qemu/target/arm/translate-m-nocp.c
Peter Maydell 5138bd0143 target/arm: Add handling for PSR.ECI/ICI
On A-profile, PSR bits [15:10][26:25] are always the IT state bits.
On M-profile, some of the reserved encodings of the IT state are used
to instead indicate partial progress through instructions that were
interrupted partway through by an exception and can be resumed.

These resumable instructions fall into two categories:

(1) load/store multiple instructions, where these bits are called
"ICI" and specify the register in the ldm/stm list where execution
should resume.  (Specifically: LDM, STM, VLDM, VSTM, VLLDM, VLSTM,
CLRM, VSCCLRM.)

(2) MVE instructions subject to beatwise execution, where these bits
are called "ECI" and specify which beats in this and possibly also
the following MVE insn have been executed.

There are also a few insns (LE, LETP, and BKPT) which do not use the
ICI/ECI bits but must leave them alone.

Otherwise, we should raise an INVSTATE UsageFault for any attempt to
execute an insn with non-zero ICI/ECI bits.

So far we have been able to ignore ECI/ICI, because the architecture
allows the IMPDEF choice of "always restart load/store multiple from
the beginning regardless of ICI state", so the only thing we have
been missing is that we don't raise the INVSTATE fault for bad guest
code.  However, MVE requires that we honour ECI bits and do not
rexecute beats of an insn that have already been executed.

Add the support in the decoder for handling ECI/ICI:
 * identify the ECI/ICI case in the CONDEXEC TB flags
 * when a load/store multiple insn succeeds, it updates the ECI/ICI
   state (both in DisasContext and in the CPU state), and sets a flag
   to say that the ECI/ICI state was handled
 * if we find that the insn we just decoded did not handle the
   ECI/ICI state, we delete all the code that we just generated for
   it and instead emit the code to raise the INVFAULT.  This allows
   us to avoid having to update every non-MVE non-LDM/STM insn to
   make it check for "is ECI/ICI set?".

We continue with our existing IMPDEF choice of not caring about the
ICI state for the load/store multiples and simply restarting them
from the beginning.  Because we don't allow interrupts in the middle
of an insn, the only way we would see this state is if the guest set
ICI manually on return from an exception handler, so it's a corner
case which doesn't merit optimisation.

ICI update for LDM/STM is simple -- it always zeroes the state.  ECI
update for MVE beatwise insns will be a little more complex, since
the ECI state may include information for the following insn.

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20210614151007.4545-5-peter.maydell@linaro.org
2021-06-16 14:33:52 +01:00

236 lines
6.9 KiB
C

/*
* ARM translation: M-profile NOCP special-case instructions
*
* Copyright (c) 2020 Linaro, Ltd.
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, see <http://www.gnu.org/licenses/>.
*/
#include "qemu/osdep.h"
#include "tcg/tcg-op.h"
#include "translate.h"
#include "translate-a32.h"
#include "decode-m-nocp.c.inc"
/*
* Decode VLLDM and VLSTM are nonstandard because:
* * if there is no FPU then these insns must NOP in
* Secure state and UNDEF in Nonsecure state
* * if there is an FPU then these insns do not have
* the usual behaviour that vfp_access_check() provides of
* being controlled by CPACR/NSACR enable bits or the
* lazy-stacking logic.
*/
static bool trans_VLLDM_VLSTM(DisasContext *s, arg_VLLDM_VLSTM *a)
{
TCGv_i32 fptr;
if (!arm_dc_feature(s, ARM_FEATURE_M) ||
!arm_dc_feature(s, ARM_FEATURE_V8)) {
return false;
}
if (a->op) {
/*
* T2 encoding ({D0-D31} reglist): v8.1M and up. We choose not
* to take the IMPDEF option to make memory accesses to the stack
* slots that correspond to the D16-D31 registers (discarding
* read data and writing UNKNOWN values), so for us the T2
* encoding behaves identically to the T1 encoding.
*/
if (!arm_dc_feature(s, ARM_FEATURE_V8_1M)) {
return false;
}
} else {
/*
* T1 encoding ({D0-D15} reglist); undef if we have 32 Dregs.
* This is currently architecturally impossible, but we add the
* check to stay in line with the pseudocode. Note that we must
* emit code for the UNDEF so it takes precedence over the NOCP.
*/
if (dc_isar_feature(aa32_simd_r32, s)) {
unallocated_encoding(s);
return true;
}
}
/*
* If not secure, UNDEF. We must emit code for this
* rather than returning false so that this takes
* precedence over the m-nocp.decode NOCP fallback.
*/
if (!s->v8m_secure) {
unallocated_encoding(s);
return true;
}
s->eci_handled = true;
/* If no fpu, NOP. */
if (!dc_isar_feature(aa32_vfp, s)) {
clear_eci_state(s);
return true;
}
fptr = load_reg(s, a->rn);
if (a->l) {
gen_helper_v7m_vlldm(cpu_env, fptr);
} else {
gen_helper_v7m_vlstm(cpu_env, fptr);
}
tcg_temp_free_i32(fptr);
clear_eci_state(s);
/* End the TB, because we have updated FP control bits */
s->base.is_jmp = DISAS_UPDATE_EXIT;
return true;
}
static bool trans_VSCCLRM(DisasContext *s, arg_VSCCLRM *a)
{
int btmreg, topreg;
TCGv_i64 zero;
TCGv_i32 aspen, sfpa;
if (!dc_isar_feature(aa32_m_sec_state, s)) {
/* Before v8.1M, fall through in decode to NOCP check */
return false;
}
/* Explicitly UNDEF because this takes precedence over NOCP */
if (!arm_dc_feature(s, ARM_FEATURE_M_MAIN) || !s->v8m_secure) {
unallocated_encoding(s);
return true;
}
s->eci_handled = true;
if (!dc_isar_feature(aa32_vfp_simd, s)) {
/* NOP if we have neither FP nor MVE */
clear_eci_state(s);
return true;
}
/*
* If FPCCR.ASPEN != 0 && CONTROL_S.SFPA == 0 then there is no
* active floating point context so we must NOP (without doing
* any lazy state preservation or the NOCP check).
*/
aspen = load_cpu_field(v7m.fpccr[M_REG_S]);
sfpa = load_cpu_field(v7m.control[M_REG_S]);
tcg_gen_andi_i32(aspen, aspen, R_V7M_FPCCR_ASPEN_MASK);
tcg_gen_xori_i32(aspen, aspen, R_V7M_FPCCR_ASPEN_MASK);
tcg_gen_andi_i32(sfpa, sfpa, R_V7M_CONTROL_SFPA_MASK);
tcg_gen_or_i32(sfpa, sfpa, aspen);
arm_gen_condlabel(s);
tcg_gen_brcondi_i32(TCG_COND_EQ, sfpa, 0, s->condlabel);
if (s->fp_excp_el != 0) {
gen_exception_insn(s, s->pc_curr, EXCP_NOCP,
syn_uncategorized(), s->fp_excp_el);
return true;
}
topreg = a->vd + a->imm - 1;
btmreg = a->vd;
/* Convert to Sreg numbers if the insn specified in Dregs */
if (a->size == 3) {
topreg = topreg * 2 + 1;
btmreg *= 2;
}
if (topreg > 63 || (topreg > 31 && !(topreg & 1))) {
/* UNPREDICTABLE: we choose to undef */
unallocated_encoding(s);
return true;
}
/* Silently ignore requests to clear D16-D31 if they don't exist */
if (topreg > 31 && !dc_isar_feature(aa32_simd_r32, s)) {
topreg = 31;
}
if (!vfp_access_check(s)) {
return true;
}
/* Zero the Sregs from btmreg to topreg inclusive. */
zero = tcg_const_i64(0);
if (btmreg & 1) {
write_neon_element64(zero, btmreg >> 1, 1, MO_32);
btmreg++;
}
for (; btmreg + 1 <= topreg; btmreg += 2) {
write_neon_element64(zero, btmreg >> 1, 0, MO_64);
}
if (btmreg == topreg) {
write_neon_element64(zero, btmreg >> 1, 0, MO_32);
btmreg++;
}
assert(btmreg == topreg + 1);
if (dc_isar_feature(aa32_mve, s)) {
TCGv_i32 z32 = tcg_const_i32(0);
store_cpu_field(z32, v7m.vpr);
}
clear_eci_state(s);
return true;
}
static bool trans_NOCP(DisasContext *s, arg_nocp *a)
{
/*
* Handle M-profile early check for disabled coprocessor:
* all we need to do here is emit the NOCP exception if
* the coprocessor is disabled. Otherwise we return false
* and the real VFP/etc decode will handle the insn.
*/
assert(arm_dc_feature(s, ARM_FEATURE_M));
if (a->cp == 11) {
a->cp = 10;
}
if (arm_dc_feature(s, ARM_FEATURE_V8_1M) &&
(a->cp == 8 || a->cp == 9 || a->cp == 14 || a->cp == 15)) {
/* in v8.1M cp 8, 9, 14, 15 also are governed by the cp10 enable */
a->cp = 10;
}
if (a->cp != 10) {
gen_exception_insn(s, s->pc_curr, EXCP_NOCP,
syn_uncategorized(), default_exception_el(s));
return true;
}
if (s->fp_excp_el != 0) {
gen_exception_insn(s, s->pc_curr, EXCP_NOCP,
syn_uncategorized(), s->fp_excp_el);
return true;
}
return false;
}
static bool trans_NOCP_8_1(DisasContext *s, arg_nocp *a)
{
/* This range needs a coprocessor check for v8.1M and later only */
if (!arm_dc_feature(s, ARM_FEATURE_V8_1M)) {
return false;
}
return trans_NOCP(s, a);
}