There are currently two problems related to the emulation of the
instruction BPOSGE32C.
The nanoMIPS instruction BPOSGE32C belongs to DSP R3 instructions
(actually, as of now, it is the only instruction of DSP R3). The
presence of DSP R3 instructions in QEMU is indicated by the flag
MIPS_HFLAG_DSP_R3 (0x20000000). This flag is currently being properly
set in CPUMIPSState's hflags (for example, for I7200 nanoMIPS CPU).
However, it is not propagated to DisasContext's hflags, since the flag
MIPS_HFLAG_DSP_R3 is not set in MIPS_HFLAG_TMASK (while similar flags
MIPS_HFLAG_DSP_R2 and MIPS_HFLAG_DSP are set in this mask, and there
is no problem in functioning check_dsp_r2(), check_dsp()). This means
the function check_dsp_r3() currently does not work properly, and the
emulation of BPOSGE32C can not work properly as well.
Change MIPS_HFLAG_TMASK from 0x1F5807FF to 0x3F5807FF (logical OR
with 0x20000000) to fix this.
Additionally, check_cp1_enabled() is currently incorrectly called
while emulating BPOSGE32C. BPOSGE32C is in the same pool (P.BR1) as
FPU branch instruction BC1EQZC and BC1NEZC, but it not a part of FPU
(CP1) instructions, and check_cp1_enabled() should not be involved
while emulating BPOSGE32C.
Rearrange invocations of check_cp1_enabled() within P.BR1 pool
handling to affect only BC1EQZC and BC1NEZC emulation, and not
BPOSGE32C emulation.
Signed-off-by: Dragan Mladjenovic <dragan.mladjenovic@syrmia.com>
Signed-off-by: Stefan Pejic <stefan.pejic@syrmia.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220504110403.613168-4-stefan.pejic@syrmia.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
The field rs in the instruction EXTRV_S.H rt, ac, rs is specified in
nanoMIPS documentation as opcode[20..16]. It is, however, erroneously
considered as opcode[25..21] in the current QEMU implementation. In
function gen_pool32axf_2_nanomips_insn(), the variable v0_t corresponds
to rt/opcode[25..21], and v1_t corresponds to rs/opcode[20..16]), and
v0_t is by mistake passed to the helper gen_helper_extr_s_h().
Use v1_t rather than v0_t in the invocation of gen_helper_extr_s_h()
to fix this.
Signed-off-by: Dragan Mladjenovic <dragan.mladjenovic@syrmia.com>
Signed-off-by: Stefan Pejic <stefan.pejic@syrmia.com>
Fixes: 8b3698b294 ("target/mips: Add emulation of DSP ASE for nanoMIPS")
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220504110403.613168-3-stefan.pejic@syrmia.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
The field ac in nanoMIPS instruction MTHLIP rs, ac is specified in
nanoMIPS documentation as opcode[15..14] (2 bits). However, in the
current QEMU code, the corresponding argument passed to the helper
gen_helper_mthlip() has the value of opcode[15..11] (5 bits). Right
shift the value of this argument by three bits to fix this.
Signed-off-by: Stefan Pejic <stefan.pejic@syrmia.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220504110403.613168-2-stefan.pejic@syrmia.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Fix the FTRUNC_S and FTRUNC_U trans helper problem.
Fixes: 5c5b64000c ("target/mips: Convert MSA 2RF instruction format to decodetree")
Signed-off-by: nihui <shuizhuyuanluo@126.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220503144241.289239-1-shuizhuyuanluo@126.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
This patch fix the issue that helper_msa_st_b() write high 64bit
data to where the low 64bit resides, leaving high 64bit undefined.
Fixes: 68ad9260e0 ("target/mips: Use 8-byte memory ops for msa load/store")
Signed-off-by: Ni Hui <shuizhuyuanluo@126.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220504023319.12923-1-shuizhuyuanluo@126.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Only for msa COPY_U/COPY_S with wd zero, we treat it as NOP.
Move this special rule into COPY_U and COPY_S trans function.
Fixes: 97fe675519 ("target/mips: Convert MSA COPY_S and INSERT opcodes to decodetree")
Signed-off-by: Ni Hui <shuizhuyuanluo@126.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220503130708.272850-4-shuizhuyuanluo@126.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Fix issue that condition of check_msa_enabled(ctx) is reversed
that causes segfault when msa elm_fn op encountered.
Fixes: 2f2745c81a ("target/mips: Convert MSA COPY_U opcode to decodetree")
Fixes: 97fe675519 ("target/mips: Convert MSA COPY_S and INSERT opcodes to decodetree")
Signed-off-by: Ni Hui <shuizhuyuanluo@126.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220503130708.272850-3-shuizhuyuanluo@126.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Actually look into dfe structure data so that df_extract_val() and
df_extract_df() can return immediate and datafield other than BYTE.
Fixes: 4701d23aef ("target/mips: Convert MSA BIT instruction format to decodetree")
Signed-off-by: Ni Hui <shuizhuyuanluo@126.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220503130708.272850-2-shuizhuyuanluo@126.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Fix the SAT_S and SAT_U trans helper confusion.
Fixes: 4701d23aef ("target/mips: Convert MSA BIT instruction format to decodetree")
Signed-off-by: Ni Hui <shuizhuyuanluo@126.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20220503130708.272850-1-shuizhuyuanluo@126.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
bit 31 (M) of WatchHiN register is a read-only register indicating
whether the next WatchHi register is present. It must not be reset
during user writes to the register.
Signed-off-by: Marcin Nowakowski <marcin.nowakowski@fungible.com>
Reviewed-by: David Daney <david.daney@fungible.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@fungible.com>
Message-Id: <20220511212953.74738-1-philmd@fungible.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Since DDI0487F.a, the RW bit is RAO/WI. When specifically
targeting such a cpu, e.g. cortex-a76, it is legitimate to
ignore the bit within the secure monitor.
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1062
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609214657.1217913-3-richard.henderson@linaro.org
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Because reset always initializes the AA64 version, SCR_EL3,
test the mode of EL3 instead of the type of the cpreg.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609214657.1217913-2-richard.henderson@linaro.org
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
We were using arm_is_secure and is_a64, which are
tests against the current EL, as opposed to
arm_el_is_aa64 and arm_is_secure_below_el3, which
can be applied to a different EL than current.
Consolidate the two tests.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609202901.1177572-24-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
This function is no longer used outside debug_helper.c.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609202901.1177572-23-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Handle the debug vs current el exception test in one place.
Leave EXCP_BKPT alone, since that treats debug < current differently.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609202901.1177572-22-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
This function is no longer used. At the same time, remove
DisasContext.secure_routed_to_el3, as it in turn becomes unused.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609202901.1177572-21-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
With the helper we can use exception_target_el at runtime,
instead of default_exception_el at translate time.
While we're at it, remove the DisasContext parameter from
gen_exception, as it is no longer used.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609202901.1177572-20-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Split out a common helper function for gen_exception_el
and gen_exception_insn_el_v.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609202901.1177572-19-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Create a new wrapper function that passes the default
exception target to gen_exception_el.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609202901.1177572-18-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609202901.1177572-17-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
This function is not required by any other translation file.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609202901.1177572-16-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
We no longer need this value during translation,
as it is now handled within the helpers.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609202901.1177572-15-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Move the computation from gen_swstep_exception into a helper.
This fixes a bug when:
- MDSCR_EL1.KDE == 1 to enable debug exceptions within EL_D itself
- we singlestep an ERET from EL_D to some lower EL
Previously we were computing 'same el' based on the EL which
executed the ERET instruction, whereas it ought to be computed
based on the EL to which ERET returned. This happens naturally
with the new helper, which runs after EL has been changed.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609202901.1177572-14-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Create a new wrapper function that passes the default
exception target to gen_exception_insn_el.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609202901.1177572-13-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609202901.1177572-12-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Create a function below gen_exception_insn that takes
the target_el as a TCGv_i32, replacing gen_exception_el.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609202901.1177572-11-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Rename to helper_exception_with_syndrome_el, to emphasize
that the target el is a parameter.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609202901.1177572-10-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
This function now now only used in debug_helper.c, so there is
no reason to have a declaration in a header.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609202901.1177572-9-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609202901.1177572-8-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Use the accessor rather than the raw structure member.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609202901.1177572-7-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Move arm_generate_debug_exceptions and its two subroutines,
{aa32,aa64}_generate_debug_exceptions into debug_helper.c,
and the one interface declaration to internals.h.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609202901.1177572-6-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Move the function to debug_helper.c, and the
declaration to internals.h.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609202901.1177572-5-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Move the function to op_helper.c, near raise_exception.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609202901.1177572-4-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
With ARMv8, this field is always RES0.
With ARMv7, targeting EL2 and TA=0, it is always 0xA.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609202901.1177572-3-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20220609202901.1177572-2-richard.henderson@linaro.org
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
When running a 32-bit guest, with a e64 vmv.v.x and vl_eq_vlmax set to
true the `tcg_debug_assert(vece <= MO_32)` will be triggered inside
tcg_gen_gvec_dup_i32().
This patch checks that condition and instead uses tcg_gen_gvec_dup_i64()
is required.
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/1028
Suggested-by: Robert Bu <robert.bu@gmail.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20220608234701.369536-1-alistair.francis@opensource.wdc.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
There are currently two types of RISC-V CPUs:
- Generic CPUs (base or any) that allow complete custimisation
- "Named" CPUs that match existing hardware
Users can use the base CPUs to custimise the extensions that they want, for
example -cpu rv64,v=true.
We originally exposed these as part of the named CPUs as well, but that was
by accident.
Exposing the CPU properties to named CPUs means that we accidently
enable extensions that don't exist on the CPUs by default. For example
the SiFive E CPU currently support the zba extension, which is a bug.
This patch instead only exposes the CPU extensions to the generic CPUs.
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Message-Id: <20220608061437.314434-1-alistair.francis@opensource.wdc.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
According to v-spec, tail agnostic behavior can be either kept as
undisturbed or set elements' bits to all 1s. To distinguish the
difference of tail policies, QEMU should be able to simulate the tail
agnostic behavior as "set tail elements' bits to all 1s".
There are multiple possibility for agnostic elements according to
v-spec. The main intent of this patch-set tries to add option that
can distinguish between tail policies. Setting agnostic elements to
all 1s allows QEMU to express this.
This commit adds option 'rvv_ta_all_1s' is added to enable the
behavior, it is default as disabled.
Signed-off-by: eop Chen <eop.chen@sifive.com>
Reviewed-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Weiwei Li <liweiwei@iscas.ac.cn>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <165449614532.19704.7000832880482980398-16@git.sr.ht>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
The tail elements in the destination mask register are updated under
a tail-agnostic policy.
Signed-off-by: eop Chen <eop.chen@sifive.com>
Reviewed-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Weiwei Li <liweiwei@iscas.ac.cn>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <165449614532.19704.7000832880482980398-14@git.sr.ht>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Compares write mask registers, and so always operate under a tail-
agnostic policy.
Signed-off-by: eop Chen <eop.chen@sifive.com>
Reviewed-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Weiwei Li <liweiwei@iscas.ac.cn>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <165449614532.19704.7000832880482980398-12@git.sr.ht>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Compares write mask registers, and so always operate under a tail-
agnostic policy.
Signed-off-by: eop Chen <eop.chen@sifive.com>
Reviewed-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Weiwei Li <liweiwei@iscas.ac.cn>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <165449614532.19704.7000832880482980398-9@git.sr.ht>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
`vmadc` and `vmsbc` produces a mask value, they always operate with
a tail agnostic policy.
Signed-off-by: eop Chen <eop.chen@sifive.com>
Reviewed-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Weiwei Li <liweiwei@iscas.ac.cn>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <165449614532.19704.7000832880482980398-7@git.sr.ht>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Destination register of unit-stride mask load and store instructions are
always written with a tail-agnostic policy.
A vector segment load / store instruction may contain fractional lmul
with nf * lmul > 1. The rest of the elements in the last register should
be treated as tail elements.
Signed-off-by: eop Chen <eop.chen@sifive.com>
Reviewed-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Weiwei Li <liweiwei@iscas.ac.cn>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <165449614532.19704.7000832880482980398-6@git.sr.ht>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
According to v-spec, tail agnostic behavior can be either kept as
undisturbed or set elements' bits to all 1s. To distinguish the
difference of tail policies, QEMU should be able to simulate the tail
agnostic behavior as "set tail elements' bits to all 1s".
There are multiple possibility for agnostic elements according to
v-spec. The main intent of this patch-set tries to add option that
can distinguish between tail policies. Setting agnostic elements to
all 1s allows QEMU to express this.
This is the first commit regarding the optional tail agnostic
behavior. Follow-up commits will add this optional behavior
for all rvv instructions.
Signed-off-by: eop Chen <eop.chen@sifive.com>
Reviewed-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Weiwei Li <liweiwei@iscas.ac.cn>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <165449614532.19704.7000832880482980398-5@git.sr.ht>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
According to v-spec (section 5.4):
When vstart ≥ vl, there are no body elements, and no elements are
updated in any destination vector register group, including that
no tail elements are updated with agnostic values.
vmsbf.m, vmsif.m, vmsof.m, viota.m, vcompress instructions themselves
require vstart to be zero. So they don't need the early exit.
Signed-off-by: eop Chen <eop.chen@sifive.com>
Reviewed-by: Frank Chang <frank.chang@sifive.com>
Reviewed-by: Weiwei Li <liweiwei@iscas.ac.cn>
Acked-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <165449614532.19704.7000832880482980398-4@git.sr.ht>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
