Skip to contentSkip to conversation

QEMU AVR 8 bit cores

PATCH Superseded Michael Rolnik 3 participants 27 messages Dec 29, 2019 – Jan 12, 2020 RISC-V

v40 of 41 Superseded Latest is v41 →

lore
Share & tools
lore
#0 lore raw

v40 cover QEMU AVR 8 bit cores

This series of patches adds 8bit AVR cores to QEMU.
All instruction, except BREAK/DES/SPM/SPMX, are implemented. Not fully tested yet. However I was able to execute simple code with functions. e.g fibonacci calculation. This series of patches include a non real, sample board.
No fuses support yet. PC is set to 0 at reset.

Following are examples of possible usages, assuming program.elf is compiled for AVR cpu
1. Continious non interrupted execution
run qemu-system-avr -kernel program.elf
2. Continious non interrupted execution with serial output into telnet window run qemu-system-avr -kernel program.elf -serial tcp::5678,server,nowait -nographic run telent localhost 5678
3. Continious non interrupted execution with serial output into stdout
run qemu-system-avr -kernel program.elf -serial stdio
4. Debugging wit GDB debugger
run qemu-system-avr -kernel program.elf -s -S
run avr-gdb program.elf and then within GDB shell target remote :1234
5. Print out executed instructions
run qemu-system-avr -kernel program.elf -d in_asm

the patches include the following
1. just a basic 8bit AVR CPU, without instruction decoding or translation
2. CPU features which allow define the following 8bit AVR cores

     avr1
     avr2 avr25
     avr3 avr31 avr35
     avr4
     avr5 avr51
     avr6
     xmega2 xmega4 xmega5 xmega6 xmega7

3. a definition of sample machine with SRAM, FLASH and CPU which allows to execute simple code
4. encoding for all AVR instructions
5. interrupt handling
6. helpers for IN, OUT, SLEEP, WBR & unsupported instructions
7. a decoder which given an opcode decides what istruction it is
8. translation of AVR instruction into TCG
9. all features together

changes since v3
1. rampD/X/Y/Z registers are encoded as 0x00ff0000 (instead of 0x000000ff) for faster address manipulaton
2. ffs changed to ctz32
3. duplicate code removed at avr_cpu_do_interrupt
4. using andc instead of not + and
5. fixing V flag calculation in varios instructions
6. freeing local variables in PUSH
7. tcg_const_local_i32 -> tcg_const_i32
8. using sextract32 instead of my implementation
9. fixing BLD instruction
10.xor(r) instead of 0xff - r at COM
11.fixing MULS/MULSU not to modify inputs' content
12.using SUB for NEG
13.fixing tcg_gen_qemu_ld/st call in XCH

changes since v4
1. target is now defined as big endian in order to optimize push_ret/pop_ret
2. all style warnings are fixed
3. adding cpu_set/get_sreg functions
4. simplifying gen_goto_tb as there is no real paging
5. env->pc -> env->pc_w
6. making flag dump more compact
7. more spacing
8. renaming CODE/DATA_INDEX -> MMU_CODE/DATA_IDX
9. removing avr_set_feature
10. SPL/SPH set bug fix
11. switching stb_phys to cpu_stb_data
12. cleaning up avr_decode
13. saving sreg, rampD/X/Y/Z, eind in HW format (savevm)
14. saving CPU features (savevm)

changes since v5
1. BLD bug fix
2. decoder generator is added

chages since v6
1. using cpu_get_sreg/cpu_set_sreg in avr_cpu_gdb_read_register/avr_cpu_gdb_write_register
2. configure the target as little endian because otherwise GDB does not work
3. fixing and testing gen_push_ret/gen_pop_ret

changes since v7
1. folding back v6
2. logging at helper_outb and helper_inb are done for non supported yet registers only
3. MAINTAINERS updated

changes since v8
1. removing hw/avr from hw/Makefile.obj as it should not be built for all
2. making linux compilable
3. testing on

    a. Mac, Apple LLVM version 7.0.0
    b. Ubuntu 12.04, gcc 4.9.2
    c. Fedora 23, gcc 5.3.1

4. folding back some patches
5. translation bug fixes for ORI, CPI, XOR instructions
6. propper handling of cpu register writes though memory

changes since v9
1. removing forward declarations of static functions
2. disabling debug prints
3. switching to case range instead of if else if ...
4. LD/ST IN/OUT accessing CPU maintainder registers are not routed to any device
5. commenst about sample board and sample IO device added
6. sample board description is more descriptive now
7. memory_region_allocate_system_memory is used to create RAM
8. now there are helper_fullrd & helper_fullwr when LD/ST try to access registers

changes since v10
1. movig back fullwr & fullrd into the commit where outb and inb were introduced
2. changing tlb_fill function signature
3. adding empty line between functions
4. adding newline on the last line of the file
5. using tb->flags to generae full access ST/LD instructions
6. fixing SBRC bug
7. folding back 10th commit
8. whenever a new file is introduced it's added to Makefile.objs

changes since v11
1. updating to v2.7.0-rc
2. removing assignment to env->fullacc from gen_intermediate_code

changes since v12
1. fixing spacing
2. fixing get/put_segment functions
3. removing target-avr/machine.h file
4. VMSTATE_SINGLE_TEST -> VMSTATE_SINGLE
5. comment spelling
6. removing hw/avr/sample_io.c
7. char const* -> const char*
8. proper ram allocation
9. fixing breakpoint functionality.
10.env1 -> env
11.fixing avr_cpu_gdb_write_register & avr_cpu_gdb_read_register functions 12.any cpu is removed
12.feature bits are not saved into vm state

changes since v13
1. rebasing to v2.7.0-rc1

changes since v14
1. I made self review with git gui tool. (I did not know such a thing exists)
2. removing all double/tripple spaces
3. removing comment reference to SampleIO
4. folding back some changes, so there is not deleted lines in my code
5. moving avr configuration, within configure file, before chris

changes since v15
1. removing IO registers cache from CPU
2. implementing CBI/SBI as read(helper_inb), modify, write(helper_outb)
3. implementing CBIC/SBIC as read(helper_inb), check, branch
4. adding missing tcg_temp_free_i32 for tcg_const_i32

changes since v16
1. removing EXT IO registers knoledge from CPU. These registers are accessible by LD/ST only. CPU has no interest in them

changes since v17 (by Richard Henderson)
This is Michael's v17, with some adjustments of my own:

1. Fix the whitespace errors reported by "git am",
2. Replace the utf-8 characters with normal ascii,
3. Ditch the separate compilation of translate.c.

I retained the two separate files that could be regenerated from the included cpugen program, but merged in translate-insn.c. Not that it matters, but the code generated is about 3k smaller.

changes since v18

1.  moving target-avr into target/avr
2.  do not call cpu_exec_initfn function from avr_cpu_initfn
3.  call cpu_exec_realizefn avr_cpu_realizefn
4.  do not fail sample machine creation if no rom is suplied
5.  add tcg_gen_exit_tb(0) for BS_BRANCH in gen_intermediate_code
6.  fix a register getters/setters in machine.c
7.  changing QEMU_ARCH_AVR from 1<<17 to 1<<18

changes since v19

1.  use decodetree.py tool to decode instructions
2.  adding USART
3.  adding 16 bit timer peripherals
4.  changing QEMU_ARCH_AVR from 1<<18 to 1<<20
5.  renaming tlb_fill to avr_cpu_tlb_fill

changes since v20

1.  use one CPU naming convention
2.  merging insn16.decode & insn32.decode files
3.  modifying skip next instruction mechanizm
4.  translate BREAK as NOP for now

changes since v21
1. Reorganize bstate.
This will make transition to <exec/translator.h> easier, and fixes a couple of bugs wrt single stepping by [email protected]
2. Drop cpc and fix page cross condition.
by [email protected]
3. Refactor checking supported/unsupported instructions
4. Add gdb-xml/avr-cpu.xml

changes since v22
1. Rebase
2. Split long comment

changes since v23

1.  remove avr_cpu_list_compare function
2.  modify avr_cpu_class_by_name function
3.  modify avr_cpu_list_entry function
4.  modify avr_cpu_list function

changes since v24
1. remove AVR_CPU_TYPE_NAME macro

changes since v25

1.  fix patches. every file belong to one patch only
2.  change copyright year from 2016 to 2019
3.  create mask device to emulate prr0/prr1

changes since v26
1. add avocado acceptence test
2. add boot serial test

changes since v27
1. list atmel2560 devices as unimplemented
2. fix sram base/size

changes since v28
1. rebase
2. fix includes & build

changes since v29
1. fixing ownership
2. using 'since' instread of 'added in'

changes since v30
1. rebase

changes since v31

1.  splitting 'Add instruction translation' commit
2.  fixing typo in qapi/machine.json sicne -> since
3.  removing unintended changes in configure file
4.  adding Richard Henderson as a co developer to 'Add instruction translation - CPU main translation funcions' commit

changes since v32
1. modify cpu_get_sreg to treat sreg C as other flags, except sreg Z

changes since v33
1. ensure flag C is always calculated as one bit
2. calculate flag Z as one bit, without using inverse logic

changes since v34
1. rebase

changes since v35

1.  rebase
2.  use NANOSECONDS_PER_SECOND instead of 1000000000 in avr_timer16.c
3.  split "target/avr: Register AVR support with the rest of QEMU" into three patches
    1.  "target/avr: Register AVR support with the rest of QEMU"
    2.  "target/avr: Update build system"
    3.  "target/avr: Update MAINTAINERS file"

4. split "target/avr: Add tests" patch into two patches
1. "target/avr: Add Avocado test"
2. "target/avr: Add boot serial test"
5. Add instruction disassembly function
6. change "since 4.2" to "since 5.0"

changes since v36
1. rebase
2. rename

    1.  NO_CPU_REGISTERS    -> NUMBER_OF_CPU_REGISTERS
    2.  NO_IO_REGISTERS     -> NUMBER_OF_IO_REGISTERS
    3.  to_A                -> to_regs_16_31_by_one
    4.  to_B                -> to_regs_16_23_by_one
    5.  to_C                -> to_regs_24_30_by_two
    6.  to_D                -> to_regs_00_30_by_two
3.  add missing licences
4.  add usage example (see above)
5.  ass Sarah Harris <[email protected]> as a reviewer to MAINTAINERS
7.  use git commit sha1 instead of `master`` in avocado test

changes since v37

1.  rebase
2.  moving all data transfer instructions into a separate commit
3.  splitting tests into 
    1.  avocado test
    2.  boot serial test
    3.  machine none test
4.  split periperals commit into 
    1.  timer
    2.  usart
    3.  mask
5.  add function `avr_flags_to_cpu_type` that translates ELF flags into core name
6.  add avrtiny, xmega3 cores
7.  remove unnecessary includes from target/avr/helper.c file
8.  remove empty lines from target/avr/translate.c file
9.  set number of interrupt lines to 64

10. determine cpu type by reading ELF flags
11. update license of all new files to be LGPL
12. udpate QEMU documentation

changes since v38

1.  rebase
2.  add examples of AVR emulator invokation into the QEMU doc
3.  reorder instructions to match AVR documentation
4.  fix elf loader function bug introduced by prev version

changes since v39

1.  rename target/avr to hw/avr for hw related commits
2.  spread instruction decoding commit
3.  add frequency parameter to AVR timer
4.  fix offset of some unimplemented registers

Michael Rolnik (21):
target/avr: Add outward facing interfaces and core CPU logic target/avr: Add instruction helpers
target/avr: Add instruction translation - Registers definition target/avr: Add instruction translation - Arithmetic and Logic
Instructions
target/avr: Add instruction translation - Branch Instructions target/avr: Add instruction translation - Data Transfer Instructions target/avr: Add instruction translation - Bit and Bit-test
Instructions
target/avr: Add instruction translation - MCU Control Instructions target/avr: Add instruction translation - CPU main translation
function
target/avr: Add instruction disassembly function
hw/avr: Add limited support for USART peripheral
hw/avr: Add limited support for 16 bit timer peripheral hw/avr: Add dummy mask device
hw/avr: Add example board configuration
target/avr: Add section about AVR into QEMU documentation target/avr: Register AVR support with the rest of QEMU target/avr: Add machine none test
target/avr: Update build system
target/avr: Add boot serial test
target/avr: Add Avocado test
target/avr: Update MAINTAINERS file

 qemu-doc.texi                    |   51 +
 configure                        |    7 +
 default-configs/avr-softmmu.mak  |    5 +
 qapi/machine.json                |    3 +-
 include/disas/dis-asm.h          |   19 +
 include/elf.h                    |    2 +
 include/hw/char/avr_usart.h      |   93 +
 include/hw/elf_ops.h             |    6 +-
 include/hw/loader.h              |    6 +-
 include/hw/misc/avr_mask.h       |   47 +
 include/hw/timer/avr_timer16.h   |   94 +
 include/sysemu/arch_init.h       |    1 +
 target/avr/cpu-param.h           |   37 +
 target/avr/cpu-qom.h             |   54 +
 target/avr/cpu.h                 |  259 +++
 target/avr/helper.h              |   29 +
 arch_init.c                      |    2 +
 hw/avr/sample.c                  |  295 +++
 hw/char/avr_usart.c              |  320 ++++
 hw/core/loader.c                 |   15 +-
 hw/misc/avr_mask.c               |  112 ++
 hw/riscv/boot.c                  |    2 +-
 hw/timer/avr_timer16.c           |  602 ++++++
 target/avr/cpu.c                 |  826 ++++++++
 target/avr/disas.c               |  245 +++
 target/avr/gdbstub.c             |   84 +
 target/avr/helper.c              |  347 ++++
 target/avr/machine.c             |  121 ++
 target/avr/translate.c           | 2997 ++++++++++++++++++++++++++++++
 tests/boot-serial-test.c         |   10 +
 tests/machine-none-test.c        |    1 +
 MAINTAINERS                      |   21 +
 gdb-xml/avr-cpu.xml              |   49 +
 hw/Kconfig                       |    1 +
 hw/avr/Kconfig                   |    6 +
 hw/avr/Makefile.objs             |    1 +
 hw/char/Kconfig                  |    3 +
 hw/char/Makefile.objs            |    1 +
 hw/misc/Kconfig                  |    3 +
 hw/misc/Makefile.objs            |    2 +
 hw/timer/Kconfig                 |    3 +
 hw/timer/Makefile.objs           |    2 +
 target/avr/Makefile.objs         |   34 +
 target/avr/insn.decode           |  183 ++
 tests/Makefile.include           |    2 +
 tests/acceptance/machine_avr6.py |   58 +
 46 files changed, 7049 insertions(+), 12 deletions(-)
 create mode 100644 default-configs/avr-softmmu.mak
 create mode 100644 include/hw/char/avr_usart.h
 create mode 100644 include/hw/misc/avr_mask.h
 create mode 100644 include/hw/timer/avr_timer16.h
 create mode 100644 target/avr/cpu-param.h
 create mode 100644 target/avr/cpu-qom.h
 create mode 100644 target/avr/cpu.h
 create mode 100644 target/avr/helper.h
 create mode 100644 hw/avr/sample.c
 create mode 100644 hw/char/avr_usart.c
 create mode 100644 hw/misc/avr_mask.c
 create mode 100644 hw/timer/avr_timer16.c
 create mode 100644 target/avr/cpu.c
 create mode 100644 target/avr/disas.c
 create mode 100644 target/avr/gdbstub.c
 create mode 100644 target/avr/helper.c
 create mode 100644 target/avr/machine.c
 create mode 100644 target/avr/translate.c
 create mode 100644 gdb-xml/avr-cpu.xml
 create mode 100644 hw/avr/Kconfig
 create mode 100644 hw/avr/Makefile.objs
 create mode 100644 target/avr/Makefile.objs
 create mode 100644 target/avr/insn.decode
 create mode 100644 tests/acceptance/machine_avr6.py
Signature · 1 line
2.17.2 (Apple Git-113)
#3 lore raw

PATCH v40 3/21 target/avr: Add instruction translation - Registers definition

---
 target/avr/translate.c | 172 +++++++++++++++++++++++++++++++++++++++++
 1 file changed, 172 insertions(+)
 create mode 100644 target/avr/translate.c
target/avr/translate.c new file +172 -0
new file mode 100644
index 0000000000..cfdff65620
@@ -0,0 +1,172 @@
+/*
+ * QEMU AVR CPU
+ *
+ * Copyright (c) 2019 Michael Rolnik
+ *
+ * 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/lgpl-2.1.html>
+ */
+
+#include "qemu/osdep.h"
+#include "qemu/qemu-print.h"
+#include "tcg/tcg.h"
+#include "cpu.h"
+#include "exec/exec-all.h"
+#include "tcg-op.h"
+#include "exec/cpu_ldst.h"
+#include "exec/helper-proto.h"
+#include "exec/helper-gen.h"
+#include "exec/log.h"
+#include "exec/translator.h"
+#include "exec/gen-icount.h"
+
+/*
+ *  Define if you want a BREAK instruction translated to a breakpoint
+ *  Active debugging connection is assumed
+ *  This is for
+ *  https://github.com/seharris/qemu-avr-tests/tree/master/instruction-tests
+ *  tests
+ */
+#undef BREAKPOINT_ON_BREAK
+
+static TCGv cpu_pc;
+
+static TCGv cpu_Cf;
+static TCGv cpu_Zf;
+static TCGv cpu_Nf;
+static TCGv cpu_Vf;
+static TCGv cpu_Sf;
+static TCGv cpu_Hf;
+static TCGv cpu_Tf;
+static TCGv cpu_If;
+
+static TCGv cpu_rampD;
+static TCGv cpu_rampX;
+static TCGv cpu_rampY;
+static TCGv cpu_rampZ;
+
+static TCGv cpu_r[NUMBER_OF_CPU_REGISTERS];
+static TCGv cpu_eind;
+static TCGv cpu_sp;
+
+static TCGv cpu_skip;
+
+static const char reg_names[NUMBER_OF_CPU_REGISTERS][8] = {
+    "r0",  "r1",  "r2",  "r3",  "r4",  "r5",  "r6",  "r7",
+    "r8",  "r9",  "r10", "r11", "r12", "r13", "r14", "r15",
+    "r16", "r17", "r18", "r19", "r20", "r21", "r22", "r23",
+    "r24", "r25", "r26", "r27", "r28", "r29", "r30", "r31",
+};
+#define REG(x) (cpu_r[x])
+
+enum {
+    DISAS_EXIT   = DISAS_TARGET_0,  /* We want return to the cpu main loop.  */
+    DISAS_LOOKUP = DISAS_TARGET_1,  /* We have a variable condition exit.  */
+    DISAS_CHAIN  = DISAS_TARGET_2,  /* We have a single condition exit.  */
+};
+
+typedef struct DisasContext DisasContext;
+
+/* This is the state at translation time. */
+struct DisasContext {
+    TranslationBlock *tb;
+
+    CPUAVRState *env;
+    CPUState *cs;
+
+    target_long npc;
+    uint32_t opcode;
+
+    /* Routine used to access memory */
+    int memidx;
+    int bstate;
+    int singlestep;
+
+    /*
+     * some AVR instructions can make the following instruction to be skipped
+     * Let's name those instructions
+     *     A   - instruction that can skip the next one
+     *     B   - instruction that can be skipped. this depends on execution of A
+     * there are two scenarios
+     * 1. A and B belong to the same translation block
+     * 2. A is the last instruction in the translation block and B is the last
+     *
+     * following variables are used to simplify the skipping logic, they are
+     * used in the following manner (sketch)
+     *
+     * TCGLabel *skip_label = NULL;
+     * if (ctx.skip_cond != TCG_COND_NEVER) {
+     *     skip_label = gen_new_label();
+     *     tcg_gen_brcond_tl(skip_cond, skip_var0, skip_var1, skip_label);
+     * }
+     *
+     * if (free_skip_var0) {
+     *     tcg_temp_free(skip_var0);
+     *     free_skip_var0 = false;
+     * }
+     *
+     * translate(&ctx);
+     *
+     * if (skip_label) {
+     *     gen_set_label(skip_label);
+     * }
+     */
+    TCGv skip_var0;
+    TCGv skip_var1;
+    TCGCond skip_cond;
+    bool free_skip_var0;
+};
+
+static int to_regs_16_31_by_one(DisasContext *ctx, int indx)
+{
+    return 16 + (indx % 16);
+}
+
+static int to_regs_16_23_by_one(DisasContext *ctx, int indx)
+{
+    return 16 + (indx % 8);
+}
+static int to_regs_24_30_by_two(DisasContext *ctx, int indx)
+{
+    return 24 + (indx % 4) * 2;
+}
+static int to_regs_00_30_by_two(DisasContext *ctx, int indx)
+{
+    return (indx % 16) * 2;
+}
+
+static uint16_t next_word(DisasContext *ctx)
+{
+    return cpu_lduw_code(ctx->env, ctx->npc++ * 2);
+}
+
+static int append_16(DisasContext *ctx, int x)
+{
+    return x << 16 | next_word(ctx);
+}
+
+
+static bool avr_have_feature(DisasContext *ctx, int feature)
+{
+    if (!avr_feature(ctx->env, feature)) {
+        gen_helper_unsupported(cpu_env);
+        ctx->bstate = DISAS_NORETURN;
+        return false;
+    }
+    return true;
+}
+
+static bool decode_insn(DisasContext *ctx, uint16_t insn);
+#include "decode_insn.inc.c"
+
Signature · 1 line
2.17.2 (Apple Git-113)
#2 lore raw

PATCH v40 2/21 target/avr: Add instruction helpers

In reply to Michael Rolnik

Stubs for unimplemented instructions and helpers for instructions that need to interact with QEMU. SPM and WDR are unimplemented because they require emulation of complex peripherals.
The implementation of SLEEP is very limited due to the lack of peripherals to generate wake interrupts. Memory access instructions are implemented here because some address ranges actually refer to CPU registers.

---
 target/avr/helper.h |  29 ++++
 target/avr/helper.c | 347 ++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 376 insertions(+)
 create mode 100644 target/avr/helper.h
 create mode 100644 target/avr/helper.c
target/avr/helper.h new file +29 -0
new file mode 100644
index 0000000000..bf087504a8
@@ -0,0 +1,29 @@
+/*
+ * QEMU AVR CPU
+ *
+ * Copyright (c) 2019 Michael Rolnik
+ *
+ * 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/lgpl-2.1.html>
+ */
+
+DEF_HELPER_1(wdr, void, env)
+DEF_HELPER_1(debug, void, env)
+DEF_HELPER_1(break, void, env)
+DEF_HELPER_1(sleep, void, env)
+DEF_HELPER_1(unsupported, void, env)
+DEF_HELPER_3(outb, void, env, i32, i32)
+DEF_HELPER_2(inb, tl, env, i32)
+DEF_HELPER_3(fullwr, void, env, i32, i32)
+DEF_HELPER_2(fullrd, tl, env, i32)
target/avr/helper.c new file +347 -0

347 lines not on this page · raw message

Signature · 1 line
2.17.2 (Apple Git-113)
#6 lore raw

PATCH v40 6/21 target/avr: Add instruction translation - Data Transfer Instructions

In reply to Michael Rolnik

This includes:
- MOV, MOVW
- LDI, LDS LDX LDY LDZ
- LDDY, LDDZ
- STS, STX STY STZ
- STDY, STDZ
- LPM, LPMX
- ELPM, ELPMX
- SPM, SPMX
- IN, OUT
- PUSH, POP
- XCH
- LAS, LAC LAT

---
 target/avr/translate.c | 986 +++++++++++++++++++++++++++++++++++++++++
 target/avr/insn.decode |  43 ++
 2 files changed, 1029 insertions(+)
target/avr/translate.c +986 -0
index a80d59464a..9c66734a19 100644
@@ -1454,3 +1454,989 @@ static bool trans_BRBS(DisasContext *ctx, arg_BRBS *a)
     return true;
 }
 
+/*
+ * Data Transfer Instructions
+ */
+
+/*
+ *  in the gen_set_addr & gen_get_addr functions
+ *  H assumed to be in 0x00ff0000 format
+ *  M assumed to be in 0x000000ff format
+ *  L assumed to be in 0x000000ff format
+ */
+static void gen_set_addr(TCGv addr, TCGv H, TCGv M, TCGv L)
+{
+
+    tcg_gen_andi_tl(L, addr, 0x000000ff);
+
+    tcg_gen_andi_tl(M, addr, 0x0000ff00);
+    tcg_gen_shri_tl(M, M, 8);
+
+    tcg_gen_andi_tl(H, addr, 0x00ff0000);
+}
+
+static void gen_set_xaddr(TCGv addr)
+{
+    gen_set_addr(addr, cpu_rampX, cpu_r[27], cpu_r[26]);
+}
+
+static void gen_set_yaddr(TCGv addr)
+{
+    gen_set_addr(addr, cpu_rampY, cpu_r[29], cpu_r[28]);
+}
+
+static void gen_set_zaddr(TCGv addr)
+{
+    gen_set_addr(addr, cpu_rampZ, cpu_r[31], cpu_r[30]);
+}
+
+static TCGv gen_get_addr(TCGv H, TCGv M, TCGv L)
+{
+    TCGv addr = tcg_temp_new_i32();
+
+    tcg_gen_deposit_tl(addr, M, H, 8, 8);
+    tcg_gen_deposit_tl(addr, L, addr, 8, 16);
+
+    return addr;
+}
+
+static TCGv gen_get_xaddr(void)
+{
+    return gen_get_addr(cpu_rampX, cpu_r[27], cpu_r[26]);
+}
+
+static TCGv gen_get_yaddr(void)
+{
+    return gen_get_addr(cpu_rampY, cpu_r[29], cpu_r[28]);
+}
+
+static TCGv gen_get_zaddr(void)
+{
+    return gen_get_addr(cpu_rampZ, cpu_r[31], cpu_r[30]);
+}
+
+/*
+ *  Load one byte indirect from data space to register and stores an clear
+ *  the bits in data space specified by the register. The instruction can only
+ *  be used towards internal SRAM.  The data location is pointed to by the Z (16
+ *  bits) Pointer Register in the Register File. Memory access is limited to the
+ *  current data segment of 64KB. To access another data segment in devices with
+ *  more than 64KB data space, the RAMPZ in register in the I/O area has to be
+ *  changed.  The Z-pointer Register is left unchanged by the operation. This
+ *  instruction is especially suited for clearing status bits stored in SRAM.
+ */
+static void gen_data_store(DisasContext *ctx, TCGv data, TCGv addr)
+{
+    if (ctx->tb->flags & TB_FLAGS_FULL_ACCESS) {
+        gen_helper_fullwr(cpu_env, data, addr);
+    } else {
+        tcg_gen_qemu_st8(data, addr, MMU_DATA_IDX); /* mem[addr] = data */
+    }
+}
+
+static void gen_data_load(DisasContext *ctx, TCGv data, TCGv addr)
+{
+    if (ctx->tb->flags & TB_FLAGS_FULL_ACCESS) {
+        gen_helper_fullrd(data, cpu_env, addr);
+    } else {
+        tcg_gen_qemu_ld8u(data, addr, MMU_DATA_IDX); /* data = mem[addr] */
+    }
+}
+
+/*
+ *  This instruction makes a copy of one register into another. The source
+ *  register Rr is left unchanged, while the destination register Rd is loaded
+ *  with a copy of Rr.
+ */
+static bool trans_MOV(DisasContext *ctx, arg_MOV *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv Rr = cpu_r[a->rr];
+
+    tcg_gen_mov_tl(Rd, Rr);
+
+    return true;
+}
+
+/*
+ *  This instruction makes a copy of one register pair into another register
+ *  pair. The source register pair Rr+1:Rr is left unchanged, while the
+ *  destination register pair Rd+1:Rd is loaded with a copy of Rr + 1:Rr.  This
+ *  instruction is not available in all devices. Refer to the device specific
+ *  instruction set summary.
+ */
+static bool trans_MOVW(DisasContext *ctx, arg_MOVW *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_MOVW)) {
+        return true;
+    }
+
+    TCGv RdL = cpu_r[a->rd];
+    TCGv RdH = cpu_r[a->rd + 1];
+    TCGv RrL = cpu_r[a->rr];
+    TCGv RrH = cpu_r[a->rr + 1];
+
+    tcg_gen_mov_tl(RdH, RrH);
+    tcg_gen_mov_tl(RdL, RrL);
+
+    return true;
+}
+
+/*
+ * Loads an 8 bit constant directly to register 16 to 31.
+ */
+static bool trans_LDI(DisasContext *ctx, arg_LDI *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    int imm = a->imm;
+
+    tcg_gen_movi_tl(Rd, imm);
+
+    return true;
+}
+
+/*
+ *  Loads one byte from the data space to a register. For parts with SRAM,
+ *  the data space consists of the Register File, I/O memory and internal SRAM
+ *  (and external SRAM if applicable). For parts without SRAM, the data space
+ *  consists of the register file only. The EEPROM has a separate address space.
+ *  A 16-bit address must be supplied. Memory access is limited to the current
+ *  data segment of 64KB. The LDS instruction uses the RAMPD Register to access
+ *  memory above 64KB. To access another data segment in devices with more than
+ *  64KB data space, the RAMPD in register in the I/O area has to be changed.
+ *  This instruction is not available in all devices. Refer to the device
+ *  specific instruction set summary.
+ */
+static bool trans_LDS(DisasContext *ctx, arg_LDS *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv addr = tcg_temp_new_i32();
+    TCGv H = cpu_rampD;
+    a->imm = next_word(ctx);
+
+    tcg_gen_mov_tl(addr, H); /* addr = H:M:L */
+    tcg_gen_shli_tl(addr, addr, 16);
+    tcg_gen_ori_tl(addr, addr, a->imm);
+
+    gen_data_load(ctx, Rd, addr);
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+/*
+ *  Loads one byte indirect from the data space to a register. For parts
+ *  with SRAM, the data space consists of the Register File, I/O memory and
+ *  internal SRAM (and external SRAM if applicable). For parts without SRAM, the
+ *  data space consists of the Register File only. In some parts the Flash
+ *  Memory has been mapped to the data space and can be read using this command.
+ *  The EEPROM has a separate address space.  The data location is pointed to by
+ *  the X (16 bits) Pointer Register in the Register File. Memory access is
+ *  limited to the current data segment of 64KB. To access another data segment
+ *  in devices with more than 64KB data space, the RAMPX in register in the I/O
+ *  area has to be changed.  The X-pointer Register can either be left unchanged
+ *  by the operation, or it can be post-incremented or predecremented.  These
+ *  features are especially suited for accessing arrays, tables, and Stack
+ *  Pointer usage of the X-pointer Register. Note that only the low byte of the
+ *  X-pointer is updated in devices with no more than 256 bytes data space. For
+ *  such devices, the high byte of the pointer is not used by this instruction
+ *  and can be used for other purposes. The RAMPX Register in the I/O area is
+ *  updated in parts with more than 64KB data space or more than 64KB Program
+ *  memory, and the increment/decrement is added to the entire 24-bit address on
+ *  such devices.  Not all variants of this instruction is available in all
+ *  devices. Refer to the device specific instruction set summary.  In the
+ *  Reduced Core tinyAVR the LD instruction can be used to achieve the same
+ *  operation as LPM since the program memory is mapped to the data memory
+ *  space.
+ */
+static bool trans_LDX1(DisasContext *ctx, arg_LDX1 *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv addr = gen_get_xaddr();
+
+    gen_data_load(ctx, Rd, addr);
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+static bool trans_LDX2(DisasContext *ctx, arg_LDX2 *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv addr = gen_get_xaddr();
+
+    gen_data_load(ctx, Rd, addr);
+    tcg_gen_addi_tl(addr, addr, 1); /* addr = addr + 1 */
+
+    gen_set_xaddr(addr);
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+static bool trans_LDX3(DisasContext *ctx, arg_LDX3 *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv addr = gen_get_xaddr();
+
+    tcg_gen_subi_tl(addr, addr, 1); /* addr = addr - 1 */
+    gen_data_load(ctx, Rd, addr);
+    gen_set_xaddr(addr);
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+/*
+ *  Loads one byte indirect with or without displacement from the data space
+ *  to a register. For parts with SRAM, the data space consists of the Register
+ *  File, I/O memory and internal SRAM (and external SRAM if applicable). For
+ *  parts without SRAM, the data space consists of the Register File only. In
+ *  some parts the Flash Memory has been mapped to the data space and can be
+ *  read using this command. The EEPROM has a separate address space.  The data
+ *  location is pointed to by the Y (16 bits) Pointer Register in the Register
+ *  File. Memory access is limited to the current data segment of 64KB. To
+ *  access another data segment in devices with more than 64KB data space, the
+ *  RAMPY in register in the I/O area has to be changed.  The Y-pointer Register
+ *  can either be left unchanged by the operation, or it can be post-incremented
+ *  or predecremented.  These features are especially suited for accessing
+ *  arrays, tables, and Stack Pointer usage of the Y-pointer Register. Note that
+ *  only the low byte of the Y-pointer is updated in devices with no more than
+ *  256 bytes data space. For such devices, the high byte of the pointer is not
+ *  used by this instruction and can be used for other purposes. The RAMPY
+ *  Register in the I/O area is updated in parts with more than 64KB data space
+ *  or more than 64KB Program memory, and the increment/decrement/displacement
+ *  is added to the entire 24-bit address on such devices.  Not all variants of
+ *  this instruction is available in all devices. Refer to the device specific
+ *  instruction set summary.  In the Reduced Core tinyAVR the LD instruction can
+ *  be used to achieve the same operation as LPM since the program memory is
+ *  mapped to the data memory space.
+ */
+static bool trans_LDY2(DisasContext *ctx, arg_LDY2 *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv addr = gen_get_yaddr();
+
+    gen_data_load(ctx, Rd, addr);
+    tcg_gen_addi_tl(addr, addr, 1); /* addr = addr + 1 */
+
+    gen_set_yaddr(addr);
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+static bool trans_LDY3(DisasContext *ctx, arg_LDY3 *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv addr = gen_get_yaddr();
+
+    tcg_gen_subi_tl(addr, addr, 1); /* addr = addr - 1 */
+    gen_data_load(ctx, Rd, addr);
+    gen_set_yaddr(addr);
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+static bool trans_LDDY(DisasContext *ctx, arg_LDDY *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv addr = gen_get_yaddr();
+
+    tcg_gen_addi_tl(addr, addr, a->imm); /* addr = addr + q */
+    gen_data_load(ctx, Rd, addr);
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+/*
+ *  Loads one byte indirect with or without displacement from the data space
+ *  to a register. For parts with SRAM, the data space consists of the Register
+ *  File, I/O memory and internal SRAM (and external SRAM if applicable). For
+ *  parts without SRAM, the data space consists of the Register File only. In
+ *  some parts the Flash Memory has been mapped to the data space and can be
+ *  read using this command. The EEPROM has a separate address space.  The data
+ *  location is pointed to by the Z (16 bits) Pointer Register in the Register
+ *  File. Memory access is limited to the current data segment of 64KB. To
+ *  access another data segment in devices with more than 64KB data space, the
+ *  RAMPZ in register in the I/O area has to be changed.  The Z-pointer Register
+ *  can either be left unchanged by the operation, or it can be post-incremented
+ *  or predecremented.  These features are especially suited for Stack Pointer
+ *  usage of the Z-pointer Register, however because the Z-pointer Register can
+ *  be used for indirect subroutine calls, indirect jumps and table lookup, it
+ *  is often more convenient to use the X or Y-pointer as a dedicated Stack
+ *  Pointer. Note that only the low byte of the Z-pointer is updated in devices
+ *  with no more than 256 bytes data space. For such devices, the high byte of
+ *  the pointer is not used by this instruction and can be used for other
+ *  purposes. The RAMPZ Register in the I/O area is updated in parts with more
+ *  than 64KB data space or more than 64KB Program memory, and the
+ *  increment/decrement/displacement is added to the entire 24-bit address on
+ *  such devices.  Not all variants of this instruction is available in all
+ *  devices. Refer to the device specific instruction set summary.  In the
+ *  Reduced Core tinyAVR the LD instruction can be used to achieve the same
+ *  operation as LPM since the program memory is mapped to the data memory
+ *  space.  For using the Z-pointer for table lookup in Program memory see the
+ *  LPM and ELPM instructions.
+ */
+static bool trans_LDZ2(DisasContext *ctx, arg_LDZ2 *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv addr = gen_get_zaddr();
+
+    gen_data_load(ctx, Rd, addr);
+    tcg_gen_addi_tl(addr, addr, 1); /* addr = addr + 1 */
+
+    gen_set_zaddr(addr);
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+static bool trans_LDZ3(DisasContext *ctx, arg_LDZ3 *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv addr = gen_get_zaddr();
+
+    tcg_gen_subi_tl(addr, addr, 1); /* addr = addr - 1 */
+    gen_data_load(ctx, Rd, addr);
+
+    gen_set_zaddr(addr);
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+static bool trans_LDDZ(DisasContext *ctx, arg_LDDZ *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv addr = gen_get_zaddr();
+
+    tcg_gen_addi_tl(addr, addr, a->imm); /* addr = addr + q */
+    gen_data_load(ctx, Rd, addr);
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+/*
+ *  Stores one byte from a Register to the data space. For parts with SRAM,
+ *  the data space consists of the Register File, I/O memory and internal SRAM
+ *  (and external SRAM if applicable). For parts without SRAM, the data space
+ *  consists of the Register File only. The EEPROM has a separate address space.
+ *  A 16-bit address must be supplied. Memory access is limited to the current
+ *  data segment of 64KB. The STS instruction uses the RAMPD Register to access
+ *  memory above 64KB. To access another data segment in devices with more than
+ *  64KB data space, the RAMPD in register in the I/O area has to be changed.
+ *  This instruction is not available in all devices. Refer to the device
+ *  specific instruction set summary.
+ */
+static bool trans_STS(DisasContext *ctx, arg_STS *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv addr = tcg_temp_new_i32();
+    TCGv H = cpu_rampD;
+    a->imm = next_word(ctx);
+
+    tcg_gen_mov_tl(addr, H); /* addr = H:M:L */
+    tcg_gen_shli_tl(addr, addr, 16);
+    tcg_gen_ori_tl(addr, addr, a->imm);
+    gen_data_store(ctx, Rd, addr);
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+/*
+ * Stores one byte indirect from a register to data space. For parts with SRAM,
+ * the data space consists of the Register File, I/O memory, and internal SRAM
+ * (and external SRAM if applicable). For parts without SRAM, the data space
+ * consists of the Register File only. The EEPROM has a separate address space.
+ *
+ * The data location is pointed to by the X (16 bits) Pointer Register in the
+ * Register File. Memory access is limited to the current data segment of 64KB.
+ * To access another data segment in devices with more than 64KB data space, the
+ * RAMPX in register in the I/O area has to be changed.
+ *
+ * The X-pointer Register can either be left unchanged by the operation, or it
+ * can be post-incremented or pre-decremented. These features are especially
+ * suited for accessing arrays, tables, and Stack Pointer usage of the
+ * X-pointer Register. Note that only the low byte of the X-pointer is updated
+ * in devices with no more than 256 bytes data space. For such devices, the high
+ * byte of the pointer is not used by this instruction and can be used for other
+ * purposes. The RAMPX Register in the I/O area is updated in parts with more
+ * than 64KB data space or more than 64KB Program memory, and the increment /
+ * decrement is added to the entire 24-bit address on such devices.
+ */
+static bool trans_STX1(DisasContext *ctx, arg_STX1 *a)
+{
+    TCGv Rd = cpu_r[a->rr];
+    TCGv addr = gen_get_xaddr();
+
+    gen_data_store(ctx, Rd, addr);
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+static bool trans_STX2(DisasContext *ctx, arg_STX2 *a)
+{
+    TCGv Rd = cpu_r[a->rr];
+    TCGv addr = gen_get_xaddr();
+
+    gen_data_store(ctx, Rd, addr);
+    tcg_gen_addi_tl(addr, addr, 1); /* addr = addr + 1 */
+    gen_set_xaddr(addr);
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+static bool trans_STX3(DisasContext *ctx, arg_STX3 *a)
+{
+    TCGv Rd = cpu_r[a->rr];
+    TCGv addr = gen_get_xaddr();
+
+    tcg_gen_subi_tl(addr, addr, 1); /* addr = addr - 1 */
+    gen_data_store(ctx, Rd, addr);
+    gen_set_xaddr(addr);
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+/*
+ * Stores one byte indirect with or without displacement from a register to data
+ * space. For parts with SRAM, the data space consists of the Register File, I/O
+ * memory, and internal SRAM (and external SRAM if applicable). For parts
+ * without SRAM, the data space consists of the Register File only. The EEPROM
+ * has a separate address space.
+ *
+ * The data location is pointed to by the Y (16 bits) Pointer Register in the
+ * Register File. Memory access is limited to the current data segment of 64KB.
+ * To access another data segment in devices with more than 64KB data space, the
+ * RAMPY in register in the I/O area has to be changed.
+ *
+ * The Y-pointer Register can either be left unchanged by the operation, or it
+ * can be post-incremented or pre-decremented. These features are especially
+ * suited for accessing arrays, tables, and Stack Pointer usage of the Y-pointer
+ * Register. Note that only the low byte of the Y-pointer is updated in devices
+ * with no more than 256 bytes data space. For such devices, the high byte of
+ * the pointer is not used by this instruction and can be used for other
+ * purposes. The RAMPY Register in the I/O area is updated in parts with more
+ * than 64KB data space or more than 64KB Program memory, and the increment /
+ * decrement / displacement is added to the entire 24-bit address on such
+ * devices.
+ */
+static bool trans_STY2(DisasContext *ctx, arg_STY2 *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv addr = gen_get_yaddr();
+
+    gen_data_store(ctx, Rd, addr);
+    tcg_gen_addi_tl(addr, addr, 1); /* addr = addr + 1 */
+    gen_set_yaddr(addr);
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+static bool trans_STY3(DisasContext *ctx, arg_STY3 *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv addr = gen_get_yaddr();
+
+    tcg_gen_subi_tl(addr, addr, 1); /* addr = addr - 1 */
+    gen_data_store(ctx, Rd, addr);
+    gen_set_yaddr(addr);
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+static bool trans_STDY(DisasContext *ctx, arg_STDY *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv addr = gen_get_yaddr();
+
+    tcg_gen_addi_tl(addr, addr, a->imm); /* addr = addr + q */
+    gen_data_store(ctx, Rd, addr);
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+/*
+ * Stores one byte indirect with or without displacement from a register to data
+ * space. For parts with SRAM, the data space consists of the Register File, I/O
+ * memory, and internal SRAM (and external SRAM if applicable). For parts
+ * without SRAM, the data space consists of the Register File only. The EEPROM
+ * has a separate address space.
+ *
+ * The data location is pointed to by the Y (16 bits) Pointer Register in the
+ * Register File. Memory access is limited to the current data segment of 64KB.
+ * To access another data segment in devices with more than 64KB data space, the
+ * RAMPY in register in the I/O area has to be changed.
+ *
+ * The Y-pointer Register can either be left unchanged by the operation, or it
+ * can be post-incremented or pre-decremented. These features are especially
+ * suited for accessing arrays, tables, and Stack Pointer usage of the Y-pointer
+ * Register. Note that only the low byte of the Y-pointer is updated in devices
+ * with no more than 256 bytes data space. For such devices, the high byte of
+ * the pointer is not used by this instruction and can be used for other
+ * purposes. The RAMPY Register in the I/O area is updated in parts with more
+ * than 64KB data space or more than 64KB Program memory, and the increment /
+ * decrement / displacement is added to the entire 24-bit address on such
+ * devices.
+ */
+static bool trans_STZ2(DisasContext *ctx, arg_STZ2 *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv addr = gen_get_zaddr();
+
+    gen_data_store(ctx, Rd, addr);
+    tcg_gen_addi_tl(addr, addr, 1); /* addr = addr + 1 */
+
+    gen_set_zaddr(addr);
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+static bool trans_STZ3(DisasContext *ctx, arg_STZ3 *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv addr = gen_get_zaddr();
+
+    tcg_gen_subi_tl(addr, addr, 1); /* addr = addr - 1 */
+    gen_data_store(ctx, Rd, addr);
+
+    gen_set_zaddr(addr);
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+static bool trans_STDZ(DisasContext *ctx, arg_STDZ *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv addr = gen_get_zaddr();
+
+    tcg_gen_addi_tl(addr, addr, a->imm); /* addr = addr + q */
+    gen_data_store(ctx, Rd, addr);
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+/*
+ *  Loads one byte pointed to by the Z-register into the destination
+ *  register Rd. This instruction features a 100% space effective constant
+ *  initialization or constant data fetch. The Program memory is organized in
+ *  16-bit words while the Z-pointer is a byte address. Thus, the least
+ *  significant bit of the Z-pointer selects either low byte (ZLSB = 0) or high
+ *  byte (ZLSB = 1). This instruction can address the first 64KB (32K words) of
+ *  Program memory. The Zpointer Register can either be left unchanged by the
+ *  operation, or it can be incremented. The incrementation does not apply to
+ *  the RAMPZ Register.
+ *
+ *  Devices with Self-Programming capability can use the LPM instruction to read
+ *  the Fuse and Lock bit values.
+ */
+static bool trans_LPM1(DisasContext *ctx, arg_LPM1 *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_LPM)) {
+        return true;
+    }
+
+    TCGv Rd = cpu_r[0];
+    TCGv addr = tcg_temp_new_i32();
+    TCGv H = cpu_r[31];
+    TCGv L = cpu_r[30];
+
+    tcg_gen_shli_tl(addr, H, 8); /* addr = H:L */
+    tcg_gen_or_tl(addr, addr, L);
+    tcg_gen_qemu_ld8u(Rd, addr, MMU_CODE_IDX); /* Rd = mem[addr] */
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+static bool trans_LPM2(DisasContext *ctx, arg_LPM2 *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_LPM)) {
+        return true;
+    }
+
+    TCGv Rd = cpu_r[a->rd];
+    TCGv addr = tcg_temp_new_i32();
+    TCGv H = cpu_r[31];
+    TCGv L = cpu_r[30];
+
+    tcg_gen_shli_tl(addr, H, 8); /* addr = H:L */
+    tcg_gen_or_tl(addr, addr, L);
+    tcg_gen_qemu_ld8u(Rd, addr, MMU_CODE_IDX); /* Rd = mem[addr] */
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+static bool trans_LPMX(DisasContext *ctx, arg_LPMX *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_LPMX)) {
+        return true;
+    }
+
+    TCGv Rd = cpu_r[a->rd];
+    TCGv addr = tcg_temp_new_i32();
+    TCGv H = cpu_r[31];
+    TCGv L = cpu_r[30];
+
+    tcg_gen_shli_tl(addr, H, 8); /* addr = H:L */
+    tcg_gen_or_tl(addr, addr, L);
+    tcg_gen_qemu_ld8u(Rd, addr, MMU_CODE_IDX); /* Rd = mem[addr] */
+    tcg_gen_addi_tl(addr, addr, 1); /* addr = addr + 1 */
+    tcg_gen_andi_tl(L, addr, 0xff);
+    tcg_gen_shri_tl(addr, addr, 8);
+    tcg_gen_andi_tl(H, addr, 0xff);
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+
+/*
+ *  Loads one byte pointed to by the Z-register and the RAMPZ Register in
+ *  the I/O space, and places this byte in the destination register Rd. This
+ *  instruction features a 100% space effective constant initialization or
+ *  constant data fetch. The Program memory is organized in 16-bit words while
+ *  the Z-pointer is a byte address. Thus, the least significant bit of the
+ *  Z-pointer selects either low byte (ZLSB = 0) or high byte (ZLSB = 1). This
+ *  instruction can address the entire Program memory space. The Z-pointer
+ *  Register can either be left unchanged by the operation, or it can be
+ *  incremented. The incrementation applies to the entire 24-bit concatenation
+ *  of the RAMPZ and Z-pointer Registers.
+ *
+ *  Devices with Self-Programming capability can use the ELPM instruction to
+ *  read the Fuse and Lock bit value.
+ */
+static bool trans_ELPM1(DisasContext *ctx, arg_ELPM1 *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_ELPM)) {
+        return true;
+    }
+
+    TCGv Rd = cpu_r[0];
+    TCGv addr = gen_get_zaddr();
+
+    tcg_gen_qemu_ld8u(Rd, addr, MMU_CODE_IDX); /* Rd = mem[addr] */
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+static bool trans_ELPM2(DisasContext *ctx, arg_ELPM2 *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_ELPM)) {
+        return true;
+    }
+
+    TCGv Rd = cpu_r[a->rd];
+    TCGv addr = gen_get_zaddr();
+
+    tcg_gen_qemu_ld8u(Rd, addr, MMU_CODE_IDX); /* Rd = mem[addr] */
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+static bool trans_ELPMX(DisasContext *ctx, arg_ELPMX *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_ELPMX)) {
+        return true;
+    }
+
+    TCGv Rd = cpu_r[a->rd];
+    TCGv addr = gen_get_zaddr();
+
+    tcg_gen_qemu_ld8u(Rd, addr, MMU_CODE_IDX); /* Rd = mem[addr] */
+    tcg_gen_addi_tl(addr, addr, 1); /* addr = addr + 1 */
+    gen_set_zaddr(addr);
+
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+/*
+ *  SPM can be used to erase a page in the Program memory, to write a page
+ *  in the Program memory (that is already erased), and to set Boot Loader Lock
+ *  bits. In some devices, the Program memory can be written one word at a time,
+ *  in other devices an entire page can be programmed simultaneously after first
+ *  filling a temporary page buffer. In all cases, the Program memory must be
+ *  erased one page at a time. When erasing the Program memory, the RAMPZ and
+ *  Z-register are used as page address. When writing the Program memory, the
+ *  RAMPZ and Z-register are used as page or word address, and the R1:R0
+ *  register pair is used as data(1). When setting the Boot Loader Lock bits,
+ *  the R1:R0 register pair is used as data. Refer to the device documentation
+ *  for detailed description of SPM usage. This instruction can address the
+ *  entire Program memory.
+ *
+ *  The SPM instruction is not available in all devices. Refer to the device
+ *  specific instruction set summary.
+ *
+ *  Note: 1. R1 determines the instruction high byte, and R0 determines the
+ *  instruction low byte.
+ */
+static bool trans_SPM(DisasContext *ctx, arg_SPM *a)
+{
+    /* TODO */
+    if (!avr_have_feature(ctx, AVR_FEATURE_SPM)) {
+        return true;
+    }
+
+    return true;
+}
+
+static bool trans_SPMX(DisasContext *ctx, arg_SPMX *a)
+{
+    /* TODO */
+    if (!avr_have_feature(ctx, AVR_FEATURE_SPMX)) {
+        return true;
+    }
+
+    return true;
+}
+
+/*
+ *  Loads data from the I/O Space (Ports, Timers, Configuration Registers,
+ *  etc.) into register Rd in the Register File.
+ */
+static bool trans_IN(DisasContext *ctx, arg_IN *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv port = tcg_const_i32(a->imm);
+
+    gen_helper_inb(Rd, cpu_env, port);
+
+    tcg_temp_free_i32(port);
+
+    return true;
+}
+
+/*
+ *  Stores data from register Rr in the Register File to I/O Space (Ports,
+ *  Timers, Configuration Registers, etc.).
+ */
+static bool trans_OUT(DisasContext *ctx, arg_OUT *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv port = tcg_const_i32(a->imm);
+
+    gen_helper_outb(cpu_env, port, Rd);
+
+    tcg_temp_free_i32(port);
+
+    return true;
+}
+
+/*
+ *  This instruction stores the contents of register Rr on the STACK. The
+ *  Stack Pointer is post-decremented by 1 after the PUSH.  This instruction is
+ *  not available in all devices. Refer to the device specific instruction set
+ *  summary.
+ */
+static bool trans_PUSH(DisasContext *ctx, arg_PUSH *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+
+    gen_data_store(ctx, Rd, cpu_sp);
+    tcg_gen_subi_tl(cpu_sp, cpu_sp, 1);
+
+    return true;
+}
+
+/*
+ *  This instruction loads register Rd with a byte from the STACK. The Stack
+ *  Pointer is pre-incremented by 1 before the POP.  This instruction is not
+ *  available in all devices. Refer to the device specific instruction set
+ *  summary.
+ */
+static bool trans_POP(DisasContext *ctx, arg_POP *a)
+{
+    /*
+     * Using a temp to work around some strange behaviour:
+     * tcg_gen_addi_tl(cpu_sp, cpu_sp, 1);
+     * gen_data_load(ctx, Rd, cpu_sp);
+     * seems to cause the add to happen twice.
+     * This doesn't happen if either the add or the load is removed.
+     */
+    TCGv t1 = tcg_temp_new_i32();
+    TCGv Rd = cpu_r[a->rd];
+
+    tcg_gen_addi_tl(t1, cpu_sp, 1);
+    gen_data_load(ctx, Rd, t1);
+    tcg_gen_mov_tl(cpu_sp, t1);
+
+    return true;
+}
+
+/*
+ *  Exchanges one byte indirect between register and data space.  The data
+ *  location is pointed to by the Z (16 bits) Pointer Register in the Register
+ *  File. Memory access is limited to the current data segment of 64KB. To
+ *  access another data segment in devices with more than 64KB data space, the
+ *  RAMPZ in register in the I/O area has to be changed.
+ *
+ *  The Z-pointer Register is left unchanged by the operation. This instruction
+ *  is especially suited for writing/reading status bits stored in SRAM.
+ */
+static bool trans_XCH(DisasContext *ctx, arg_XCH *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_RMW)) {
+        return true;
+    }
+
+    TCGv Rd = cpu_r[a->rd];
+    TCGv t0 = tcg_temp_new_i32();
+    TCGv addr = gen_get_zaddr();
+
+    gen_data_load(ctx, t0, addr);
+    gen_data_store(ctx, Rd, addr);
+    tcg_gen_mov_tl(Rd, t0);
+
+    tcg_temp_free_i32(t0);
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+/*
+ *  Load one byte indirect from data space to register and set bits in data
+ *  space specified by the register. The instruction can only be used towards
+ *  internal SRAM.  The data location is pointed to by the Z (16 bits) Pointer
+ *  Register in the Register File. Memory access is limited to the current data
+ *  segment of 64KB. To access another data segment in devices with more than
+ *  64KB data space, the RAMPZ in register in the I/O area has to be changed.
+ *
+ *  The Z-pointer Register is left unchanged by the operation. This instruction
+ *  is especially suited for setting status bits stored in SRAM.
+ */
+static bool trans_LAS(DisasContext *ctx, arg_LAS *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_RMW)) {
+        return true;
+    }
+
+    TCGv Rr = cpu_r[a->rd];
+    TCGv addr = gen_get_zaddr();
+    TCGv t0 = tcg_temp_new_i32();
+    TCGv t1 = tcg_temp_new_i32();
+
+    gen_data_load(ctx, t0, addr); /* t0 = mem[addr] */
+    tcg_gen_or_tl(t1, t0, Rr);
+    tcg_gen_mov_tl(Rr, t0); /* Rr = t0 */
+    gen_data_store(ctx, t1, addr); /* mem[addr] = t1 */
+
+    tcg_temp_free_i32(t1);
+    tcg_temp_free_i32(t0);
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+/*
+ *  Load one byte indirect from data space to register and stores and clear
+ *  the bits in data space specified by the register. The instruction can
+ *  only be used towards internal SRAM.  The data location is pointed to by
+ *  the Z (16 bits) Pointer Register in the Register File. Memory access is
+ *  limited to the current data segment of 64KB. To access another data
+ *  segment in devices with more than 64KB data space, the RAMPZ in register
+ *  in the I/O area has to be changed.
+ *
+ *  The Z-pointer Register is left unchanged by the operation. This instruction
+ *  is especially suited for clearing status bits stored in SRAM.
+ */
+static bool trans_LAC(DisasContext *ctx, arg_LAC *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_RMW)) {
+        return true;
+    }
+
+    TCGv Rr = cpu_r[a->rd];
+    TCGv addr = gen_get_zaddr();
+    TCGv t0 = tcg_temp_new_i32();
+    TCGv t1 = tcg_temp_new_i32();
+
+    gen_data_load(ctx, t0, addr); /* t0 = mem[addr] */
+    tcg_gen_andc_tl(t1, t0, Rr); /* t1 = t0 & (0xff - Rr) = t0 & ~Rr */
+    tcg_gen_mov_tl(Rr, t0); /* Rr = t0 */
+    gen_data_store(ctx, t1, addr); /* mem[addr] = t1 */
+
+    tcg_temp_free_i32(t1);
+    tcg_temp_free_i32(t0);
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
+
+
+/*
+ *  Load one byte indirect from data space to register and toggles bits in
+ *  the data space specified by the register.  The instruction can only be used
+ *  towards SRAM.  The data location is pointed to by the Z (16 bits) Pointer
+ *  Register in the Register File. Memory access is limited to the current data
+ *  segment of 64KB. To access another data segment in devices with more than
+ *  64KB data space, the RAMPZ in register in the I/O area has to be changed.
+ *
+ *  The Z-pointer Register is left unchanged by the operation. This instruction
+ *  is especially suited for changing status bits stored in SRAM.
+ */
+static bool trans_LAT(DisasContext *ctx, arg_LAT *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_RMW)) {
+        return true;
+    }
+
+    TCGv Rd = cpu_r[a->rd];
+    TCGv addr = gen_get_zaddr();
+    TCGv t0 = tcg_temp_new_i32();
+    TCGv t1 = tcg_temp_new_i32();
+
+    gen_data_load(ctx, t0, addr); /* t0 = mem[addr] */
+    tcg_gen_xor_tl(t1, t0, Rd);
+    tcg_gen_mov_tl(Rd, t0); /* Rd = t0 */
+    gen_data_store(ctx, t1, addr); /* mem[addr] = t1 */
+
+    tcg_temp_free_i32(t1);
+    tcg_temp_free_i32(t0);
+    tcg_temp_free_i32(addr);
+
+    return true;
+}
target/avr/insn.decode +43 -0

46 lines not on this page · raw message

Signature · 1 line
2.17.2 (Apple Git-113)
#1 lore raw

PATCH v40 1/21 target/avr: Add outward facing interfaces and core CPU logic

In reply to Michael Rolnik

This includes:
- CPU data structures
- object model classes and functions
- migration functions
- GDB hooks

---
 target/avr/cpu-param.h |  37 ++
 target/avr/cpu-qom.h   |  54 +++
 target/avr/cpu.h       | 258 +++++++++++++
 target/avr/cpu.c       | 826 +++++++++++++++++++++++++++++++++++++++++
 target/avr/gdbstub.c   |  84 +++++
 target/avr/machine.c   | 121 ++++++
 gdb-xml/avr-cpu.xml    |  49 +++
 7 files changed, 1429 insertions(+)
 create mode 100644 target/avr/cpu-param.h
 create mode 100644 target/avr/cpu-qom.h
 create mode 100644 target/avr/cpu.h
 create mode 100644 target/avr/cpu.c
 create mode 100644 target/avr/gdbstub.c
 create mode 100644 target/avr/machine.c
 create mode 100644 gdb-xml/avr-cpu.xml
target/avr/cpu-param.h new file +37 -0
new file mode 100644
index 0000000000..0c29ce4223
@@ -0,0 +1,37 @@
+/*
+ * QEMU AVR CPU
+ *
+ * Copyright (c) 2019 Michael Rolnik
+ *
+ * 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/lgpl-2.1.html>
+ */
+
+#ifndef AVR_CPU_PARAM_H
+#define AVR_CPU_PARAM_H
+
+#define TARGET_LONG_BITS 32
+/*
+ * TARGET_PAGE_BITS cannot be more than 8 bits because
+ * 1.  all IO registers occupy [0x0000 .. 0x00ff] address range, and they
+ *     should be implemented as a device and not memory
+ * 2.  SRAM starts at the address 0x0100
+ */
+#define TARGET_PAGE_BITS 8
+#define TARGET_PHYS_ADDR_SPACE_BITS 24
+#define TARGET_VIRT_ADDR_SPACE_BITS 24
+#define NB_MMU_MODES 2
+
+
+#endif
target/avr/cpu-qom.h new file +54 -0

54 lines not on this page · raw message

target/avr/cpu.h new file +258 -0

258 lines not on this page · raw message

target/avr/cpu.c new file +826 -0

826 lines not on this page · raw message

target/avr/gdbstub.c new file +84 -0

84 lines not on this page · raw message

target/avr/machine.c new file +121 -0

121 lines not on this page · raw message

gdb-xml/avr-cpu.xml new file +49 -0

49 lines not on this page · raw message

Signature · 1 line
2.17.2 (Apple Git-113)
#8 lore raw

PATCH v40 7/21 target/avr: Add instruction translation - Bit and Bit-test Instructions

In reply to Michael Rolnik

This includes:
- LSR, ROR
- ASR
- SWAP
- SBI, CBI
- BST, BLD
- BSET, BCLR

---
 target/avr/translate.c | 241 +++++++++++++++++++++++++++++++++++++++++
 target/avr/insn.decode |  14 +++
 2 files changed, 255 insertions(+)
target/avr/translate.c +241 -0
index 9c66734a19..55706c6b29 100644
@@ -2440,3 +2440,244 @@ static bool trans_LAT(DisasContext *ctx, arg_LAT *a)
 
     return true;
 }
+
+/*
+ * Bit and Bit-test Instructions
+ */
+static void gen_rshift_ZNVSf(TCGv R)
+{
+    tcg_gen_setcondi_tl(TCG_COND_EQ, cpu_Zf, R, 0); /* Zf = R == 0 */
+    tcg_gen_shri_tl(cpu_Nf, R, 7); /* Nf = R(7) */
+    tcg_gen_xor_tl(cpu_Vf, cpu_Nf, cpu_Cf);
+    tcg_gen_xor_tl(cpu_Sf, cpu_Nf, cpu_Vf); /* Sf = Nf ^ Vf */
+}
+
+/*
+ *  Shifts all bits in Rd one place to the right. Bit 7 is cleared. Bit 0 is
+ *  loaded into the C Flag of the SREG. This operation effectively divides an
+ *  unsigned value by two. The C Flag can be used to round the result.
+ */
+static bool trans_LSR(DisasContext *ctx, arg_LSR *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+
+    tcg_gen_andi_tl(cpu_Cf, Rd, 1);
+    tcg_gen_shri_tl(Rd, Rd, 1);
+    /* update status register */
+    tcg_gen_setcondi_tl(TCG_COND_EQ, cpu_Zf, Rd, 0); /* Zf = Rd == 0 */
+    tcg_gen_movi_tl(cpu_Nf, 0);
+    tcg_gen_mov_tl(cpu_Vf, cpu_Cf);
+    tcg_gen_mov_tl(cpu_Sf, cpu_Vf);
+
+    return true;
+}
+
+/*
+ *  Shifts all bits in Rd one place to the right. The C Flag is shifted into
+ *  bit 7 of Rd. Bit 0 is shifted into the C Flag.  This operation, combined
+ *  with ASR, effectively divides multi-byte signed values by two. Combined with
+ *  LSR it effectively divides multi-byte unsigned values by two. The Carry Flag
+ *  can be used to round the result.
+ */
+static bool trans_ROR(DisasContext *ctx, arg_ROR *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv t0 = tcg_temp_new_i32();
+
+    tcg_gen_shli_tl(t0, cpu_Cf, 7);
+    /* update status register */
+    tcg_gen_andi_tl(cpu_Cf, Rd, 1);
+    /* update output register */
+    tcg_gen_shri_tl(Rd, Rd, 1);
+    tcg_gen_or_tl(Rd, Rd, t0);
+    /* update status register */
+    gen_rshift_ZNVSf(Rd);
+
+    tcg_temp_free_i32(t0);
+
+    return true;
+}
+
+/*
+ *  Shifts all bits in Rd one place to the right. Bit 7 is held constant. Bit 0
+ *  is loaded into the C Flag of the SREG. This operation effectively divides a
+ *  signed value by two without changing its sign. The Carry Flag can be used to
+ *  round the result.
+ */
+static bool trans_ASR(DisasContext *ctx, arg_ASR *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv t0 = tcg_temp_new_i32();
+
+    /* update status register */
+    tcg_gen_andi_tl(cpu_Cf, Rd, 1); /* Cf = Rd(0) */
+    /* update output register */
+    tcg_gen_andi_tl(t0, Rd, 0x80); /* Rd = (Rd & 0x80) | (Rd >> 1) */
+    tcg_gen_shri_tl(Rd, Rd, 1);
+    tcg_gen_or_tl(Rd, Rd, t0);
+    /* update status register */
+    gen_rshift_ZNVSf(Rd);
+
+    tcg_temp_free_i32(t0);
+
+    return true;
+}
+
+/*
+ *  Swaps high and low nibbles in a register.
+ */
+static bool trans_SWAP(DisasContext *ctx, arg_SWAP *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv t0 = tcg_temp_new_i32();
+    TCGv t1 = tcg_temp_new_i32();
+
+    tcg_gen_andi_tl(t0, Rd, 0x0f);
+    tcg_gen_shli_tl(t0, t0, 4);
+    tcg_gen_andi_tl(t1, Rd, 0xf0);
+    tcg_gen_shri_tl(t1, t1, 4);
+    tcg_gen_or_tl(Rd, t0, t1);
+
+    tcg_temp_free_i32(t1);
+    tcg_temp_free_i32(t0);
+
+    return true;
+}
+
+/*
+ *  Sets a specified bit in an I/O Register. This instruction operates on
+ *  the lower 32 I/O Registers -- addresses 0-31.
+ */
+static bool trans_SBI(DisasContext *ctx, arg_SBI *a)
+{
+    TCGv data = tcg_temp_new_i32();
+    TCGv port = tcg_const_i32(a->reg);
+
+    gen_helper_inb(data, cpu_env, port);
+    tcg_gen_ori_tl(data, data, 1 << a->bit);
+    gen_helper_outb(cpu_env, port, data);
+
+    tcg_temp_free_i32(port);
+    tcg_temp_free_i32(data);
+
+    return true;
+}
+
+/*
+ *  Clears a specified bit in an I/O Register. This instruction operates on
+ *  the lower 32 I/O Registers -- addresses 0-31.
+ */
+static bool trans_CBI(DisasContext *ctx, arg_CBI *a)
+{
+    TCGv data = tcg_temp_new_i32();
+    TCGv port = tcg_const_i32(a->reg);
+
+    gen_helper_inb(data, cpu_env, port);
+    tcg_gen_andi_tl(data, data, ~(1 << a->bit));
+    gen_helper_outb(cpu_env, port, data);
+
+    tcg_temp_free_i32(data);
+    tcg_temp_free_i32(port);
+
+    return true;
+}
+
+/*
+ *  Stores bit b from Rd to the T Flag in SREG (Status Register).
+ */
+static bool trans_BST(DisasContext *ctx, arg_BST *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+
+    tcg_gen_andi_tl(cpu_Tf, Rd, 1 << a->bit);
+    tcg_gen_shri_tl(cpu_Tf, cpu_Tf, a->bit);
+
+    return true;
+}
+
+/*
+ *  Copies the T Flag in the SREG (Status Register) to bit b in register Rd.
+ */
+static bool trans_BLD(DisasContext *ctx, arg_BLD *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv t1 = tcg_temp_new_i32();
+
+    tcg_gen_andi_tl(Rd, Rd, ~(1u << a->bit)); /* clear bit */
+    tcg_gen_shli_tl(t1, cpu_Tf, a->bit); /* create mask */
+    tcg_gen_or_tl(Rd, Rd, t1);
+
+    tcg_temp_free_i32(t1);
+
+    return true;
+}
+
+/*
+ *  Sets a single Flag or bit in SREG.
+ */
+static bool trans_BSET(DisasContext *ctx, arg_BSET *a)
+{
+    switch (a->bit) {
+    case 0x00:
+        tcg_gen_movi_tl(cpu_Cf, 0x01);
+        break;
+    case 0x01:
+        tcg_gen_movi_tl(cpu_Zf, 0x01);
+        break;
+    case 0x02:
+        tcg_gen_movi_tl(cpu_Nf, 0x01);
+        break;
+    case 0x03:
+        tcg_gen_movi_tl(cpu_Vf, 0x01);
+        break;
+    case 0x04:
+        tcg_gen_movi_tl(cpu_Sf, 0x01);
+        break;
+    case 0x05:
+        tcg_gen_movi_tl(cpu_Hf, 0x01);
+        break;
+    case 0x06:
+        tcg_gen_movi_tl(cpu_Tf, 0x01);
+        break;
+    case 0x07:
+        tcg_gen_movi_tl(cpu_If, 0x01);
+        break;
+    }
+
+    return true;
+}
+
+/*
+ *  Clears a single Flag in SREG.
+ */
+static bool trans_BCLR(DisasContext *ctx, arg_BCLR *a)
+{
+    switch (a->bit) {
+    case 0x00:
+        tcg_gen_movi_tl(cpu_Cf, 0x00);
+        break;
+    case 0x01:
+        tcg_gen_movi_tl(cpu_Zf, 0x00);
+        break;
+    case 0x02:
+        tcg_gen_movi_tl(cpu_Nf, 0x00);
+        break;
+    case 0x03:
+        tcg_gen_movi_tl(cpu_Vf, 0x00);
+        break;
+    case 0x04:
+        tcg_gen_movi_tl(cpu_Sf, 0x00);
+        break;
+    case 0x05:
+        tcg_gen_movi_tl(cpu_Hf, 0x00);
+        break;
+    case 0x06:
+        tcg_gen_movi_tl(cpu_Tf, 0x00);
+        break;
+    case 0x07:
+        tcg_gen_movi_tl(cpu_If, 0x00);
+        break;
+    }
+
+    return true;
+}
target/avr/insn.decode +14 -0
index 3f9304f8b0..4ee55862b2 100644
@@ -158,3 +158,17 @@ XCH             1001 001 rd:5 0100
 LAC             1001 001 rd:5 0110
 LAS             1001 001 rd:5 0101
 LAT             1001 001 rd:5 0111
+
+#
+# Bit and Bit-test Instructions
+#
+LSR             1001 010 rd:5 0110
+ROR             1001 010 rd:5 0111
+ASR             1001 010 rd:5 0101
+SWAP            1001 010 rd:5 0010
+SBI             1001 1010 reg:5 bit:3
+CBI             1001 1000 reg:5 bit:3
+BST             1111 101 rd:5 0 bit:3
+BLD             1111 100 rd:5 0 bit:3
+BSET            1001 0100 0 bit:3 1000
+BCLR            1001 0100 1 bit:3 1000
Signature · 1 line
2.17.2 (Apple Git-113)
#4 lore raw

PATCH v40 4/21 target/avr: Add instruction translation - Arithmetic and Logic Instructions

In reply to Michael Rolnik

This includes:
- ADD, ADC, ADIW
- SBIW, SUB, SUBI, SBC, SBCI
- AND, ANDI
- OR, ORI, EOR
- COM, NEG
- INC, DEC
- MUL, MULS, MULSU
- FMUL, FMULS, FMULSU
- DES

---
 target/avr/translate.c | 751 +++++++++++++++++++++++++++++++++++++++++
 target/avr/insn.decode |  93 +++++
 2 files changed, 844 insertions(+)
 create mode 100644 target/avr/insn.decode
target/avr/translate.c +751 -0
index cfdff65620..2b6411bbed 100644
@@ -170,3 +170,754 @@ static bool avr_have_feature(DisasContext *ctx, int feature)
 static bool decode_insn(DisasContext *ctx, uint16_t insn);
 #include "decode_insn.inc.c"
 
+/*
+ * Arithmetic Instructions
+ */
+
+static void gen_add_CHf(TCGv R, TCGv Rd, TCGv Rr)
+{
+    TCGv t1 = tcg_temp_new_i32();
+    TCGv t2 = tcg_temp_new_i32();
+    TCGv t3 = tcg_temp_new_i32();
+
+    tcg_gen_and_tl(t1, Rd, Rr); /* t1 = Rd & Rr */
+    tcg_gen_andc_tl(t2, Rd, R); /* t2 = Rd & ~R */
+    tcg_gen_andc_tl(t3, Rr, R); /* t3 = Rr & ~R */
+    tcg_gen_or_tl(t1, t1, t2); /* t1 = t1 | t2 | t3 */
+    tcg_gen_or_tl(t1, t1, t3);
+    tcg_gen_shri_tl(cpu_Cf, t1, 7); /* Cf = t1(7) */
+    tcg_gen_shri_tl(cpu_Hf, t1, 3); /* Hf = t1(3) */
+    tcg_gen_andi_tl(cpu_Hf, cpu_Hf, 1);
+
+    tcg_temp_free_i32(t3);
+    tcg_temp_free_i32(t2);
+    tcg_temp_free_i32(t1);
+}
+
+
+static void gen_add_Vf(TCGv R, TCGv Rd, TCGv Rr)
+{
+    TCGv t1 = tcg_temp_new_i32();
+    TCGv t2 = tcg_temp_new_i32();
+
+    /* t1 = Rd & Rr & ~R | ~Rd & ~Rr & R */
+    /*    = (Rd ^ R) & ~(Rd ^ Rr) */
+    tcg_gen_xor_tl(t1, Rd, R);
+    tcg_gen_xor_tl(t2, Rd, Rr);
+    tcg_gen_andc_tl(t1, t1, t2);
+    tcg_gen_shri_tl(cpu_Vf, t1, 7); /* Vf = t1(7) */
+
+    tcg_temp_free_i32(t2);
+    tcg_temp_free_i32(t1);
+}
+
+
+static void gen_sub_CHf(TCGv R, TCGv Rd, TCGv Rr)
+{
+    TCGv t1 = tcg_temp_new_i32();
+    TCGv t2 = tcg_temp_new_i32();
+    TCGv t3 = tcg_temp_new_i32();
+
+    tcg_gen_not_tl(t1, Rd); /* t1 = ~Rd */
+    tcg_gen_and_tl(t2, t1, Rr); /* t2 = ~Rd & Rr */
+    tcg_gen_or_tl(t3, t1, Rr); /* t3 = (~Rd | Rr) & R */
+    tcg_gen_and_tl(t3, t3, R);
+    tcg_gen_or_tl(t2, t2, t3); /* t2 = ~Rd & Rr | ~Rd & R | R & Rr */
+    tcg_gen_shri_tl(cpu_Cf, t2, 7); /* Cf = t2(7) */
+    tcg_gen_shri_tl(cpu_Hf, t2, 3); /* Hf = t2(3) */
+    tcg_gen_andi_tl(cpu_Hf, cpu_Hf, 1);
+
+    tcg_temp_free_i32(t3);
+    tcg_temp_free_i32(t2);
+    tcg_temp_free_i32(t1);
+}
+
+
+static void gen_sub_Vf(TCGv R, TCGv Rd, TCGv Rr)
+{
+    TCGv t1 = tcg_temp_new_i32();
+    TCGv t2 = tcg_temp_new_i32();
+
+    /* t1 = Rd & ~Rr & ~R | ~Rd & Rr & R */
+    /*    = (Rd ^ R) & (Rd ^ R) */
+    tcg_gen_xor_tl(t1, Rd, R);
+    tcg_gen_xor_tl(t2, Rd, Rr);
+    tcg_gen_and_tl(t1, t1, t2);
+    tcg_gen_shri_tl(cpu_Vf, t1, 7); /* Vf = t1(7) */
+
+    tcg_temp_free_i32(t2);
+    tcg_temp_free_i32(t1);
+}
+
+
+static void gen_NSf(TCGv R)
+{
+    tcg_gen_shri_tl(cpu_Nf, R, 7); /* Nf = R(7) */
+    tcg_gen_xor_tl(cpu_Sf, cpu_Nf, cpu_Vf); /* Sf = Nf ^ Vf */
+}
+
+
+static void gen_ZNSf(TCGv R)
+{
+    tcg_gen_setcondi_tl(TCG_COND_EQ, cpu_Zf, R, 0); /* Zf = R == 0 */
+    tcg_gen_shri_tl(cpu_Nf, R, 7); /* Nf = R(7) */
+    tcg_gen_xor_tl(cpu_Sf, cpu_Nf, cpu_Vf); /* Sf = Nf ^ Vf */
+}
+
+/*
+ *  Adds two registers without the C Flag and places the result in the
+ *  destination register Rd.
+ */
+static bool trans_ADD(DisasContext *ctx, arg_ADD *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv Rr = cpu_r[a->rr];
+    TCGv R = tcg_temp_new_i32();
+
+    tcg_gen_add_tl(R, Rd, Rr); /* Rd = Rd + Rr */
+    tcg_gen_andi_tl(R, R, 0xff); /* make it 8 bits */
+    /* update status register */
+    gen_add_CHf(R, Rd, Rr);
+    gen_add_Vf(R, Rd, Rr);
+    gen_ZNSf(R);
+    /* update output registers */
+    tcg_gen_mov_tl(Rd, R);
+
+    tcg_temp_free_i32(R);
+
+    return true;
+}
+
+/*
+ *  Adds two registers and the contents of the C Flag and places the result in
+ *  the destination register Rd.
+ */
+static bool trans_ADC(DisasContext *ctx, arg_ADC *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv Rr = cpu_r[a->rr];
+    TCGv R = tcg_temp_new_i32();
+
+    tcg_gen_add_tl(R, Rd, Rr); /* R = Rd + Rr + Cf */
+    tcg_gen_add_tl(R, R, cpu_Cf);
+    tcg_gen_andi_tl(R, R, 0xff); /* make it 8 bits */
+    /* update status register */
+    gen_add_CHf(R, Rd, Rr);
+    gen_add_Vf(R, Rd, Rr);
+    gen_ZNSf(R);
+    /* update output registers */
+    tcg_gen_mov_tl(Rd, R);
+
+    tcg_temp_free_i32(R);
+
+    return true;
+}
+
+/*
+ *  Adds an immediate value (0 - 63) to a register pair and places the result
+ *  in the register pair. This instruction operates on the upper four register
+ *  pairs, and is well suited for operations on the pointer registers.  This
+ *  instruction is not available in all devices. Refer to the device specific
+ *  instruction set summary.
+ */
+static bool trans_ADIW(DisasContext *ctx, arg_ADIW *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_ADIW_SBIW)) {
+        return true;
+    }
+
+    TCGv RdL = cpu_r[a->rd];
+    TCGv RdH = cpu_r[a->rd + 1];
+    int Imm = (a->imm);
+    TCGv R = tcg_temp_new_i32();
+    TCGv Rd = tcg_temp_new_i32();
+
+    tcg_gen_deposit_tl(Rd, RdL, RdH, 8, 8); /* Rd = RdH:RdL */
+    tcg_gen_addi_tl(R, Rd, Imm); /* R = Rd + Imm */
+    tcg_gen_andi_tl(R, R, 0xffff); /* make it 16 bits */
+    /* update status register */
+    tcg_gen_andc_tl(cpu_Cf, Rd, R); /* Cf = Rd & ~R */
+    tcg_gen_shri_tl(cpu_Cf, cpu_Cf, 15);
+    tcg_gen_andc_tl(cpu_Vf, R, Rd); /* Vf = R & ~Rd */
+    tcg_gen_shri_tl(cpu_Vf, cpu_Vf, 15);
+    tcg_gen_setcondi_tl(TCG_COND_EQ, cpu_Zf, R, 0); /* Zf = R == 0 */
+    tcg_gen_shri_tl(cpu_Nf, R, 15); /* Nf = R(15) */
+    tcg_gen_xor_tl(cpu_Sf, cpu_Nf, cpu_Vf);/* Sf = Nf ^ Vf */
+    /* update output registers */
+    tcg_gen_andi_tl(RdL, R, 0xff);
+    tcg_gen_shri_tl(RdH, R, 8);
+
+    tcg_temp_free_i32(Rd);
+    tcg_temp_free_i32(R);
+
+    return true;
+}
+
+/*
+ *  Subtracts two registers and places the result in the destination
+ *  register Rd.
+ */
+static bool trans_SUB(DisasContext *ctx, arg_SUB *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv Rr = cpu_r[a->rr];
+    TCGv R = tcg_temp_new_i32();
+
+    tcg_gen_sub_tl(R, Rd, Rr); /* R = Rd - Rr */
+    tcg_gen_andi_tl(R, R, 0xff); /* make it 8 bits */
+    /* update status register */
+    tcg_gen_andc_tl(cpu_Cf, Rd, R); /* Cf = Rd & ~R */
+    gen_sub_CHf(R, Rd, Rr);
+    gen_sub_Vf(R, Rd, Rr);
+    gen_ZNSf(R);
+    /* update output registers */
+    tcg_gen_mov_tl(Rd, R);
+
+    tcg_temp_free_i32(R);
+
+    return true;
+}
+
+/*
+ *  Subtracts a register and a constant and places the result in the
+ *  destination register Rd. This instruction is working on Register R16 to R31
+ *  and is very well suited for operations on the X, Y, and Z-pointers.
+ */
+static bool trans_SUBI(DisasContext *ctx, arg_SUBI *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv Rr = tcg_const_i32(a->imm);
+    TCGv R = tcg_temp_new_i32();
+
+    tcg_gen_sub_tl(R, Rd, Rr); /* R = Rd - Imm */
+    tcg_gen_andi_tl(R, R, 0xff); /* make it 8 bits */
+    /* update status register */
+    gen_sub_CHf(R, Rd, Rr);
+    gen_sub_Vf(R, Rd, Rr);
+    gen_ZNSf(R);
+    /* update output registers */
+    tcg_gen_mov_tl(Rd, R);
+
+    tcg_temp_free_i32(R);
+    tcg_temp_free_i32(Rr);
+
+    return true;
+}
+
+/*
+ *  Subtracts two registers and subtracts with the C Flag and places the
+ *  result in the destination register Rd.
+ */
+static bool trans_SBC(DisasContext *ctx, arg_SBC *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv Rr = cpu_r[a->rr];
+    TCGv R = tcg_temp_new_i32();
+    TCGv zero = tcg_const_i32(0);
+
+    tcg_gen_sub_tl(R, Rd, Rr); /* R = Rd - Rr - Cf */
+    tcg_gen_sub_tl(R, R, cpu_Cf);
+    tcg_gen_andi_tl(R, R, 0xff); /* make it 8 bits */
+    /* update status register */
+    gen_sub_CHf(R, Rd, Rr);
+    gen_sub_Vf(R, Rd, Rr);
+    gen_NSf(R);
+
+    /*
+     * Previous value remains unchanged when the result is zero;
+     * cleared otherwise.
+     */
+    tcg_gen_movcond_tl(TCG_COND_EQ, cpu_Zf, R, zero, cpu_Zf, zero);
+    /* update output registers */
+    tcg_gen_mov_tl(Rd, R);
+
+    tcg_temp_free_i32(zero);
+    tcg_temp_free_i32(R);
+
+    return true;
+}
+
+/*
+ *  SBCI -- Subtract Immediate with Carry
+ */
+static bool trans_SBCI(DisasContext *ctx, arg_SBCI *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv Rr = tcg_const_i32(a->imm);
+    TCGv R = tcg_temp_new_i32();
+    TCGv zero = tcg_const_i32(0);
+
+    tcg_gen_sub_tl(R, Rd, Rr); /* R = Rd - Rr - Cf */
+    tcg_gen_sub_tl(R, R, cpu_Cf);
+    tcg_gen_andi_tl(R, R, 0xff); /* make it 8 bits */
+    /* update status register */
+    gen_sub_CHf(R, Rd, Rr);
+    gen_sub_Vf(R, Rd, Rr);
+    gen_NSf(R);
+
+    /*
+     * Previous value remains unchanged when the result is zero;
+     * cleared otherwise.
+     */
+    tcg_gen_movcond_tl(TCG_COND_EQ, cpu_Zf, R, zero, cpu_Zf, zero);
+    /* update output registers */
+    tcg_gen_mov_tl(Rd, R);
+
+    tcg_temp_free_i32(zero);
+    tcg_temp_free_i32(R);
+    tcg_temp_free_i32(Rr);
+
+    return true;
+}
+
+/*
+ *  Subtracts an immediate value (0-63) from a register pair and places the
+ *  result in the register pair. This instruction operates on the upper four
+ *  register pairs, and is well suited for operations on the Pointer Registers.
+ *  This instruction is not available in all devices. Refer to the device
+ *  specific instruction set summary.
+ */
+static bool trans_SBIW(DisasContext *ctx, arg_SBIW *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_ADIW_SBIW)) {
+        return true;
+    }
+
+    TCGv RdL = cpu_r[a->rd];
+    TCGv RdH = cpu_r[a->rd + 1];
+    int Imm = (a->imm);
+    TCGv R = tcg_temp_new_i32();
+    TCGv Rd = tcg_temp_new_i32();
+
+    tcg_gen_deposit_tl(Rd, RdL, RdH, 8, 8); /* Rd = RdH:RdL */
+    tcg_gen_subi_tl(R, Rd, Imm); /* R = Rd - Imm */
+    tcg_gen_andi_tl(R, R, 0xffff); /* make it 16 bits */
+    /* update status register */
+    tcg_gen_andc_tl(cpu_Cf, R, Rd);
+    tcg_gen_shri_tl(cpu_Cf, cpu_Cf, 15); /* Cf = R & ~Rd */
+    tcg_gen_andc_tl(cpu_Vf, Rd, R);
+    tcg_gen_shri_tl(cpu_Vf, cpu_Vf, 15); /* Vf = Rd & ~R */
+    tcg_gen_setcondi_tl(TCG_COND_EQ, cpu_Zf, R, 0); /* Zf = R == 0 */
+    tcg_gen_shri_tl(cpu_Nf, R, 15); /* Nf = R(15) */
+    tcg_gen_xor_tl(cpu_Sf, cpu_Nf, cpu_Vf); /* Sf = Nf ^ Vf */
+    /* update output registers */
+    tcg_gen_andi_tl(RdL, R, 0xff);
+    tcg_gen_shri_tl(RdH, R, 8);
+
+    tcg_temp_free_i32(Rd);
+    tcg_temp_free_i32(R);
+
+    return true;
+}
+
+/*
+ *  Performs the logical AND between the contents of register Rd and register
+ *  Rr and places the result in the destination register Rd.
+ */
+static bool trans_AND(DisasContext *ctx, arg_AND *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv Rr = cpu_r[a->rr];
+    TCGv R = tcg_temp_new_i32();
+
+    tcg_gen_and_tl(R, Rd, Rr); /* Rd = Rd and Rr */
+    /* update status register */
+    tcg_gen_movi_tl(cpu_Vf, 0); /* Vf = 0 */
+    tcg_gen_setcondi_tl(TCG_COND_EQ, cpu_Zf, R, 0); /* Zf = R == 0 */
+    gen_ZNSf(R);
+    /* update output registers */
+    tcg_gen_mov_tl(Rd, R);
+
+    tcg_temp_free_i32(R);
+
+    return true;
+}
+
+/*
+ *  Performs the logical AND between the contents of register Rd and a constant
+ *  and places the result in the destination register Rd.
+ */
+static bool trans_ANDI(DisasContext *ctx, arg_ANDI *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    int Imm = (a->imm);
+
+    tcg_gen_andi_tl(Rd, Rd, Imm); /* Rd = Rd & Imm */
+    /* update status register */
+    tcg_gen_movi_tl(cpu_Vf, 0x00); /* Vf = 0 */
+    gen_ZNSf(Rd);
+
+    return true;
+}
+
+/*
+ *  Performs the logical OR between the contents of register Rd and register
+ *  Rr and places the result in the destination register Rd.
+ */
+static bool trans_OR(DisasContext *ctx, arg_OR *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv Rr = cpu_r[a->rr];
+    TCGv R = tcg_temp_new_i32();
+
+    tcg_gen_or_tl(R, Rd, Rr);
+    /* update status register */
+    tcg_gen_movi_tl(cpu_Vf, 0);
+    gen_ZNSf(R);
+    /* update output registers */
+    tcg_gen_mov_tl(Rd, R);
+
+    tcg_temp_free_i32(R);
+
+    return true;
+}
+
+/*
+ *  Performs the logical OR between the contents of register Rd and a
+ *  constant and places the result in the destination register Rd.
+ */
+static bool trans_ORI(DisasContext *ctx, arg_ORI *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    int Imm = (a->imm);
+
+    tcg_gen_ori_tl(Rd, Rd, Imm); /* Rd = Rd | Imm */
+    /* update status register */
+    tcg_gen_movi_tl(cpu_Vf, 0x00); /* Vf = 0 */
+    gen_ZNSf(Rd);
+
+    return true;
+}
+
+/*
+ *  Performs the logical EOR between the contents of register Rd and
+ *  register Rr and places the result in the destination register Rd.
+ */
+static bool trans_EOR(DisasContext *ctx, arg_EOR *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv Rr = cpu_r[a->rr];
+
+    tcg_gen_xor_tl(Rd, Rd, Rr);
+    /* update status register */
+    tcg_gen_movi_tl(cpu_Vf, 0);
+    gen_ZNSf(Rd);
+
+    return true;
+}
+
+/*
+ *  Clears the specified bits in register Rd. Performs the logical AND
+ *  between the contents of register Rd and the complement of the constant mask
+ *  K. The result will be placed in register Rd.
+ */
+static bool trans_COM(DisasContext *ctx, arg_COM *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv R = tcg_temp_new_i32();
+
+    tcg_gen_xori_tl(Rd, Rd, 0xff);
+    /* update status register */
+    tcg_gen_movi_tl(cpu_Cf, 1); /* Cf = 1 */
+    tcg_gen_movi_tl(cpu_Vf, 0); /* Vf = 0 */
+    gen_ZNSf(Rd);
+
+    tcg_temp_free_i32(R);
+
+    return true;
+}
+
+/*
+ *  Replaces the contents of register Rd with its two's complement; the
+ *  value $80 is left unchanged.
+ */
+static bool trans_NEG(DisasContext *ctx, arg_NEG *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv t0 = tcg_const_i32(0);
+    TCGv R = tcg_temp_new_i32();
+
+    tcg_gen_sub_tl(R, t0, Rd); /* R = 0 - Rd */
+    tcg_gen_andi_tl(R, R, 0xff); /* make it 8 bits */
+    /* update status register */
+    gen_sub_CHf(R, t0, Rd);
+    gen_sub_Vf(R, t0, Rd);
+    gen_ZNSf(R);
+    /* update output registers */
+    tcg_gen_mov_tl(Rd, R);
+
+    tcg_temp_free_i32(t0);
+    tcg_temp_free_i32(R);
+
+    return true;
+}
+
+/*
+ *  Adds one -1- to the contents of register Rd and places the result in the
+ *  destination register Rd.  The C Flag in SREG is not affected by the
+ *  operation, thus allowing the INC instruction to be used on a loop counter in
+ *  multiple-precision computations.  When operating on unsigned numbers, only
+ *  BREQ and BRNE branches can be expected to perform consistently. When
+ *  operating on two's complement values, all signed branches are available.
+ */
+static bool trans_INC(DisasContext *ctx, arg_INC *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+
+    tcg_gen_addi_tl(Rd, Rd, 1);
+    tcg_gen_andi_tl(Rd, Rd, 0xff);
+    /* update status register */
+    tcg_gen_setcondi_tl(TCG_COND_EQ, cpu_Vf, Rd, 0x80); /* Vf = Rd == 0x80 */
+    gen_ZNSf(Rd);
+
+    return true;
+}
+
+/*
+ *  Subtracts one -1- from the contents of register Rd and places the result
+ *  in the destination register Rd.  The C Flag in SREG is not affected by the
+ *  operation, thus allowing the DEC instruction to be used on a loop counter in
+ *  multiple-precision computations.  When operating on unsigned values, only
+ *  BREQ and BRNE branches can be expected to perform consistently.  When
+ *  operating on two's complement values, all signed branches are available.
+ */
+static bool trans_DEC(DisasContext *ctx, arg_DEC *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+
+    tcg_gen_subi_tl(Rd, Rd, 1); /* Rd = Rd - 1 */
+    tcg_gen_andi_tl(Rd, Rd, 0xff); /* make it 8 bits */
+    /* update status register */
+    tcg_gen_setcondi_tl(TCG_COND_EQ, cpu_Vf, Rd, 0x7f); /* Vf = Rd == 0x7f */
+    gen_ZNSf(Rd);
+
+    return true;
+}
+
+/*
+ *  This instruction performs 8-bit x 8-bit -> 16-bit unsigned multiplication.
+ */
+static bool trans_MUL(DisasContext *ctx, arg_MUL *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_MUL)) {
+        return true;
+    }
+
+    TCGv R0 = cpu_r[0];
+    TCGv R1 = cpu_r[1];
+    TCGv Rd = cpu_r[a->rd];
+    TCGv Rr = cpu_r[a->rr];
+    TCGv R = tcg_temp_new_i32();
+
+    tcg_gen_mul_tl(R, Rd, Rr); /* R = Rd * Rr */
+    tcg_gen_andi_tl(R0, R, 0xff);
+    tcg_gen_shri_tl(R1, R, 8);
+    /* update status register */
+    tcg_gen_shri_tl(cpu_Cf, R, 15); /* Cf = R(15) */
+    tcg_gen_setcondi_tl(TCG_COND_EQ, cpu_Zf, R, 0); /* Zf = R == 0 */
+
+    tcg_temp_free_i32(R);
+
+    return true;
+}
+
+/*
+ *  This instruction performs 8-bit x 8-bit -> 16-bit signed multiplication.
+ */
+static bool trans_MULS(DisasContext *ctx, arg_MULS *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_MUL)) {
+        return true;
+    }
+
+    TCGv R0 = cpu_r[0];
+    TCGv R1 = cpu_r[1];
+    TCGv Rd = cpu_r[a->rd];
+    TCGv Rr = cpu_r[a->rr];
+    TCGv R = tcg_temp_new_i32();
+    TCGv t0 = tcg_temp_new_i32();
+    TCGv t1 = tcg_temp_new_i32();
+
+    tcg_gen_ext8s_tl(t0, Rd); /* make Rd full 32 bit signed */
+    tcg_gen_ext8s_tl(t1, Rr); /* make Rr full 32 bit signed */
+    tcg_gen_mul_tl(R, t0, t1); /* R = Rd * Rr */
+    tcg_gen_andi_tl(R, R, 0xffff); /* make it 16 bits */
+    tcg_gen_andi_tl(R0, R, 0xff);
+    tcg_gen_shri_tl(R1, R, 8);
+    /* update status register */
+    tcg_gen_shri_tl(cpu_Cf, R, 15); /* Cf = R(15) */
+    tcg_gen_setcondi_tl(TCG_COND_EQ, cpu_Zf, R, 0); /* Zf = R == 0 */
+
+    tcg_temp_free_i32(t1);
+    tcg_temp_free_i32(t0);
+    tcg_temp_free_i32(R);
+
+    return true;
+}
+
+/*
+ *  This instruction performs 8-bit x 8-bit -> 16-bit multiplication of a
+ *  signed and an unsigned number.
+ */
+static bool trans_MULSU(DisasContext *ctx, arg_MULSU *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_MUL)) {
+        return true;
+    }
+
+    TCGv R0 = cpu_r[0];
+    TCGv R1 = cpu_r[1];
+    TCGv Rd = cpu_r[a->rd];
+    TCGv Rr = cpu_r[a->rr];
+    TCGv R = tcg_temp_new_i32();
+    TCGv t0 = tcg_temp_new_i32();
+
+    tcg_gen_ext8s_tl(t0, Rd); /* make Rd full 32 bit signed */
+    tcg_gen_mul_tl(R, t0, Rr); /* R = Rd * Rr */
+    tcg_gen_andi_tl(R, R, 0xffff); /* make R 16 bits */
+    tcg_gen_andi_tl(R0, R, 0xff);
+    tcg_gen_shri_tl(R1, R, 8);
+    /* update status register */
+    tcg_gen_shri_tl(cpu_Cf, R, 15); /* Cf = R(15) */
+    tcg_gen_setcondi_tl(TCG_COND_EQ, cpu_Zf, R, 0); /* Zf = R == 0 */
+
+    tcg_temp_free_i32(t0);
+    tcg_temp_free_i32(R);
+
+    return true;
+}
+
+/*
+ *  This instruction performs 8-bit x 8-bit -> 16-bit unsigned
+ *  multiplication and shifts the result one bit left.
+ */
+static bool trans_FMUL(DisasContext *ctx, arg_FMUL *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_MUL)) {
+        return true;
+    }
+
+    TCGv R0 = cpu_r[0];
+    TCGv R1 = cpu_r[1];
+    TCGv Rd = cpu_r[a->rd];
+    TCGv Rr = cpu_r[a->rr];
+    TCGv R = tcg_temp_new_i32();
+
+    tcg_gen_mul_tl(R, Rd, Rr); /* R = Rd * Rr */
+    /* update status register */
+    tcg_gen_shri_tl(cpu_Cf, R, 15); /* Cf = R(15) */
+    tcg_gen_setcondi_tl(TCG_COND_EQ, cpu_Zf, R, 0); /* Zf = R == 0 */
+    /* update output registers */
+    tcg_gen_shli_tl(R, R, 1);
+    tcg_gen_andi_tl(R0, R, 0xff);
+    tcg_gen_shri_tl(R1, R, 8);
+    tcg_gen_andi_tl(R1, R1, 0xff);
+
+
+    tcg_temp_free_i32(R);
+
+    return true;
+}
+
+/*
+ *  This instruction performs 8-bit x 8-bit -> 16-bit signed multiplication
+ *  and shifts the result one bit left.
+ */
+static bool trans_FMULS(DisasContext *ctx, arg_FMULS *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_MUL)) {
+        return true;
+    }
+
+    TCGv R0 = cpu_r[0];
+    TCGv R1 = cpu_r[1];
+    TCGv Rd = cpu_r[a->rd];
+    TCGv Rr = cpu_r[a->rr];
+    TCGv R = tcg_temp_new_i32();
+    TCGv t0 = tcg_temp_new_i32();
+    TCGv t1 = tcg_temp_new_i32();
+
+    tcg_gen_ext8s_tl(t0, Rd); /* make Rd full 32 bit signed */
+    tcg_gen_ext8s_tl(t1, Rr); /* make Rr full 32 bit signed */
+    tcg_gen_mul_tl(R, t0, t1); /* R = Rd * Rr */
+    tcg_gen_andi_tl(R, R, 0xffff); /* make it 16 bits */
+    /* update status register */
+    tcg_gen_shri_tl(cpu_Cf, R, 15); /* Cf = R(15) */
+    tcg_gen_setcondi_tl(TCG_COND_EQ, cpu_Zf, R, 0); /* Zf = R == 0 */
+    /* update output registers */
+    tcg_gen_shli_tl(R, R, 1);
+    tcg_gen_andi_tl(R0, R, 0xff);
+    tcg_gen_shri_tl(R1, R, 8);
+    tcg_gen_andi_tl(R1, R1, 0xff);
+
+    tcg_temp_free_i32(t1);
+    tcg_temp_free_i32(t0);
+    tcg_temp_free_i32(R);
+
+    return true;
+}
+
+/*
+ *  This instruction performs 8-bit x 8-bit -> 16-bit signed multiplication
+ *  and shifts the result one bit left.
+ */
+static bool trans_FMULSU(DisasContext *ctx, arg_FMULSU *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_MUL)) {
+        return true;
+    }
+
+    TCGv R0 = cpu_r[0];
+    TCGv R1 = cpu_r[1];
+    TCGv Rd = cpu_r[a->rd];
+    TCGv Rr = cpu_r[a->rr];
+    TCGv R = tcg_temp_new_i32();
+    TCGv t0 = tcg_temp_new_i32();
+
+    tcg_gen_ext8s_tl(t0, Rd); /* make Rd full 32 bit signed */
+    tcg_gen_mul_tl(R, t0, Rr); /* R = Rd * Rr */
+    tcg_gen_andi_tl(R, R, 0xffff); /* make it 16 bits */
+    /* update status register */
+    tcg_gen_shri_tl(cpu_Cf, R, 15); /* Cf = R(15) */
+    tcg_gen_setcondi_tl(TCG_COND_EQ, cpu_Zf, R, 0); /* Zf = R == 0 */
+    /* update output registers */
+    tcg_gen_shli_tl(R, R, 1);
+    tcg_gen_andi_tl(R0, R, 0xff);
+    tcg_gen_shri_tl(R1, R, 8);
+    tcg_gen_andi_tl(R1, R1, 0xff);
+
+    tcg_temp_free_i32(t0);
+    tcg_temp_free_i32(R);
+
+    return true;
+}
+
+/*
+ *  The module is an instruction set extension to the AVR CPU, performing
+ *  DES iterations. The 64-bit data block (plaintext or ciphertext) is placed in
+ *  the CPU register file, registers R0-R7, where LSB of data is placed in LSB
+ *  of R0 and MSB of data is placed in MSB of R7. The full 64-bit key (including
+ *  parity bits) is placed in registers R8- R15, organized in the register file
+ *  with LSB of key in LSB of R8 and MSB of key in MSB of R15. Executing one DES
+ *  instruction performs one round in the DES algorithm. Sixteen rounds must be
+ *  executed in increasing order to form the correct DES ciphertext or
+ *  plaintext. Intermediate results are stored in the register file (R0-R15)
+ *  after each DES instruction. The instruction's operand (K) determines which
+ *  round is executed, and the half carry flag (H) determines whether encryption
+ *  or decryption is performed.  The DES algorithm is described in
+ *  "Specifications for the Data Encryption Standard" (Federal Information
+ *  Processing Standards Publication 46). Intermediate results in this
+ *  implementation differ from the standard because the initial permutation and
+ *  the inverse initial permutation are performed each iteration. This does not
+ *  affect the result in the final ciphertext or plaintext, but reduces
+ *  execution time.
+ */
+static bool trans_DES(DisasContext *ctx, arg_DES *a)
+{
+    /* TODO */
+    if (!avr_have_feature(ctx, AVR_FEATURE_DES)) {
+        return true;
+    }
+
+    return true;
+}
target/avr/insn.decode new file +93 -0

93 lines not on this page · raw message

Signature · 1 line
2.17.2 (Apple Git-113)
#12 lore raw

PATCH v40 10/21 target/avr: Add instruction disassembly function

In reply to Michael Rolnik

Provide function disassembles executed instruction when -d in_asm is provided

Example:
./avr-softmmu/qemu-system-avr -bios free-rtos/Demo/AVR_ATMega2560_GCC/demo.elf -d in_asm will produce something like the following

```
    ...
    IN:
    0x0000014a:  CALL      0x3808

    IN: main
    0x00003808:  CALL      0x4b4

    IN: vParTestInitialise
    0x000004b4:  LDI       r24, 255
    0x000004b6:  STS       r24, 0
    0x000004b8:  MULS      r16, r20
    0x000004ba:  OUT       $1, r24
    0x000004bc:  LDS       r24, 0
    0x000004be:  MULS      r16, r20
    0x000004c0:  OUT       $2, r24
    0x000004c2:  RET
    ...
```
---
 target/avr/cpu.h       |   1 +
 target/avr/cpu.c       |   2 +-
 target/avr/disas.c     | 245 +++++++++++++++++++++++++++++++++++++++++
 target/avr/translate.c |  12 ++
 4 files changed, 259 insertions(+), 1 deletion(-)
 create mode 100644 target/avr/disas.c
target/avr/cpu.h +1 -0
index b74bcf01ae..af89b6611e 100644
@@ -160,6 +160,7 @@ bool avr_cpu_exec_interrupt(CPUState *cpu, int int_req);
 hwaddr avr_cpu_get_phys_page_debug(CPUState *cpu, vaddr addr);
 int avr_cpu_gdb_read_register(CPUState *cpu, uint8_t *buf, int reg);
 int avr_cpu_gdb_write_register(CPUState *cpu, uint8_t *buf, int reg);
+int avr_print_insn(bfd_vma addr, disassemble_info *info);
 
 static inline int avr_feature(CPUAVRState *env, AVRFeature feature)
 {
target/avr/cpu.c +1 -1
index c74c5106fe..fa51f771c0 100644
@@ -84,7 +84,7 @@ static void avr_cpu_reset(CPUState *cs)
 static void avr_cpu_disas_set_info(CPUState *cpu, disassemble_info *info)
 {
     info->mach = bfd_arch_avr;
-    info->print_insn = NULL;
+    info->print_insn = avr_print_insn;
 }
 
 static void avr_cpu_realizefn(DeviceState *dev, Error **errp)
target/avr/disas.c new file +245 -0

245 lines not on this page · raw message

target/avr/translate.c +12 -0
index 5cbdf80a6e..7124b04a97 100644
@@ -2976,6 +2976,18 @@ done_generating:
 
     tb->size = (ctx.npc - pc_start) * 2;
     tb->icount = num_insns;
+
+#ifdef DEBUG_DISAS
+    if (qemu_loglevel_mask(CPU_LOG_TB_IN_ASM)
+        && qemu_log_in_addr_range(tb->pc)) {
+        FILE *fd;
+        fd = qemu_log_lock();
+        qemu_log("IN: %s\n", lookup_symbol(tb->pc));
+        log_target_disas(cs, tb->pc, tb->size);
+        qemu_log("\n");
+        qemu_log_unlock(fd);
+    }
+#endif
 }
 
 void restore_state_to_opc(CPUAVRState *env, TranslationBlock *tb,
Signature · 1 line
2.17.2 (Apple Git-113)
#11 lore raw

PATCH v40 9/21 target/avr: Add instruction translation - CPU main translation function

In reply to Michael Rolnik
---
 target/avr/translate.c | 234 +++++++++++++++++++++++++++++++++++++++++
 1 file changed, 234 insertions(+)
target/avr/translate.c +234 -0
index 246f9e7e47..5cbdf80a6e 100644
@@ -2749,3 +2749,237 @@ static bool trans_WDR(DisasContext *ctx, arg_WDR *a)
 
     return true;
 }
+
+
+void avr_cpu_tcg_init(void)
+{
+    int i;
+
+#define AVR_REG_OFFS(x) offsetof(CPUAVRState, x)
+    cpu_pc = tcg_global_mem_new_i32(cpu_env, AVR_REG_OFFS(pc_w), "pc");
+    cpu_Cf = tcg_global_mem_new_i32(cpu_env, AVR_REG_OFFS(sregC), "Cf");
+    cpu_Zf = tcg_global_mem_new_i32(cpu_env, AVR_REG_OFFS(sregZ), "Zf");
+    cpu_Nf = tcg_global_mem_new_i32(cpu_env, AVR_REG_OFFS(sregN), "Nf");
+    cpu_Vf = tcg_global_mem_new_i32(cpu_env, AVR_REG_OFFS(sregV), "Vf");
+    cpu_Sf = tcg_global_mem_new_i32(cpu_env, AVR_REG_OFFS(sregS), "Sf");
+    cpu_Hf = tcg_global_mem_new_i32(cpu_env, AVR_REG_OFFS(sregH), "Hf");
+    cpu_Tf = tcg_global_mem_new_i32(cpu_env, AVR_REG_OFFS(sregT), "Tf");
+    cpu_If = tcg_global_mem_new_i32(cpu_env, AVR_REG_OFFS(sregI), "If");
+    cpu_rampD = tcg_global_mem_new_i32(cpu_env, AVR_REG_OFFS(rampD), "rampD");
+    cpu_rampX = tcg_global_mem_new_i32(cpu_env, AVR_REG_OFFS(rampX), "rampX");
+    cpu_rampY = tcg_global_mem_new_i32(cpu_env, AVR_REG_OFFS(rampY), "rampY");
+    cpu_rampZ = tcg_global_mem_new_i32(cpu_env, AVR_REG_OFFS(rampZ), "rampZ");
+    cpu_eind = tcg_global_mem_new_i32(cpu_env, AVR_REG_OFFS(eind), "eind");
+    cpu_sp = tcg_global_mem_new_i32(cpu_env, AVR_REG_OFFS(sp), "sp");
+    cpu_skip = tcg_global_mem_new_i32(cpu_env, AVR_REG_OFFS(skip), "skip");
+
+    for (i = 0; i < NUMBER_OF_CPU_REGISTERS; i++) {
+        cpu_r[i] = tcg_global_mem_new_i32(cpu_env, AVR_REG_OFFS(r[i]),
+                                          reg_names[i]);
+    }
+#undef AVR_REG_OFFS
+}
+
+static void translate(DisasContext *ctx)
+{
+    uint32_t opcode = next_word(ctx);
+
+    if (!decode_insn(ctx, opcode)) {
+        gen_helper_unsupported(cpu_env);
+        ctx->bstate = DISAS_NORETURN;
+    }
+}
+
+/* Standardize the cpu_skip condition to NE.  */
+static bool canonicalize_skip(DisasContext *ctx)
+{
+    switch (ctx->skip_cond) {
+    case TCG_COND_NEVER:
+        /* Normal case: cpu_skip is known to be false.  */
+        return false;
+
+    case TCG_COND_ALWAYS:
+        /*
+         * Breakpoint case: cpu_skip is known to be true, via TB_FLAGS_SKIP.
+         * The breakpoint is on the instruction being skipped, at the start
+         * of the TranslationBlock.  No need to update.
+         */
+        return false;
+
+    case TCG_COND_NE:
+        if (ctx->skip_var1 == NULL) {
+            tcg_gen_mov_tl(cpu_skip, ctx->skip_var0);
+        } else {
+            tcg_gen_xor_tl(cpu_skip, ctx->skip_var0, ctx->skip_var1);
+            ctx->skip_var1 = NULL;
+        }
+        break;
+
+    default:
+        /* Convert to a NE condition vs 0. */
+        if (ctx->skip_var1 == NULL) {
+            tcg_gen_setcondi_tl(ctx->skip_cond, cpu_skip, ctx->skip_var0, 0);
+        } else {
+            tcg_gen_setcond_tl(ctx->skip_cond, cpu_skip,
+                               ctx->skip_var0, ctx->skip_var1);
+            ctx->skip_var1 = NULL;
+        }
+        ctx->skip_cond = TCG_COND_NE;
+        break;
+    }
+    if (ctx->free_skip_var0) {
+        tcg_temp_free(ctx->skip_var0);
+        ctx->free_skip_var0 = false;
+    }
+    ctx->skip_var0 = cpu_skip;
+    return true;
+}
+
+void gen_intermediate_code(CPUState *cs, TranslationBlock *tb, int max_insns)
+{
+    CPUAVRState *env = cs->env_ptr;
+    DisasContext ctx = {
+        .tb = tb,
+        .cs = cs,
+        .env = env,
+        .memidx = 0,
+        .bstate = DISAS_NEXT,
+        .skip_cond = TCG_COND_NEVER,
+        .singlestep = cs->singlestep_enabled,
+    };
+    target_ulong pc_start = tb->pc / 2;
+    int num_insns = 0;
+
+    if (tb->flags & TB_FLAGS_FULL_ACCESS) {
+        /*
+         * This flag is set by ST/LD instruction we will regenerate it ONLY
+         * with mem/cpu memory access instead of mem access
+         */
+        max_insns = 1;
+    }
+    if (ctx.singlestep) {
+        max_insns = 1;
+    }
+
+    gen_tb_start(tb);
+
+    ctx.npc = pc_start;
+    if (tb->flags & TB_FLAGS_SKIP) {
+        ctx.skip_cond = TCG_COND_ALWAYS;
+        ctx.skip_var0 = cpu_skip;
+    }
+
+    do {
+        TCGLabel *skip_label = NULL;
+
+        /* translate current instruction */
+        tcg_gen_insn_start(ctx.npc);
+        num_insns++;
+
+        /*
+         * this is due to some strange GDB behavior
+         * let's assume main has address 0x100
+         * b main   - sets breakpoint at address 0x00000100 (code)
+         * b *0x100 - sets breakpoint at address 0x00800100 (data)
+         */
+        if (unlikely(!ctx.singlestep &&
+                (cpu_breakpoint_test(cs, OFFSET_CODE + ctx.npc * 2, BP_ANY) ||
+                 cpu_breakpoint_test(cs, OFFSET_DATA + ctx.npc * 2, BP_ANY)))) {
+            canonicalize_skip(&ctx);
+            tcg_gen_movi_tl(cpu_pc, ctx.npc);
+            gen_helper_debug(cpu_env);
+            goto done_generating;
+        }
+
+        /* Conditionally skip the next instruction, if indicated.  */
+        if (ctx.skip_cond != TCG_COND_NEVER) {
+            skip_label = gen_new_label();
+            if (ctx.skip_var0 == cpu_skip) {
+                /*
+                 * Copy cpu_skip so that we may zero it before the branch.
+                 * This ensures that cpu_skip is non-zero after the label
+                 * if and only if the skipped insn itself sets a skip.
+                 */
+                ctx.free_skip_var0 = true;
+                ctx.skip_var0 = tcg_temp_new();
+                tcg_gen_mov_tl(ctx.skip_var0, cpu_skip);
+                tcg_gen_movi_tl(cpu_skip, 0);
+            }
+            if (ctx.skip_var1 == NULL) {
+                tcg_gen_brcondi_tl(ctx.skip_cond, ctx.skip_var0, 0, skip_label);
+            } else {
+                tcg_gen_brcond_tl(ctx.skip_cond, ctx.skip_var0,
+                                  ctx.skip_var1, skip_label);
+                ctx.skip_var1 = NULL;
+            }
+            if (ctx.free_skip_var0) {
+                tcg_temp_free(ctx.skip_var0);
+                ctx.free_skip_var0 = false;
+            }
+            ctx.skip_cond = TCG_COND_NEVER;
+            ctx.skip_var0 = NULL;
+        }
+
+        translate(&ctx);
+
+        if (skip_label) {
+            canonicalize_skip(&ctx);
+            gen_set_label(skip_label);
+            if (ctx.bstate == DISAS_NORETURN) {
+                ctx.bstate = DISAS_CHAIN;
+            }
+        }
+    } while (ctx.bstate == DISAS_NEXT
+             && num_insns < max_insns
+             && (ctx.npc - pc_start) * 2 < TARGET_PAGE_SIZE - 4
+             && !tcg_op_buf_full());
+
+    if (tb->cflags & CF_LAST_IO) {
+        gen_io_end();
+    }
+
+    bool nonconst_skip = canonicalize_skip(&ctx);
+
+    switch (ctx.bstate) {
+    case DISAS_NORETURN:
+        assert(!nonconst_skip);
+        break;
+    case DISAS_NEXT:
+    case DISAS_TOO_MANY:
+    case DISAS_CHAIN:
+        if (!nonconst_skip) {
+            /* Note gen_goto_tb checks singlestep.  */
+            gen_goto_tb(&ctx, 1, ctx.npc);
+            break;
+        }
+        tcg_gen_movi_tl(cpu_pc, ctx.npc);
+        /* fall through */
+    case DISAS_LOOKUP:
+        if (!ctx.singlestep) {
+            tcg_gen_lookup_and_goto_ptr();
+            break;
+        }
+        /* fall through */
+    case DISAS_EXIT:
+        if (ctx.singlestep) {
+            gen_helper_debug(cpu_env);
+        } else {
+            tcg_gen_exit_tb(NULL, 0);
+        }
+        break;
+    default:
+        g_assert_not_reached();
+    }
+
+done_generating:
+    gen_tb_end(tb, num_insns);
+
+    tb->size = (ctx.npc - pc_start) * 2;
+    tb->icount = num_insns;
+}
+
+void restore_state_to_opc(CPUAVRState *env, TranslationBlock *tb,
+                            target_ulong *data)
+{
+    env->pc_w = data[0];
+}
Signature · 1 line
2.17.2 (Apple Git-113)
#5 lore raw

PATCH v40 5/21 target/avr: Add instruction translation - Branch Instructions

In reply to Michael Rolnik

This includes:
- RJMP, IJMP, EIJMP, JMP
- RCALL, ICALL, EICALL, CALL
- RET, RETI
- CPSE, CP, CPC, CPI
- SBRC, SBRS, SBIC, SBIS
- BRBC, BRBS

---
 target/avr/translate.c | 533 +++++++++++++++++++++++++++++++++++++++++
 target/avr/insn.decode |  24 ++
 2 files changed, 557 insertions(+)
target/avr/translate.c +533 -0
index 2b6411bbed..a80d59464a 100644
@@ -921,3 +921,536 @@ static bool trans_DES(DisasContext *ctx, arg_DES *a)
 
     return true;
 }
+
+/*
+ * Branch Instructions
+ */
+static void gen_jmp_ez(DisasContext *ctx)
+{
+    tcg_gen_deposit_tl(cpu_pc, cpu_r[30], cpu_r[31], 8, 8);
+    tcg_gen_or_tl(cpu_pc, cpu_pc, cpu_eind);
+    ctx->bstate = DISAS_LOOKUP;
+}
+
+static void gen_jmp_z(DisasContext *ctx)
+{
+    tcg_gen_deposit_tl(cpu_pc, cpu_r[30], cpu_r[31], 8, 8);
+    ctx->bstate = DISAS_LOOKUP;
+}
+
+static void gen_push_ret(DisasContext *ctx, int ret)
+{
+    if (avr_feature(ctx->env, AVR_FEATURE_1_BYTE_PC)) {
+
+        TCGv t0 = tcg_const_i32((ret & 0x0000ff));
+
+        tcg_gen_qemu_st_tl(t0, cpu_sp, MMU_DATA_IDX, MO_UB);
+        tcg_gen_subi_tl(cpu_sp, cpu_sp, 1);
+
+        tcg_temp_free_i32(t0);
+    } else if (avr_feature(ctx->env, AVR_FEATURE_2_BYTE_PC)) {
+
+        TCGv t0 = tcg_const_i32((ret & 0x00ffff));
+
+        tcg_gen_subi_tl(cpu_sp, cpu_sp, 1);
+        tcg_gen_qemu_st_tl(t0, cpu_sp, MMU_DATA_IDX, MO_BEUW);
+        tcg_gen_subi_tl(cpu_sp, cpu_sp, 1);
+
+        tcg_temp_free_i32(t0);
+
+    } else if (avr_feature(ctx->env, AVR_FEATURE_3_BYTE_PC)) {
+
+        TCGv lo = tcg_const_i32((ret & 0x0000ff));
+        TCGv hi = tcg_const_i32((ret & 0xffff00) >> 8);
+
+        tcg_gen_qemu_st_tl(lo, cpu_sp, MMU_DATA_IDX, MO_UB);
+        tcg_gen_subi_tl(cpu_sp, cpu_sp, 2);
+        tcg_gen_qemu_st_tl(hi, cpu_sp, MMU_DATA_IDX, MO_BEUW);
+        tcg_gen_subi_tl(cpu_sp, cpu_sp, 1);
+
+        tcg_temp_free_i32(lo);
+        tcg_temp_free_i32(hi);
+    }
+}
+
+static void gen_pop_ret(DisasContext *ctx, TCGv ret)
+{
+    if (avr_feature(ctx->env, AVR_FEATURE_1_BYTE_PC)) {
+        tcg_gen_addi_tl(cpu_sp, cpu_sp, 1);
+        tcg_gen_qemu_ld_tl(ret, cpu_sp, MMU_DATA_IDX, MO_UB);
+    } else if (avr_feature(ctx->env, AVR_FEATURE_2_BYTE_PC)) {
+        tcg_gen_addi_tl(cpu_sp, cpu_sp, 1);
+        tcg_gen_qemu_ld_tl(ret, cpu_sp, MMU_DATA_IDX, MO_BEUW);
+        tcg_gen_addi_tl(cpu_sp, cpu_sp, 1);
+    } else if (avr_feature(ctx->env, AVR_FEATURE_3_BYTE_PC)) {
+        TCGv lo = tcg_temp_new_i32();
+        TCGv hi = tcg_temp_new_i32();
+
+        tcg_gen_addi_tl(cpu_sp, cpu_sp, 1);
+        tcg_gen_qemu_ld_tl(hi, cpu_sp, MMU_DATA_IDX, MO_BEUW);
+
+        tcg_gen_addi_tl(cpu_sp, cpu_sp, 2);
+        tcg_gen_qemu_ld_tl(lo, cpu_sp, MMU_DATA_IDX, MO_UB);
+
+        tcg_gen_deposit_tl(ret, lo, hi, 8, 16);
+
+        tcg_temp_free_i32(lo);
+        tcg_temp_free_i32(hi);
+    }
+}
+
+static void gen_goto_tb(DisasContext *ctx, int n, target_ulong dest)
+{
+    TranslationBlock *tb = ctx->tb;
+
+    if (ctx->singlestep == 0) {
+        tcg_gen_goto_tb(n);
+        tcg_gen_movi_i32(cpu_pc, dest);
+        tcg_gen_exit_tb(tb, n);
+    } else {
+        tcg_gen_movi_i32(cpu_pc, dest);
+        gen_helper_debug(cpu_env);
+        tcg_gen_exit_tb(NULL, 0);
+    }
+    ctx->bstate = DISAS_NORETURN;
+}
+
+/*
+ *  Relative jump to an address within PC - 2K +1 and PC + 2K (words). For
+ *  AVR microcontrollers with Program memory not exceeding 4K words (8KB) this
+ *  instruction can address the entire memory from every address location. See
+ *  also JMP.
+ */
+static bool trans_RJMP(DisasContext *ctx, arg_RJMP *a)
+{
+    int dst = ctx->npc + a->imm;
+
+    gen_goto_tb(ctx, 0, dst);
+
+    return true;
+}
+
+/*
+ *  Indirect jump to the address pointed to by the Z (16 bits) Pointer
+ *  Register in the Register File. The Z-pointer Register is 16 bits wide and
+ *  allows jump within the lowest 64K words (128KB) section of Program memory.
+ *  This instruction is not available in all devices. Refer to the device
+ *  specific instruction set summary.
+ */
+static bool trans_IJMP(DisasContext *ctx, arg_IJMP *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_IJMP_ICALL)) {
+        return true;
+    }
+
+    gen_jmp_z(ctx);
+
+    return true;
+}
+
+/*
+ *  Indirect jump to the address pointed to by the Z (16 bits) Pointer
+ *  Register in the Register File and the EIND Register in the I/O space. This
+ *  instruction allows for indirect jumps to the entire 4M (words) Program
+ *  memory space. See also IJMP.  This instruction is not available in all
+ *  devices. Refer to the device specific instruction set summary.
+ */
+static bool trans_EIJMP(DisasContext *ctx, arg_EIJMP *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_EIJMP_EICALL)) {
+        return true;
+    }
+
+    gen_jmp_ez(ctx);
+    return true;
+}
+
+/*
+ *  Jump to an address within the entire 4M (words) Program memory. See also
+ *  RJMP.  This instruction is not available in all devices. Refer to the device
+ *  specific instruction set summary.0
+ */
+static bool trans_JMP(DisasContext *ctx, arg_JMP *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_JMP_CALL)) {
+        return true;
+    }
+
+    gen_goto_tb(ctx, 0, a->imm);
+
+    return true;
+}
+
+/*
+ *  Relative call to an address within PC - 2K + 1 and PC + 2K (words). The
+ *  return address (the instruction after the RCALL) is stored onto the Stack.
+ *  See also CALL. For AVR microcontrollers with Program memory not exceeding 4K
+ *  words (8KB) this instruction can address the entire memory from every
+ *  address location. The Stack Pointer uses a post-decrement scheme during
+ *  RCALL.
+ */
+static bool trans_RCALL(DisasContext *ctx, arg_RCALL *a)
+{
+    int ret = ctx->npc;
+    int dst = ctx->npc + a->imm;
+
+    gen_push_ret(ctx, ret);
+    gen_goto_tb(ctx, 0, dst);
+
+    return true;
+}
+
+/*
+ *  Calls to a subroutine within the entire 4M (words) Program memory. The
+ *  return address (to the instruction after the CALL) will be stored onto the
+ *  Stack. See also RCALL. The Stack Pointer uses a post-decrement scheme during
+ *  CALL.  This instruction is not available in all devices. Refer to the device
+ *  specific instruction set summary.
+ */
+static bool trans_ICALL(DisasContext *ctx, arg_ICALL *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_IJMP_ICALL)) {
+        return true;
+    }
+
+    int ret = ctx->npc;
+
+    gen_push_ret(ctx, ret);
+    gen_jmp_z(ctx);
+
+    return true;
+}
+
+/*
+ *  Indirect call of a subroutine pointed to by the Z (16 bits) Pointer
+ *  Register in the Register File and the EIND Register in the I/O space. This
+ *  instruction allows for indirect calls to the entire 4M (words) Program
+ *  memory space. See also ICALL. The Stack Pointer uses a post-decrement scheme
+ *  during EICALL.  This instruction is not available in all devices. Refer to
+ *  the device specific instruction set summary.
+ */
+static bool trans_EICALL(DisasContext *ctx, arg_EICALL *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_EIJMP_EICALL)) {
+        return true;
+    }
+
+    int ret = ctx->npc;
+
+    gen_push_ret(ctx, ret);
+    gen_jmp_ez(ctx);
+    return true;
+}
+
+/*
+ *  Calls to a subroutine within the entire Program memory. The return
+ *  address (to the instruction after the CALL) will be stored onto the Stack.
+ *  (See also RCALL). The Stack Pointer uses a post-decrement scheme during
+ *  CALL.  This instruction is not available in all devices. Refer to the device
+ *  specific instruction set summary.
+ */
+static bool trans_CALL(DisasContext *ctx, arg_CALL *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_JMP_CALL)) {
+        return true;
+    }
+
+    int Imm = a->imm;
+    int ret = ctx->npc;
+
+    gen_push_ret(ctx, ret);
+    gen_goto_tb(ctx, 0, Imm);
+
+    return true;
+}
+
+/*
+ *  Returns from subroutine. The return address is loaded from the STACK.
+ *  The Stack Pointer uses a preincrement scheme during RET.
+ */
+static bool trans_RET(DisasContext *ctx, arg_RET *a)
+{
+    gen_pop_ret(ctx, cpu_pc);
+
+    ctx->bstate = DISAS_LOOKUP;
+    return true;
+}
+
+/*
+ *  Returns from interrupt. The return address is loaded from the STACK and
+ *  the Global Interrupt Flag is set.  Note that the Status Register is not
+ *  automatically stored when entering an interrupt routine, and it is not
+ *  restored when returning from an interrupt routine. This must be handled by
+ *  the application program. The Stack Pointer uses a pre-increment scheme
+ *  during RETI.
+ */
+static bool trans_RETI(DisasContext *ctx, arg_RETI *a)
+{
+    gen_pop_ret(ctx, cpu_pc);
+    tcg_gen_movi_tl(cpu_If, 1);
+
+    /* Need to return to main loop to re-evaluate interrupts.  */
+    ctx->bstate = DISAS_EXIT;
+    return true;
+}
+
+/*
+ *  This instruction performs a compare between two registers Rd and Rr, and
+ *  skips the next instruction if Rd = Rr.
+ */
+static bool trans_CPSE(DisasContext *ctx, arg_CPSE *a)
+{
+    ctx->skip_cond = TCG_COND_EQ;
+    ctx->skip_var0 = cpu_r[a->rd];
+    ctx->skip_var1 = cpu_r[a->rr];
+    return true;
+}
+
+/*
+ *  This instruction performs a compare between two registers Rd and Rr.
+ *  None of the registers are changed. All conditional branches can be used
+ *  after this instruction.
+ */
+static bool trans_CP(DisasContext *ctx, arg_CP *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv Rr = cpu_r[a->rr];
+    TCGv R = tcg_temp_new_i32();
+
+    tcg_gen_sub_tl(R, Rd, Rr); /* R = Rd - Rr */
+    tcg_gen_andi_tl(R, R, 0xff); /* make it 8 bits */
+
+    /* update status register */
+    gen_sub_CHf(R, Rd, Rr);
+    gen_sub_Vf(R, Rd, Rr);
+    gen_ZNSf(R);
+
+    tcg_temp_free_i32(R);
+
+    return true;
+}
+
+/*
+ *  This instruction performs a compare between two registers Rd and Rr and
+ *  also takes into account the previous carry. None of the registers are
+ *  changed. All conditional branches can be used after this instruction.
+ */
+static bool trans_CPC(DisasContext *ctx, arg_CPC *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    TCGv Rr = cpu_r[a->rr];
+    TCGv R = tcg_temp_new_i32();
+    TCGv zero = tcg_const_i32(0);
+
+    tcg_gen_sub_tl(R, Rd, Rr); /* R = Rd - Rr - Cf */
+    tcg_gen_sub_tl(R, R, cpu_Cf);
+    tcg_gen_andi_tl(R, R, 0xff); /* make it 8 bits */
+    /* update status register */
+    gen_sub_CHf(R, Rd, Rr);
+    gen_sub_Vf(R, Rd, Rr);
+    gen_NSf(R);
+
+    /*
+     * Previous value remains unchanged when the result is zero;
+     * cleared otherwise.
+     */
+    tcg_gen_movcond_tl(TCG_COND_EQ, cpu_Zf, R, zero, cpu_Zf, zero);
+
+    tcg_temp_free_i32(zero);
+    tcg_temp_free_i32(R);
+
+    return true;
+}
+
+/*
+ *  This instruction performs a compare between register Rd and a constant.
+ *  The register is not changed. All conditional branches can be used after this
+ *  instruction.
+ */
+static bool trans_CPI(DisasContext *ctx, arg_CPI *a)
+{
+    TCGv Rd = cpu_r[a->rd];
+    int Imm = a->imm;
+    TCGv Rr = tcg_const_i32(Imm);
+    TCGv R = tcg_temp_new_i32();
+
+    tcg_gen_sub_tl(R, Rd, Rr); /* R = Rd - Rr */
+    tcg_gen_andi_tl(R, R, 0xff); /* make it 8 bits */
+    /* update status register */
+    gen_sub_CHf(R, Rd, Rr);
+    gen_sub_Vf(R, Rd, Rr);
+    gen_ZNSf(R);
+
+    tcg_temp_free_i32(R);
+    tcg_temp_free_i32(Rr);
+
+    return true;
+}
+
+/*
+ *  This instruction tests a single bit in a register and skips the next
+ *  instruction if the bit is cleared.
+ */
+static bool trans_SBRC(DisasContext *ctx, arg_SBRC *a)
+{
+    TCGv Rr = cpu_r[a->rr];
+
+    ctx->skip_cond = TCG_COND_EQ;
+    ctx->skip_var0 = tcg_temp_new();
+    ctx->free_skip_var0 = true;
+
+    tcg_gen_andi_tl(ctx->skip_var0, Rr, 1 << a->bit);
+    return true;
+}
+
+/*
+ *  This instruction tests a single bit in a register and skips the next
+ *  instruction if the bit is set.
+ */
+static bool trans_SBRS(DisasContext *ctx, arg_SBRS *a)
+{
+    TCGv Rr = cpu_r[a->rr];
+
+    ctx->skip_cond = TCG_COND_NE;
+    ctx->skip_var0 = tcg_temp_new();
+    ctx->free_skip_var0 = true;
+
+    tcg_gen_andi_tl(ctx->skip_var0, Rr, 1 << a->bit);
+    return true;
+}
+
+/*
+ *  This instruction tests a single bit in an I/O Register and skips the
+ *  next instruction if the bit is cleared. This instruction operates on the
+ *  lower 32 I/O Registers -- addresses 0-31.
+ */
+static bool trans_SBIC(DisasContext *ctx, arg_SBIC *a)
+{
+    TCGv temp = tcg_const_i32(a->reg);
+
+    gen_helper_inb(temp, cpu_env, temp);
+    tcg_gen_andi_tl(temp, temp, 1 << a->bit);
+    ctx->skip_cond = TCG_COND_EQ;
+    ctx->skip_var0 = temp;
+    ctx->free_skip_var0 = true;
+
+    return true;
+}
+
+/*
+ *  This instruction tests a single bit in an I/O Register and skips the
+ *  next instruction if the bit is set. This instruction operates on the lower
+ *  32 I/O Registers -- addresses 0-31.
+ */
+static bool trans_SBIS(DisasContext *ctx, arg_SBIS *a)
+{
+    TCGv temp = tcg_const_i32(a->reg);
+
+    gen_helper_inb(temp, cpu_env, temp);
+    tcg_gen_andi_tl(temp, temp, 1 << a->bit);
+    ctx->skip_cond = TCG_COND_NE;
+    ctx->skip_var0 = temp;
+    ctx->free_skip_var0 = true;
+
+    return true;
+}
+
+/*
+ *  Conditional relative branch. Tests a single bit in SREG and branches
+ *  relatively to PC if the bit is cleared. This instruction branches relatively
+ *  to PC in either direction (PC - 63 < = destination <= PC + 64). The
+ *  parameter k is the offset from PC and is represented in two's complement
+ *  form.
+ */
+static bool trans_BRBC(DisasContext *ctx, arg_BRBC *a)
+{
+    TCGLabel *not_taken = gen_new_label();
+
+    TCGv var;
+
+    switch (a->bit) {
+    case 0x00:
+        var = cpu_Cf;
+        break;
+    case 0x01:
+        var = cpu_Zf;
+        break;
+    case 0x02:
+        var = cpu_Nf;
+        break;
+    case 0x03:
+        var = cpu_Vf;
+        break;
+    case 0x04:
+        var = cpu_Sf;
+        break;
+    case 0x05:
+        var = cpu_Hf;
+        break;
+    case 0x06:
+        var = cpu_Tf;
+        break;
+    case 0x07:
+        var = cpu_If;
+        break;
+    default:
+        g_assert_not_reached();
+    }
+
+    tcg_gen_brcondi_i32(TCG_COND_NE, var, 0, not_taken);
+    gen_goto_tb(ctx, 0, ctx->npc + a->imm);
+    gen_set_label(not_taken);
+
+    ctx->bstate = DISAS_CHAIN;
+    return true;
+}
+
+/*
+ *  Conditional relative branch. Tests a single bit in SREG and branches
+ *  relatively to PC if the bit is set. This instruction branches relatively to
+ *  PC in either direction (PC - 63 < = destination <= PC + 64). The parameter k
+ *  is the offset from PC and is represented in two's complement form.
+ */
+static bool trans_BRBS(DisasContext *ctx, arg_BRBS *a)
+{
+    TCGLabel *not_taken = gen_new_label();
+
+    TCGv var;
+
+    switch (a->bit) {
+    case 0x00:
+        var = cpu_Cf;
+        break;
+    case 0x01:
+        var = cpu_Zf;
+        break;
+    case 0x02:
+        var = cpu_Nf;
+        break;
+    case 0x03:
+        var = cpu_Vf;
+        break;
+    case 0x04:
+        var = cpu_Sf;
+        break;
+    case 0x05:
+        var = cpu_Hf;
+        break;
+    case 0x06:
+        var = cpu_Tf;
+        break;
+    case 0x07:
+        var = cpu_If;
+        break;
+    default:
+        g_assert_not_reached();
+    }
+
+    tcg_gen_brcondi_i32(TCG_COND_EQ, var, 0, not_taken);
+    gen_goto_tb(ctx, 0, ctx->npc + a->imm);
+    gen_set_label(not_taken);
+
+    ctx->bstate = DISAS_CHAIN;
+    return true;
+}
+
target/avr/insn.decode +24 -0
index 9c71ed6b2f..32034c10d2 100644
@@ -91,3 +91,27 @@ FMUL            0000 0011 0 ... 1 ...       @fmul
 FMULS           0000 0011 1 ... 0 ...       @fmul
 FMULSU          0000 0011 1 ... 1 ...       @fmul
 DES             1001 0100 imm:4 1011
+
+#
+# Branch Instructions
+#
+RJMP            1100 imm:s12
+IJMP            1001 0100 0000 1001
+EIJMP           1001 0100 0001 1001
+JMP             1001 010 ..... 110 .        imm=%imm_call
+RCALL           1101 imm:s12
+ICALL           1001 0101 0000 1001
+EICALL          1001 0101 0001 1001
+CALL            1001 010 ..... 111 .        imm=%imm_call
+RET             1001 0101 0000 1000
+RETI            1001 0101 0001 1000
+CPSE            0001 00 . ..... ....        @op_rd_rr
+CP              0001 01 . ..... ....        @op_rd_rr
+CPC             0000 01 . ..... ....        @op_rd_rr
+CPI             0011 .... .... ....         @op_rd_imm8
+SBRC            1111 110 rr:5 0 bit:3
+SBRS            1111 111 rr:5 0 bit:3
+SBIC            1001 1001 reg:5 bit:3
+SBIS            1001 1011 reg:5 bit:3
+BRBS            1111 00 ....... ...         @op_bit_imm
+BRBC            1111 01 ....... ...         @op_bit_imm
Signature · 1 line
2.17.2 (Apple Git-113)
#13 lore raw

PATCH v40 11/21 hw/avr: Add limited support for USART peripheral

In reply to Michael Rolnik

These were designed to facilitate testing but should provide enough function to be useful in other contexts.
Only a subset of the functions of each peripheral is implemented, mainly due to the lack of a standard way to handle electrical connections (like GPIO pins).

---
 include/hw/char/avr_usart.h |  93 +++++++++++
 hw/char/avr_usart.c         | 320 ++++++++++++++++++++++++++++++++++++
 hw/char/Kconfig             |   3 +
 hw/char/Makefile.objs       |   1 +
 4 files changed, 417 insertions(+)
 create mode 100644 include/hw/char/avr_usart.h
 create mode 100644 hw/char/avr_usart.c
include/hw/char/avr_usart.h new file +93 -0
new file mode 100644
index 0000000000..467e97e8c0
@@ -0,0 +1,93 @@
+/*
+ * AVR USART
+ *
+ * Copyright (c) 2018 University of Kent
+ * Author: Sarah Harris
+ *
+ * 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/lgpl-2.1.html>
+ */
+
+#ifndef HW_AVR_USART_H
+#define HW_AVR_USART_H
+
+#include "hw/sysbus.h"
+#include "chardev/char-fe.h"
+#include "hw/hw.h"
+
+/* Offsets of registers. */
+#define USART_DR   0x06
+#define USART_CSRA  0x00
+#define USART_CSRB  0x01
+#define USART_CSRC  0x02
+#define USART_BRRH 0x05
+#define USART_BRRL 0x04
+
+/* Relevant bits in regiters. */
+#define USART_CSRA_RXC    (1 << 7)
+#define USART_CSRA_TXC    (1 << 6)
+#define USART_CSRA_DRE    (1 << 5)
+#define USART_CSRA_MPCM   (1 << 0)
+
+#define USART_CSRB_RXCIE  (1 << 7)
+#define USART_CSRB_TXCIE  (1 << 6)
+#define USART_CSRB_DREIE  (1 << 5)
+#define USART_CSRB_RXEN   (1 << 4)
+#define USART_CSRB_TXEN   (1 << 3)
+#define USART_CSRB_CSZ2   (1 << 2)
+#define USART_CSRB_RXB8   (1 << 1)
+#define USART_CSRB_TXB8   (1 << 0)
+
+#define USART_CSRC_MSEL1  (1 << 7)
+#define USART_CSRC_MSEL0  (1 << 6)
+#define USART_CSRC_PM1    (1 << 5)
+#define USART_CSRC_PM0    (1 << 4)
+#define USART_CSRC_CSZ1   (1 << 2)
+#define USART_CSRC_CSZ0   (1 << 1)
+
+#define TYPE_AVR_USART "avr-usart"
+#define AVR_USART(obj) \
+    OBJECT_CHECK(AVRUsartState, (obj), TYPE_AVR_USART)
+
+typedef struct {
+    /* <private> */
+    SysBusDevice parent_obj;
+
+    /* <public> */
+    MemoryRegion mmio;
+
+    CharBackend chr;
+
+    bool enabled;
+
+    uint8_t data;
+    bool data_valid;
+    uint8_t char_mask;
+    /* Control and Status Registers */
+    uint8_t csra;
+    uint8_t csrb;
+    uint8_t csrc;
+    /* Baud Rate Registers (low/high byte) */
+    uint8_t brrh;
+    uint8_t brrl;
+
+    /* Receive Complete */
+    qemu_irq rxc_irq;
+    /* Transmit Complete */
+    qemu_irq txc_irq;
+    /* Data Register Empty */
+    qemu_irq dre_irq;
+} AVRUsartState;
+
+#endif /* HW_AVR_USART_H */
hw/char/avr_usart.c new file +320 -0

320 lines not on this page · raw message

hw/char/Kconfig +3 -0
index 40e7a8b8bb..331b20983f 100644
@@ -46,3 +46,6 @@ config SCLPCONSOLE
 
 config TERMINAL3270
     bool
+
+config AVR_USART
+    bool
hw/char/Makefile.objs +1 -0
index 02d8a66925..f05c1f5667 100644
@@ -21,6 +21,7 @@ obj-$(CONFIG_PSERIES) += spapr_vty.o
 obj-$(CONFIG_DIGIC) += digic-uart.o
 obj-$(CONFIG_STM32F2XX_USART) += stm32f2xx_usart.o
 obj-$(CONFIG_RASPI) += bcm2835_aux.o
+common-obj-$(CONFIG_AVR_USART) += avr_usart.o
 
 common-obj-$(CONFIG_CMSDK_APB_UART) += cmsdk-apb-uart.o
 common-obj-$(CONFIG_ETRAXFS) += etraxfs_ser.o
Signature · 1 line
2.17.2 (Apple Git-113)
#20 lore raw

PATCH v40 18/21 target/avr: Update build system

In reply to Michael Rolnik

Make AVR support buildable

---
 configure                       |  7 +++++++
 default-configs/avr-softmmu.mak |  5 +++++
 target/avr/Makefile.objs        | 34 +++++++++++++++++++++++++++++++++
 3 files changed, 46 insertions(+)
 create mode 100644 default-configs/avr-softmmu.mak
 create mode 100644 target/avr/Makefile.objs
configure +7 -0
index 940bf9e87a..249b7adff0 100755
@@ -7613,6 +7613,10 @@ case "$target_name" in
     mttcg="yes"
     gdb_xml_files="aarch64-core.xml aarch64-fpu.xml arm-core.xml arm-vfp.xml arm-vfp3.xml arm-neon.xml"
   ;;
+  avr)
+    gdb_xml_files="avr-cpu.xml"
+    target_compiler=$cross_cc_avr
+  ;;
   cris)
   ;;
   hppa)
@@ -7832,6 +7836,9 @@ for i in $ARCH $TARGET_BASE_ARCH ; do
       disas_config "ARM_A64"
     fi
   ;;
+  avr)
+    disas_config "AVR"
+  ;;
   cris)
     disas_config "CRIS"
   ;;
default-configs/avr-softmmu.mak new file +5 -0
new file mode 100644
index 0000000000..d1e1c28118
@@ -0,0 +1,5 @@
+# Default configuration for avr-softmmu
+
+# Boards:
+#
+CONFIG_AVR_SAMPLE=y
target/avr/Makefile.objs new file +34 -0
new file mode 100644
index 0000000000..7523e0c6e2
@@ -0,0 +1,34 @@
+#
+#  QEMU AVR CPU
+#
+#  Copyright (c) 2019 Michael Rolnik
+#
+#  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/lgpl-2.1.html>
+#
+
+DECODETREE = $(SRC_PATH)/scripts/decodetree.py
+decode-y = $(SRC_PATH)/target/avr/insn.decode
+
+target/avr/decode_insn.inc.c: $(decode-y) $(DECODETREE)
+	$(call quiet-command, \
+	  $(PYTHON) $(DECODETREE) -o $@ --decode decode_insn --insnwidth 16 $<, \
+	  "GEN", $(TARGET_DIR)$@)
+
+target/avr/translate.o: target/avr/decode_insn.inc.c
+
+obj-y += translate.o cpu.o helper.o
+obj-y += gdbstub.o
+obj-y += disas.o
+obj-$(CONFIG_SOFTMMU) += machine.o
Signature · 1 line
2.17.2 (Apple Git-113)
#10 lore raw

PATCH v40 8/21 target/avr: Add instruction translation - MCU Control Instructions

In reply to Michael Rolnik

This includes:
- BREAK
- NOP
- SLEEP
- WDR

---
 target/avr/translate.c | 68 ++++++++++++++++++++++++++++++++++++++++++
 target/avr/insn.decode |  9 ++++++
 2 files changed, 77 insertions(+)
target/avr/translate.c +68 -0
index 55706c6b29..246f9e7e47 100644
@@ -2681,3 +2681,71 @@ static bool trans_BCLR(DisasContext *ctx, arg_BCLR *a)
 
     return true;
 }
+
+/*
+ * MCU Control Instructions
+ */
+
+/*
+ *  The BREAK instruction is used by the On-chip Debug system, and is
+ *  normally not used in the application software. When the BREAK instruction is
+ *  executed, the AVR CPU is set in the Stopped Mode. This gives the On-chip
+ *  Debugger access to internal resources.  If any Lock bits are set, or either
+ *  the JTAGEN or OCDEN Fuses are unprogrammed, the CPU will treat the BREAK
+ *  instruction as a NOP and will not enter the Stopped mode.  This instruction
+ *  is not available in all devices. Refer to the device specific instruction
+ *  set summary.
+ */
+static bool trans_BREAK(DisasContext *ctx, arg_BREAK *a)
+{
+    if (!avr_have_feature(ctx, AVR_FEATURE_BREAK)) {
+        return true;
+    }
+
+#ifdef BREAKPOINT_ON_BREAK
+    tcg_gen_movi_tl(cpu_pc, ctx->npc - 1);
+    gen_helper_debug(cpu_env);
+    ctx->bstate = DISAS_EXIT;
+#else
+    /* NOP */
+#endif
+
+    return true;
+}
+
+
+/*
+ *  This instruction performs a single cycle No Operation.
+ */
+static bool trans_NOP(DisasContext *ctx, arg_NOP *a)
+{
+
+    /* NOP */
+
+    return true;
+}
+
+
+/*
+ *  This instruction sets the circuit in sleep mode defined by the MCU
+ *  Control Register.
+ */
+static bool trans_SLEEP(DisasContext *ctx, arg_SLEEP *a)
+{
+    gen_helper_sleep(cpu_env);
+    ctx->bstate = DISAS_NORETURN;
+    return true;
+}
+
+
+/*
+ *  This instruction resets the Watchdog Timer. This instruction must be
+ *  executed within a limited time given by the WD prescaler. See the Watchdog
+ *  Timer hardware specification.
+ */
+static bool trans_WDR(DisasContext *ctx, arg_WDR *a)
+{
+    gen_helper_wdr(cpu_env);
+
+    return true;
+}
target/avr/insn.decode +9 -0
index 4ee55862b2..0e4ec9ddf0 100644
@@ -172,3 +172,12 @@ BST             1111 101 rd:5 0 bit:3
 BLD             1111 100 rd:5 0 bit:3
 BSET            1001 0100 0 bit:3 1000
 BCLR            1001 0100 1 bit:3 1000
+
+#
+# MCU Control Instructions
+#
+BREAK           1001 0101 1001 1000
+NOP             0000 0000 0000 0000
+SLEEP           1001 0101 1000 1000
+WDR             1001 0101 1010 1000
+
Signature · 1 line
2.17.2 (Apple Git-113)
#17 lore raw

PATCH v40 15/21 target/avr: Add section about AVR into QEMU documentation

In reply to Michael Rolnik
---
 qemu-doc.texi | 51 +++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 51 insertions(+)
qemu-doc.texi +51 -0
index 39f950471f..515aacfae9 100644
@@ -1741,6 +1741,7 @@ differences are mentioned in the following sections.
 * Microblaze System emulator::
 * SH4 System emulator::
 * Xtensa System emulator::
+* AVR System emulator::
 @end menu
 
 @node PowerPC System emulator
@@ -2514,6 +2515,56 @@ so should only be used with trusted guest OS.
 
 @c man end
 
+@node AVR System emulator
+@section AVR System emulator
+@cindex system emulation (AVR)
+
+Use the executable @file{qemu-system-avr} to emulates a AVR 8 bit based machine having one for the following cores: avr1, avr2, avr25, avr3, avr31, avr35, avr4, avr5, avr51, avr6, avrtiny, xmega2, xmega3, xmega4, xmega5, xmega6 and xmega7.
+
+As for now it does not support any real MCUs. However, it does support a "sample" board for educational and testing purposes. This "sample" board hosts USART & 16 bit timer devices and it's enought to run FreeRTOS based applicaton (like this @url{https://github.com/seharris/qemu-avr-tests/blob/master/free-rtos/Demo/AVR_ATMega2560_GCC/demo.elf,,demo})
+
+Following are examples of possible usages, assuming program.elf is compiled for AVR cpu
+@itemize
+
+@item Continious non interrupted execution
+@example
+qemu-system-avr -kernel program.elf
+@end example
+
+@item Continious non interrupted execution with serial output into telnet window
+@example
+qemu-system-avr -kernel program.elf -serial tcp::5678,server,nowait -nographic
+@end example
+and then in another shell
+@example
+telent localhost 5678
+@end example
+
+@item Continious non interrupted execution with serial output into stdout
+@example
+qemu-system-avr -kernel program.elf -serial stdio
+@end example
+
+@item Debugging wit GDB debugger
+@example
+qemu-system-avr -kernel program.elf -s -S
+@end example
+and then in another shell
+@example
+avr-gdb program.elf
+@end example
+and then within GDB shell
+@example
+target remote :1234
+@end example
+
+@item Print out executed instructions
+@example
+qemu-system-avr -kernel program.elf -d in_asm
+@end example
+
+@end itemize
+
 @node QEMU User space emulator
 @chapter QEMU User space emulator
 
Signature · 1 line
2.17.2 (Apple Git-113)
#15 lore raw

PATCH v40 13/21 hw/avr: Add dummy mask device

In reply to Michael Rolnik

This is a simple device of just one register, whenver this register is written it calls qemu_set_irq function for each of 8 bits/IRQs.. It is used to implement AVR Power Reduction

---
 include/hw/misc/avr_mask.h |  47 ++++++++++++++++
 hw/misc/avr_mask.c         | 112 +++++++++++++++++++++++++++++++++++++
 hw/misc/Kconfig            |   3 +
 hw/misc/Makefile.objs      |   2 +
 4 files changed, 164 insertions(+)
 create mode 100644 include/hw/misc/avr_mask.h
 create mode 100644 hw/misc/avr_mask.c
include/hw/misc/avr_mask.h new file +47 -0
new file mode 100644
index 0000000000..d3e21972d8
@@ -0,0 +1,47 @@
+/*
+ * AVR Power Reduction
+ *
+ * Copyright (c) 2019 Michael Rolnik
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#ifndef HW_avr_mask_H
+#define HW_avr_mask_H
+
+#include "hw/sysbus.h"
+#include "chardev/char-fe.h"
+#include "hw/hw.h"
+
+
+#define TYPE_AVR_MASK "avr-mask"
+#define AVR_MASK(obj) OBJECT_CHECK(AVRMaskState, (obj), TYPE_AVR_MASK)
+
+typedef struct {
+    /* <private> */
+    SysBusDevice parent_obj;
+
+    /* <public> */
+    MemoryRegion iomem;
+
+    uint8_t val;
+    qemu_irq irq[8];
+} AVRMaskState;
+
+#endif /* HW_avr_mask_H */
hw/misc/avr_mask.c new file +112 -0

112 lines not on this page · raw message

hw/misc/Kconfig +3 -0
index 2164646553..e79841e3a4 100644
@@ -125,4 +125,7 @@ config MAC_VIA
     select MOS6522
     select ADB
 
+config AVR_MASK
+    bool
+
 source macio/Kconfig
hw/misc/Makefile.objs +2 -0
index ba898a5781..3a8093be6a 100644
@@ -82,3 +82,5 @@ common-obj-$(CONFIG_NRF51_SOC) += nrf51_rng.o
 obj-$(CONFIG_MAC_VIA) += mac_via.o
 
 common-obj-$(CONFIG_GRLIB) += grlib_ahb_apb_pnp.o
+
+obj-$(CONFIG_AVR_MASK) += avr_mask.o
Signature · 1 line
2.17.2 (Apple Git-113)
#19 lore raw

PATCH v40 17/21 target/avr: Add machine none test

In reply to Michael Rolnik
---
 tests/machine-none-test.c | 1 +
 1 file changed, 1 insertion(+)
tests/machine-none-test.c +1 -0
index 5953d31755..3e5c74e73e 100644
@@ -27,6 +27,7 @@ static struct arch2cpu cpus_map[] = {
     /* tested targets list */
     { "arm", "cortex-a15" },
     { "aarch64", "cortex-a57" },
+    { "avr", "avr6-avr-cpu" },
     { "x86_64", "qemu64,apic-id=0" },
     { "i386", "qemu32,apic-id=0" },
     { "alpha", "ev67" },
Signature · 1 line
2.17.2 (Apple Git-113)
#14 lore raw

PATCH v40 12/21 hw/avr: Add limited support for 16 bit timer peripheral

In reply to Michael Rolnik

These were designed to facilitate testing but should provide enough function to be useful in other contexts.
Only a subset of the functions of each peripheral is implemented, mainly due to the lack of a standard way to handle electrical connections (like GPIO pins).

---
 include/hw/timer/avr_timer16.h |  94 +++++
 hw/timer/avr_timer16.c         | 602 +++++++++++++++++++++++++++++++++
 hw/timer/Kconfig               |   3 +
 hw/timer/Makefile.objs         |   2 +
 4 files changed, 701 insertions(+)
 create mode 100644 include/hw/timer/avr_timer16.h
 create mode 100644 hw/timer/avr_timer16.c
include/hw/timer/avr_timer16.h new file +94 -0
new file mode 100644
index 0000000000..4ae0c64a34
@@ -0,0 +1,94 @@
+/*
+ * AVR 16 bit timer
+ *
+ * Copyright (c) 2018 University of Kent
+ * Author: Ed Robbins
+ *
+ * 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/lgpl-2.1.html>
+ */
+
+/*
+ * Driver for 16 bit timers on 8 bit AVR devices.
+ * Note:
+ * On ATmega640/V-1280/V-1281/V-2560/V-2561/V timers 1, 3, 4 and 5 are 16 bit
+ */
+
+#ifndef AVR_TIMER16_H
+#define AVR_TIMER16_H
+
+#include "hw/sysbus.h"
+#include "qemu/timer.h"
+#include "hw/hw.h"
+
+enum NextInterrupt {
+    OVERFLOW,
+    COMPA,
+    COMPB,
+    COMPC,
+    CAPT
+};
+
+#define TYPE_AVR_TIMER16 "avr-timer16"
+#define AVR_TIMER16(obj) \
+    OBJECT_CHECK(AVRTimer16State, (obj), TYPE_AVR_TIMER16)
+
+typedef struct AVRTimer16State {
+    /* <private> */
+    SysBusDevice parent_obj;
+
+    /* <public> */
+    MemoryRegion iomem;
+    MemoryRegion imsk_iomem;
+    MemoryRegion ifr_iomem;
+    QEMUTimer *timer;
+    qemu_irq capt_irq;
+    qemu_irq compa_irq;
+    qemu_irq compb_irq;
+    qemu_irq compc_irq;
+    qemu_irq ovf_irq;
+
+    bool enabled;
+
+    /* registers */
+    uint8_t cra;
+    uint8_t crb;
+    uint8_t crc;
+    uint8_t cntl;
+    uint8_t cnth;
+    uint8_t icrl;
+    uint8_t icrh;
+    uint8_t ocral;
+    uint8_t ocrah;
+    uint8_t ocrbl;
+    uint8_t ocrbh;
+    uint8_t ocrcl;
+    uint8_t ocrch;
+    /*
+     * Reads and writes to CNT and ICR utilise a bizarre temporary
+     * register, which we emulate
+     */
+    uint8_t rtmp;
+    uint8_t imsk;
+    uint8_t ifr;
+
+    uint8_t id;
+    uint64_t cpu_freq_hz;
+    uint64_t freq_hz;
+    uint64_t period_ns;
+    uint64_t reset_time_ns;
+    enum NextInterrupt next_interrupt;
+} AVRTimer16State;
+
+#endif /* AVR_TIMER16_H */
hw/timer/avr_timer16.c new file +602 -0

602 lines not on this page · raw message

hw/timer/Kconfig +3 -0
index a990f9fe35..4343bc23f3 100644
@@ -34,3 +34,6 @@ config CMSDK_APB_TIMER
 config CMSDK_APB_DUALTIMER
     bool
     select PTIMER
+
+config AVR_TIMER16
+    bool
hw/timer/Makefile.objs +2 -0
index dece235fd7..af0913ca3b 100644
@@ -35,3 +35,5 @@ common-obj-$(CONFIG_CMSDK_APB_TIMER) += cmsdk-apb-timer.o
 common-obj-$(CONFIG_CMSDK_APB_DUALTIMER) += cmsdk-apb-dualtimer.o
 common-obj-$(CONFIG_MSF2) += mss-timer.o
 common-obj-$(CONFIG_RASPI) += bcm2835_systmr.o
+
+obj-$(CONFIG_AVR_TIMER16) += avr_timer16.o
Signature · 1 line
2.17.2 (Apple Git-113)
#18 lore raw

PATCH v40 16/21 target/avr: Register AVR support with the rest of QEMU

In reply to Michael Rolnik

Add AVR related definitions into QEMU

include/disas/dis-asm.h

---
 qapi/machine.json          |  3 ++-
 include/disas/dis-asm.h    | 19 +++++++++++++++++++
 include/sysemu/arch_init.h |  1 +
 arch_init.c                |  2 ++
 4 files changed, 24 insertions(+), 1 deletion(-)
qapi/machine.json +2 -1
index ca26779f1a..8c6df54921 100644
@@ -21,11 +21,12 @@
 #        is true even for "qemu-system-x86_64".
 #
 # ppcemb: dropped in 3.1
+# avr: since 5.0
 #
 # Since: 3.0
 ##
 { 'enum' : 'SysEmuTarget',
-  'data' : [ 'aarch64', 'alpha', 'arm', 'cris', 'hppa', 'i386', 'lm32',
+  'data' : [ 'aarch64', 'alpha', 'arm', 'avr', 'cris', 'hppa', 'i386', 'lm32',
              'm68k', 'microblaze', 'microblazeel', 'mips', 'mips64',
              'mips64el', 'mipsel', 'moxie', 'nios2', 'or1k', 'ppc',
              'ppc64', 'riscv32', 'riscv64', 's390x', 'sh4',
include/disas/dis-asm.h +19 -0
index e9c7dd8eb4..79bbc8b498 100644
@@ -211,6 +211,25 @@ enum bfd_architecture
 #define bfd_mach_m32r          0  /* backwards compatibility */
   bfd_arch_mn10200,    /* Matsushita MN10200 */
   bfd_arch_mn10300,    /* Matsushita MN10300 */
+  bfd_arch_avr,       /* Atmel AVR microcontrollers.  */
+#define bfd_mach_avr1       1
+#define bfd_mach_avr2       2
+#define bfd_mach_avr25      25
+#define bfd_mach_avr3       3
+#define bfd_mach_avr31      31
+#define bfd_mach_avr35      35
+#define bfd_mach_avr4       4
+#define bfd_mach_avr5       5
+#define bfd_mach_avr51      51
+#define bfd_mach_avr6       6
+#define bfd_mach_avrtiny    100
+#define bfd_mach_avrxmega1  101
+#define bfd_mach_avrxmega2  102
+#define bfd_mach_avrxmega3  103
+#define bfd_mach_avrxmega4  104
+#define bfd_mach_avrxmega5  105
+#define bfd_mach_avrxmega6  106
+#define bfd_mach_avrxmega7  107
   bfd_arch_cris,       /* Axis CRIS */
 #define bfd_mach_cris_v0_v10   255
 #define bfd_mach_cris_v32      32
include/sysemu/arch_init.h +1 -0
index 62c6fe4cf1..893df26ce2 100644
@@ -24,6 +24,7 @@ enum {
     QEMU_ARCH_NIOS2 = (1 << 17),
     QEMU_ARCH_HPPA = (1 << 18),
     QEMU_ARCH_RISCV = (1 << 19),
+    QEMU_ARCH_AVR = (1 << 20),
 };
 
 extern const uint32_t arch_type;
arch_init.c +2 -0
index 705d0b94ad..6a741165b2 100644
@@ -89,6 +89,8 @@ int graphic_depth = 32;
 #define QEMU_ARCH QEMU_ARCH_UNICORE32
 #elif defined(TARGET_XTENSA)
 #define QEMU_ARCH QEMU_ARCH_XTENSA
+#elif defined(TARGET_AVR)
+#define QEMU_ARCH QEMU_ARCH_AVR
 #endif
 
 const uint32_t arch_type = QEMU_ARCH;
Signature · 1 line
2.17.2 (Apple Git-113)
author NAK
#16 lore raw

PATCH v40 14/21 hw/avr: Add example board configuration

In reply to Michael Rolnik

A simple board setup that configures an AVR CPU to run a given firmware image.
This is all that's useful to implement without peripheral emulation as AVR CPUs include a lot of on-board peripherals.

NOTE: this is not a real board !!!!
NOTE: it's used for CPU testing!!!!

---
 include/elf.h        |   2 +
 include/hw/elf_ops.h |   6 +-
 include/hw/loader.h  |   6 +-
 hw/avr/sample.c      | 295 +++++++++++++++++++++++++++++++++++++++++++
 hw/core/loader.c     |  15 ++-
 hw/riscv/boot.c      |   2 +-
 hw/Kconfig           |   1 +
 hw/avr/Kconfig       |   6 +
 hw/avr/Makefile.objs |   1 +
 9 files changed, 323 insertions(+), 11 deletions(-)
 create mode 100644 hw/avr/sample.c
 create mode 100644 hw/avr/Kconfig
 create mode 100644 hw/avr/Makefile.objs
include/elf.h +2 -0
index 3501e0c8d0..53cdfa23b7 100644
@@ -202,6 +202,8 @@ typedef struct mips_elf_abiflags_v0 {
 #define EM_MOXIE           223     /* Moxie processor family */
 #define EM_MOXIE_OLD       0xFEED
 
+#define EM_AVR 83 /* AVR 8-bit microcontroller */
+
 /* This is the info that is needed to parse the dynamic section of the file */
 #define DT_NULL		0
 #define DT_NEEDED	1
include/hw/elf_ops.h +5 -1
index e07d276df7..70de85fa72 100644
@@ -316,7 +316,8 @@ static int glue(load_elf, SZ)(const char *name, int fd,
                               void *translate_opaque,
                               int must_swab, uint64_t *pentry,
                               uint64_t *lowaddr, uint64_t *highaddr,
-                              int elf_machine, int clear_lsb, int data_swab,
+                              uint32_t *pe_flags, int elf_machine,
+                              int clear_lsb, int data_swab,
                               AddressSpace *as, bool load_rom,
                               symbol_fn_t sym_cb)
 {
@@ -594,6 +595,9 @@ static int glue(load_elf, SZ)(const char *name, int fd,
         }
     }
 
+    if (pe_flags) {
+        *pe_flags = (uint32_t)(elf_sword)ehdr.e_flags;
+    }
     if (lowaddr)
         *lowaddr = (uint64_t)(elf_sword)low;
     if (highaddr)
include/hw/loader.h +4 -2
index 48a96cd559..22b59e15ba 100644
@@ -101,6 +101,7 @@ const char *load_elf_strerror(int error);
  * @pentry: Populated with program entry point. Ignored if NULL.
  * @lowaddr: Populated with lowest loaded address. Ignored if NULL.
  * @highaddr: Populated with highest loaded address. Ignored if NULL.
+ * @pe_flags: Populated with e_flags. Ignore if NULL.
  * @bigendian: Expected ELF endianness. 0 for LE otherwise BE
  * @elf_machine: Expected ELF machine type
  * @clear_lsb: Set to mask off LSB of addresses (Some architectures use
@@ -131,8 +132,9 @@ int load_elf_ram_sym(const char *filename,
                      uint64_t (*elf_note_fn)(void *, void *, bool),
                      uint64_t (*translate_fn)(void *, uint64_t),
                      void *translate_opaque, uint64_t *pentry,
-                     uint64_t *lowaddr, uint64_t *highaddr, int big_endian,
-                     int elf_machine, int clear_lsb, int data_swab,
+                     uint64_t *lowaddr, uint64_t *highaddr, uint32_t *pe_flags,
+                     int big_endian, int elf_machine,
+                     int clear_lsb, int data_swab,
                      AddressSpace *as, bool load_rom, symbol_fn_t sym_cb);
 
 /** load_elf_ram:
hw/avr/sample.c new file +295 -0

295 lines not on this page · raw message

hw/core/loader.c +8 -7
index 5099f27dc8..9961b4423b 100644
@@ -438,7 +438,7 @@ int load_elf_ram(const char *filename,
 {
     return load_elf_ram_sym(filename, elf_note_fn,
                             translate_fn, translate_opaque,
-                            pentry, lowaddr, highaddr, big_endian,
+                            pentry, lowaddr, highaddr, NULL, big_endian,
                             elf_machine, clear_lsb, data_swab, as,
                             load_rom, NULL);
 }
@@ -448,8 +448,9 @@ int load_elf_ram_sym(const char *filename,
                      uint64_t (*elf_note_fn)(void *, void *, bool),
                      uint64_t (*translate_fn)(void *, uint64_t),
                      void *translate_opaque, uint64_t *pentry,
-                     uint64_t *lowaddr, uint64_t *highaddr, int big_endian,
-                     int elf_machine, int clear_lsb, int data_swab,
+                     uint64_t *lowaddr, uint64_t *highaddr, uint32_t *pe_flags,
+                     int big_endian, int elf_machine,
+                     int clear_lsb, int data_swab,
                      AddressSpace *as, bool load_rom, symbol_fn_t sym_cb)
 {
     int fd, data_order, target_data_order, must_swab, ret = ELF_LOAD_FAILED;
@@ -490,13 +491,13 @@ int load_elf_ram_sym(const char *filename,
     if (e_ident[EI_CLASS] == ELFCLASS64) {
         ret = load_elf64(filename, fd, elf_note_fn,
                          translate_fn, translate_opaque, must_swab,
-                         pentry, lowaddr, highaddr, elf_machine, clear_lsb,
-                         data_swab, as, load_rom, sym_cb);
+                         pentry, lowaddr, highaddr, pe_flags, elf_machine,
+                         clear_lsb, data_swab, as, load_rom, sym_cb);
     } else {
         ret = load_elf32(filename, fd, elf_note_fn,
                          translate_fn, translate_opaque, must_swab,
-                         pentry, lowaddr, highaddr, elf_machine, clear_lsb,
-                         data_swab, as, load_rom, sym_cb);
+                         pentry, lowaddr, highaddr, pe_flags, elf_machine,
+                         clear_lsb, data_swab, as, load_rom, sym_cb);
     }
 
  fail:
hw/riscv/boot.c +1 -1
index 027303d2a3..746ca1f795 100644
@@ -119,7 +119,7 @@ target_ulong riscv_load_kernel(const char *kernel_filename, symbol_fn_t sym_cb)
     uint64_t kernel_entry, kernel_high;
 
     if (load_elf_ram_sym(kernel_filename, NULL, NULL, NULL,
-                         &kernel_entry, NULL, &kernel_high, 0,
+                         &kernel_entry, NULL, &kernel_high, NULL, 0,
                          EM_RISCV, 1, 0, NULL, true, sym_cb) > 0) {
         return kernel_entry;
     }
hw/Kconfig +1 -0
index ecf491bf04..f80dff3b75 100644
@@ -43,6 +43,7 @@ source watchdog/Kconfig
 # arch Kconfig
 source arm/Kconfig
 source alpha/Kconfig
+source avr/Kconfig
 source cris/Kconfig
 source hppa/Kconfig
 source i386/Kconfig
hw/avr/Kconfig new file +6 -0
new file mode 100644
index 0000000000..92aa1e6afb
@@ -0,0 +1,6 @@
+config AVR_SAMPLE
+    bool
+    select AVR_TIMER16
+    select AVR_USART
+    select AVR_MASK
+    select UNIMP
hw/avr/Makefile.objs new file +1 -0
new file mode 100644
index 0000000000..626b7064b3
@@ -0,0 +1 @@
+obj-y += sample.o
Signature · 1 line
2.17.2 (Apple Git-113)
#21 lore raw

PATCH v40 19/21 target/avr: Add boot serial test

In reply to Michael Rolnik

Print out 'T' through serial port

---
 tests/boot-serial-test.c | 10 ++++++++++
 tests/Makefile.include   |  2 ++
 2 files changed, 12 insertions(+)
tests/boot-serial-test.c +10 -0
index 05c7f44457..e556f09db8 100644
@@ -16,6 +16,15 @@
 #include "qemu/osdep.h"
 #include "libqtest.h"
 
+static const uint8_t bios_avr[] = {
+    0x88, 0xe0,             /* ldi r24, 0x08   */
+    0x80, 0x93, 0xc1, 0x00, /* sts 0x00C1, r24 ; Enable tx */
+    0x86, 0xe0,             /* ldi r24, 0x06   */
+    0x80, 0x93, 0xc2, 0x00, /* sts 0x00C2, r24 ; Set the data bits to 8 */
+    0x84, 0xe5,             /* ldi r24, 0x54   */
+    0x80, 0x93, 0xc6, 0x00, /* sts 0x00C6, r24 ; Output 'T' */
+};
+
 static const uint8_t kernel_mcf5208[] = {
     0x41, 0xf9, 0xfc, 0x06, 0x00, 0x00,     /* lea 0xfc060000,%a0 */
     0x10, 0x3c, 0x00, 0x54,                 /* move.b #'T',%d0 */
@@ -103,6 +112,7 @@ typedef struct testdef {
 
 static testdef_t tests[] = {
     { "alpha", "clipper", "", "PCI:" },
+    { "avr", "sample", "", "T", sizeof(bios_avr), NULL, bios_avr },
     { "ppc", "ppce500", "", "U-Boot" },
     { "ppc", "40p", "-vga none -boot d", "Trying cd:," },
     { "ppc", "g3beige", "", "PowerPC,750" },
tests/Makefile.include +2 -0
index 31b86674e5..554a5c3084 100644
@@ -206,6 +206,8 @@ check-qtest-i386-y += tests/test-x86-cpuid-compat$(EXESUF)
 check-qtest-i386-y += tests/numa-test$(EXESUF)
 check-qtest-x86_64-y += $(check-qtest-i386-y)
 
+check-qtest-avr-y += tests/boot-serial-test$(EXESUF)
+
 check-qtest-alpha-y += tests/boot-serial-test$(EXESUF)
 check-qtest-alpha-$(CONFIG_VGA) += tests/display-vga-test$(EXESUF)
 
Signature · 1 line
2.17.2 (Apple Git-113)
#22 lore raw

PATCH v40 20/21 target/avr: Add Avocado test

In reply to Michael Rolnik

The test is based on
https://github.com/seharris/qemu-avr-tests/tree/master/free-rtos/Demo
demo which. If working correctly, prints 'ABCDEFGHIJKLMNOPQRSTUVWX' out. it also demostrates that timer and IRQ are working

---
 tests/acceptance/machine_avr6.py | 58 ++++++++++++++++++++++++++++++++
 1 file changed, 58 insertions(+)
 create mode 100644 tests/acceptance/machine_avr6.py
tests/acceptance/machine_avr6.py new file +58 -0
new file mode 100644
index 0000000000..7a7d8afc29
@@ -0,0 +1,58 @@
+#
+# QEMU AVR
+#
+# Copyright (c) 2019 Michael Rolnik <[email protected]>
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program 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 General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+#
+
+import logging
+import time
+import distutils.spawn
+
+from avocado import skipUnless
+from avocado_qemu import Test
+from avocado.utils import process
+
+class AVR6Machine(Test):
+    timeout = 5
+
+    def test_freertos(self):
+        """
+        :avocado: tags=arch:avr
+        :avocado: tags=machine:sample
+        """
+        """
+        https://github.com/seharris/qemu-avr-tests/raw/master/free-rtos/Demo/AVR_ATMega2560_GCC/demo.elf
+        constantly prints out 'ABCDEFGHIJKLMNOPQRSTUVWXABCDEFGHIJKLMNOPQRSTUVWX'
+        """
+        rom_url = 'https://github.com/seharris/qemu-avr-tests'
+        rom_sha1= '36c3e67b8755dcf37e06af6730ef5d477b8ed16d'
+        rom_url += '/raw/'
+        rom_url += rom_sha1
+        rom_url += '/free-rtos/Demo/AVR_ATMega2560_GCC/demo.elf'
+        rom_hash = '7eb521f511ca8f2622e0a3c5e8dd686efbb911d4'
+        rom_path = self.fetch_asset(rom_url, asset_hash=rom_hash)
+
+        self.vm.set_machine('sample')
+        self.vm.add_args('-bios', rom_path)
+        self.vm.add_args('-nographic')
+        self.vm.launch()
+
+        time.sleep(2)
+        self.vm.shutdown()
+
+        match = 'ABCDEFGHIJKLMNOPQRSTUVWXABCDEFGHIJKLMNOPQRSTUVWX'
+
+        self.assertIn(match, self.vm.get_log())
Signature · 1 line
2.17.2 (Apple Git-113)
#25 lore raw

PATCH v40 21/21 target/avr: Update MAINTAINERS file

In reply to Michael Rolnik

Include AVR maintaners in MAINTAINERS file

---
 MAINTAINERS | 21 +++++++++++++++++++++
 1 file changed, 21 insertions(+)
MAINTAINERS +21 -0
index 387879aebc..90c0a0c27f 100644
@@ -163,6 +163,15 @@ S: Maintained
 F: hw/arm/smmu*
 F: include/hw/arm/smmu*
 
+AVR TCG CPUs
+M: Michael Rolnik <[email protected]>
+R: Sarah Harris <[email protected]>
+S: Maintained
+F: target/avr/
+F: tests/acceptance/machine_avr6.py
+F: default-configs/avr-softmmu.mak
+F: gdb-xml/avr-cpu.xml
+
 CRIS TCG CPUs
 M: Edgar E. Iglesias <[email protected]>
 S: Maintained
@@ -481,6 +490,18 @@ F: hw/*/allwinner*
 F: include/hw/*/allwinner*
 F: hw/arm/cubieboard.c
 
+AVR Machines
+M: Michael Rolnik <[email protected]>
+R: Sarah Harris <[email protected]>
+S: Maintained
+F: hw/avr/
+F: hw/char/avr_usart.c
+F: include/hw/char/avr_usart.h
+F: hw/timer/avr_timer16.c
+F: include/hw/timer/avr_timer16.h
+F: hw/misc/avr_mask.c
+F: include/hw/misc/avr_mask.h
+
 ARM PrimeCell and CMSDK devices
 M: Peter Maydell <[email protected]>
 L: [email protected]
Signature · 1 line
2.17.2 (Apple Git-113)
Tested-by on v40 6/21
#7 lore raw

Re: [PATCH v40 06/21] target/avr: Add instruction translation - Data Transfer Instructions

In reply to Michael Rolnik
Quoted 14 lines · Michael Rolnik

This includes:
- MOV, MOVW
- LDI, LDS LDX LDY LDZ
- LDDY, LDDZ
- STS, STX STY STZ
- STDY, STDZ
- LPM, LPMX
- ELPM, ELPMX
- SPM, SPMX
- IN, OUT
- PUSH, POP
- XCH
- LAS, LAC LAT

Quoted 1 line
Quoted 4 lines
---
  target/avr/translate.c | 986 +++++++++++++++++++++++++++++++++++++++++
  target/avr/insn.decode |  43 ++
  2 files changed, 1029 insertions(+)
Tested-by on v40 7/21
#9 lore raw

Re: [PATCH v40 07/21] target/avr: Add instruction translation - Bit and Bit-test Instructions

In reply to Michael Rolnik
Quoted 9 lines · Michael Rolnik

This includes:
- LSR, ROR
- ASR
- SWAP
- SBI, CBI
- BST, BLD
- BSET, BCLR

Quoted 4 lines
---
  target/avr/translate.c | 241 +++++++++++++++++++++++++++++++++++++++++
  target/avr/insn.decode |  14 +++
  2 files changed, 255 insertions(+)
#23 lore raw

Re: [PATCH v40 20/21] target/avr: Add Avocado test

In reply to Michael Rolnik

Hi Michael,

Quoted 36 lines · Michael Rolnik

The test is based on
https://github.com/seharris/qemu-avr-tests/tree/master/free-rtos/Demo
demo which. If working correctly, prints 'ABCDEFGHIJKLMNOPQRSTUVWX' out. it also demostrates that timer and IRQ are working

---
  tests/acceptance/machine_avr6.py | 58 ++++++++++++++++++++++++++++++++
  1 file changed, 58 insertions(+)

create mode 100644 tests/acceptance/machine_avr6.py

tests/acceptance/machine_avr6.py new file +16 -0
new file mode 100644
index 0000000000..7a7d8afc29
@@ -0,0 +1,58 @@
+#
+# QEMU AVR
+#
+# Copyright (c) 2019 Michael Rolnik <[email protected]>
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program 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 General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License

tests/acceptance/machine_avr6.py line 26

Quoted 10 lines
tests/acceptance/machine_avr6.py new file +10 -0
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+#
+
+import logging
+import time
+import distutils.spawn
+
+from avocado import skipUnless
+from avocado_qemu import Test
+from avocado.utils import process

Please remove unused imports: logging, distutils.spawn, skipUnless and process.

tests/acceptance/machine_avr6.py line 28

Quoted 2 lines
tests/acceptance/machine_avr6.py new file +2 -0
+
+class AVR6Machine(Test):

It helps others reading this if you document the test purpose here. Besides it makes the pylinter happier. ;)

Quoted 16 lines
tests/acceptance/machine_avr6.py new file +16 -0
+    timeout = 5
+
+    def test_freertos(self):
+        """
+        :avocado: tags=arch:avr
+        :avocado: tags=machine:sample
+        """
+        """
+        https://github.com/seharris/qemu-avr-tests/raw/master/free-rtos/Demo/AVR_ATMega2560_GCC/demo.elf
+        constantly prints out 'ABCDEFGHIJKLMNOPQRSTUVWXABCDEFGHIJKLMNOPQRSTUVWX'
+        """
+        rom_url = 'https://github.com/seharris/qemu-avr-tests'
+        rom_sha1= '36c3e67b8755dcf37e06af6730ef5d477b8ed16d'
+        rom_url += '/raw/'
+        rom_url += rom_sha1
+        rom_url += '/free-rtos/Demo/AVR_ATMega2560_GCC/demo.elf'

tests/acceptance/machine_avr6.py line 54

Quoted 10 lines
tests/acceptance/machine_avr6.py new file +10 -0
+        rom_hash = '7eb521f511ca8f2622e0a3c5e8dd686efbb911d4'
+        rom_path = self.fetch_asset(rom_url, asset_hash=rom_hash)
+
+        self.vm.set_machine('sample')
+        self.vm.add_args('-bios', rom_path)
+        self.vm.add_args('-nographic')
+        self.vm.launch()
+
+        time.sleep(2)
+        self.vm.shutdown()

Do you really need to shutdown the VM here? Because it will be shut down later on avocado_qemu.Test.tearDown() anyway.

tests/acceptance/machine_avr6.py line 58

Quoted 4 lines
tests/acceptance/machine_avr6.py new file +4 -0
+
+        match = 'ABCDEFGHIJKLMNOPQRSTUVWXABCDEFGHIJKLMNOPQRSTUVWX'
+
+        self.assertIn(match, self.vm.get_log())

It is a matter of taste, but I would simply do:

self.assertIn('ABCDEFGHIJKLMNOPQRSTUVWXABCDEFGHIJKLMNOPQRSTUVWX',

                     self.vm.get_log())

Thanks for writing this acceptance test!

- Wainer

#24 lore raw

Re: [PATCH v40 20/21] target/avr: Add Avocado test

Hi Wainer.

thanks for reviewing.

1. when self.vm.shutdown() is not called self.vm.get_log() returns ERROR: argument of type 'NoneType' is not iterable
2. I will remove the unnecessary imports

Thanks,
Michael Rolnik

Quoted 108 lines · Wainer dos Santos Moschetta

Hi Michael,

Quoted 9 lines · Michael Rolnik

The test is based on
https://github.com/seharris/qemu-avr-tests/tree/master/free-rtos/Demo
demo which. If working correctly, prints 'ABCDEFGHIJKLMNOPQRSTUVWX' out. it also demostrates that timer and IRQ are working

Review

Quoted 6 lines
---
  tests/acceptance/machine_avr6.py | 58 ++++++++++++++++++++++++++++++++
  1 file changed, 58 insertions(+)

create mode 100644 tests/acceptance/machine_avr6.py

(quoted diff) +0 -0

b/tests/acceptance/machine_avr6.py

Quoted 21 lines

new file mode 100644
index 0000000000..7a7d8afc29

tests/acceptance/machine_avr6.py new file +16 -0
@@ -0,0 +1,58 @@
+#
+# QEMU AVR
+#
+# Copyright (c) 2019 Michael Rolnik <[email protected]>
+#
+# This program is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program 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 General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License

tests/acceptance/machine_avr6.py line 26

Quoted 10 lines
tests/acceptance/machine_avr6.py new file +10 -0
+# along with this program.  If not, see <http://www.gnu.org/licenses/>.
+#
+
+import logging
+import time
+import distutils.spawn
+
+from avocado import skipUnless
+from avocado_qemu import Test
+from avocado.utils import process

Please remove unused imports: logging, distutils.spawn, skipUnless and process.

tests/acceptance/machine_avr6.py line 28

Quoted 2 lines
tests/acceptance/machine_avr6.py new file +2 -0
+
+class AVR6Machine(Test):

It helps others reading this if you document the test purpose here. Besides it makes the pylinter happier. ;)

tests/acceptance/machine_avr6.py line 47

Quoted 9 lines
tests/acceptance/machine_avr6.py new file +9 -0
+    timeout = 5
+
+    def test_freertos(self):
+        """
+        :avocado: tags=arch:avr
+        :avocado: tags=machine:sample
+        """
+        """
+

https://github.com/seharris/qemu-avr-tests/raw/master/free-rtos/Demo/AVR_ATMega2560_GCC/demo.elf

tests/acceptance/machine_avr6.py line 38

Quoted 1 line
tests/acceptance/machine_avr6.py new file +1 -0
+        constantly prints out

'ABCDEFGHIJKLMNOPQRSTUVWXABCDEFGHIJKLMNOPQRSTUVWX'

Quoted 6 lines
tests/acceptance/machine_avr6.py new file +6 -0
+        """
+        rom_url = 'https://github.com/seharris/qemu-avr-tests'
+        rom_sha1= '36c3e67b8755dcf37e06af6730ef5d477b8ed16d'
+        rom_url += '/raw/'
+        rom_url += rom_sha1
+        rom_url += '/free-rtos/Demo/AVR_ATMega2560_GCC/demo.elf'

tests/acceptance/machine_avr6.py line 54

Quoted 10 lines
tests/acceptance/machine_avr6.py new file +10 -0
+        rom_hash = '7eb521f511ca8f2622e0a3c5e8dd686efbb911d4'
+        rom_path = self.fetch_asset(rom_url, asset_hash=rom_hash)
+
+        self.vm.set_machine('sample')
+        self.vm.add_args('-bios', rom_path)
+        self.vm.add_args('-nographic')
+        self.vm.launch()
+
+        time.sleep(2)
+        self.vm.shutdown()

Do you really need to shutdown the VM here? Because it will be shut down later on avocado_qemu.Test.tearDown() anyway.

tests/acceptance/machine_avr6.py line 58

Quoted 4 lines
tests/acceptance/machine_avr6.py new file +4 -0
+
+        match = 'ABCDEFGHIJKLMNOPQRSTUVWXABCDEFGHIJKLMNOPQRSTUVWX'
+
+        self.assertIn(match, self.vm.get_log())

It is a matter of taste, but I would simply do:

self.assertIn('ABCDEFGHIJKLMNOPQRSTUVWXABCDEFGHIJKLMNOPQRSTUVWX',

                      self.vm.get_log())

Thanks for writing this acceptance test!

- Wainer

Signature · 2 lines
Best Regards,
Michael Rolnik
#26 lore raw
In reply to Michael Rolnik

Hi all,

any news?

Regards,
Michael Rolnik

Quoted 431 lines · Michael Rolnik

This series of patches adds 8bit AVR cores to QEMU.
All instruction, except BREAK/DES/SPM/SPMX, are implemented. Not fully tested yet.
However I was able to execute simple code with functions. e.g fibonacci calculation.
This series of patches include a non real, sample board.
No fuses support yet. PC is set to 0 at reset.

Following are examples of possible usages, assuming program.elf is compiled for AVR cpu
1. Continious non interrupted execution
run qemu-system-avr -kernel program.elf
2. Continious non interrupted execution with serial output into telnet window
run `qemu-system-avr -kernel program.elf -serial
tcp::5678,server,nowait -nographic `
run telent localhost 5678
3. Continious non interrupted execution with serial output into stdout run qemu-system-avr -kernel program.elf -serial stdio
4. Debugging wit GDB debugger
run qemu-system-avr -kernel program.elf -s -S
run avr-gdb program.elf and then within GDB shell `target remote
:1234`
5. Print out executed instructions
run qemu-system-avr -kernel program.elf -d in_asm

the patches include the following
1. just a basic 8bit AVR CPU, without instruction decoding or translation
2. CPU features which allow define the following 8bit AVR cores

     avr1
     avr2 avr25
     avr3 avr31 avr35
     avr4
     avr5 avr51
     avr6
     xmega2 xmega4 xmega5 xmega6 xmega7

3. a definition of sample machine with SRAM, FLASH and CPU which allows to execute simple code
4. encoding for all AVR instructions
5. interrupt handling
6. helpers for IN, OUT, SLEEP, WBR & unsupported instructions
7. a decoder which given an opcode decides what istruction it is
8. translation of AVR instruction into TCG
9. all features together

changes since v3
1. rampD/X/Y/Z registers are encoded as 0x00ff0000 (instead of 0x000000ff) for faster address manipulaton
2. ffs changed to ctz32
3. duplicate code removed at avr_cpu_do_interrupt
4. using andc instead of not + and
5. fixing V flag calculation in varios instructions
6. freeing local variables in PUSH
7. tcg_const_local_i32 -> tcg_const_i32
8. using sextract32 instead of my implementation
9. fixing BLD instruction
10.xor(r) instead of 0xff - r at COM
11.fixing MULS/MULSU not to modify inputs' content
12.using SUB for NEG
13.fixing tcg_gen_qemu_ld/st call in XCH

changes since v4
1. target is now defined as big endian in order to optimize push_ret/pop_ret
2. all style warnings are fixed
3. adding cpu_set/get_sreg functions
4. simplifying gen_goto_tb as there is no real paging
5. env->pc -> env->pc_w
6. making flag dump more compact
7. more spacing
8. renaming CODE/DATA_INDEX -> MMU_CODE/DATA_IDX
9. removing avr_set_feature
10. SPL/SPH set bug fix
11. switching stb_phys to cpu_stb_data
12. cleaning up avr_decode
13. saving sreg, rampD/X/Y/Z, eind in HW format (savevm)
14. saving CPU features (savevm)

changes since v5
1. BLD bug fix
2. decoder generator is added

chages since v6
1. using cpu_get_sreg/cpu_set_sreg in avr_cpu_gdb_read_register/avr_cpu_gdb_write_register
2. configure the target as little endian because otherwise GDB does not work
3. fixing and testing gen_push_ret/gen_pop_ret

changes since v7
1. folding back v6
2. logging at helper_outb and helper_inb are done for non supported yet registers only
3. MAINTAINERS updated

changes since v8
1. removing hw/avr from hw/Makefile.obj as it should not be built for all
2. making linux compilable
3. testing on

    a. Mac, Apple LLVM version 7.0.0
    b. Ubuntu 12.04, gcc 4.9.2
    c. Fedora 23, gcc 5.3.1

4. folding back some patches
5. translation bug fixes for ORI, CPI, XOR instructions
6. propper handling of cpu register writes though memory

changes since v9
1. removing forward declarations of static functions
2. disabling debug prints
3. switching to case range instead of if else if ...
4. LD/ST IN/OUT accessing CPU maintainder registers are not routed to any device
5. commenst about sample board and sample IO device added
6. sample board description is more descriptive now
7. memory_region_allocate_system_memory is used to create RAM
8. now there are helper_fullrd & helper_fullwr when LD/ST try to access registers

changes since v10
1. movig back fullwr & fullrd into the commit where outb and inb were introduced
2. changing tlb_fill function signature
3. adding empty line between functions
4. adding newline on the last line of the file
5. using tb->flags to generae full access ST/LD instructions
6. fixing SBRC bug
7. folding back 10th commit
8. whenever a new file is introduced it's added to Makefile.objs

changes since v11
1. updating to v2.7.0-rc
2. removing assignment to env->fullacc from gen_intermediate_code

changes since v12
1. fixing spacing
2. fixing get/put_segment functions
3. removing target-avr/machine.h file
4. VMSTATE_SINGLE_TEST -> VMSTATE_SINGLE
5. comment spelling
6. removing hw/avr/sample_io.c
7. char const* -> const char*
8. proper ram allocation
9. fixing breakpoint functionality.
10.env1 -> env
11.fixing avr_cpu_gdb_write_register & avr_cpu_gdb_read_register functions 12.any cpu is removed
12.feature bits are not saved into vm state

changes since v13
1. rebasing to v2.7.0-rc1

changes since v14
1. I made self review with git gui tool. (I did not know such a thing exists)
2. removing all double/tripple spaces
3. removing comment reference to SampleIO
4. folding back some changes, so there is not deleted lines in my code
5. moving avr configuration, within configure file, before chris

changes since v15
1. removing IO registers cache from CPU
2. implementing CBI/SBI as read(helper_inb), modify, write(helper_outb)
3. implementing CBIC/SBIC as read(helper_inb), check, branch
4. adding missing tcg_temp_free_i32 for tcg_const_i32

changes since v16
1. removing EXT IO registers knoledge from CPU. These registers are accessible
by LD/ST only. CPU has no interest in them

changes since v17 (by Richard Henderson)
This is Michael's v17, with some adjustments of my own:

1. Fix the whitespace errors reported by "git am",
2. Replace the utf-8 characters with normal ascii,
3. Ditch the separate compilation of translate.c.

I retained the two separate files that could be regenerated from the included cpugen program, but merged in translate-insn.c. Not that it matters, but the code generated is about 3k smaller.

changes since v18

1.  moving target-avr into target/avr
2.  do not call cpu_exec_initfn function from avr_cpu_initfn
3.  call cpu_exec_realizefn avr_cpu_realizefn
4.  do not fail sample machine creation if no rom is suplied
5.  add tcg_gen_exit_tb(0) for BS_BRANCH in gen_intermediate_code
6.  fix a register getters/setters in machine.c
7.  changing QEMU_ARCH_AVR from 1<<17 to 1<<18

changes since v19

1.  use decodetree.py tool to decode instructions
2.  adding USART
3.  adding 16 bit timer peripherals
4.  changing QEMU_ARCH_AVR from 1<<18 to 1<<20
5.  renaming tlb_fill to avr_cpu_tlb_fill

changes since v20

1.  use one CPU naming convention
2.  merging insn16.decode & insn32.decode files
3.  modifying skip next instruction mechanizm
4.  translate BREAK as NOP for now

changes since v21
1. Reorganize bstate.
This will make transition to <exec/translator.h> easier, and fixes a
couple of bugs wrt single stepping
by [email protected]
2. Drop cpc and fix page cross condition.
by [email protected]
3. Refactor checking supported/unsupported instructions
4. Add gdb-xml/avr-cpu.xml

changes since v22
1. Rebase
2. Split long comment

changes since v23

1.  remove avr_cpu_list_compare function
2.  modify avr_cpu_class_by_name function
3.  modify avr_cpu_list_entry function
4.  modify avr_cpu_list function

changes since v24
1. remove AVR_CPU_TYPE_NAME macro

changes since v25

1.  fix patches. every file belong to one patch only
2.  change copyright year from 2016 to 2019
3.  create mask device to emulate prr0/prr1

changes since v26
1. add avocado acceptence test
2. add boot serial test

changes since v27
1. list atmel2560 devices as unimplemented
2. fix sram base/size

changes since v28
1. rebase
2. fix includes & build

changes since v29
1. fixing ownership
2. using 'since' instread of 'added in'

changes since v30
1. rebase

changes since v31

1.  splitting 'Add instruction translation' commit
2.  fixing typo in qapi/machine.json sicne -> since
3.  removing unintended changes in configure file
4.  adding Richard Henderson as a co developer to 'Add instruction

translation - CPU main translation funcions' commit

changes since v32
1. modify cpu_get_sreg to treat sreg C as other flags, except sreg Z

changes since v33
1. ensure flag C is always calculated as one bit
2. calculate flag Z as one bit, without using inverse logic

changes since v34
1. rebase

changes since v35

1.  rebase
2.  use NANOSECONDS_PER_SECOND instead of 1000000000 in avr_timer16.c
3.  split "target/avr: Register AVR support with the rest of QEMU" into
three patches
    1.  "target/avr: Register AVR support with the rest of QEMU"
    2.  "target/avr: Update build system"
    3.  "target/avr: Update MAINTAINERS file"

4. split "target/avr: Add tests" patch into two patches
1. "target/avr: Add Avocado test"
2. "target/avr: Add boot serial test"
5. Add instruction disassembly function
6. change "since 4.2" to "since 5.0"

changes since v36
1. rebase
2. rename

    1.  NO_CPU_REGISTERS    -> NUMBER_OF_CPU_REGISTERS
    2.  NO_IO_REGISTERS     -> NUMBER_OF_IO_REGISTERS
    3.  to_A                -> to_regs_16_31_by_one
    4.  to_B                -> to_regs_16_23_by_one
    5.  to_C                -> to_regs_24_30_by_two
    6.  to_D                -> to_regs_00_30_by_two
3.  add missing licences
4.  add usage example (see above)
5.  ass Sarah Harris <[email protected]> as a reviewer to MAINTAINERS
7.  use git commit sha1 instead of `master`` in avocado test

changes since v37

1.  rebase
2.  moving all data transfer instructions into a separate commit
3.  splitting tests into
    1.  avocado test
    2.  boot serial test
    3.  machine none test
4.  split periperals commit into
    1.  timer
    2.  usart
    3.  mask
5.  add function `avr_flags_to_cpu_type` that translates ELF flags into
core name
6.  add avrtiny, xmega3 cores
7.  remove unnecessary includes from target/avr/helper.c file
8.  remove empty lines from target/avr/translate.c file
9.  set number of interrupt lines to 64

10. determine cpu type by reading ELF flags
11. update license of all new files to be LGPL
12. udpate QEMU documentation

changes since v38

1.  rebase
2.  add examples of AVR emulator invokation into the QEMU doc
3.  reorder instructions to match AVR documentation
4.  fix elf loader function bug introduced by prev version

changes since v39

1.  rename target/avr to hw/avr for hw related commits
2.  spread instruction decoding commit
3.  add frequency parameter to AVR timer
4.  fix offset of some unimplemented registers

Michael Rolnik (21):
target/avr: Add outward facing interfaces and core CPU logic target/avr: Add instruction helpers
target/avr: Add instruction translation - Registers definition target/avr: Add instruction translation - Arithmetic and Logic
Instructions
target/avr: Add instruction translation - Branch Instructions target/avr: Add instruction translation - Data Transfer Instructions target/avr: Add instruction translation - Bit and Bit-test
Instructions
target/avr: Add instruction translation - MCU Control Instructions target/avr: Add instruction translation - CPU main translation
function
target/avr: Add instruction disassembly function
hw/avr: Add limited support for USART peripheral
hw/avr: Add limited support for 16 bit timer peripheral hw/avr: Add dummy mask device
hw/avr: Add example board configuration
target/avr: Add section about AVR into QEMU documentation target/avr: Register AVR support with the rest of QEMU target/avr: Add machine none test
target/avr: Update build system
target/avr: Add boot serial test
target/avr: Add Avocado test
target/avr: Update MAINTAINERS file

 qemu-doc.texi                    |   51 +
 configure                        |    7 +
 default-configs/avr-softmmu.mak  |    5 +
 qapi/machine.json                |    3 +-
 include/disas/dis-asm.h          |   19 +
 include/elf.h                    |    2 +
 include/hw/char/avr_usart.h      |   93 +
 include/hw/elf_ops.h             |    6 +-
 include/hw/loader.h              |    6 +-
 include/hw/misc/avr_mask.h       |   47 +
 include/hw/timer/avr_timer16.h   |   94 +
 include/sysemu/arch_init.h       |    1 +
 target/avr/cpu-param.h           |   37 +
 target/avr/cpu-qom.h             |   54 +
 target/avr/cpu.h                 |  259 +++
 target/avr/helper.h              |   29 +
 arch_init.c                      |    2 +
 hw/avr/sample.c                  |  295 +++
 hw/char/avr_usart.c              |  320 ++++
 hw/core/loader.c                 |   15 +-
 hw/misc/avr_mask.c               |  112 ++
 hw/riscv/boot.c                  |    2 +-
 hw/timer/avr_timer16.c           |  602 ++++++
 target/avr/cpu.c                 |  826 ++++++++
 target/avr/disas.c               |  245 +++
 target/avr/gdbstub.c             |   84 +
 target/avr/helper.c              |  347 ++++
 target/avr/machine.c             |  121 ++
 target/avr/translate.c           | 2997 ++++++++++++++++++++++++++++++
 tests/boot-serial-test.c         |   10 +
 tests/machine-none-test.c        |    1 +
 MAINTAINERS                      |   21 +
 gdb-xml/avr-cpu.xml              |   49 +
 hw/Kconfig                       |    1 +
 hw/avr/Kconfig                   |    6 +
 hw/avr/Makefile.objs             |    1 +
 hw/char/Kconfig                  |    3 +
 hw/char/Makefile.objs            |    1 +
 hw/misc/Kconfig                  |    3 +
 hw/misc/Makefile.objs            |    2 +
 hw/timer/Kconfig                 |    3 +
 hw/timer/Makefile.objs           |    2 +
 target/avr/Makefile.objs         |   34 +
 target/avr/insn.decode           |  183 ++
 tests/Makefile.include           |    2 +
 tests/acceptance/machine_avr6.py |   58 +
 46 files changed, 7049 insertions(+), 12 deletions(-)
 create mode 100644 default-configs/avr-softmmu.mak
 create mode 100644 include/hw/char/avr_usart.h
 create mode 100644 include/hw/misc/avr_mask.h
 create mode 100644 include/hw/timer/avr_timer16.h
 create mode 100644 target/avr/cpu-param.h
 create mode 100644 target/avr/cpu-qom.h
 create mode 100644 target/avr/cpu.h
 create mode 100644 target/avr/helper.h
 create mode 100644 hw/avr/sample.c
 create mode 100644 hw/char/avr_usart.c
 create mode 100644 hw/misc/avr_mask.c
 create mode 100644 hw/timer/avr_timer16.c
 create mode 100644 target/avr/cpu.c
 create mode 100644 target/avr/disas.c
 create mode 100644 target/avr/gdbstub.c
 create mode 100644 target/avr/helper.c
 create mode 100644 target/avr/machine.c
 create mode 100644 target/avr/translate.c
 create mode 100644 gdb-xml/avr-cpu.xml
 create mode 100644 hw/avr/Kconfig
 create mode 100644 hw/avr/Makefile.objs
 create mode 100644 target/avr/Makefile.objs
 create mode 100644 target/avr/insn.decode
 create mode 100644 tests/acceptance/machine_avr6.py

--
2.17.2 (Apple Git-113)

Signature · 2 lines
Best Regards,
Michael Rolnik