kernel_optimize_test/arch/ia64/kernel/pal.S
Zou Nan hai acb15c85de [IA64] Do not assume output registers be reservered.
We found an issue in pal.S.

According to the software runtime SPEC,
The caller's output registers do not need to be preserved for
caller. The callee may reuse input registers for any other
purpose within the procedure.

in ia64_pal_call_phys_stacked,

input registers are copied to output registers before call
into ia64_switch_mode_phys, then used to call into PAL. This
assumes output registers are preserved in ia64_switch_mode_phys,
which may not be true.

In this particular case, ia64_switch_mode_phys alloc a null frame
, and mask off psr.i.
If an interrupt comes at this small window,
or an MCA comes inside the procedure, output registers
maybe changed,
then the pal call may got some staled input registers.

This patch moves the copies from input to output
after ia64_switch_mode_phys to follow the software
runtime convention.

It  also removed some unused labels in
ia64_pal_call_phys_stacked.

Signed-off-by: Zou Nan hai <nanhai.zou@intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
2006-08-02 16:12:08 -07:00

303 lines
7.5 KiB
ArmAsm

/*
* PAL Firmware support
* IA-64 Processor Programmers Reference Vol 2
*
* Copyright (C) 1999 Don Dugger <don.dugger@intel.com>
* Copyright (C) 1999 Walt Drummond <drummond@valinux.com>
* Copyright (C) 1999-2001, 2003 Hewlett-Packard Co
* David Mosberger <davidm@hpl.hp.com>
* Stephane Eranian <eranian@hpl.hp.com>
*
* 05/22/2000 eranian Added support for stacked register calls
* 05/24/2000 eranian Added support for physical mode static calls
*/
#include <asm/asmmacro.h>
#include <asm/processor.h>
.data
pal_entry_point:
data8 ia64_pal_default_handler
.text
/*
* Set the PAL entry point address. This could be written in C code, but we do it here
* to keep it all in one module (besides, it's so trivial that it's
* not a big deal).
*
* in0 Address of the PAL entry point (text address, NOT a function descriptor).
*/
GLOBAL_ENTRY(ia64_pal_handler_init)
alloc r3=ar.pfs,1,0,0,0
movl r2=pal_entry_point
;;
st8 [r2]=in0
br.ret.sptk.many rp
END(ia64_pal_handler_init)
/*
* Default PAL call handler. This needs to be coded in assembly because it uses
* the static calling convention, i.e., the RSE may not be used and calls are
* done via "br.cond" (not "br.call").
*/
GLOBAL_ENTRY(ia64_pal_default_handler)
mov r8=-1
br.cond.sptk.many rp
END(ia64_pal_default_handler)
/*
* Make a PAL call using the static calling convention.
*
* in0 Index of PAL service
* in1 - in3 Remaining PAL arguments
* in4 1 ==> clear psr.ic, 0 ==> don't clear psr.ic
*
*/
GLOBAL_ENTRY(ia64_pal_call_static)
.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(5)
alloc loc1 = ar.pfs,5,5,0,0
movl loc2 = pal_entry_point
1: {
mov r28 = in0
mov r29 = in1
mov r8 = ip
}
;;
ld8 loc2 = [loc2] // loc2 <- entry point
tbit.nz p6,p7 = in4, 0
adds r8 = 1f-1b,r8
mov loc4=ar.rsc // save RSE configuration
;;
mov ar.rsc=0 // put RSE in enforced lazy, LE mode
mov loc3 = psr
mov loc0 = rp
.body
mov r30 = in2
(p6) rsm psr.i | psr.ic
mov r31 = in3
mov b7 = loc2
(p7) rsm psr.i
;;
(p6) srlz.i
mov rp = r8
br.cond.sptk.many b7
1: mov psr.l = loc3
mov ar.rsc = loc4 // restore RSE configuration
mov ar.pfs = loc1
mov rp = loc0
;;
srlz.d // seralize restoration of psr.l
br.ret.sptk.many b0
END(ia64_pal_call_static)
/*
* Make a PAL call using the stacked registers calling convention.
*
* Inputs:
* in0 Index of PAL service
* in2 - in3 Remaning PAL arguments
*/
GLOBAL_ENTRY(ia64_pal_call_stacked)
.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(4)
alloc loc1 = ar.pfs,4,4,4,0
movl loc2 = pal_entry_point
mov r28 = in0 // Index MUST be copied to r28
mov out0 = in0 // AND in0 of PAL function
mov loc0 = rp
.body
;;
ld8 loc2 = [loc2] // loc2 <- entry point
mov out1 = in1
mov out2 = in2
mov out3 = in3
mov loc3 = psr
;;
rsm psr.i
mov b7 = loc2
;;
br.call.sptk.many rp=b7 // now make the call
.ret0: mov psr.l = loc3
mov ar.pfs = loc1
mov rp = loc0
;;
srlz.d // serialize restoration of psr.l
br.ret.sptk.many b0
END(ia64_pal_call_stacked)
/*
* Make a physical mode PAL call using the static registers calling convention.
*
* Inputs:
* in0 Index of PAL service
* in2 - in3 Remaning PAL arguments
*
* PSR_LP, PSR_TB, PSR_ID, PSR_DA are never set by the kernel.
* So we don't need to clear them.
*/
#define PAL_PSR_BITS_TO_CLEAR \
(IA64_PSR_I | IA64_PSR_IT | IA64_PSR_DT | IA64_PSR_DB | IA64_PSR_RT | \
IA64_PSR_DD | IA64_PSR_SS | IA64_PSR_RI | IA64_PSR_ED | \
IA64_PSR_DFL | IA64_PSR_DFH)
#define PAL_PSR_BITS_TO_SET \
(IA64_PSR_BN)
GLOBAL_ENTRY(ia64_pal_call_phys_static)
.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(4)
alloc loc1 = ar.pfs,4,7,0,0
movl loc2 = pal_entry_point
1: {
mov r28 = in0 // copy procedure index
mov r8 = ip // save ip to compute branch
mov loc0 = rp // save rp
}
.body
;;
ld8 loc2 = [loc2] // loc2 <- entry point
mov r29 = in1 // first argument
mov r30 = in2 // copy arg2
mov r31 = in3 // copy arg3
;;
mov loc3 = psr // save psr
adds r8 = 1f-1b,r8 // calculate return address for call
;;
mov loc4=ar.rsc // save RSE configuration
dep.z loc2=loc2,0,61 // convert pal entry point to physical
tpa r8=r8 // convert rp to physical
;;
mov b7 = loc2 // install target to branch reg
mov ar.rsc=0 // put RSE in enforced lazy, LE mode
movl r16=PAL_PSR_BITS_TO_CLEAR
movl r17=PAL_PSR_BITS_TO_SET
;;
or loc3=loc3,r17 // add in psr the bits to set
;;
andcm r16=loc3,r16 // removes bits to clear from psr
br.call.sptk.many rp=ia64_switch_mode_phys
.ret1: mov rp = r8 // install return address (physical)
mov loc5 = r19
mov loc6 = r20
br.cond.sptk.many b7
1:
mov ar.rsc=0 // put RSE in enforced lazy, LE mode
mov r16=loc3 // r16= original psr
mov r19=loc5
mov r20=loc6
br.call.sptk.many rp=ia64_switch_mode_virt // return to virtual mode
.ret2:
mov psr.l = loc3 // restore init PSR
mov ar.pfs = loc1
mov rp = loc0
;;
mov ar.rsc=loc4 // restore RSE configuration
srlz.d // seralize restoration of psr.l
br.ret.sptk.many b0
END(ia64_pal_call_phys_static)
/*
* Make a PAL call using the stacked registers in physical mode.
*
* Inputs:
* in0 Index of PAL service
* in2 - in3 Remaning PAL arguments
*/
GLOBAL_ENTRY(ia64_pal_call_phys_stacked)
.prologue ASM_UNW_PRLG_RP|ASM_UNW_PRLG_PFS, ASM_UNW_PRLG_GRSAVE(5)
alloc loc1 = ar.pfs,5,7,4,0
movl loc2 = pal_entry_point
1: {
mov r28 = in0 // copy procedure index
mov loc0 = rp // save rp
}
.body
;;
ld8 loc2 = [loc2] // loc2 <- entry point
mov loc3 = psr // save psr
;;
mov loc4=ar.rsc // save RSE configuration
dep.z loc2=loc2,0,61 // convert pal entry point to physical
;;
mov ar.rsc=0 // put RSE in enforced lazy, LE mode
movl r16=PAL_PSR_BITS_TO_CLEAR
movl r17=PAL_PSR_BITS_TO_SET
;;
or loc3=loc3,r17 // add in psr the bits to set
mov b7 = loc2 // install target to branch reg
;;
andcm r16=loc3,r16 // removes bits to clear from psr
br.call.sptk.many rp=ia64_switch_mode_phys
mov out0 = in0 // first argument
mov out1 = in1 // copy arg2
mov out2 = in2 // copy arg3
mov out3 = in3 // copy arg3
mov loc5 = r19
mov loc6 = r20
br.call.sptk.many rp=b7 // now make the call
mov ar.rsc=0 // put RSE in enforced lazy, LE mode
mov r16=loc3 // r16= original psr
mov r19=loc5
mov r20=loc6
br.call.sptk.many rp=ia64_switch_mode_virt // return to virtual mode
mov psr.l = loc3 // restore init PSR
mov ar.pfs = loc1
mov rp = loc0
;;
mov ar.rsc=loc4 // restore RSE configuration
srlz.d // seralize restoration of psr.l
br.ret.sptk.many b0
END(ia64_pal_call_phys_stacked)
/*
* Save scratch fp scratch regs which aren't saved in pt_regs already (fp10-fp15).
*
* NOTE: We need to do this since firmware (SAL and PAL) may use any of the scratch
* regs fp-low partition.
*
* Inputs:
* in0 Address of stack storage for fp regs
*/
GLOBAL_ENTRY(ia64_save_scratch_fpregs)
alloc r3=ar.pfs,1,0,0,0
add r2=16,in0
;;
stf.spill [in0] = f10,32
stf.spill [r2] = f11,32
;;
stf.spill [in0] = f12,32
stf.spill [r2] = f13,32
;;
stf.spill [in0] = f14,32
stf.spill [r2] = f15,32
br.ret.sptk.many rp
END(ia64_save_scratch_fpregs)
/*
* Load scratch fp scratch regs (fp10-fp15)
*
* Inputs:
* in0 Address of stack storage for fp regs
*/
GLOBAL_ENTRY(ia64_load_scratch_fpregs)
alloc r3=ar.pfs,1,0,0,0
add r2=16,in0
;;
ldf.fill f10 = [in0],32
ldf.fill f11 = [r2],32
;;
ldf.fill f12 = [in0],32
ldf.fill f13 = [r2],32
;;
ldf.fill f14 = [in0],32
ldf.fill f15 = [r2],32
br.ret.sptk.many rp
END(ia64_load_scratch_fpregs)