kernel_optimize_test/arch/sparc/kernel/winfixup.S
Babu Moger 0ae2d26ffe arch/sparc: Avoid DCTI Couples
Avoid un-intended DCTI Couples. Use of DCTI couples is deprecated.
Also address the "Programming Note" for optimal performance.

Here is the complete text from Oracle SPARC Architecture Specs.

6.3.4.7 DCTI Couples
"A delayed control transfer instruction (DCTI) in the delay slot of
another DCTI is referred to as a “DCTI couple”. The use of DCTI couples
is deprecated in the Oracle SPARC Architecture; no new software should
place a DCTI in the delay slot of another DCTI, because on future Oracle
SPARC Architecture implementations DCTI couples may execute either
slowly or differently than the programmer assumes it will.

SPARC V8 and SPARC V9 Compatibility Note
The SPARC V8 architecture left behavior undefined for a DCTI couple. The
SPARC V9 architecture defined behavior in that case, but as of
UltraSPARC Architecture 2005, use of DCTI couples was deprecated.
Software should not expect high performance from DCTI couples, and
performance of DCTI couples should be expected to decline further in
future processors.

Programming Note
As noted in TABLE 6-5 on page 115, an annulled branch-always
(branch-always with a = 1) instruction is not architecturally a DCTI.
However, since not all implementations make that distinction, for
optimal performance, a DCTI should not be placed in the instruction word
immediately following an annulled branch-always instruction (BA,A or
BPA,A)."

Signed-off-by: Babu Moger <babu.moger@oracle.com>
Reviewed-by: Rob Gardner <rob.gardner@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2017-03-27 21:51:40 -07:00

160 lines
3.9 KiB
ArmAsm

/* winfixup.S: Handle cases where user stack pointer is found to be bogus.
*
* Copyright (C) 1997, 2006 David S. Miller (davem@davemloft.net)
*/
#include <asm/asi.h>
#include <asm/head.h>
#include <asm/page.h>
#include <asm/ptrace.h>
#include <asm/processor.h>
#include <asm/spitfire.h>
#include <asm/thread_info.h>
.text
/* It used to be the case that these register window fault
* handlers could run via the save and restore instructions
* done by the trap entry and exit code. They now do the
* window spill/fill by hand, so that case no longer can occur.
*/
.align 32
fill_fixup:
TRAP_LOAD_THREAD_REG(%g6, %g1)
rdpr %tstate, %g1
and %g1, TSTATE_CWP, %g1
or %g4, FAULT_CODE_WINFIXUP, %g4
stb %g4, [%g6 + TI_FAULT_CODE]
stx %g5, [%g6 + TI_FAULT_ADDR]
wrpr %g1, %cwp
ba,pt %xcc, etrap
rd %pc, %g7
call do_sparc64_fault
add %sp, PTREGS_OFF, %o0
ba,a,pt %xcc, rtrap
/* Be very careful about usage of the trap globals here.
* You cannot touch %g5 as that has the fault information.
*/
spill_fixup:
spill_fixup_mna:
spill_fixup_dax:
TRAP_LOAD_THREAD_REG(%g6, %g1)
ldx [%g6 + TI_FLAGS], %g1
andcc %sp, 0x1, %g0
movne %icc, 0, %g1
andcc %g1, _TIF_32BIT, %g0
ldub [%g6 + TI_WSAVED], %g1
sll %g1, 3, %g3
add %g6, %g3, %g3
stx %sp, [%g3 + TI_RWIN_SPTRS]
sll %g1, 7, %g3
bne,pt %xcc, 1f
add %g6, %g3, %g3
stx %l0, [%g3 + TI_REG_WINDOW + 0x00]
stx %l1, [%g3 + TI_REG_WINDOW + 0x08]
stx %l2, [%g3 + TI_REG_WINDOW + 0x10]
stx %l3, [%g3 + TI_REG_WINDOW + 0x18]
stx %l4, [%g3 + TI_REG_WINDOW + 0x20]
stx %l5, [%g3 + TI_REG_WINDOW + 0x28]
stx %l6, [%g3 + TI_REG_WINDOW + 0x30]
stx %l7, [%g3 + TI_REG_WINDOW + 0x38]
stx %i0, [%g3 + TI_REG_WINDOW + 0x40]
stx %i1, [%g3 + TI_REG_WINDOW + 0x48]
stx %i2, [%g3 + TI_REG_WINDOW + 0x50]
stx %i3, [%g3 + TI_REG_WINDOW + 0x58]
stx %i4, [%g3 + TI_REG_WINDOW + 0x60]
stx %i5, [%g3 + TI_REG_WINDOW + 0x68]
stx %i6, [%g3 + TI_REG_WINDOW + 0x70]
ba,pt %xcc, 2f
stx %i7, [%g3 + TI_REG_WINDOW + 0x78]
1: stw %l0, [%g3 + TI_REG_WINDOW + 0x00]
stw %l1, [%g3 + TI_REG_WINDOW + 0x04]
stw %l2, [%g3 + TI_REG_WINDOW + 0x08]
stw %l3, [%g3 + TI_REG_WINDOW + 0x0c]
stw %l4, [%g3 + TI_REG_WINDOW + 0x10]
stw %l5, [%g3 + TI_REG_WINDOW + 0x14]
stw %l6, [%g3 + TI_REG_WINDOW + 0x18]
stw %l7, [%g3 + TI_REG_WINDOW + 0x1c]
stw %i0, [%g3 + TI_REG_WINDOW + 0x20]
stw %i1, [%g3 + TI_REG_WINDOW + 0x24]
stw %i2, [%g3 + TI_REG_WINDOW + 0x28]
stw %i3, [%g3 + TI_REG_WINDOW + 0x2c]
stw %i4, [%g3 + TI_REG_WINDOW + 0x30]
stw %i5, [%g3 + TI_REG_WINDOW + 0x34]
stw %i6, [%g3 + TI_REG_WINDOW + 0x38]
stw %i7, [%g3 + TI_REG_WINDOW + 0x3c]
2: add %g1, 1, %g1
stb %g1, [%g6 + TI_WSAVED]
rdpr %tstate, %g1
andcc %g1, TSTATE_PRIV, %g0
saved
be,pn %xcc, 1f
and %g1, TSTATE_CWP, %g1
retry
1: mov FAULT_CODE_WRITE | FAULT_CODE_DTLB | FAULT_CODE_WINFIXUP, %g4
stb %g4, [%g6 + TI_FAULT_CODE]
stx %g5, [%g6 + TI_FAULT_ADDR]
wrpr %g1, %cwp
ba,pt %xcc, etrap
rd %pc, %g7
call do_sparc64_fault
add %sp, PTREGS_OFF, %o0
ba,a,pt %xcc, rtrap
winfix_mna:
andn %g3, 0x7f, %g3
add %g3, 0x78, %g3
wrpr %g3, %tnpc
done
fill_fixup_mna:
rdpr %tstate, %g1
and %g1, TSTATE_CWP, %g1
wrpr %g1, %cwp
ba,pt %xcc, etrap
rd %pc, %g7
sethi %hi(tlb_type), %g1
lduw [%g1 + %lo(tlb_type)], %g1
cmp %g1, 3
bne,pt %icc, 1f
add %sp, PTREGS_OFF, %o0
mov %l4, %o2
call sun4v_do_mna
mov %l5, %o1
ba,a,pt %xcc, rtrap
1: mov %l4, %o1
mov %l5, %o2
call mem_address_unaligned
nop
ba,a,pt %xcc, rtrap
winfix_dax:
andn %g3, 0x7f, %g3
add %g3, 0x74, %g3
wrpr %g3, %tnpc
done
fill_fixup_dax:
rdpr %tstate, %g1
and %g1, TSTATE_CWP, %g1
wrpr %g1, %cwp
ba,pt %xcc, etrap
rd %pc, %g7
sethi %hi(tlb_type), %g1
mov %l4, %o1
lduw [%g1 + %lo(tlb_type)], %g1
mov %l5, %o2
cmp %g1, 3
bne,pt %icc, 1f
add %sp, PTREGS_OFF, %o0
call sun4v_data_access_exception
nop
ba,a,pt %xcc, rtrap
nop
1: call spitfire_data_access_exception
nop
ba,a,pt %xcc, rtrap
nop