linux/arch/ia64/kernel/entry.S

1809 lines
48 KiB
ArmAsm
Raw Normal View History

/*
* arch/ia64/kernel/entry.S
*
* Kernel entry points.
*
* Copyright (C) 1998-2003, 2005 Hewlett-Packard Co
* David Mosberger-Tang <davidm@hpl.hp.com>
* Copyright (C) 1999, 2002-2003
* Asit Mallick <Asit.K.Mallick@intel.com>
* Don Dugger <Don.Dugger@intel.com>
* Suresh Siddha <suresh.b.siddha@intel.com>
* Fenghua Yu <fenghua.yu@intel.com>
* Copyright (C) 1999 VA Linux Systems
* Copyright (C) 1999 Walt Drummond <drummond@valinux.com>
*/
/*
* ia64_switch_to now places correct virtual mapping in in TR2 for
* kernel stack. This allows us to handle interrupts without changing
* to physical mode.
*
* Jonathan Nicklin <nicklin@missioncriticallinux.com>
* Patrick O'Rourke <orourke@missioncriticallinux.com>
* 11/07/2000
*/
/*
* Copyright (c) 2008 Isaku Yamahata <yamahata at valinux co jp>
* VA Linux Systems Japan K.K.
* pv_ops.
*/
/*
* Global (preserved) predicate usage on syscall entry/exit path:
*
* pKStk: See entry.h.
* pUStk: See entry.h.
* pSys: See entry.h.
* pNonSys: !pSys
*/
#include <asm/asmmacro.h>
#include <asm/cache.h>
#include <asm/errno.h>
#include <asm/kregs.h>
#include <asm/asm-offsets.h>
#include <asm/pgtable.h>
#include <asm/percpu.h>
#include <asm/processor.h>
#include <asm/thread_info.h>
#include <asm/unistd.h>
#include <asm/ftrace.h>
#include "minstate.h"
#ifdef __IA64_ASM_PARAVIRTUALIZED_NATIVE
/*
* execve() is special because in case of success, we need to
* setup a null register window frame.
*/
ENTRY(ia64_execve)
/*
* Allocate 8 input registers since ptrace() may clobber them
*/
.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(8)
alloc loc1=ar.pfs,8,2,4,0
mov loc0=rp
.body
mov out0=in0 // filename
;; // stop bit between alloc and call
mov out1=in1 // argv
mov out2=in2 // envp
add out3=16,sp // regs
br.call.sptk.many rp=sys_execve
.ret0:
#ifdef CONFIG_IA32_SUPPORT
/*
* Check if we're returning to ia32 mode. If so, we need to restore ia32 registers
* from pt_regs.
*/
adds r16=PT(CR_IPSR)+16,sp
;;
ld8 r16=[r16]
#endif
cmp4.ge p6,p7=r8,r0
mov ar.pfs=loc1 // restore ar.pfs
sxt4 r8=r8 // return 64-bit result
;;
stf.spill [sp]=f0
(p6) cmp.ne pKStk,pUStk=r0,r0 // a successful execve() lands us in user-mode...
mov rp=loc0
(p6) mov ar.pfs=r0 // clear ar.pfs on success
(p7) br.ret.sptk.many rp
/*
* In theory, we'd have to zap this state only to prevent leaking of
* security sensitive state (e.g., if current->mm->dumpable is zero). However,
* this executes in less than 20 cycles even on Itanium, so it's not worth
* optimizing for...).
*/
mov ar.unat=0; mov ar.lc=0
mov r4=0; mov f2=f0; mov b1=r0
mov r5=0; mov f3=f0; mov b2=r0
mov r6=0; mov f4=f0; mov b3=r0
mov r7=0; mov f5=f0; mov b4=r0
ldf.fill f12=[sp]; mov f13=f0; mov b5=r0
ldf.fill f14=[sp]; ldf.fill f15=[sp]; mov f16=f0
ldf.fill f17=[sp]; ldf.fill f18=[sp]; mov f19=f0
ldf.fill f20=[sp]; ldf.fill f21=[sp]; mov f22=f0
ldf.fill f23=[sp]; ldf.fill f24=[sp]; mov f25=f0
ldf.fill f26=[sp]; ldf.fill f27=[sp]; mov f28=f0
ldf.fill f29=[sp]; ldf.fill f30=[sp]; mov f31=f0
#ifdef CONFIG_IA32_SUPPORT
tbit.nz p6,p0=r16, IA64_PSR_IS_BIT
movl loc0=ia64_ret_from_ia32_execve
;;
(p6) mov rp=loc0
#endif
br.ret.sptk.many rp
END(ia64_execve)
/*
* sys_clone2(u64 flags, u64 ustack_base, u64 ustack_size, u64 parent_tidptr, u64 child_tidptr,
* u64 tls)
*/
GLOBAL_ENTRY(sys_clone2)
/*
* Allocate 8 input registers since ptrace() may clobber them
*/
.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(8)
alloc r16=ar.pfs,8,2,6,0
DO_SAVE_SWITCH_STACK
adds r2=PT(R16)+IA64_SWITCH_STACK_SIZE+16,sp
mov loc0=rp
mov loc1=r16 // save ar.pfs across do_fork
.body
mov out1=in1
mov out3=in2
tbit.nz p6,p0=in0,CLONE_SETTLS_BIT
mov out4=in3 // parent_tidptr: valid only w/CLONE_PARENT_SETTID
;;
(p6) st8 [r2]=in5 // store TLS in r16 for copy_thread()
mov out5=in4 // child_tidptr: valid only w/CLONE_CHILD_SETTID or CLONE_CHILD_CLEARTID
adds out2=IA64_SWITCH_STACK_SIZE+16,sp // out2 = &regs
mov out0=in0 // out0 = clone_flags
br.call.sptk.many rp=do_fork
.ret1: .restore sp
adds sp=IA64_SWITCH_STACK_SIZE,sp // pop the switch stack
mov ar.pfs=loc1
mov rp=loc0
br.ret.sptk.many rp
END(sys_clone2)
/*
* sys_clone(u64 flags, u64 ustack_base, u64 parent_tidptr, u64 child_tidptr, u64 tls)
* Deprecated. Use sys_clone2() instead.
*/
GLOBAL_ENTRY(sys_clone)
/*
* Allocate 8 input registers since ptrace() may clobber them
*/
.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(8)
alloc r16=ar.pfs,8,2,6,0
DO_SAVE_SWITCH_STACK
adds r2=PT(R16)+IA64_SWITCH_STACK_SIZE+16,sp
mov loc0=rp
mov loc1=r16 // save ar.pfs across do_fork
.body
mov out1=in1
mov out3=16 // stacksize (compensates for 16-byte scratch area)
tbit.nz p6,p0=in0,CLONE_SETTLS_BIT
mov out4=in2 // parent_tidptr: valid only w/CLONE_PARENT_SETTID
;;
(p6) st8 [r2]=in4 // store TLS in r13 (tp)
mov out5=in3 // child_tidptr: valid only w/CLONE_CHILD_SETTID or CLONE_CHILD_CLEARTID
adds out2=IA64_SWITCH_STACK_SIZE+16,sp // out2 = &regs
mov out0=in0 // out0 = clone_flags
br.call.sptk.many rp=do_fork
.ret2: .restore sp
adds sp=IA64_SWITCH_STACK_SIZE,sp // pop the switch stack
mov ar.pfs=loc1
mov rp=loc0
br.ret.sptk.many rp
END(sys_clone)
#endif /* __IA64_ASM_PARAVIRTUALIZED_NATIVE */
/*
* prev_task <- ia64_switch_to(struct task_struct *next)
* With Ingo's new scheduler, interrupts are disabled when this routine gets
* called. The code starting at .map relies on this. The rest of the code
* doesn't care about the interrupt masking status.
*/
GLOBAL_ENTRY(__paravirt_switch_to)
.prologue
alloc r16=ar.pfs,1,0,0,0
DO_SAVE_SWITCH_STACK
.body
adds r22=IA64_TASK_THREAD_KSP_OFFSET,r13
movl r25=init_task
mov r27=IA64_KR(CURRENT_STACK)
adds r21=IA64_TASK_THREAD_KSP_OFFSET,in0
dep r20=0,in0,61,3 // physical address of "next"
;;
st8 [r22]=sp // save kernel stack pointer of old task
shr.u r26=r20,IA64_GRANULE_SHIFT
cmp.eq p7,p6=r25,in0
;;
/*
* If we've already mapped this task's page, we can skip doing it again.
*/
(p6) cmp.eq p7,p6=r26,r27
(p6) br.cond.dpnt .map
;;
.done:
ld8 sp=[r21] // load kernel stack pointer of new task
MOV_TO_KR(CURRENT, in0, r8, r9) // update "current" application register
mov r8=r13 // return pointer to previously running task
mov r13=in0 // set "current" pointer
;;
DO_LOAD_SWITCH_STACK
#ifdef CONFIG_SMP
sync.i // ensure "fc"s done by this CPU are visible on other CPUs
#endif
br.ret.sptk.many rp // boogie on out in new context
.map:
RSM_PSR_IC(r25) // interrupts (psr.i) are already disabled here
movl r25=PAGE_KERNEL
;;
srlz.d
or r23=r25,r20 // construct PA | page properties
mov r25=IA64_GRANULE_SHIFT<<2
;;
MOV_TO_ITIR(p0, r25, r8)
MOV_TO_IFA(in0, r8) // VA of next task...
;;
mov r25=IA64_TR_CURRENT_STACK
MOV_TO_KR(CURRENT_STACK, r26, r8, r9) // remember last page we mapped...
;;
itr.d dtr[r25]=r23 // wire in new mapping...
SSM_PSR_IC_AND_SRLZ_D(r8, r9) // reenable the psr.ic bit
br.cond.sptk .done
END(__paravirt_switch_to)
#ifdef __IA64_ASM_PARAVIRTUALIZED_NATIVE
/*
* Note that interrupts are enabled during save_switch_stack and load_switch_stack. This
* means that we may get an interrupt with "sp" pointing to the new kernel stack while
* ar.bspstore is still pointing to the old kernel backing store area. Since ar.rsc,
* ar.rnat, ar.bsp, and ar.bspstore are all preserved by interrupts, this is not a
* problem. Also, we don't need to specify unwind information for preserved registers
* that are not modified in save_switch_stack as the right unwind information is already
* specified at the call-site of save_switch_stack.
*/
/*
* save_switch_stack:
* - r16 holds ar.pfs
* - b7 holds address to return to
* - rp (b0) holds return address to save
*/
GLOBAL_ENTRY(save_switch_stack)
.prologue
.altrp b7
flushrs // flush dirty regs to backing store (must be first in insn group)
.save @priunat,r17
mov r17=ar.unat // preserve caller's
.body
#ifdef CONFIG_ITANIUM
adds r2=16+128,sp
adds r3=16+64,sp
adds r14=SW(R4)+16,sp
;;
st8.spill [r14]=r4,16 // spill r4
lfetch.fault.excl.nt1 [r3],128
;;
lfetch.fault.excl.nt1 [r2],128
lfetch.fault.excl.nt1 [r3],128
;;
lfetch.fault.excl [r2]
lfetch.fault.excl [r3]
adds r15=SW(R5)+16,sp
#else
add r2=16+3*128,sp
add r3=16,sp
add r14=SW(R4)+16,sp
;;
st8.spill [r14]=r4,SW(R6)-SW(R4) // spill r4 and prefetch offset 0x1c0
lfetch.fault.excl.nt1 [r3],128 // prefetch offset 0x010
;;
lfetch.fault.excl.nt1 [r3],128 // prefetch offset 0x090
lfetch.fault.excl.nt1 [r2],128 // prefetch offset 0x190
;;
lfetch.fault.excl.nt1 [r3] // prefetch offset 0x110
lfetch.fault.excl.nt1 [r2] // prefetch offset 0x210
adds r15=SW(R5)+16,sp
#endif
;;
st8.spill [r15]=r5,SW(R7)-SW(R5) // spill r5
mov.m ar.rsc=0 // put RSE in mode: enforced lazy, little endian, pl 0
add r2=SW(F2)+16,sp // r2 = &sw->f2
;;
st8.spill [r14]=r6,SW(B0)-SW(R6) // spill r6
mov.m r18=ar.fpsr // preserve fpsr
add r3=SW(F3)+16,sp // r3 = &sw->f3
;;
stf.spill [r2]=f2,32
mov.m r19=ar.rnat
mov r21=b0
stf.spill [r3]=f3,32
st8.spill [r15]=r7,SW(B2)-SW(R7) // spill r7
mov r22=b1
;;
// since we're done with the spills, read and save ar.unat:
mov.m r29=ar.unat
mov.m r20=ar.bspstore
mov r23=b2
stf.spill [r2]=f4,32
stf.spill [r3]=f5,32
mov r24=b3
;;
st8 [r14]=r21,SW(B1)-SW(B0) // save b0
st8 [r15]=r23,SW(B3)-SW(B2) // save b2
mov r25=b4
mov r26=b5
;;
st8 [r14]=r22,SW(B4)-SW(B1) // save b1
st8 [r15]=r24,SW(AR_PFS)-SW(B3) // save b3
mov r21=ar.lc // I-unit
stf.spill [r2]=f12,32
stf.spill [r3]=f13,32
;;
st8 [r14]=r25,SW(B5)-SW(B4) // save b4
st8 [r15]=r16,SW(AR_LC)-SW(AR_PFS) // save ar.pfs
stf.spill [r2]=f14,32
stf.spill [r3]=f15,32
;;
st8 [r14]=r26 // save b5
st8 [r15]=r21 // save ar.lc
stf.spill [r2]=f16,32
stf.spill [r3]=f17,32
;;
stf.spill [r2]=f18,32
stf.spill [r3]=f19,32
;;
stf.spill [r2]=f20,32
stf.spill [r3]=f21,32
;;
stf.spill [r2]=f22,32
stf.spill [r3]=f23,32
;;
stf.spill [r2]=f24,32
stf.spill [r3]=f25,32
;;
stf.spill [r2]=f26,32
stf.spill [r3]=f27,32
;;
stf.spill [r2]=f28,32
stf.spill [r3]=f29,32
;;
stf.spill [r2]=f30,SW(AR_UNAT)-SW(F30)
stf.spill [r3]=f31,SW(PR)-SW(F31)
add r14=SW(CALLER_UNAT)+16,sp
;;
st8 [r2]=r29,SW(AR_RNAT)-SW(AR_UNAT) // save ar.unat
st8 [r14]=r17,SW(AR_FPSR)-SW(CALLER_UNAT) // save caller_unat
mov r21=pr
;;
st8 [r2]=r19,SW(AR_BSPSTORE)-SW(AR_RNAT) // save ar.rnat
st8 [r3]=r21 // save predicate registers
;;
st8 [r2]=r20 // save ar.bspstore
st8 [r14]=r18 // save fpsr
mov ar.rsc=3 // put RSE back into eager mode, pl 0
br.cond.sptk.many b7
END(save_switch_stack)
/*
* load_switch_stack:
* - "invala" MUST be done at call site (normally in DO_LOAD_SWITCH_STACK)
* - b7 holds address to return to
* - must not touch r8-r11
*/
GLOBAL_ENTRY(load_switch_stack)
.prologue
.altrp b7
.body
lfetch.fault.nt1 [sp]
adds r2=SW(AR_BSPSTORE)+16,sp
adds r3=SW(AR_UNAT)+16,sp
mov ar.rsc=0 // put RSE into enforced lazy mode
adds r14=SW(CALLER_UNAT)+16,sp
adds r15=SW(AR_FPSR)+16,sp
;;
ld8 r27=[r2],(SW(B0)-SW(AR_BSPSTORE)) // bspstore
ld8 r29=[r3],(SW(B1)-SW(AR_UNAT)) // unat
;;
ld8 r21=[r2],16 // restore b0
ld8 r22=[r3],16 // restore b1
;;
ld8 r23=[r2],16 // restore b2
ld8 r24=[r3],16 // restore b3
;;
ld8 r25=[r2],16 // restore b4
ld8 r26=[r3],16 // restore b5
;;
ld8 r16=[r2],(SW(PR)-SW(AR_PFS)) // ar.pfs
ld8 r17=[r3],(SW(AR_RNAT)-SW(AR_LC)) // ar.lc
;;
ld8 r28=[r2] // restore pr
ld8 r30=[r3] // restore rnat
;;
ld8 r18=[r14],16 // restore caller's unat
ld8 r19=[r15],24 // restore fpsr
;;
ldf.fill f2=[r14],32
ldf.fill f3=[r15],32
;;
ldf.fill f4=[r14],32
ldf.fill f5=[r15],32
;;
ldf.fill f12=[r14],32
ldf.fill f13=[r15],32
;;
ldf.fill f14=[r14],32
ldf.fill f15=[r15],32
;;
ldf.fill f16=[r14],32
ldf.fill f17=[r15],32
;;
ldf.fill f18=[r14],32
ldf.fill f19=[r15],32
mov b0=r21
;;
ldf.fill f20=[r14],32
ldf.fill f21=[r15],32
mov b1=r22
;;
ldf.fill f22=[r14],32
ldf.fill f23=[r15],32
mov b2=r23
;;
mov ar.bspstore=r27
mov ar.unat=r29 // establish unat holding the NaT bits for r4-r7
mov b3=r24
;;
ldf.fill f24=[r14],32
ldf.fill f25=[r15],32
mov b4=r25
;;
ldf.fill f26=[r14],32
ldf.fill f27=[r15],32
mov b5=r26
;;
ldf.fill f28=[r14],32
ldf.fill f29=[r15],32
mov ar.pfs=r16
;;
ldf.fill f30=[r14],32
ldf.fill f31=[r15],24
mov ar.lc=r17
;;
ld8.fill r4=[r14],16
ld8.fill r5=[r15],16
mov pr=r28,-1
;;
ld8.fill r6=[r14],16
ld8.fill r7=[r15],16
mov ar.unat=r18 // restore caller's unat
mov ar.rnat=r30 // must restore after bspstore but before rsc!
mov ar.fpsr=r19 // restore fpsr
mov ar.rsc=3 // put RSE back into eager mode, pl 0
br.cond.sptk.many b7
END(load_switch_stack)
GLOBAL_ENTRY(prefetch_stack)
add r14 = -IA64_SWITCH_STACK_SIZE, sp
add r15 = IA64_TASK_THREAD_KSP_OFFSET, in0
;;
ld8 r16 = [r15] // load next's stack pointer
lfetch.fault.excl [r14], 128
;;
lfetch.fault.excl [r14], 128
lfetch.fault [r16], 128
;;
lfetch.fault.excl [r14], 128
lfetch.fault [r16], 128
;;
lfetch.fault.excl [r14], 128
lfetch.fault [r16], 128
;;
lfetch.fault.excl [r14], 128
lfetch.fault [r16], 128
;;
lfetch.fault [r16], 128
br.ret.sptk.many rp
END(prefetch_stack)
[PATCH] rename the provided execve functions to kernel_execve Some architectures provide an execve function that does not set errno, but instead returns the result code directly. Rename these to kernel_execve to get the right semantics there. Moreover, there is no reasone for any of these architectures to still provide __KERNEL_SYSCALLS__ or _syscallN macros, so remove these right away. [akpm@osdl.org: build fix] [bunk@stusta.de: build fix] Signed-off-by: Arnd Bergmann <arnd@arndb.de> Cc: Andi Kleen <ak@muc.de> Acked-by: Paul Mackerras <paulus@samba.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Richard Henderson <rth@twiddle.net> Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru> Cc: Russell King <rmk@arm.linux.org.uk> Cc: Ian Molton <spyro@f2s.com> Cc: Mikael Starvik <starvik@axis.com> Cc: David Howells <dhowells@redhat.com> Cc: Yoshinori Sato <ysato@users.sourceforge.jp> Cc: Hirokazu Takata <takata.hirokazu@renesas.com> Cc: Ralf Baechle <ralf@linux-mips.org> Cc: Kyle McMartin <kyle@mcmartin.ca> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Paul Mundt <lethal@linux-sh.org> Cc: Kazumoto Kojima <kkojima@rr.iij4u.or.jp> Cc: Richard Curnow <rc@rc0.org.uk> Cc: William Lee Irwin III <wli@holomorphy.com> Cc: "David S. Miller" <davem@davemloft.net> Cc: Jeff Dike <jdike@addtoit.com> Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Cc: Miles Bader <uclinux-v850@lsi.nec.co.jp> Cc: Chris Zankel <chris@zankel.net> Cc: "Luck, Tony" <tony.luck@intel.com> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Roman Zippel <zippel@linux-m68k.org> Signed-off-by: Adrian Bunk <bunk@stusta.de> Cc: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-10-02 09:18:31 +00:00
GLOBAL_ENTRY(kernel_execve)
rum psr.ac
mov r15=__NR_execve // put syscall number in place
break __BREAK_SYSCALL
br.ret.sptk.many rp
[PATCH] rename the provided execve functions to kernel_execve Some architectures provide an execve function that does not set errno, but instead returns the result code directly. Rename these to kernel_execve to get the right semantics there. Moreover, there is no reasone for any of these architectures to still provide __KERNEL_SYSCALLS__ or _syscallN macros, so remove these right away. [akpm@osdl.org: build fix] [bunk@stusta.de: build fix] Signed-off-by: Arnd Bergmann <arnd@arndb.de> Cc: Andi Kleen <ak@muc.de> Acked-by: Paul Mackerras <paulus@samba.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Richard Henderson <rth@twiddle.net> Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru> Cc: Russell King <rmk@arm.linux.org.uk> Cc: Ian Molton <spyro@f2s.com> Cc: Mikael Starvik <starvik@axis.com> Cc: David Howells <dhowells@redhat.com> Cc: Yoshinori Sato <ysato@users.sourceforge.jp> Cc: Hirokazu Takata <takata.hirokazu@renesas.com> Cc: Ralf Baechle <ralf@linux-mips.org> Cc: Kyle McMartin <kyle@mcmartin.ca> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Paul Mundt <lethal@linux-sh.org> Cc: Kazumoto Kojima <kkojima@rr.iij4u.or.jp> Cc: Richard Curnow <rc@rc0.org.uk> Cc: William Lee Irwin III <wli@holomorphy.com> Cc: "David S. Miller" <davem@davemloft.net> Cc: Jeff Dike <jdike@addtoit.com> Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Cc: Miles Bader <uclinux-v850@lsi.nec.co.jp> Cc: Chris Zankel <chris@zankel.net> Cc: "Luck, Tony" <tony.luck@intel.com> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Roman Zippel <zippel@linux-m68k.org> Signed-off-by: Adrian Bunk <bunk@stusta.de> Cc: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-10-02 09:18:31 +00:00
END(kernel_execve)
GLOBAL_ENTRY(clone)
mov r15=__NR_clone // put syscall number in place
break __BREAK_SYSCALL
br.ret.sptk.many rp
END(clone)
/*
* Invoke a system call, but do some tracing before and after the call.
* We MUST preserve the current register frame throughout this routine
* because some system calls (such as ia64_execve) directly
* manipulate ar.pfs.
*/
GLOBAL_ENTRY(ia64_trace_syscall)
PT_REGS_UNWIND_INFO(0)
/*
* We need to preserve the scratch registers f6-f11 in case the system
* call is sigreturn.
*/
adds r16=PT(F6)+16,sp
adds r17=PT(F7)+16,sp
;;
stf.spill [r16]=f6,32
stf.spill [r17]=f7,32
;;
stf.spill [r16]=f8,32
stf.spill [r17]=f9,32
;;
stf.spill [r16]=f10
stf.spill [r17]=f11
br.call.sptk.many rp=syscall_trace_enter // give parent a chance to catch syscall args
cmp.lt p6,p0=r8,r0 // check tracehook
adds r2=PT(R8)+16,sp // r2 = &pt_regs.r8
adds r3=PT(R10)+16,sp // r3 = &pt_regs.r10
mov r10=0
(p6) br.cond.sptk strace_error // syscall failed ->
adds r16=PT(F6)+16,sp
adds r17=PT(F7)+16,sp
;;
ldf.fill f6=[r16],32
ldf.fill f7=[r17],32
;;
ldf.fill f8=[r16],32
ldf.fill f9=[r17],32
;;
ldf.fill f10=[r16]
ldf.fill f11=[r17]
// the syscall number may have changed, so re-load it and re-calculate the
// syscall entry-point:
adds r15=PT(R15)+16,sp // r15 = &pt_regs.r15 (syscall #)
;;
ld8 r15=[r15]
mov r3=NR_syscalls - 1
;;
adds r15=-1024,r15
movl r16=sys_call_table
;;
shladd r20=r15,3,r16 // r20 = sys_call_table + 8*(syscall-1024)
cmp.leu p6,p7=r15,r3
;;
(p6) ld8 r20=[r20] // load address of syscall entry point
(p7) movl r20=sys_ni_syscall
;;
mov b6=r20
br.call.sptk.many rp=b6 // do the syscall
.strace_check_retval:
cmp.lt p6,p0=r8,r0 // syscall failed?
adds r2=PT(R8)+16,sp // r2 = &pt_regs.r8
adds r3=PT(R10)+16,sp // r3 = &pt_regs.r10
mov r10=0
(p6) br.cond.sptk strace_error // syscall failed ->
;; // avoid RAW on r10
.strace_save_retval:
.mem.offset 0,0; st8.spill [r2]=r8 // store return value in slot for r8
.mem.offset 8,0; st8.spill [r3]=r10 // clear error indication in slot for r10
br.call.sptk.many rp=syscall_trace_leave // give parent a chance to catch return value
.ret3:
(pUStk) cmp.eq.unc p6,p0=r0,r0 // p6 <- pUStk
[IA64] disable interrupts on exit of ia64_trace_syscall While testing with CONFIG_VIRT_CPU_ACCOUNTING=y, I found that I occasionally get very huge system time in some threads. So I dug the issue and finally noticed that it was caused because of an interrupt which interrupt in the following window: > [arch/ia64/kernel/entry.S: (!CONFIG_PREEMPT && CONFIG_VIRT_CPU_ACCOUNTING)] > > ENTRY(ia64_leave_syscall) > : > (pUStk) rsm psr.i > cmp.eq pLvSys,p0=r0,r0 // pLvSys=1: leave from syscall > (pUStk) cmp.eq.unc p6,p0=r0,r0 // p6 <- pUStk > .work_processed_syscall: > adds r2=PT(LOADRS)+16,r12 > (pUStk) mov.m r22=ar.itc // fetch time at leave > adds r18=TI_FLAGS+IA64_TASK_SIZE,r13 > ;; > <<< window: from here >>> > (p6) ld4 r31=[r18] // load current_thread_info()->flags > ld8 r19=[r2],PT(B6)-PT(LOADRS) > adds r3=PT(AR_BSPSTORE)+16,r12 > ;; > mov r16=ar.bsp > ld8 r18=[r2],PT(R9)-PT(B6) > (p6) and r15=TIF_WORK_MASK,r31 // any work other than TIF_SYSCALL_TRACE? > ;; > ld8 r23=[r3],PT(R11)-PT(AR_BSPSTORE) > (p6) cmp4.ne.unc p6,p0=r15, r0 // any special work pending? > (p6) br.cond.spnt .work_pending_syscall > ;; > ld8 r9=[r2],PT(CR_IPSR)-PT(R9) > ld8 r11=[r3],PT(CR_IIP)-PT(R11) > (pNonSys) break 0 // bug check: we shouldn't be here if pNonSys is TRUE! > ;; > invala > <<< window: to here >>> > rsm psr.i | psr.ic // turn off interrupts and interruption collection If pUStk is true, it means we are going to return user mode, hence we fetch ar.itc to get time at leave from system. It seems that it is not possible to interrupt the window if pUStk is true, because interrupts are disabled early. And also disabling interrupt makes sense because it is safe for referring current_thread_info()->flags. However interrupting the window while pUStk is true was possible. The route was: ia64_trace_syscall -> .work_pending_syscall_end -> .work_processed_syscall Only in case entering the window from this route, interrupts are enabled during in the window even if pUStk is true. I suppose interrupts must be disabled here anyway if pUStk is true. I'm not sure but afraid that what kind of bad effect were there, other than crazy system time which I found. FYI, there was a commit 6f6d75825dc49b082906b84537b4df28293c2977 that points out a bug at same point(exit of ia64_trace_syscall) in 2006. It can be said that there was an another bug. Signed-off-by: Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com> Signed-off-by: Tony Luck <tony.luck@intel.com>
2008-04-21 21:34:39 +00:00
(pUStk) rsm psr.i // disable interrupts
br.cond.sptk ia64_work_pending_syscall_end
strace_error:
ld8 r3=[r2] // load pt_regs.r8
sub r9=0,r8 // negate return value to get errno value
;;
cmp.ne p6,p0=r3,r0 // is pt_regs.r8!=0?
adds r3=16,r2 // r3=&pt_regs.r10
;;
(p6) mov r10=-1
(p6) mov r8=r9
br.cond.sptk .strace_save_retval
END(ia64_trace_syscall)
/*
* When traced and returning from sigreturn, we invoke syscall_trace but then
* go straight to ia64_leave_kernel rather than ia64_leave_syscall.
*/
GLOBAL_ENTRY(ia64_strace_leave_kernel)
PT_REGS_UNWIND_INFO(0)
{ /*
* Some versions of gas generate bad unwind info if the first instruction of a
* procedure doesn't go into the first slot of a bundle. This is a workaround.
*/
nop.m 0
nop.i 0
br.call.sptk.many rp=syscall_trace_leave // give parent a chance to catch return value
}
.ret4: br.cond.sptk ia64_leave_kernel
END(ia64_strace_leave_kernel)
GLOBAL_ENTRY(ia64_ret_from_clone)
PT_REGS_UNWIND_INFO(0)
{ /*
* Some versions of gas generate bad unwind info if the first instruction of a
* procedure doesn't go into the first slot of a bundle. This is a workaround.
*/
nop.m 0
nop.i 0
/*
* We need to call schedule_tail() to complete the scheduling process.
* Called by ia64_switch_to() after do_fork()->copy_thread(). r8 contains the
* address of the previously executing task.
*/
br.call.sptk.many rp=ia64_invoke_schedule_tail
}
.ret8:
adds r2=TI_FLAGS+IA64_TASK_SIZE,r13
;;
ld4 r2=[r2]
;;
mov r8=0
and r2=_TIF_SYSCALL_TRACEAUDIT,r2
;;
cmp.ne p6,p0=r2,r0
(p6) br.cond.spnt .strace_check_retval
;; // added stop bits to prevent r8 dependency
END(ia64_ret_from_clone)
// fall through
GLOBAL_ENTRY(ia64_ret_from_syscall)
PT_REGS_UNWIND_INFO(0)
cmp.ge p6,p7=r8,r0 // syscall executed successfully?
adds r2=PT(R8)+16,sp // r2 = &pt_regs.r8
mov r10=r0 // clear error indication in r10
(p7) br.cond.spnt handle_syscall_error // handle potential syscall failure
#ifdef CONFIG_PARAVIRT
;;
br.cond.sptk.few ia64_leave_syscall
;;
#endif /* CONFIG_PARAVIRT */
END(ia64_ret_from_syscall)
#ifndef CONFIG_PARAVIRT
// fall through
#endif
#endif /* __IA64_ASM_PARAVIRTUALIZED_NATIVE */
/*
* ia64_leave_syscall(): Same as ia64_leave_kernel, except that it doesn't
* need to switch to bank 0 and doesn't restore the scratch registers.
* To avoid leaking kernel bits, the scratch registers are set to
* the following known-to-be-safe values:
*
* r1: restored (global pointer)
* r2: cleared
* r3: 1 (when returning to user-level)
* r8-r11: restored (syscall return value(s))
* r12: restored (user-level stack pointer)
* r13: restored (user-level thread pointer)
* r14: set to __kernel_syscall_via_epc
* r15: restored (syscall #)
* r16-r17: cleared
* r18: user-level b6
* r19: cleared
* r20: user-level ar.fpsr
* r21: user-level b0
* r22: cleared
* r23: user-level ar.bspstore
* r24: user-level ar.rnat
* r25: user-level ar.unat
* r26: user-level ar.pfs
* r27: user-level ar.rsc
* r28: user-level ip
* r29: user-level psr
* r30: user-level cfm
* r31: user-level pr
* f6-f11: cleared
* pr: restored (user-level pr)
* b0: restored (user-level rp)
* b6: restored
* b7: set to __kernel_syscall_via_epc
* ar.unat: restored (user-level ar.unat)
* ar.pfs: restored (user-level ar.pfs)
* ar.rsc: restored (user-level ar.rsc)
* ar.rnat: restored (user-level ar.rnat)
* ar.bspstore: restored (user-level ar.bspstore)
* ar.fpsr: restored (user-level ar.fpsr)
* ar.ccv: cleared
* ar.csd: cleared
* ar.ssd: cleared
*/
GLOBAL_ENTRY(__paravirt_leave_syscall)
PT_REGS_UNWIND_INFO(0)
/*
* work.need_resched etc. mustn't get changed by this CPU before it returns to
* user- or fsys-mode, hence we disable interrupts early on.
*
* p6 controls whether current_thread_info()->flags needs to be check for
* extra work. We always check for extra work when returning to user-level.
* With CONFIG_PREEMPT, we also check for extra work when the preempt_count
* is 0. After extra work processing has been completed, execution
* resumes at ia64_work_processed_syscall with p6 set to 1 if the extra-work-check
* needs to be redone.
*/
#ifdef CONFIG_PREEMPT
RSM_PSR_I(p0, r2, r18) // disable interrupts
cmp.eq pLvSys,p0=r0,r0 // pLvSys=1: leave from syscall
(pKStk) adds r20=TI_PRE_COUNT+IA64_TASK_SIZE,r13
;;
.pred.rel.mutex pUStk,pKStk
(pKStk) ld4 r21=[r20] // r21 <- preempt_count
(pUStk) mov r21=0 // r21 <- 0
;;
cmp.eq p6,p0=r21,r0 // p6 <- pUStk || (preempt_count == 0)
#else /* !CONFIG_PREEMPT */
RSM_PSR_I(pUStk, r2, r18)
cmp.eq pLvSys,p0=r0,r0 // pLvSys=1: leave from syscall
(pUStk) cmp.eq.unc p6,p0=r0,r0 // p6 <- pUStk
#endif
.global __paravirt_work_processed_syscall;
__paravirt_work_processed_syscall:
#ifdef CONFIG_VIRT_CPU_ACCOUNTING
adds r2=PT(LOADRS)+16,r12
MOV_FROM_ITC(pUStk, p9, r22, r19) // fetch time at leave
adds r18=TI_FLAGS+IA64_TASK_SIZE,r13
;;
(p6) ld4 r31=[r18] // load current_thread_info()->flags
ld8 r19=[r2],PT(B6)-PT(LOADRS) // load ar.rsc value for "loadrs"
adds r3=PT(AR_BSPSTORE)+16,r12 // deferred
;;
#else
adds r2=PT(LOADRS)+16,r12
adds r3=PT(AR_BSPSTORE)+16,r12
adds r18=TI_FLAGS+IA64_TASK_SIZE,r13
;;
(p6) ld4 r31=[r18] // load current_thread_info()->flags
ld8 r19=[r2],PT(B6)-PT(LOADRS) // load ar.rsc value for "loadrs"
nop.i 0
;;
#endif
mov r16=ar.bsp // M2 get existing backing store pointer
ld8 r18=[r2],PT(R9)-PT(B6) // load b6
(p6) and r15=TIF_WORK_MASK,r31 // any work other than TIF_SYSCALL_TRACE?
;;
ld8 r23=[r3],PT(R11)-PT(AR_BSPSTORE) // load ar.bspstore (may be garbage)
(p6) cmp4.ne.unc p6,p0=r15, r0 // any special work pending?
(p6) br.cond.spnt .work_pending_syscall
;;
// start restoring the state saved on the kernel stack (struct pt_regs):
ld8 r9=[r2],PT(CR_IPSR)-PT(R9)
ld8 r11=[r3],PT(CR_IIP)-PT(R11)
(pNonSys) break 0 // bug check: we shouldn't be here if pNonSys is TRUE!
;;
invala // M0|1 invalidate ALAT
RSM_PSR_I_IC(r28, r29, r30) // M2 turn off interrupts and interruption collection
cmp.eq p9,p0=r0,r0 // A set p9 to indicate that we should restore cr.ifs
ld8 r29=[r2],16 // M0|1 load cr.ipsr
ld8 r28=[r3],16 // M0|1 load cr.iip
#ifdef CONFIG_VIRT_CPU_ACCOUNTING
(pUStk) add r14=TI_AC_LEAVE+IA64_TASK_SIZE,r13
;;
ld8 r30=[r2],16 // M0|1 load cr.ifs
ld8 r25=[r3],16 // M0|1 load ar.unat
(pUStk) add r15=IA64_TASK_THREAD_ON_USTACK_OFFSET,r13
;;
#else
mov r22=r0 // A clear r22
;;
ld8 r30=[r2],16 // M0|1 load cr.ifs
ld8 r25=[r3],16 // M0|1 load ar.unat
(pUStk) add r14=IA64_TASK_THREAD_ON_USTACK_OFFSET,r13
;;
#endif
ld8 r26=[r2],PT(B0)-PT(AR_PFS) // M0|1 load ar.pfs
MOV_FROM_PSR(pKStk, r22, r21) // M2 read PSR now that interrupts are disabled
nop 0
;;
ld8 r21=[r2],PT(AR_RNAT)-PT(B0) // M0|1 load b0
ld8 r27=[r3],PT(PR)-PT(AR_RSC) // M0|1 load ar.rsc
mov f6=f0 // F clear f6
;;
ld8 r24=[r2],PT(AR_FPSR)-PT(AR_RNAT) // M0|1 load ar.rnat (may be garbage)
ld8 r31=[r3],PT(R1)-PT(PR) // M0|1 load predicates
mov f7=f0 // F clear f7
;;
ld8 r20=[r2],PT(R12)-PT(AR_FPSR) // M0|1 load ar.fpsr
ld8.fill r1=[r3],16 // M0|1 load r1
(pUStk) mov r17=1 // A
;;
#ifdef CONFIG_VIRT_CPU_ACCOUNTING
(pUStk) st1 [r15]=r17 // M2|3
#else
(pUStk) st1 [r14]=r17 // M2|3
#endif
ld8.fill r13=[r3],16 // M0|1
mov f8=f0 // F clear f8
;;
ld8.fill r12=[r2] // M0|1 restore r12 (sp)
ld8.fill r15=[r3] // M0|1 restore r15
mov b6=r18 // I0 restore b6
[IA64] remove per-cpu ia64_phys_stacked_size_p8 It's not efficient to use a per-cpu variable just to store how many physical stack register a cpu has. Ever since the incarnation of ia64 up till upcoming Montecito processor, that variable has "glued" to 96. Having a variable in memory means that the kernel is burning an extra cacheline access on every syscall and kernel exit path. Such "static" value is better served with the instruction patching utility exists today. Convert ia64_phys_stacked_size_p8 into dynamic insn patching. This also has a pleasant side effect of eliminating access to per-cpu area while psr.ic=0 in the kernel exit path. (fixable for per-cpu DTC work, but why bother?) There are some concerns with the default value that the instruc- tion encoded in the kernel image. It shouldn't be concerned. The reasons are: (1) cpu_init() is called at CPU initialization. In there, we find out physical stack register size from PAL and patch two instructions in kernel exit code. The code in question can not be executed before the patching is done. (2) current implementation stores zero in ia64_phys_stacked_size_p8, and that's what the current kernel exit path loads the value with. With the new code, it is equivalent that we store reg size 96 in ia64_phys_stacked_size_p8, thus creating a better safety net. Given (1) above can never fail, having (2) is just a bonus. All in all, this patch allow one less memory reference in the kernel exit path, thus reducing syscall and interrupt return latency; and avoid polluting potential useful data in the CPU cache. Signed-off-by: Ken Chen <kenneth.w.chen@intel.com> Signed-off-by: Tony Luck <tony.luck@intel.com>
2006-10-13 17:05:45 +00:00
LOAD_PHYS_STACK_REG_SIZE(r17)
mov f9=f0 // F clear f9
(pKStk) br.cond.dpnt.many skip_rbs_switch // B
srlz.d // M0 ensure interruption collection is off (for cover)
shr.u r18=r19,16 // I0|1 get byte size of existing "dirty" partition
COVER // B add current frame into dirty partition & set cr.ifs
;;
#ifdef CONFIG_VIRT_CPU_ACCOUNTING
mov r19=ar.bsp // M2 get new backing store pointer
st8 [r14]=r22 // M save time at leave
mov f10=f0 // F clear f10
mov r22=r0 // A clear r22
movl r14=__kernel_syscall_via_epc // X
;;
#else
mov r19=ar.bsp // M2 get new backing store pointer
mov f10=f0 // F clear f10
nop.m 0
movl r14=__kernel_syscall_via_epc // X
;;
#endif
mov.m ar.csd=r0 // M2 clear ar.csd
mov.m ar.ccv=r0 // M2 clear ar.ccv
mov b7=r14 // I0 clear b7 (hint with __kernel_syscall_via_epc)
mov.m ar.ssd=r0 // M2 clear ar.ssd
mov f11=f0 // F clear f11
br.cond.sptk.many rbs_switch // B
END(__paravirt_leave_syscall)
#ifdef __IA64_ASM_PARAVIRTUALIZED_NATIVE
#ifdef CONFIG_IA32_SUPPORT
GLOBAL_ENTRY(ia64_ret_from_ia32_execve)
PT_REGS_UNWIND_INFO(0)
adds r2=PT(R8)+16,sp // r2 = &pt_regs.r8
adds r3=PT(R10)+16,sp // r3 = &pt_regs.r10
;;
.mem.offset 0,0
st8.spill [r2]=r8 // store return value in slot for r8 and set unat bit
.mem.offset 8,0
st8.spill [r3]=r0 // clear error indication in slot for r10 and set unat bit
#ifdef CONFIG_PARAVIRT
;;
// don't fall through, ia64_leave_kernel may be #define'd
br.cond.sptk.few ia64_leave_kernel
;;
#endif /* CONFIG_PARAVIRT */
END(ia64_ret_from_ia32_execve)
#ifndef CONFIG_PARAVIRT
// fall through
#endif
#endif /* CONFIG_IA32_SUPPORT */
#endif /* __IA64_ASM_PARAVIRTUALIZED_NATIVE */
GLOBAL_ENTRY(__paravirt_leave_kernel)
PT_REGS_UNWIND_INFO(0)
/*
* work.need_resched etc. mustn't get changed by this CPU before it returns to
* user- or fsys-mode, hence we disable interrupts early on.
*
* p6 controls whether current_thread_info()->flags needs to be check for
* extra work. We always check for extra work when returning to user-level.
* With CONFIG_PREEMPT, we also check for extra work when the preempt_count
* is 0. After extra work processing has been completed, execution
* resumes at .work_processed_syscall with p6 set to 1 if the extra-work-check
* needs to be redone.
*/
#ifdef CONFIG_PREEMPT
RSM_PSR_I(p0, r17, r31) // disable interrupts
cmp.eq p0,pLvSys=r0,r0 // pLvSys=0: leave from kernel
(pKStk) adds r20=TI_PRE_COUNT+IA64_TASK_SIZE,r13
;;
.pred.rel.mutex pUStk,pKStk
(pKStk) ld4 r21=[r20] // r21 <- preempt_count
(pUStk) mov r21=0 // r21 <- 0
;;
cmp.eq p6,p0=r21,r0 // p6 <- pUStk || (preempt_count == 0)
#else
RSM_PSR_I(pUStk, r17, r31)
cmp.eq p0,pLvSys=r0,r0 // pLvSys=0: leave from kernel
(pUStk) cmp.eq.unc p6,p0=r0,r0 // p6 <- pUStk
#endif
.work_processed_kernel:
adds r17=TI_FLAGS+IA64_TASK_SIZE,r13
;;
(p6) ld4 r31=[r17] // load current_thread_info()->flags
adds r21=PT(PR)+16,r12
;;
lfetch [r21],PT(CR_IPSR)-PT(PR)
adds r2=PT(B6)+16,r12
adds r3=PT(R16)+16,r12
;;
lfetch [r21]
ld8 r28=[r2],8 // load b6
adds r29=PT(R24)+16,r12
ld8.fill r16=[r3],PT(AR_CSD)-PT(R16)
adds r30=PT(AR_CCV)+16,r12
(p6) and r19=TIF_WORK_MASK,r31 // any work other than TIF_SYSCALL_TRACE?
;;
ld8.fill r24=[r29]
ld8 r15=[r30] // load ar.ccv
(p6) cmp4.ne.unc p6,p0=r19, r0 // any special work pending?
;;
ld8 r29=[r2],16 // load b7
ld8 r30=[r3],16 // load ar.csd
(p6) br.cond.spnt .work_pending
;;
ld8 r31=[r2],16 // load ar.ssd
ld8.fill r8=[r3],16
;;
ld8.fill r9=[r2],16
ld8.fill r10=[r3],PT(R17)-PT(R10)
;;
ld8.fill r11=[r2],PT(R18)-PT(R11)
ld8.fill r17=[r3],16
;;
ld8.fill r18=[r2],16
ld8.fill r19=[r3],16
;;
ld8.fill r20=[r2],16
ld8.fill r21=[r3],16
mov ar.csd=r30
mov ar.ssd=r31
;;
RSM_PSR_I_IC(r23, r22, r25) // initiate turning off of interrupt and interruption collection
invala // invalidate ALAT
;;
ld8.fill r22=[r2],24
ld8.fill r23=[r3],24
mov b6=r28
;;
ld8.fill r25=[r2],16
ld8.fill r26=[r3],16
mov b7=r29
;;
ld8.fill r27=[r2],16
ld8.fill r28=[r3],16
;;
ld8.fill r29=[r2],16
ld8.fill r30=[r3],24
;;
ld8.fill r31=[r2],PT(F9)-PT(R31)
adds r3=PT(F10)-PT(F6),r3
;;
ldf.fill f9=[r2],PT(F6)-PT(F9)
ldf.fill f10=[r3],PT(F8)-PT(F10)
;;
ldf.fill f6=[r2],PT(F7)-PT(F6)
;;
ldf.fill f7=[r2],PT(F11)-PT(F7)
ldf.fill f8=[r3],32
;;
srlz.d // ensure that inter. collection is off (VHPT is don't care, since text is pinned)
mov ar.ccv=r15
;;
ldf.fill f11=[r2]
BSW_0(r2, r3, r15) // switch back to bank 0 (no stop bit required beforehand...)
;;
(pUStk) mov r18=IA64_KR(CURRENT)// M2 (12 cycle read latency)
adds r16=PT(CR_IPSR)+16,r12
adds r17=PT(CR_IIP)+16,r12
#ifdef CONFIG_VIRT_CPU_ACCOUNTING
.pred.rel.mutex pUStk,pKStk
MOV_FROM_PSR(pKStk, r22, r29) // M2 read PSR now that interrupts are disabled
MOV_FROM_ITC(pUStk, p9, r22, r29) // M fetch time at leave
nop.i 0
;;
#else
MOV_FROM_PSR(pKStk, r22, r29) // M2 read PSR now that interrupts are disabled
nop.i 0
nop.i 0
;;
#endif
ld8 r29=[r16],16 // load cr.ipsr
ld8 r28=[r17],16 // load cr.iip
;;
ld8 r30=[r16],16 // load cr.ifs
ld8 r25=[r17],16 // load ar.unat
;;
ld8 r26=[r16],16 // load ar.pfs
ld8 r27=[r17],16 // load ar.rsc
cmp.eq p9,p0=r0,r0 // set p9 to indicate that we should restore cr.ifs
;;
ld8 r24=[r16],16 // load ar.rnat (may be garbage)
ld8 r23=[r17],16 // load ar.bspstore (may be garbage)
;;
ld8 r31=[r16],16 // load predicates
ld8 r21=[r17],16 // load b0
;;
ld8 r19=[r16],16 // load ar.rsc value for "loadrs"
ld8.fill r1=[r17],16 // load r1
;;
ld8.fill r12=[r16],16
ld8.fill r13=[r17],16
#ifdef CONFIG_VIRT_CPU_ACCOUNTING
(pUStk) adds r3=TI_AC_LEAVE+IA64_TASK_SIZE,r18
#else
(pUStk) adds r18=IA64_TASK_THREAD_ON_USTACK_OFFSET,r18
#endif
;;
ld8 r20=[r16],16 // ar.fpsr
ld8.fill r15=[r17],16
#ifdef CONFIG_VIRT_CPU_ACCOUNTING
(pUStk) adds r18=IA64_TASK_THREAD_ON_USTACK_OFFSET,r18 // deferred
#endif
;;
ld8.fill r14=[r16],16
ld8.fill r2=[r17]
(pUStk) mov r17=1
;;
#ifdef CONFIG_VIRT_CPU_ACCOUNTING
// mmi_ : ld8 st1 shr;; mmi_ : st8 st1 shr;;
// mib : mov add br -> mib : ld8 add br
// bbb_ : br nop cover;; mbb_ : mov br cover;;
//
// no one require bsp in r16 if (pKStk) branch is selected.
(pUStk) st8 [r3]=r22 // save time at leave
(pUStk) st1 [r18]=r17 // restore current->thread.on_ustack
shr.u r18=r19,16 // get byte size of existing "dirty" partition
;;
ld8.fill r3=[r16] // deferred
LOAD_PHYS_STACK_REG_SIZE(r17)
(pKStk) br.cond.dpnt skip_rbs_switch
mov r16=ar.bsp // get existing backing store pointer
#else
ld8.fill r3=[r16]
(pUStk) st1 [r18]=r17 // restore current->thread.on_ustack
shr.u r18=r19,16 // get byte size of existing "dirty" partition
;;
mov r16=ar.bsp // get existing backing store pointer
[IA64] remove per-cpu ia64_phys_stacked_size_p8 It's not efficient to use a per-cpu variable just to store how many physical stack register a cpu has. Ever since the incarnation of ia64 up till upcoming Montecito processor, that variable has "glued" to 96. Having a variable in memory means that the kernel is burning an extra cacheline access on every syscall and kernel exit path. Such "static" value is better served with the instruction patching utility exists today. Convert ia64_phys_stacked_size_p8 into dynamic insn patching. This also has a pleasant side effect of eliminating access to per-cpu area while psr.ic=0 in the kernel exit path. (fixable for per-cpu DTC work, but why bother?) There are some concerns with the default value that the instruc- tion encoded in the kernel image. It shouldn't be concerned. The reasons are: (1) cpu_init() is called at CPU initialization. In there, we find out physical stack register size from PAL and patch two instructions in kernel exit code. The code in question can not be executed before the patching is done. (2) current implementation stores zero in ia64_phys_stacked_size_p8, and that's what the current kernel exit path loads the value with. With the new code, it is equivalent that we store reg size 96 in ia64_phys_stacked_size_p8, thus creating a better safety net. Given (1) above can never fail, having (2) is just a bonus. All in all, this patch allow one less memory reference in the kernel exit path, thus reducing syscall and interrupt return latency; and avoid polluting potential useful data in the CPU cache. Signed-off-by: Ken Chen <kenneth.w.chen@intel.com> Signed-off-by: Tony Luck <tony.luck@intel.com>
2006-10-13 17:05:45 +00:00
LOAD_PHYS_STACK_REG_SIZE(r17)
(pKStk) br.cond.dpnt skip_rbs_switch
#endif
/*
* Restore user backing store.
*
* NOTE: alloc, loadrs, and cover can't be predicated.
*/
(pNonSys) br.cond.dpnt dont_preserve_current_frame
COVER // add current frame into dirty partition and set cr.ifs
;;
mov r19=ar.bsp // get new backing store pointer
rbs_switch:
sub r16=r16,r18 // krbs = old bsp - size of dirty partition
cmp.ne p9,p0=r0,r0 // clear p9 to skip restore of cr.ifs
;;
sub r19=r19,r16 // calculate total byte size of dirty partition
add r18=64,r18 // don't force in0-in7 into memory...
;;
shl r19=r19,16 // shift size of dirty partition into loadrs position
;;
dont_preserve_current_frame:
/*
* To prevent leaking bits between the kernel and user-space,
* we must clear the stacked registers in the "invalid" partition here.
* Not pretty, but at least it's fast (3.34 registers/cycle on Itanium,
* 5 registers/cycle on McKinley).
*/
# define pRecurse p6
# define pReturn p7
#ifdef CONFIG_ITANIUM
# define Nregs 10
#else
# define Nregs 14
#endif
alloc loc0=ar.pfs,2,Nregs-2,2,0
shr.u loc1=r18,9 // RNaTslots <= floor(dirtySize / (64*8))
sub r17=r17,r18 // r17 = (physStackedSize + 8) - dirtySize
;;
mov ar.rsc=r19 // load ar.rsc to be used for "loadrs"
shladd in0=loc1,3,r17
mov in1=0
;;
TEXT_ALIGN(32)
rse_clear_invalid:
#ifdef CONFIG_ITANIUM
// cycle 0
{ .mii
alloc loc0=ar.pfs,2,Nregs-2,2,0
cmp.lt pRecurse,p0=Nregs*8,in0 // if more than Nregs regs left to clear, (re)curse
add out0=-Nregs*8,in0
}{ .mfb
add out1=1,in1 // increment recursion count
nop.f 0
nop.b 0 // can't do br.call here because of alloc (WAW on CFM)
;;
}{ .mfi // cycle 1
mov loc1=0
nop.f 0
mov loc2=0
}{ .mib
mov loc3=0
mov loc4=0
(pRecurse) br.call.sptk.many b0=rse_clear_invalid
}{ .mfi // cycle 2
mov loc5=0
nop.f 0
cmp.ne pReturn,p0=r0,in1 // if recursion count != 0, we need to do a br.ret
}{ .mib
mov loc6=0
mov loc7=0
(pReturn) br.ret.sptk.many b0
}
#else /* !CONFIG_ITANIUM */
alloc loc0=ar.pfs,2,Nregs-2,2,0
cmp.lt pRecurse,p0=Nregs*8,in0 // if more than Nregs regs left to clear, (re)curse
add out0=-Nregs*8,in0
add out1=1,in1 // increment recursion count
mov loc1=0
mov loc2=0
;;
mov loc3=0
mov loc4=0
mov loc5=0
mov loc6=0
mov loc7=0
(pRecurse) br.call.dptk.few b0=rse_clear_invalid
;;
mov loc8=0
mov loc9=0
cmp.ne pReturn,p0=r0,in1 // if recursion count != 0, we need to do a br.ret
mov loc10=0
mov loc11=0
(pReturn) br.ret.dptk.many b0
#endif /* !CONFIG_ITANIUM */
# undef pRecurse
# undef pReturn
;;
alloc r17=ar.pfs,0,0,0,0 // drop current register frame
;;
loadrs
;;
skip_rbs_switch:
mov ar.unat=r25 // M2
(pKStk) extr.u r22=r22,21,1 // I0 extract current value of psr.pp from r22
(pLvSys)mov r19=r0 // A clear r19 for leave_syscall, no-op otherwise
;;
(pUStk) mov ar.bspstore=r23 // M2
(pKStk) dep r29=r22,r29,21,1 // I0 update ipsr.pp with psr.pp
(pLvSys)mov r16=r0 // A clear r16 for leave_syscall, no-op otherwise
;;
MOV_TO_IPSR(p0, r29, r25) // M2
mov ar.pfs=r26 // I0
(pLvSys)mov r17=r0 // A clear r17 for leave_syscall, no-op otherwise
MOV_TO_IFS(p9, r30, r25)// M2
mov b0=r21 // I0
(pLvSys)mov r18=r0 // A clear r18 for leave_syscall, no-op otherwise
mov ar.fpsr=r20 // M2
MOV_TO_IIP(r28, r25) // M2
nop 0
;;
(pUStk) mov ar.rnat=r24 // M2 must happen with RSE in lazy mode
nop 0
(pLvSys)mov r2=r0
mov ar.rsc=r27 // M2
mov pr=r31,-1 // I0
RFI // B
/*
* On entry:
* r20 = &current->thread_info->pre_count (if CONFIG_PREEMPT)
* r31 = current->thread_info->flags
* On exit:
* p6 = TRUE if work-pending-check needs to be redone
*
* Interrupts are disabled on entry, reenabled depend on work, and
* disabled on exit.
*/
.work_pending_syscall:
add r2=-8,r2
add r3=-8,r3
;;
st8 [r2]=r8
st8 [r3]=r10
.work_pending:
tbit.z p6,p0=r31,TIF_NEED_RESCHED // is resched not needed?
(p6) br.cond.sptk.few .notify
#ifdef CONFIG_PREEMPT
(pKStk) dep r21=-1,r0,PREEMPT_ACTIVE_BIT,1
;;
(pKStk) st4 [r20]=r21
#endif
SSM_PSR_I(p0, p6, r2) // enable interrupts
br.call.spnt.many rp=schedule
.ret9: cmp.eq p6,p0=r0,r0 // p6 <- 1 (re-check)
RSM_PSR_I(p0, r2, r20) // disable interrupts
;;
#ifdef CONFIG_PREEMPT
(pKStk) adds r20=TI_PRE_COUNT+IA64_TASK_SIZE,r13
;;
(pKStk) st4 [r20]=r0 // preempt_count() <- 0
#endif
(pLvSys)br.cond.sptk.few __paravirt_pending_syscall_end
br.cond.sptk.many .work_processed_kernel
.notify:
(pUStk) br.call.spnt.many rp=notify_resume_user
.ret10: cmp.ne p6,p0=r0,r0 // p6 <- 0 (don't re-check)
(pLvSys)br.cond.sptk.few __paravirt_pending_syscall_end
br.cond.sptk.many .work_processed_kernel
.global __paravirt_pending_syscall_end;
__paravirt_pending_syscall_end:
adds r2=PT(R8)+16,r12
adds r3=PT(R10)+16,r12
;;
ld8 r8=[r2]
ld8 r10=[r3]
br.cond.sptk.many __paravirt_work_processed_syscall_target
END(__paravirt_leave_kernel)
#ifdef __IA64_ASM_PARAVIRTUALIZED_NATIVE
ENTRY(handle_syscall_error)
/*
* Some system calls (e.g., ptrace, mmap) can return arbitrary values which could
* lead us to mistake a negative return value as a failed syscall. Those syscall
* must deposit a non-zero value in pt_regs.r8 to indicate an error. If
* pt_regs.r8 is zero, we assume that the call completed successfully.
*/
PT_REGS_UNWIND_INFO(0)
ld8 r3=[r2] // load pt_regs.r8
;;
cmp.eq p6,p7=r3,r0 // is pt_regs.r8==0?
;;
(p7) mov r10=-1
(p7) sub r8=0,r8 // negate return value to get errno
br.cond.sptk ia64_leave_syscall
END(handle_syscall_error)
/*
* Invoke schedule_tail(task) while preserving in0-in7, which may be needed
* in case a system call gets restarted.
*/
GLOBAL_ENTRY(ia64_invoke_schedule_tail)
.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(8)
alloc loc1=ar.pfs,8,2,1,0
mov loc0=rp
mov out0=r8 // Address of previous task
;;
br.call.sptk.many rp=schedule_tail
.ret11: mov ar.pfs=loc1
mov rp=loc0
br.ret.sptk.many rp
END(ia64_invoke_schedule_tail)
/*
* Setup stack and call do_notify_resume_user(), keeping interrupts
* disabled.
*
* Note that pSys and pNonSys need to be set up by the caller.
* We declare 8 input registers so the system call args get preserved,
* in case we need to restart a system call.
*/
GLOBAL_ENTRY(notify_resume_user)
.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(8)
alloc loc1=ar.pfs,8,2,3,0 // preserve all eight input regs in case of syscall restart!
mov r9=ar.unat
mov loc0=rp // save return address
mov out0=0 // there is no "oldset"
adds out1=8,sp // out1=&sigscratch->ar_pfs
(pSys) mov out2=1 // out2==1 => we're in a syscall
;;
(pNonSys) mov out2=0 // out2==0 => not a syscall
.fframe 16
.spillsp ar.unat, 16
st8 [sp]=r9,-16 // allocate space for ar.unat and save it
st8 [out1]=loc1,-8 // save ar.pfs, out1=&sigscratch
.body
br.call.sptk.many rp=do_notify_resume_user
.ret15: .restore sp
adds sp=16,sp // pop scratch stack space
;;
ld8 r9=[sp] // load new unat from sigscratch->scratch_unat
mov rp=loc0
;;
mov ar.unat=r9
mov ar.pfs=loc1
br.ret.sptk.many rp
END(notify_resume_user)
ENTRY(sys_rt_sigreturn)
PT_REGS_UNWIND_INFO(0)
/*
* Allocate 8 input registers since ptrace() may clobber them
*/
alloc r2=ar.pfs,8,0,1,0
.prologue
PT_REGS_SAVES(16)
adds sp=-16,sp
.body
cmp.eq pNonSys,pSys=r0,r0 // sigreturn isn't a normal syscall...
;;
/*
* leave_kernel() restores f6-f11 from pt_regs, but since the streamlined
* syscall-entry path does not save them we save them here instead. Note: we
* don't need to save any other registers that are not saved by the stream-lined
* syscall path, because restore_sigcontext() restores them.
*/
adds r16=PT(F6)+32,sp
adds r17=PT(F7)+32,sp
;;
stf.spill [r16]=f6,32
stf.spill [r17]=f7,32
;;
stf.spill [r16]=f8,32
stf.spill [r17]=f9,32
;;
stf.spill [r16]=f10
stf.spill [r17]=f11
adds out0=16,sp // out0 = &sigscratch
br.call.sptk.many rp=ia64_rt_sigreturn
.ret19: .restore sp,0
adds sp=16,sp
;;
ld8 r9=[sp] // load new ar.unat
mov.sptk b7=r8,ia64_native_leave_kernel
;;
mov ar.unat=r9
br.many b7
END(sys_rt_sigreturn)
GLOBAL_ENTRY(ia64_prepare_handle_unaligned)
.prologue
/*
* r16 = fake ar.pfs, we simply need to make sure privilege is still 0
*/
mov r16=r0
DO_SAVE_SWITCH_STACK
br.call.sptk.many rp=ia64_handle_unaligned // stack frame setup in ivt
.ret21: .body
DO_LOAD_SWITCH_STACK
br.cond.sptk.many rp // goes to ia64_leave_kernel
END(ia64_prepare_handle_unaligned)
//
// unw_init_running(void (*callback)(info, arg), void *arg)
//
# define EXTRA_FRAME_SIZE ((UNW_FRAME_INFO_SIZE+15)&~15)
GLOBAL_ENTRY(unw_init_running)
.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(2)
alloc loc1=ar.pfs,2,3,3,0
;;
ld8 loc2=[in0],8
mov loc0=rp
mov r16=loc1
DO_SAVE_SWITCH_STACK
.body
.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(2)
.fframe IA64_SWITCH_STACK_SIZE+EXTRA_FRAME_SIZE
SWITCH_STACK_SAVES(EXTRA_FRAME_SIZE)
adds sp=-EXTRA_FRAME_SIZE,sp
.body
;;
adds out0=16,sp // &info
mov out1=r13 // current
adds out2=16+EXTRA_FRAME_SIZE,sp // &switch_stack
br.call.sptk.many rp=unw_init_frame_info
1: adds out0=16,sp // &info
mov b6=loc2
mov loc2=gp // save gp across indirect function call
;;
ld8 gp=[in0]
mov out1=in1 // arg
br.call.sptk.many rp=b6 // invoke the callback function
1: mov gp=loc2 // restore gp
// For now, we don't allow changing registers from within
// unw_init_running; if we ever want to allow that, we'd
// have to do a load_switch_stack here:
.restore sp
adds sp=IA64_SWITCH_STACK_SIZE+EXTRA_FRAME_SIZE,sp
mov ar.pfs=loc1
mov rp=loc0
br.ret.sptk.many rp
END(unw_init_running)
#ifdef CONFIG_FUNCTION_TRACER
#ifdef CONFIG_DYNAMIC_FTRACE
GLOBAL_ENTRY(_mcount)
br ftrace_stub
END(_mcount)
.here:
br.ret.sptk.many b0
GLOBAL_ENTRY(ftrace_caller)
alloc out0 = ar.pfs, 8, 0, 4, 0
mov out3 = r0
;;
mov out2 = b0
add r3 = 0x20, r3
mov out1 = r1;
br.call.sptk.many b0 = ftrace_patch_gp
//this might be called from module, so we must patch gp
ftrace_patch_gp:
movl gp=__gp
mov b0 = r3
;;
.global ftrace_call;
ftrace_call:
{
.mlx
nop.m 0x0
movl r3 = .here;;
}
alloc loc0 = ar.pfs, 4, 4, 2, 0
;;
mov loc1 = b0
mov out0 = b0
mov loc2 = r8
mov loc3 = r15
;;
adds out0 = -MCOUNT_INSN_SIZE, out0
mov out1 = in2
mov b6 = r3
br.call.sptk.many b0 = b6
;;
mov ar.pfs = loc0
mov b0 = loc1
mov r8 = loc2
mov r15 = loc3
br ftrace_stub
;;
END(ftrace_caller)
#else
GLOBAL_ENTRY(_mcount)
movl r2 = ftrace_stub
movl r3 = ftrace_trace_function;;
ld8 r3 = [r3];;
ld8 r3 = [r3];;
cmp.eq p7,p0 = r2, r3
(p7) br.sptk.many ftrace_stub
;;
alloc loc0 = ar.pfs, 4, 4, 2, 0
;;
mov loc1 = b0
mov out0 = b0
mov loc2 = r8
mov loc3 = r15
;;
adds out0 = -MCOUNT_INSN_SIZE, out0
mov out1 = in2
mov b6 = r3
br.call.sptk.many b0 = b6
;;
mov ar.pfs = loc0
mov b0 = loc1
mov r8 = loc2
mov r15 = loc3
br ftrace_stub
;;
END(_mcount)
#endif
GLOBAL_ENTRY(ftrace_stub)
mov r3 = b0
movl r2 = _mcount_ret_helper
;;
mov b6 = r2
mov b7 = r3
br.ret.sptk.many b6
_mcount_ret_helper:
mov b0 = r42
mov r1 = r41
mov ar.pfs = r40
br b7
END(ftrace_stub)
#endif /* CONFIG_FUNCTION_TRACER */
.rodata
.align 8
.globl sys_call_table
sys_call_table:
data8 sys_ni_syscall // This must be sys_ni_syscall! See ivt.S.
data8 sys_exit // 1025
data8 sys_read
data8 sys_write
data8 sys_open
data8 sys_close
data8 sys_creat // 1030
data8 sys_link
data8 sys_unlink
data8 ia64_execve
data8 sys_chdir
data8 sys_fchdir // 1035
data8 sys_utimes
data8 sys_mknod
data8 sys_chmod
data8 sys_chown
data8 sys_lseek // 1040
data8 sys_getpid
data8 sys_getppid
data8 sys_mount
data8 sys_umount
data8 sys_setuid // 1045
data8 sys_getuid
data8 sys_geteuid
data8 sys_ptrace
data8 sys_access
data8 sys_sync // 1050
data8 sys_fsync
data8 sys_fdatasync
data8 sys_kill
data8 sys_rename
data8 sys_mkdir // 1055
data8 sys_rmdir
data8 sys_dup
data8 sys_ia64_pipe
data8 sys_times
data8 ia64_brk // 1060
data8 sys_setgid
data8 sys_getgid
data8 sys_getegid
data8 sys_acct
data8 sys_ioctl // 1065
data8 sys_fcntl
data8 sys_umask
data8 sys_chroot
data8 sys_ustat
data8 sys_dup2 // 1070
data8 sys_setreuid
data8 sys_setregid
data8 sys_getresuid
data8 sys_setresuid
data8 sys_getresgid // 1075
data8 sys_setresgid
data8 sys_getgroups
data8 sys_setgroups
data8 sys_getpgid
data8 sys_setpgid // 1080
data8 sys_setsid
data8 sys_getsid
data8 sys_sethostname
data8 sys_setrlimit
data8 sys_getrlimit // 1085
data8 sys_getrusage
data8 sys_gettimeofday
data8 sys_settimeofday
data8 sys_select
data8 sys_poll // 1090
data8 sys_symlink
data8 sys_readlink
data8 sys_uselib
data8 sys_swapon
data8 sys_swapoff // 1095
data8 sys_reboot
data8 sys_truncate
data8 sys_ftruncate
data8 sys_fchmod
data8 sys_fchown // 1100
data8 ia64_getpriority
data8 sys_setpriority
data8 sys_statfs
data8 sys_fstatfs
data8 sys_gettid // 1105
data8 sys_semget
data8 sys_semop
data8 sys_semctl
data8 sys_msgget
data8 sys_msgsnd // 1110
data8 sys_msgrcv
data8 sys_msgctl
data8 sys_shmget
data8 sys_shmat
data8 sys_shmdt // 1115
data8 sys_shmctl
data8 sys_syslog
data8 sys_setitimer
data8 sys_getitimer
data8 sys_ni_syscall // 1120 /* was: ia64_oldstat */
data8 sys_ni_syscall /* was: ia64_oldlstat */
data8 sys_ni_syscall /* was: ia64_oldfstat */
data8 sys_vhangup
data8 sys_lchown
data8 sys_remap_file_pages // 1125
data8 sys_wait4
data8 sys_sysinfo
data8 sys_clone
data8 sys_setdomainname
data8 sys_newuname // 1130
data8 sys_adjtimex
data8 sys_ni_syscall /* was: ia64_create_module */
data8 sys_init_module
data8 sys_delete_module
data8 sys_ni_syscall // 1135 /* was: sys_get_kernel_syms */
data8 sys_ni_syscall /* was: sys_query_module */
data8 sys_quotactl
data8 sys_bdflush
data8 sys_sysfs
data8 sys_personality // 1140
data8 sys_ni_syscall // sys_afs_syscall
data8 sys_setfsuid
data8 sys_setfsgid
data8 sys_getdents
data8 sys_flock // 1145
data8 sys_readv
data8 sys_writev
data8 sys_pread64
data8 sys_pwrite64
data8 sys_sysctl // 1150
data8 sys_mmap
data8 sys_munmap
data8 sys_mlock
data8 sys_mlockall
data8 sys_mprotect // 1155
data8 ia64_mremap
data8 sys_msync
data8 sys_munlock
data8 sys_munlockall
data8 sys_sched_getparam // 1160
data8 sys_sched_setparam
data8 sys_sched_getscheduler
data8 sys_sched_setscheduler
data8 sys_sched_yield
data8 sys_sched_get_priority_max // 1165
data8 sys_sched_get_priority_min
data8 sys_sched_rr_get_interval
data8 sys_nanosleep
data8 sys_nfsservctl
data8 sys_prctl // 1170
data8 sys_getpagesize
data8 sys_mmap2
data8 sys_pciconfig_read
data8 sys_pciconfig_write
data8 sys_perfmonctl // 1175
data8 sys_sigaltstack
data8 sys_rt_sigaction
data8 sys_rt_sigpending
data8 sys_rt_sigprocmask
data8 sys_rt_sigqueueinfo // 1180
data8 sys_rt_sigreturn
data8 sys_rt_sigsuspend
data8 sys_rt_sigtimedwait
data8 sys_getcwd
data8 sys_capget // 1185
data8 sys_capset
data8 sys_sendfile64
data8 sys_ni_syscall // sys_getpmsg (STREAMS)
data8 sys_ni_syscall // sys_putpmsg (STREAMS)
data8 sys_socket // 1190
data8 sys_bind
data8 sys_connect
data8 sys_listen
data8 sys_accept
data8 sys_getsockname // 1195
data8 sys_getpeername
data8 sys_socketpair
data8 sys_send
data8 sys_sendto
data8 sys_recv // 1200
data8 sys_recvfrom
data8 sys_shutdown
data8 sys_setsockopt
data8 sys_getsockopt
data8 sys_sendmsg // 1205
data8 sys_recvmsg
data8 sys_pivot_root
data8 sys_mincore
data8 sys_madvise
data8 sys_newstat // 1210
data8 sys_newlstat
data8 sys_newfstat
data8 sys_clone2
data8 sys_getdents64
data8 sys_getunwind // 1215
data8 sys_readahead
data8 sys_setxattr
data8 sys_lsetxattr
data8 sys_fsetxattr
data8 sys_getxattr // 1220
data8 sys_lgetxattr
data8 sys_fgetxattr
data8 sys_listxattr
data8 sys_llistxattr
data8 sys_flistxattr // 1225
data8 sys_removexattr
data8 sys_lremovexattr
data8 sys_fremovexattr
data8 sys_tkill
data8 sys_futex // 1230
data8 sys_sched_setaffinity
data8 sys_sched_getaffinity
data8 sys_set_tid_address
data8 sys_fadvise64_64
data8 sys_tgkill // 1235
data8 sys_exit_group
data8 sys_lookup_dcookie
data8 sys_io_setup
data8 sys_io_destroy
data8 sys_io_getevents // 1240
data8 sys_io_submit
data8 sys_io_cancel
data8 sys_epoll_create
data8 sys_epoll_ctl
data8 sys_epoll_wait // 1245
data8 sys_restart_syscall
data8 sys_semtimedop
data8 sys_timer_create
data8 sys_timer_settime
data8 sys_timer_gettime // 1250
data8 sys_timer_getoverrun
data8 sys_timer_delete
data8 sys_clock_settime
data8 sys_clock_gettime
data8 sys_clock_getres // 1255
data8 sys_clock_nanosleep
data8 sys_fstatfs64
data8 sys_statfs64
data8 sys_mbind
data8 sys_get_mempolicy // 1260
data8 sys_set_mempolicy
data8 sys_mq_open
data8 sys_mq_unlink
data8 sys_mq_timedsend
data8 sys_mq_timedreceive // 1265
data8 sys_mq_notify
data8 sys_mq_getsetattr
data8 sys_kexec_load
data8 sys_ni_syscall // reserved for vserver
data8 sys_waitid // 1270
data8 sys_add_key
data8 sys_request_key
data8 sys_keyctl
data8 sys_ioprio_set
data8 sys_ioprio_get // 1275
[PATCH] page migration: sys_move_pages(): support moving of individual pages move_pages() is used to move individual pages of a process. The function can be used to determine the location of pages and to move them onto the desired node. move_pages() returns status information for each page. long move_pages(pid, number_of_pages_to_move, addresses_of_pages[], nodes[] or NULL, status[], flags); The addresses of pages is an array of void * pointing to the pages to be moved. The nodes array contains the node numbers that the pages should be moved to. If a NULL is passed instead of an array then no pages are moved but the status array is updated. The status request may be used to determine the page state before issuing another move_pages() to move pages. The status array will contain the state of all individual page migration attempts when the function terminates. The status array is only valid if move_pages() completed successfullly. Possible page states in status[]: 0..MAX_NUMNODES The page is now on the indicated node. -ENOENT Page is not present -EACCES Page is mapped by multiple processes and can only be moved if MPOL_MF_MOVE_ALL is specified. -EPERM The page has been mlocked by a process/driver and cannot be moved. -EBUSY Page is busy and cannot be moved. Try again later. -EFAULT Invalid address (no VMA or zero page). -ENOMEM Unable to allocate memory on target node. -EIO Unable to write back page. The page must be written back in order to move it since the page is dirty and the filesystem does not provide a migration function that would allow the moving of dirty pages. -EINVAL A dirty page cannot be moved. The filesystem does not provide a migration function and has no ability to write back pages. The flags parameter indicates what types of pages to move: MPOL_MF_MOVE Move pages that are only mapped by the process. MPOL_MF_MOVE_ALL Also move pages that are mapped by multiple processes. Requires sufficient capabilities. Possible return codes from move_pages() -ENOENT No pages found that would require moving. All pages are either already on the target node, not present, had an invalid address or could not be moved because they were mapped by multiple processes. -EINVAL Flags other than MPOL_MF_MOVE(_ALL) specified or an attempt to migrate pages in a kernel thread. -EPERM MPOL_MF_MOVE_ALL specified without sufficient priviledges. or an attempt to move a process belonging to another user. -EACCES One of the target nodes is not allowed by the current cpuset. -ENODEV One of the target nodes is not online. -ESRCH Process does not exist. -E2BIG Too many pages to move. -ENOMEM Not enough memory to allocate control array. -EFAULT Parameters could not be accessed. A test program for move_pages() may be found with the patches on ftp.kernel.org:/pub/linux/kernel/people/christoph/pmig/patches-2.6.17-rc4-mm3 From: Christoph Lameter <clameter@sgi.com> Detailed results for sys_move_pages() Pass a pointer to an integer to get_new_page() that may be used to indicate where the completion status of a migration operation should be placed. This allows sys_move_pags() to report back exactly what happened to each page. Wish there would be a better way to do this. Looks a bit hacky. Signed-off-by: Christoph Lameter <clameter@sgi.com> Cc: Hugh Dickins <hugh@veritas.com> Cc: Jes Sorensen <jes@trained-monkey.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Lee Schermerhorn <lee.schermerhorn@hp.com> Cc: Andi Kleen <ak@muc.de> Cc: Michael Kerrisk <mtk-manpages@gmx.net> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-06-23 09:03:55 +00:00
data8 sys_move_pages
data8 sys_inotify_init
data8 sys_inotify_add_watch
data8 sys_inotify_rm_watch
[PATCH] Swap Migration V5: sys_migrate_pages interface sys_migrate_pages implementation using swap based page migration This is the original API proposed by Ray Bryant in his posts during the first half of 2005 on linux-mm@kvack.org and linux-kernel@vger.kernel.org. The intent of sys_migrate is to migrate memory of a process. A process may have migrated to another node. Memory was allocated optimally for the prior context. sys_migrate_pages allows to shift the memory to the new node. sys_migrate_pages is also useful if the processes available memory nodes have changed through cpuset operations to manually move the processes memory. Paul Jackson is working on an automated mechanism that will allow an automatic migration if the cpuset of a process is changed. However, a user may decide to manually control the migration. This implementation is put into the policy layer since it uses concepts and functions that are also needed for mbind and friends. The patch also provides a do_migrate_pages function that may be useful for cpusets to automatically move memory. sys_migrate_pages does not modify policies in contrast to Ray's implementation. The current code here is based on the swap based page migration capability and thus is not able to preserve the physical layout relative to it containing nodeset (which may be a cpuset). When direct page migration becomes available then the implementation needs to be changed to do a isomorphic move of pages between different nodesets. The current implementation simply evicts all pages in source nodeset that are not in the target nodeset. Patch supports ia64, i386 and x86_64. Signed-off-by: Christoph Lameter <clameter@sgi.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-08 09:00:51 +00:00
data8 sys_migrate_pages // 1280
data8 sys_openat
data8 sys_mkdirat
data8 sys_mknodat
data8 sys_fchownat
data8 sys_futimesat // 1285
data8 sys_newfstatat
data8 sys_unlinkat
data8 sys_renameat
data8 sys_linkat
data8 sys_symlinkat // 1290
data8 sys_readlinkat
data8 sys_fchmodat
data8 sys_faccessat
data8 sys_pselect6
data8 sys_ppoll // 1295
data8 sys_unshare
data8 sys_splice
data8 sys_set_robust_list
data8 sys_get_robust_list
data8 sys_sync_file_range // 1300
data8 sys_tee
data8 sys_vmsplice
data8 sys_fallocate
data8 sys_getcpu
data8 sys_epoll_pwait // 1305
data8 sys_utimensat
data8 sys_signalfd
timerfd: new timerfd API This is the new timerfd API as it is implemented by the following patch: int timerfd_create(int clockid, int flags); int timerfd_settime(int ufd, int flags, const struct itimerspec *utmr, struct itimerspec *otmr); int timerfd_gettime(int ufd, struct itimerspec *otmr); The timerfd_create() API creates an un-programmed timerfd fd. The "clockid" parameter can be either CLOCK_MONOTONIC or CLOCK_REALTIME. The timerfd_settime() API give new settings by the timerfd fd, by optionally retrieving the previous expiration time (in case the "otmr" parameter is not NULL). The time value specified in "utmr" is absolute, if the TFD_TIMER_ABSTIME bit is set in the "flags" parameter. Otherwise it's a relative time. The timerfd_gettime() API returns the next expiration time of the timer, or {0, 0} if the timerfd has not been set yet. Like the previous timerfd API implementation, read(2) and poll(2) are supported (with the same interface). Here's a simple test program I used to exercise the new timerfd APIs: http://www.xmailserver.org/timerfd-test2.c [akpm@linux-foundation.org: coding-style cleanups] [akpm@linux-foundation.org: fix ia64 build] [akpm@linux-foundation.org: fix m68k build] [akpm@linux-foundation.org: fix mips build] [akpm@linux-foundation.org: fix alpha, arm, blackfin, cris, m68k, s390, sparc and sparc64 builds] [heiko.carstens@de.ibm.com: fix s390] [akpm@linux-foundation.org: fix powerpc build] [akpm@linux-foundation.org: fix sparc64 more] Signed-off-by: Davide Libenzi <davidel@xmailserver.org> Cc: Michael Kerrisk <mtk-manpages@gmx.net> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Davide Libenzi <davidel@xmailserver.org> Cc: Michael Kerrisk <mtk-manpages@gmx.net> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: Michael Kerrisk <mtk.manpages@gmail.com> Cc: Davide Libenzi <davidel@xmailserver.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-05 06:27:26 +00:00
data8 sys_ni_syscall
data8 sys_eventfd
data8 sys_timerfd_create // 1310
data8 sys_timerfd_settime
data8 sys_timerfd_gettime
data8 sys_signalfd4
data8 sys_eventfd2
data8 sys_epoll_create1 // 1315
data8 sys_dup3
data8 sys_pipe2
data8 sys_inotify_init1
.org sys_call_table + 8*NR_syscalls // guard against failures to increase NR_syscalls
#endif /* __IA64_ASM_PARAVIRTUALIZED_NATIVE */