The following improvements are made for predicated HVX instructions
During gen_commit_hvx, unconditionally move the "new" value into
the dest
Don't set slot_cancelled
Remove runtime bookkeeping of which registers were updated
Reduce the cases where gen_log_vreg_write[_pair] is called
It's only needed for special operands VxxV and VyV
Remove gen_log_qreg_write
Signed-off-by: Taylor Simpson <tsimpson@quicinc.com>
Reviewed-by: Anton Johansson <anjo@rev.ng>
Message-Id: <20230307025828.1612809-15-tsimpson@quicinc.com>
We only need to track slot for predicated stores and predicated HVX
instructions.
Add arguments to the probe helper functions to indicate if the slot
is predicated.
Here is a simple example of the differences in the TCG code generated:
IN:
0x00400094: 0xf900c102 { if (P0) R2 = and(R0,R1) }
BEFORE
---- 00400094
mov_i32 slot_cancelled,$0x0
mov_i32 new_r2,r2
and_i32 tmp0,p0,$0x1
brcond_i32 tmp0,$0x0,eq,$L1
and_i32 tmp0,r0,r1
mov_i32 new_r2,tmp0
br $L2
set_label $L1
or_i32 slot_cancelled,slot_cancelled,$0x8
set_label $L2
mov_i32 r2,new_r2
AFTER
---- 00400094
mov_i32 new_r2,r2
and_i32 tmp0,p0,$0x1
brcond_i32 tmp0,$0x0,eq,$L1
and_i32 tmp0,r0,r1
mov_i32 new_r2,tmp0
br $L2
set_label $L1
set_label $L2
mov_i32 r2,new_r2
Signed-off-by: Taylor Simpson <tsimpson@quicinc.com>
Reviewed-by: Anton Johansson <anjo@rev.ng>
Message-Id: <20230307025828.1612809-14-tsimpson@quicinc.com>
We assign the instruction destination register to hex_new_value[num]
instead of a TCG temp that gets copied back to hex_new_value[num].
We introduce new functions get_result_gpr[_pair] to facilitate getting
the proper destination register.
Since we preload hex_new_value for predicated instructions, we don't
need the check for slot_cancelled. So, we call gen_log_reg_write instead.
We update the helper function generation and gen_tcg.h to maintain the
disable-hexagon-idef-parser configuration.
Here is a simple example of the differences in the TCG code generated:
IN:
0x00400094: 0xf900c102 { if (P0) R2 = and(R0,R1) }
BEFORE
---- 00400094
mov_i32 slot_cancelled,$0x0
mov_i32 new_r2,r2
mov_i32 loc2,$0x0
and_i32 tmp0,p0,$0x1
brcond_i32 tmp0,$0x0,eq,$L1
and_i32 tmp0,r0,r1
mov_i32 loc2,tmp0
br $L2
set_label $L1
or_i32 slot_cancelled,slot_cancelled,$0x8
set_label $L2
and_i32 tmp0,slot_cancelled,$0x8
movcond_i32 new_r2,tmp0,$0x0,loc2,new_r2,eq
mov_i32 r2,new_r2
AFTER
---- 00400094
mov_i32 slot_cancelled,$0x0
mov_i32 new_r2,r2
and_i32 tmp0,p0,$0x1
brcond_i32 tmp0,$0x0,eq,$L1
and_i32 tmp0,r0,r1
mov_i32 new_r2,tmp0
br $L2
set_label $L1
or_i32 slot_cancelled,slot_cancelled,$0x8
set_label $L2
mov_i32 r2,new_r2
We'll remove the unnecessary manipulation of slot_cancelled in a
subsequent patch.
Signed-off-by: Taylor Simpson <tsimpson@quicinc.com>
Reviewed-by: Anton Johansson <anjo@rev.ng>
Message-Id: <20230307025828.1612809-13-tsimpson@quicinc.com>
The F2_sffms instruction [r0 -= sfmpy(r1, r2)] doesn't properly
handle -0. Previously we would negate the input operand by subtracting
from zero. Instead, we negate by changing the sign bit.
Test case added to tests/tcg/hexagon/fpstuff.c
Signed-off-by: Taylor Simpson <tsimpson@quicinc.com>
Reviewed-by: Anton Johansson <anjo@rev.ng>
Message-Id: <20230307025828.1612809-12-tsimpson@quicinc.com>
Extend the analyze_<tag> functions for HVX vector and predicate writes
Remove calls to ctx_log_vreg_write[_pair] from gen_tcg_funcs.py
During gen_start_packet, reload the predicated HVX registers into
fugure_VRegs and tmp_VRegs
Signed-off-by: Taylor Simpson <tsimpson@quicinc.com>
Reviewed-by: Anton Johansson <anjo@rev.ng>
Message-Id: <20230307025828.1612809-8-tsimpson@quicinc.com>
The pkt_has_store_s1 field in CPUHexagonState is only needed in generated
helpers for scalar load instructions. See check_noshuf and mem_load[1248]
in op_helper.c.
We add logic in gen_analyze_funcs.py to set need_pkt_has_store_s1 in
DisasContext when it is needed at runtime.
Signed-off-by: Taylor Simpson <tsimpson@quicinc.com>
Reviewed-by: Anton Johansson <anjo@rev.ng>
Message-Id: <20230307025828.1612809-7-tsimpson@quicinc.com>
We create a new generator that creates an analyze_<tag> function for
each instruction. Currently, these functions record the writes to
R, P, and C registers by calling ctx_log_reg_write[_pair] or
ctx_log_pred_write.
During gen_start_packet, we invoke the analyze_<tag> function for
each instruction in the packet, and we mark the implicit register
and predicate writes.
Doing the analysis up front has several advantages
- We remove calls to ctx_log_* from gen_tcg_funcs.py and genptr.c
- After the analysis is performed, we can initialize hex_new_value
for each of the predicated assignments rather than during TCG
generation for the instructions
- This is a stepping stone for future work where the analysis will
include the set of registers that are read. In cases where
the packet doesn't have an overlap between the registers that are
written and registers that are read, we can avoid the intermediate
step of writing to hex_new_value. Note that other checks will also
be needed (e.g., no instructions can raise an exception).
Signed-off-by: Taylor Simpson <tsimpson@quicinc.com>
Reviewed-by: Anton Johansson <anjo@rev.ng>
Message-Id: <20230307025828.1612809-6-tsimpson@quicinc.com>
These instructions perform a deallocframe+return (jumpr r31)
Add overrides for
L4_return
SL2_return
L4_return_t
L4_return_f
L4_return_tnew_pt
L4_return_fnew_pt
L4_return_tnew_pnt
L4_return_fnew_pnt
SL2_return_t
SL2_return_f
SL2_return_tnew
SL2_return_fnew
This patch eliminates the last helper that uses write_new_pc, so we
remove it from op_helper.c
Signed-off-by: Taylor Simpson <tsimpson@quicinc.com>
Reviewed-by: Anton Johansson <anjo@rev.ng>
Message-Id: <20230307025828.1612809-5-tsimpson@quicinc.com>
The --disable-hexagon-idef-parser configuration was broken by this patch
2feacf60c23ba6 (target/hexagon: Drop tcg_temp_free from C code)
That config is not tested by CI
Fix is simple: Mark a few TCGv variables as unused
Signed-off-by: Taylor Simpson <tsimpson@quicinc.com>
Reviewed-by: Anton Johansson <anjo@rev.ng>
Message-Id: <20230306172515.346813-1-tsimpson@quicinc.com>
We already have a temporary, res, which we can use for the intermediate
shift result. Simplify the constant to -1 instead of 0xf*f.
This was the last use of gen_tmp_value, so remove it.
Reviewed-by: Taylor Simpson <tsimpson@quicinc.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The allocation is immediately followed by either tcg_gen_mov_i32
or gen_read_preg (which contains tcg_gen_mov_i32), so the zero
initialization is immediately discarded.
Reviewed-by: Taylor Simpson <tsimpson@quicinc.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The allocation is immediately followed by tcg_gen_mov_i32,
so the initial assignment of zero is discarded.
Reviewed-by: Taylor Simpson <tsimpson@quicinc.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The GET_USR_FIELD macro initializes the output, so the initial assignment
of zero is discarded. This is the only use of get_tmp_value outside of
parser-helper.c, so make it static.
Reviewed-by: Taylor Simpson <tsimpson@quicinc.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Translators are no longer required to free tcg temporaries.
This removes gen_rvalue_free, gen_rvalue_free_manual and
free_variables, whose only purpose was to emit tcg_temp_free.
Reviewed-by: Taylor Simpson <tsimpson@quicinc.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
In preparation for returning the number of insns generated
via the same pointer. Adjust only the prototypes so far.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
This commit was created with scripts/clean-includes.
All .c should include qemu/osdep.h first. The script performs three
related cleanups:
* Ensure .c files include qemu/osdep.h first.
* Including it in a .h is redundant, since the .c already includes
it. Drop such inclusions.
* Likewise, including headers qemu/osdep.h includes is redundant.
Drop these, too.
Changes to standalone programs dropped, because these intentionally
don't use qemu/osdep.h:
target/hexagon/gen_dectree_import.c
target/hexagon/gen_semantics.c
target/hexagon/idef-parser/idef-parser.h
target/hexagon/idef-parser/parser-helpers.c
target/hexagon/idef-parser/parser-helpers.h
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Taylor Simpson <tsimpson@quicinc.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Message-Id: <20230202133830.2152150-14-armbru@redhat.com>
This patch manually suppresses a warning for an unused variable
(yynerrs) emitted by bison.
This warning has been triggered for the first time by clang 15.
This patch also disables `-Wextra`, which is not usually adopted in
QEMU. However, clang 15 triggers the warning fixed in this patch even in
absence of `-Wextra`.
Signed-off-by: Alessandro Di Federico <ale@rev.ng>
Signed-off-by: Taylor Simpson <tsimpson@quicinc.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Tested-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Taylor Simpson <tsimpson@quicinc.com>
Tested-by: Taylor Simpson <tsimpson@quicinc.com>
Message-Id: <20221221155327.1504117-1-ale@rev.ng>
The patch renames the identifier of the 32bit register
HEX_REG_P3_0 to HEX_REG_P3_0_ALIASED.
This change is to intended to provide some warning that
HEX_REG_P3_0 is an aliased register which has multiple
representations in CPU state and therefore might require
special handling in some contexts. The hope is to prevent
accidental misuse of this register e.g the issue reported
for the signals tests failure [here][1].
[1]: https://lists.gnu.org/archive/html/qemu-devel/2021-11/msg01102.html
Signed-off-by: Mukilan Thiyagarajan <quic_mthiyaga@quicinc.com>
Signed-off-by: Taylor Simpson <tsimpson@quicinc.com>
Reviewed-by: Taylor Simpson <tsimpson@quicinc.com>
Message-Id: <20221229092006.10709-3-quic_mthiyaga@quicinc.com>
1)
Performance improvement
Add pkt and insn to DisasContext
Many functions need information from all 3 structures, so merge
them together.
2)
Bug fix
Fix predicated assignment to .tmp and .cur
3)
Performance improvement
Add overrides for S2_asr_r_r_sat/S2_asl_r_r_sat
These functions will not be handled by idef-parser
4-11)
The final 8 patches improve change-of-flow handling.
Currently, we set the PC to a new address before exiting a TB. The
ultimate goal is to use direct block chaining. However, several steps
are needed along the way.
4)
When a packet has more than one change-of-flow (COF) instruction, only
the first one taken is considered. The runtime bookkeeping is only
needed when there is more than one COF instruction in a packet.
5, 6)
Remove PC and next_PC from the runtime state and always use a
translation-time constant. Note that next_PC is used by call instructions
to set LR and by conditional COF instructions to set the fall-through
address.
7, 8, 9)
Add helper overrides for COF instructions. In particular, we must
distinguish those that use a PC-relative address for the destination.
These are candidates for direct block chaining later.
10)
Use direct block chaining for packets that have a single PC-relative
COF instruction. Instead of generating the code while processing the
instruction, we record the effect in DisasContext and generate the code
during gen_end_tb.
11)
Use direct block chaining for tight loops. We look for TBs that end
with an endloop0 that will branch back to the TB start address.
12-21)
Instruction definition parser (idef-parser) from rev.ng
Parses the instruction semantics and generates TCG
# gpg: Signature made Fri 16 Dec 2022 20:41:53 GMT
# gpg: using RSA key 3635C788CE62B91FD4C59AB47B0244FB12DE4422
# gpg: Good signature from "Taylor Simpson (Rock on) <tsimpson@quicinc.com>" [undefined]
# gpg: WARNING: This key is not certified with a trusted signature!
# gpg: There is no indication that the signature belongs to the owner.
# Primary key fingerprint: 3635 C788 CE62 B91F D4C5 9AB4 7B02 44FB 12DE 4422
* tag 'pull-hex-20221216-1' of https://github.com/quic/qemu: (21 commits)
target/hexagon: import additional tests
target/hexagon: call idef-parser functions
target/hexagon: import parser for idef-parser
target/hexagon: import lexer for idef-parser
target/hexagon: prepare input for the idef-parser
target/hexagon: introduce new helper functions
target/hexagon: make helper functions non-static
target/hexagon: make slot number an unsigned
target/hexagon: import README for idef-parser
target/hexagon: update MAINTAINERS for idef-parser
Hexagon (target/hexagon) Use direct block chaining for tight loops
Hexagon (target/hexagon) Use direct block chaining for direct jump/branch
Hexagon (target/hexagon) Add overrides for various forms of jump
Hexagon (target/hexagon) Add overrides for compound compare and jump
Hexagon (target/hexagon) Add overrides for direct call instructions
Hexagon (target/hexagon) Remove next_PC from runtime state
Hexagon (target/hexagon) Remove PC from the runtime state
Hexagon (target/hexagon) Only use branch_taken when packet has multi cof
Hexagon (target/hexagon) Add overrides for S2_asr_r_r_sat/S2_asl_r_r_sat
Hexagon (target/hexagon) Fix predicated assignment to .tmp and .cur
...
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Introduce infrastructure necessary to produce a file suitable for being
parsed by the idef-parser. A build option is also added to fully disable
the output of idef-parser, which is useful for debugging.
Signed-off-by: Alessandro Di Federico <ale@rev.ng>
Signed-off-by: Anton Johansson <anjo@rev.ng>
Signed-off-by: Taylor Simpson <tsimpson@quicinc.com>
Reviewed-by: Taylor Simpson <tsimpson@quicinc.com>
Message-Id: <20220923173831.227551-8-anjo@rev.ng>
These helpers will be employed by the idef-parser generated code, to
correctly implement instruction semantics. "Helper" functions, in the
context of this patch, refers to functions which provide a manual TCG
implementation of certain features.
Signed-off-by: Alessandro Di Federico <ale@rev.ng>
Signed-off-by: Niccolò Izzo <nizzo@rev.ng>
Signed-off-by: Anton Johansson <anjo@rev.ng>
Signed-off-by: Taylor Simpson <tsimpson@quicinc.com>
Reviewed-by: Taylor Simpson <tsimpson@quicinc.com>
Message-Id: <20220923173831.227551-6-anjo@rev.ng>
Make certain helper functions non-static, making them available outside
genptr.c. These functions are required by code generated by the
idef-parser.
This commit also makes some functions in op_helper.c non-static in order
to avoid having them marked as unused when using the idef-parser
generated code.
Signed-off-by: Alessandro Di Federico <ale@rev.ng>
Signed-off-by: Paolo Montesel <babush@rev.ng>
Signed-off-by: Taylor Simpson <tsimpson@quicinc.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Taylor Simpson <tsimpson@quicinc.com>
Message-Id: <20220923173831.227551-5-anjo@rev.ng>
Direct block chaining is documented here
https://qemu.readthedocs.io/en/latest/devel/tcg.html#direct-block-chaining
Hexagon inner loops end with the endloop0 instruction
To go back to the beginning of the loop, this instructions writes to PC
from register SA0 (start address 0). To use direct block chaining, we
have to assign PC with a constant value. So, we specialize the code
generation when the start of the translation block is equal to SA0.
When this is the case, we defer the compare/branch from endloop0 to
gen_end_tb. When this is done, we can assign the start address of the TB
to PC.
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Taylor Simpson <tsimpson@quicinc.com>
Message-Id: <20221108162906.3166-12-tsimpson@quicinc.com>
Direct block chaining is documented here
https://qemu.readthedocs.io/en/latest/devel/tcg.html#direct-block-chaining
Recall that Hexagon allows packets with multiple jumps where only the
first one with a true predicate will actually jump. We can use
tcg_gen_goto_tb/tcg_gen_exit_tb when the packet contains a single
PC-relative branch or jump. If not, we use tcg_gen_lookup_and_goto_ptr.
We add the following to DisasContext in order to delay the branching
until the end of packet commit (in gen_end_tb)
branch_cond
The TCGCond condition under which the branch is taken
When branch_cond == TCG_COND_NEVER, there isn't a single
direct branch in this packet.
When branch_cond != TCG_COND_ALWAYS, the value is in
hex_branch_taken
branch_dest
The destination of the branch
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Taylor Simpson <tsimpson@quicinc.com>
Message-Id: <20221108162906.3166-11-tsimpson@quicinc.com>
When a packet has more than one change-of-flow instruction, only the first
one to branch is considered. We use the branch_taken variable to keep
track of this.
However, when there is a single cof instruction, we don't need the same
amount of bookkeeping.
We add the pkt_has_multi_cof member to the Packet structure, and pass this
information to the needed functions.
When there is a generated helper function with cof, the generator will
pass this pkt_has_multi_cof as a runtime value.
Acked-by: Richard Henderson <richard.henderson@linaro.org>
Signed-off-by: Taylor Simpson <tsimpson@quicinc.com>
Message-Id: <20221108162906.3166-5-tsimpson@quicinc.com>
