace_hardirqs_on
nop
/* Do not actually set the %pil here. We will do that
* below after we clear PSTATE_IE in the %pstate register.
* If we re-enable interrupts here, we can recurse down
* the hardirq stack potentially endlessly, causing a
* stack overflow.
*
* It is tempting to put this test and trace_hardirqs_on
* call at the 'rt_continue' label, but that will not work
* as that path hits unconditionally and we do not want to
* execute this in NMI return paths, for example.
*/
#endif
rtrap_no_irq_enable:
andcc %l1, TSTATE_PRIV, %l3
bne,pn %icc, to_kernel
nop
/* We must hold IRQs off and atomically test schedule+signal
* state, then hold them off all the way back to userspace.
* If we are returning to kernel, none of this matters. Note
* that we are disabling interrupts via PSTATE_IE, not using
* %pil.
*
* If we do not do this, there is a window where we would do
* the tests, later the signal/resched event arrives but we do
* not process it since we are still in kernel mode. It would
* take until the next local IRQ before the signal/resched
* event would be handled.
*
* This also means that if we have to deal with user
* windows, we have to redo all of these sched+signal checks
* with IRQs disabled.
*/
to_user: wrpr %g0, RTRAP_PSTATE_IRQOFF, %pstate
wrpr 0, %pil
__handle_preemption_continue:
ldx [%g6 + TI_FLAGS], %l0
sethi %hi(_TIF_USER_WORK_MASK), %o0
or %o0, %lo(_TIF_USER_WORK_MASK), %o0
andcc %l0, %o0, %g0
sethi %hi(TSTATE_PEF), %o0
be,pt %xcc, user_nowork
andcc %l1, %o0, %g0
andcc %l0, _TIF_NEED_RESCHED, %g0
bne,pn %xcc, __handle_preemption
andcc %l0, _TIF_DO_NOTIFY_RESUME_MASK, %g0
bne,pn %xcc, __handle_signal
__handle_signal_continue:
ldub [%g6 + TI_WSAVED], %o2
brnz,pn %o2, __handle_user_windows
nop
__handle_user_windows_continue:
sethi %hi(TSTATE_PEF), %o0
andcc %l1, %o0, %g0
/* This fpdepth clear is necessary for non-syscall rtraps only */
user_nowork:
bne,pn %xcc, __handle_userfpu
stb %g0, [%g6 + TI_FPDEPTH]
__handle_userfpu_continue:
rt_continue: ldx [%sp + PTREGS_OFF + PT_V9_G1], %g1
ldx [%sp + PTREGS_OFF + PT_V9_G2], %g2
ldx [%sp + PTREGS_OFF + PT_V9_G3], %g3
ldx [%sp + PTREGS_OFF + PT_V9_G4], %g4
ldx [%sp + PTREGS_OFF + PT_V9_G5], %g5
brz,pt %l3, 1f
mov %g6, %l2
/* Must do this before thread reg is clobbered below. */
LOAD_PER_CPU_BASE(%g5, %g6, %i0, %i1, %i2)
1:
ldx [%sp + PTREGS_OFF + PT_V9_G6], %g6
ldx [%sp + PTREGS_OFF + PT_V9_G7], %g7
/* Normal globals are restored, go to trap globals. */
661: wrpr %g0, RTRAP_PSTATE_AG_IRQOFF, %pstate
nop
.section .sun4v_2insn_patch, "ax"
.word 661b
wrpr %g0, RTRAP_PSTATE_IRQOFF, %pstate
SET_GL(1)
.previous
mov %l2, %g6
ldx [%sp + PTREGS_OFF + PT_V9_I0], %i0
ldx [%sp + PTREGS_OFF + PT_V9_I1], %i1
ldx [%sp + PTREGS_OFF + PT_V9_I2], %i2
ldx [%sp + PTREGS_OFF + PT_V9_I3], %i3
ldx [%sp + PTREGS_OFF + PT_V9_I4], %i4
ldx [%sp + PTREGS_OFF + PT_V9_I5], %i5
ldx [%sp + PTREGS_OFF + PT_V9_I6], %i6
ldx [%sp + PTREGS_OFF + PT_V9_I7], %i7
ldx [%sp + PTREGS_OFF + PT_V9_TPC], %l2
ldx [%sp + PTREGS_OFF + PT_V9_TNPC], %o2
ld [%sp + PTREGS_OFF + PT_V9_Y], %o3
wr %o3, %g0, %y
wrpr %l4, 0x0, %pil
wrpr %g0, 0x1, %tl
andn %l1, TSTATE_SYSCALL, %l1
wrpr %l1, %g0, %tstate
wrpr %l2, %g0, %tpc
wrpr %o2, %g0, %tnpc
brnz,pn %l3, kern_rtt
mov PRIMARY_CONTEXT, %l7
661: ldxa [%l7 + %l7] ASI_DMMU, %l0
.section .sun4v_1insn_patch, "ax"
.word 661b
ldxa [%l7 + %l7] ASI_MMU, %l0
.previous
sethi %hi(sparc64_kern_pri_nuc_bits), %l1
ldx [%l1 + %lo(sparc64_kern_pri_nuc_bits)], %l1
or %l0, %l1, %l0
661: stxa %l0, [%l7] ASI_DMMU
.section .sun4v_1insn_patch, "ax"
.word 661b
stxa %l0, [%l7] ASI_MMU
.previous
sethi %hi(KERNBASE), %l7
flush %l7
rdpr %wstate, %l1
rdpr %otherwin, %l2
srl %l1, 3, %l1
wrpr %l2, %g0, %canrestore
wrpr %l1, %g0, %wstate
brnz,pt %l2, user_rtt_restore
wrpr %g0, %g0, %otherwin
ldx [%g6 + TI_FLAGS], %g3
wr %g0, ASI_AIUP, %asi
rdpr %cwp, %g1
andcc %g3, _TIF_32BIT, %g0
sub %g1, 1, %g1
bne,pt %xcc, user_rtt_fill_32bit
wrpr %g1, %cwp
ba,a,pt %xcc, user_rtt_fill_64bit
user_rtt_fill_fixup:
rdpr %cwp, %g1
add %g1, 1, %g1
wrpr %g1, 0x0, %cwp
rdpr %wstate, %g2
sll %g2, 3, %g2
wrpr %g2, 0x0, %wstate
/* We know %canrestore and %otherwin are both zero. */
sethi %hi(sparc64_kern_pri_context), %g2
ldx [%g2 + %lo(sparc64_kern_pri_context)], %g2
mov PRIMARY_CONTEXT, %g1
661: stxa %g2, [%g1] ASI_DMMU
.section .sun4v_1insn_patch, "ax"
.word 661b
stxa %g2, [%g1] ASI_MMU
.previous
sethi %hi(KERNBASE), %g1
flush %g1
or %g4, FAULT_CODE_WINFIXUP, %g4
stb %g4, [%g6 + TI_FAULT_CODE]
stx %g5, [%g6 + TI_FAULT_ADDR]
mov %g6, %l1
wrpr %g0, 0x0, %tl
661: nop
.section .sun4v_1insn_patch, "ax"
.word 661b
SET_GL(0)
.previous
wrpr %g0, RTRAP_PSTATE, %pstate
mov %l1, %g6
ldx [%g6 + TI_TASK], %g4
LOAD_PER_CPU_BASE(%g5, %g6, %g1, %g2, %g3)
call do_sparc64_fault
add %sp, PTREGS_OFF, %o0
ba,pt %xcc, rtrap
nop
user_rtt_pre_restore:
add %g1, 1, %g1
wrpr %g1, 0x0, %cwp
user_rtt_restore:
restore
rdpr %canrestore, %g1
wrpr %g1, 0x0, %cleanwin
retry
nop
kern_rtt: rdpr %canrestore, %g1
brz,pn %g1, kern_rtt_fill
nop
kern_rtt_restore:
stw %g0, [%sp + PTREGS_OFF + PT_V9_MAGIC]
restore
retry
to_kernel:
#ifdef CONFIG_PREEMPT
ldsw [%g6 + TI_PRE_COUNT], %l5
brnz %l5, kern_fpucheck
ldx [%g6 + TI_FLAGS], %l5
andcc %l5, _TIF_NEED_RESCHED, %g0
be,pt %xcc, kern_fpucheck
nop
cmp %l4, 0
bne,pn %xcc, kern_fpucheck
sethi %hi(PREEMPT_ACTIVE), %l6
stw %l6, [%g6 + TI_PRE_COUNT]
call schedule
nop
ba,pt %xcc, rtrap
stw %g0, [%g6 + TI_PRE_COUNT]
#endif
kern_fpucheck: ldub [%g6 + TI_FPDEPTH], %l5
brz,pt %l5, rt_continue
srl %l5, 1, %o0
add %g6, TI_FPSAVED, %l6
ldub [%l6 + %o0], %l2
sub %l5, 2, %l5
add %g6, TI_GSR, %o1
andcc %l2, (FPRS_FEF|FPRS_DU), %g0
be,pt %icc, 2f
and %l2, FPRS_DL, %l6
andcc %l2, FPRS_FEF, %g0
be,pn %icc, 5f
sll %o0, 3, %o5
rd %fprs, %g1
wr %g1, FPRS_FEF, %fprs
ldx [%o1 + %o5], %g1
add %g6, TI_XFSR, %o1
sll %o0, 8, %o2
add %g6, TI_FPREGS, %o3
brz,pn %l6, 1f
add %g6, TI_FPREGS+0x40, %o4
membar #Sync
ldda [%o3 + %o2] ASI_BLK_P, %f0
ldda [%o4 + %o2] ASI_BLK_P, %f16
membar #Sync
1: andcc %l2, FPRS_DU, %g0
be,pn %icc, 1f
wr %g1, 0, %gsr
add %o2, 0x80, %o2
membar #Sync
ldda [%o3 + %o2] ASI_BLK_P, %f32
ldda [%o4 + %o2] ASI_BLK_P, %f48
1: membar #Sync
ldx [%o1 + %o5], %fsr
2: stb %l5, [%g6 + TI_FPDEPTH]
ba,pt %xcc, rt_continue
nop
5: wr %g0, FPRS_FEF, %fprs
sll %o0, 8, %o2
add %g6, TI_FPREGS+0x80, %o3
add %g6, TI_FPREGS+0xc0, %o4
membar #Sync
ldda [%o3 + %o2] ASI_BLK_P, %f32
ldda [%o4 + %o2] ASI_BLK_P, %f48
membar #Sync
wr %g0, FPRS_DU, %fprs
ba,pt %xcc, rt_continue
stb %l5, [%g6 + TI_FPDEPTH]