aboutsummaryrefslogtreecommitdiff
path: root/target/hexagon
AgeCommit message (Collapse)AuthorFilesLines
2024-02-03include/exec: Implement cpu_mmu_index genericallyRichard Henderson1-9/+0
For user-only mode, use MMU_USER_IDX. For system mode, use CPUClass.mmu_index. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2024-01-29include/qemu: Add TCGCPUOps typedef to typedefs.hRichard Henderson1-1/+1
QEMU coding style recommends using structure typedefs. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2024-01-29include/exec: Use vaddr in DisasContextBase for virtual addressesAnton Johansson1-1/+2
Updates target/ QEMU_LOG macros to use VADDR_PRIx for printing updated DisasContextBase fields. Signed-off-by: Anton Johansson <anjo@rev.ng> Message-Id: <20240119144024.14289-10-anjo@rev.ng> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2024-01-29target: Use vaddr in gen_intermediate_codeAnton Johansson1-1/+1
Makes gen_intermediate_code() signature target agnostic so the function can be called from accel/tcg/translate-all.c without target specifics. Signed-off-by: Anton Johansson <anjo@rev.ng> Message-Id: <20240119144024.14289-9-anjo@rev.ng> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2024-01-21target/hexagon: reduce scope of def_regnum, remove dead assignmentBrian Cain1-3/+1
This is intended to address a coverity finding: CID 1527408. Signed-off-by: Brian Cain <bcain@quicinc.com> Reviewed-by: Matheus Tavares Bernardino <quic_mathbern@quicinc.com> Message-Id: <20240114234453.4114587-1-bcain@quicinc.com>
2024-01-21Hexagon (target/hexagon) Remove old dectree.pyTaylor Simpson6-839/+0
Now that we are using QEMU decodetree.py, remove the old decoder Signed-off-by: Taylor Simpson <ltaylorsimpson@gmail.com> Reviewed-by: Brian Cain <bcain@quicinc.com> Message-Id: <20240115221443.365287-4-ltaylorsimpson@gmail.com> Signed-off-by: Brian Cain <bcain@quicinc.com>
2024-01-21Hexagon (target/hexagon) Use QEMU decodetree (16-bit instructions)Taylor Simpson5-12/+188
Section 10.3 of the Hexagon V73 Programmer's Reference Manual A duplex is encoded as a 32-bit instruction with bits [15:14] set to 00. The sub-instructions that comprise a duplex are encoded as 13-bit fields in the duplex. Create a decoder for each subinstruction class (a, l1, l2, s1, s2). Extend gen_trans_funcs.py to handle all instructions rather than filter by instruction class. There is a g_assert_not_reached() in decode_insns() in decode.c to verify we never try to use the old decoder on 16-bit instructions. Signed-off-by: Taylor Simpson <ltaylorsimpson@gmail.com> Reviewed-by: Brian Cain <bcain@quicinc.com> Message-Id: <20240115221443.365287-3-ltaylorsimpson@gmail.com> Signed-off-by: Brian Cain <bcain@quicinc.com>
2024-01-21Hexagon (target/hexagon) Use QEMU decodetree (32-bit instructions)Taylor Simpson7-14/+439
The Decodetree Specification can be found here https://www.qemu.org/docs/master/devel/decodetree.html Covers all 32-bit instructions, including HVX We generate separate decoders for each instruction class. The reason will be more apparent in the next patch in this series. We add 2 new scripts gen_decodetree.py Generate the input to decodetree.py gen_trans_funcs.py Generate the trans_* functions used by the output of decodetree.py Since the functions generated by decodetree.py take DisasContext * as an argument, we add the argument to a couple of functions that didn't need it previously. We also set the insn field in DisasContext during decode because it is used by the trans_* functions. There is a g_assert_not_reached() in decode_insns() in decode.c to verify we never try to use the old decoder on 32-bit instructions Signed-off-by: Taylor Simpson <ltaylorsimpson@gmail.com> Reviewed-by: Brian Cain <bcain@quicinc.com> Message-Id: <20240115221443.365287-2-ltaylorsimpson@gmail.com> Signed-off-by: Brian Cain <bcain@quicinc.com>
2024-01-21Hexagon (target/hexagon) Remove dead functions from hex_common.pyTaylor Simpson1-51/+0
These functions are no longer used after making the generators object oriented. Signed-off-by: Taylor Simpson <ltaylorsimpson@gmail.com> Reviewed-by: Brian Cain <bcain@quicinc.com> Message-Id: <20231210220712.491494-10-ltaylorsimpson@gmail.com> Signed-off-by: Brian Cain <bcain@quicinc.com>
2024-01-21Hexagon (target/hexagon) Remove unused WRITES_PRED_REG attributeTaylor Simpson2-12/+0
This is the only remaining use of the is_written function. We will remove it in the subsequent commit. Signed-off-by: Taylor Simpson <ltaylorsimpson@gmail.com> Reviewed-by: Brian Cain <bcain@quicinc.com> Message-Id: <20231210220712.491494-9-ltaylorsimpson@gmail.com> Signed-off-by: Brian Cain <bcain@quicinc.com>
2024-01-21Hexagon (target/hexagon) Make generators object oriented - gen_analyze_funcsTaylor Simpson2-157/+157
This patch conflicts with https://lists.gnu.org/archive/html/qemu-devel/2023-11/msg00729.html If that series goes in first, we'll rework this patch and vice versa. Signed-off-by: Taylor Simpson <ltaylorsimpson@gmail.com> Reviewed-by: Brian Cain <bcain@quicinc.com> Message-Id: <20231210220712.491494-8-ltaylorsimpson@gmail.com> Signed-off-by: Brian Cain <bcain@quicinc.com>
2024-01-21Hexagon (target/hexagon) Make generators object oriented - gen_op_regsTaylor Simpson1-2/+4
Reviewed-by: Brian Cain <bcain@quicinc.com> Signed-off-by: Taylor Simpson <ltaylorsimpson@gmail.com> Message-Id: <20231210220712.491494-7-ltaylorsimpson@gmail.com> Signed-off-by: Brian Cain <bcain@quicinc.com>
2024-01-21Hexagon (target/hexagon) Make generators object oriented - gen_idef_parser_funcsTaylor Simpson1-16/+4
Signed-off-by: Taylor Simpson <ltaylorsimpson@gmail.com> Reviewed-by: Brian Cain <bcain@quicinc.com> Message-Id: <20231210220712.491494-6-ltaylorsimpson@gmail.com> Signed-off-by: Brian Cain <bcain@quicinc.com>
2024-01-21Hexagon (target/hexagon) Make generators object oriented - gen_helper_funcsTaylor Simpson2-313/+103
Signed-off-by: Taylor Simpson <ltaylorsimpson@gmail.com> Reviewed-by: Brian Cain <bcain@quicinc.com> Message-Id: <20231210220712.491494-5-ltaylorsimpson@gmail.com> Signed-off-by: Brian Cain <bcain@quicinc.com>
2024-01-21Hexagon (target/hexagon) Make generators object oriented - gen_helper_protosTaylor Simpson2-148/+8
Signed-off-by: Taylor Simpson <ltaylorsimpson@gmail.com> Reviewed-by: Brian Cain <bcain@quicinc.com> Message-Id: <20231210220712.491494-4-ltaylorsimpson@gmail.com> Signed-off-by: Brian Cain <bcain@quicinc.com>
2024-01-21Hexagon (target/hexagon) Make generators object oriented - gen_tcg_funcsTaylor Simpson2-547/+683
The generators are generally a bunch of Python if-then-else statements based on the regtype and regid. Encapsulate regtype/regid into a class hierarchy. Clients lookup the register and invoke methods. This has several advantages for making the code easier to read, understand, and maintain - The class name makes it more clear what the operand does - All the methods for a given type of operand are together - Don't need hex_common.bad_register If a regtype/regid is missing, the lookup in hex_common.get_register will fail - We can remove the functions in hex_common that use regtype/regid (e.g., is_read) This patch creates the class hierarchy in hex_common and converts gen_tcg_funcs.py. The other scripts will be converted in subsequent patches in this series. Signed-off-by: Taylor Simpson <ltaylorsimpson@gmail.com> Reviewed-by: Brian Cain <bcain@quicinc.com> Message-Id: <20231210220712.491494-3-ltaylorsimpson@gmail.com> Signed-off-by: Brian Cain <bcain@quicinc.com>
2024-01-21Hexagon (target/hexagon) Clean up handling of modifier registersTaylor Simpson4-16/+17
Currently, the register number (MuN) for modifier registers is the modifier register number rather than the index into hex_gpr. This patch changes MuN to the hex_gpr index, which is consistent with the handling of control registers. Note that HELPER(fcircadd) needs the CS register corresponding to the modifier register specified in the instruction. We create a TCGv variable "CS" to hold the value to pass to the helper. Reviewed-by: Brian Cain <bcain@quicinc.com> Signed-off-by: Taylor Simpson <ltaylorsimpson@gmail.com> Message-Id: <20231210220712.491494-2-ltaylorsimpson@gmail.com> Signed-off-by: Brian Cain <bcain@quicinc.com>
2024-01-21Hexagon (target/hexagon) Fix shadow variable when idef-parser is offTaylor Simpson1-3/+3
Adding -Werror=shadow=compatible-local causes Hexagon not to build when idef-parser is off. The "label" variable in CHECK_NOSHUF_PRED shadows a variable in the surrounding code. Signed-off-by: Taylor Simpson <ltaylorsimpson@gmail.com> Reviewed-by: Brian Cain <bcain@quicinc.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Message-Id: <20231130183955.54314-1-ltaylorsimpson@gmail.com> Signed-off-by: Brian Cain <bcain@quicinc.com>
2024-01-05target/hexagon: Use generic cpu_list()Gavin Shan2-23/+0
No changes in the output from the following command. [gshan@gshan q]$ ./build/qemu-hexagon -cpu ? Available CPUs: v67 v68 v69 v71 v73 Signed-off-by: Gavin Shan <gshan@redhat.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Message-ID: <20231114235628.534334-12-gshan@redhat.com> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
2024-01-05cpu: Call object_class_dynamic_cast() once in cpu_class_by_name()Philippe Mathieu-Daudé1-3/+1
For all targets, the CPU class returned from CPUClass::class_by_name() and object_class_dynamic_cast(oc, CPU_RESOLVING_TYPE) need to be compatible. Lets apply the check in cpu_class_by_name() for once, instead of having the check in CPUClass::class_by_name() for individual target. Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Gavin Shan <gshan@redhat.com> Reviewed-by: Igor Mammedov <imammedo@redhat.com> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Gavin Shan <gshan@redhat.com> Message-ID: <20231114235628.534334-4-gshan@redhat.com>
2023-11-28target/hexagon/idef-parser/prepare: use env to invoke bashSamuel Tardieu1-1/+1
This file is the only one involved in the compilation process which still uses the /bin/bash path. Signed-off-by: Samuel Tardieu <sam@rfc1149.net> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Anton Johansson <anjo@rev.ng> Message-ID: <20231123211506.636533-1-sam@rfc1149.net> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
2023-11-07hw/cpu: Call object_class_is_abstract() once in cpu_class_by_name()Philippe Mathieu-Daudé1-2/+1
Let CPUClass::class_by_name() handlers to return abstract classes, and filter them once in the public cpu_class_by_name() method. Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-Id: <20230908112235.75914-3-philmd@linaro.org>
2023-11-07target: Move ArchCPUClass definition to 'cpu.h'Philippe Mathieu-Daudé1-1/+0
The OBJECT_DECLARE_CPU_TYPE() macro forward-declares each ArchCPUClass type. These forward declarations are sufficient for code in hw/ to use the QOM definitions. No need to expose these structure definitions. Keep each local to their target/ by moving them to the corresponding "cpu.h" header. Suggested-by: Richard Henderson <richard.henderson@linaro.org> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-Id: <20231013140116.255-13-philmd@linaro.org>
2023-11-07target/hexagon: Declare QOM definitions in 'cpu-qom.h'Philippe Mathieu-Daudé2-14/+29
"target/foo/cpu.h" contains the target specific declarations. A heterogeneous setup need to access target agnostic declarations (at least the QOM ones, to instantiate the objects). Our convention is to add such target agnostic QOM declarations in the "target/foo/cpu-qom.h" header. Add a comment clarifying that in the header. Extract QOM definitions from "cpu.h" to "cpu-qom.h". Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Brian Cain <bcain@quicinc.com> Message-Id: <20231013140116.255-8-philmd@linaro.org>
2023-11-07target: Unify QOM stylePhilippe Mathieu-Daudé1-4/+1
Enforce the style described by commit 067109a11c ("docs/devel: mention the spacing requirement for QOM"): The first declaration of a storage or class structure should always be the parent and leave a visual space between that declaration and the new code. It is also useful to separate backing for properties (options driven by the user) and internal state to make navigation easier. Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Reviewed-by: Zhao Liu <zhao1.liu@intel.com> Message-Id: <20231013140116.255-2-philmd@linaro.org>
2023-10-18target/hexagon: fix some occurrences of -Wshadow=localBrian Cain4-15/+12
Of the changes in this commit, the changes in `HELPER(commit_hvx_stores)()` are less obvious. They are required because of some macro invocations like SCATTER_OP_WRITE_TO_MEM(). e.g.: In file included from ../target/hexagon/op_helper.c:31: ../target/hexagon/mmvec/macros.h:205:18: error: declaration of ‘i’ shadows a previous local [-Werror=shadow=compatible-local] 205 | for (int i = 0; i < sizeof(MMVector); i += sizeof(TYPE)) { \ | ^ ../target/hexagon/op_helper.c:157:17: note: in expansion of macro ‘SCATTER_OP_WRITE_TO_MEM’ 157 | SCATTER_OP_WRITE_TO_MEM(uint16_t); | ^~~~~~~~~~~~~~~~~~~~~~~ ../target/hexagon/op_helper.c:135:9: note: shadowed declaration is here 135 | int i; | ^ In file included from ../target/hexagon/op_helper.c:31: ../target/hexagon/mmvec/macros.h:204:19: error: declaration of ‘ra’ shadows a previous local [-Werror=shadow=compatible-local] 204 | uintptr_t ra = GETPC(); \ | ^~ ../target/hexagon/op_helper.c:160:17: note: in expansion of macro ‘SCATTER_OP_WRITE_TO_MEM’ 160 | SCATTER_OP_WRITE_TO_MEM(uint32_t); | ^~~~~~~~~~~~~~~~~~~~~~~ ../target/hexagon/op_helper.c:134:15: note: shadowed declaration is here 134 | uintptr_t ra = GETPC(); | ^~ Reviewed-by: Matheus Tavares Bernardino <quic_mathbern@quicinc.com> Signed-off-by: Brian Cain <bcain@quicinc.com> Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Message-Id: <20231008220945.983643-3-bcain@quicinc.com>
2023-10-18target/hexagon: move GETPC() calls to top level helpersMatheus Tavares Bernardino3-65/+38
As docs/devel/loads-stores.rst states: ``GETPC()`` should be used with great care: calling it in other functions that are *not* the top level ``HELPER(foo)`` will cause unexpected behavior. Instead, the value of ``GETPC()`` should be read from the helper and passed if needed to the functions that the helper calls. Let's fix the GETPC() usage in Hexagon, making sure it's always called from top level helpers and passed down to the places where it's needed. There are a few snippets where that is not currently the case: - probe_store(), which is only called from two helpers, so it's easy to move GETPC() up. - mem_load*() functions, which are also called directly from helpers, but through the MEM_LOAD*() set of macros. Note that this are only used when compiling with --disable-hexagon-idef-parser. In this case, we also take this opportunity to simplify the code, unifying the mem_load*() functions. - HELPER(probe_hvx_stores), when called from another helper, ends up using its own GETPC() expansion instead of the top level caller. Signed-off-by: Matheus Tavares Bernardino <quic_mathbern@quicinc.com> Reviewed-by: Taylor Simpson <ltaylorsimpson@gmail.com> Message-Id: <2c74c3696946edba7cc5b2942cf296a5af532052.1689070412.git.quic_mathbern@quicinc.com>-ne Reviewed-by: Brian Cain <bcain@quicinc.com> Signed-off-by: Brian Cain <bcain@quicinc.com> Message-Id: <20231008220945.983643-2-bcain@quicinc.com>
2023-10-17target/hexagon: avoid invalid escape in Python stringPaolo Bonzini1-2/+2
This is an error in Python 3.12; fix it by using a raw string literal. Cc: qemu-stable@nongnu.org Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2023-10-04accel/tcg: Remove cpu_set_cpustate_pointersRichard Henderson1-3/+0
This function is now empty, so remove it. In the case of m68k and tricore, this empties the class instance initfn, so remove those as well. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2023-10-04accel/tcg: Replace CPUState.env_ptr with cpu_env()Richard Henderson1-2/+2
Reviewed-by: Anton Johansson <anjo@rev.ng> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2023-10-03tcg: Rename cpu_env to tcg_envRichard Henderson8-130/+130
Allow the name 'cpu_env' to be used for something else. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2023-10-03accel/tcg: Move CPUNegativeOffsetState into CPUStateRichard Henderson1-1/+1
Retain the separate structure to emphasize its importance. Enforce CPUArchState always follows CPUState without padding. Reviewed-by: Anton Johansson <anjo@rev.ng> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2023-10-03target/*: Add instance_align to all cpu base classesRichard Henderson1-0/+1
The omission of alignment has technically been wrong since 269bd5d8f61, where QEMU_ALIGNED was added to CPUTLBDescFast. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2023-09-08hexagon: spelling fixesMichael Tokarev8-17/+17
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru> Reviewed-by: Brian Cain <bcain@quicinc.com>
2023-08-31target/translate: Remove unnecessary 'exec/cpu_ldst.h' headerPhilippe Mathieu-Daudé1-1/+0
All these files only access the translator_ld/st API declared in "exec/translator.h". The CPU ld/st API from declared in "exec/cpu_ldst.h" is not used, remove it. Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-Id: <20230828221314.18435-5-philmd@linaro.org>
2023-06-26target: Widen pc/cs_base in cpu_get_tb_cpu_stateAnton Johansson1-2/+2
Signed-off-by: Anton Johansson <anjo@rev.ng> Reviewed-by: Richard Henderson <richard.henderson@linaro.org> Message-Id: <20230621135633.1649-4-anjo@rev.ng> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2023-06-05target/*: Add missing includes of exec/translation-block.hRichard Henderson1-0/+1
This had been pulled in via exec/exec-all.h, via exec/translator.h, but the include of exec-all.h will be removed. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2023-06-05tcg: Split helper-proto.hRichard Henderson1-0/+1
Create helper-proto-common.h without the target specific portion. Use that in tcg-op-common.h. Include helper-proto.h in target/arm and target/hexagon before helper-info.c.inc; all other targets are already correct in this regard. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2023-06-05tcg: Pass TCGHelperInfo to tcg_gen_callNRichard Henderson1-0/+4
In preparation for compiling tcg/ only once, eliminate the all_helpers array. Instantiate the info structs for the generic helpers in accel/tcg/, and the structs for the target-specific helpers in each translate.c. Since we don't see all of the info structs at startup, initialize at first use, using g_once_init_* to make sure we don't race while doing so. Reviewed-by: Anton Johansson <anjo@rev.ng> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2023-06-05target/hexagon: Include helper-gen.h where neededRichard Henderson3-1/+4
This had been included via tcg-op-common.h via tcg-op.h, but that is going away. In idef-parser.y, shuffle some tcg related includes into a more logical order. Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org> Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
2023-05-26Hexagon: fix outdated `hex_new_*` commentsMatheus Tavares Bernardino2-15/+13
Some code comments refer to hex_new_value and hex_new_pred_value, which have been transferred to DisasContext and, in the case of hex_new_value, should now be accessed through get_result_gpr(). In order to fix this outdated comments and also avoid having to tweak them whenever we make a variable name change in the future, let's replace them with pseudocode. Suggested-by: Taylor Simpson <tsimpson@quicinc.com> Signed-off-by: Matheus Tavares Bernardino <quic_mathbern@quicinc.com> Reviewed-by: Taylor Simpson <tsimpson@quicinc.com> Signed-off-by: Taylor Simpson <tsimpson@quicinc.com> Message-Id: <8e1689e28dd7b1318369b55127cf47b82ab75921.1684939078.git.quic_mathbern@quicinc.com>
2023-05-26target/hexagon/*.py: clean up used 'toss' and 'numregs' varsMatheus Tavares Bernardino7-70/+70
Many Hexagon python scripts call hex_common.get_tagregs(), but only one call site use the full reg structure given by this function. To make the code cleaner, let's make get_tagregs() filter out the unused fields (i.e. 'toss' and 'numregs'), properly removed the unused variables at the call sites. The hex_common.bad_register() function is also adjusted to work exclusively with 'regtype' and 'regid' args. For the single call site that does use toss/numregs, we provide an optional parameter to get_tagregs() which will restore the old full behavior. Suggested-by: Taylor Simpson <tsimpson@quicinc.com> Signed-off-by: Matheus Tavares Bernardino <quic_mathbern@quicinc.com> Reviewed-by: Taylor Simpson <tsimpson@quicinc.com> Tested-by: Taylor Simpson <tsimpson@quicinc.com> Signed-off-by: Taylor Simpson <tsimpson@quicinc.com> Message-Id: <3ffd4ccb972879f57f499705c624e8eaba7f8b52.1684939078.git.quic_mathbern@quicinc.com>
2023-05-26Hexagon (target/hexagon) Fix assignment to tmp registersMarco Liebel1-4/+4
The order in which instructions are generated by gen_insn() influences assignment to tmp registers. During generation, tmp instructions (e.g. generate_V6_vassign_tmp) use vreg_src_off() to determine what kind of register to use as source. If some instruction (e.g. generate_V6_vmpyowh_64_acc) uses a tmp register but is generated prior to the corresponding tmp instruction, the vregs_updated_tmp bit map isn't updated in time. Exmple: { v14.tmp = v16; v25 = v14 } This works properly because generate_V6_vassign_tmp is generated before generate_V6_vassign and the bit map is updated. { v15:14.tmp = vcombine(v21, v16); v25:24 += vmpyo(v18.w,v14.h) } This does not work properly because vmpyo is generated before vcombine and therefore the bit map does not yet know that there's a tmp register. The parentheses in the decoding function were in the wrong place. Moving them to the correct location makes shuffling of .tmp vector registers work as expected. Signed-off-by: Marco Liebel <quic_mliebel@quicinc.com> Reviewed-by: Taylor Simpson <tsimpson@quicinc.com> Tested-by: Taylor Simpson <tsimpson@quicinc.com> Signed-off-by: Taylor Simpson <tsimpson@quicinc.com> Reviewed-by: Brian Cain <bcain@quicinc.com> Message-Id: <20230522174708.464197-1-quic_mliebel@quicinc.com>
2023-05-18Hexagon (gdbstub): add HVX supportTaylor Simpson3-0/+76
Signed-off-by: Taylor Simpson <tsimpson@quicinc.com> Co-authored-by: Brian Cain <bcain@quicinc.com> Signed-off-by: Brian Cain <bcain@quicinc.com> Co-authored-by: Matheus Tavares Bernardino <quic_mathbern@quicinc.com> Signed-off-by: Matheus Tavares Bernardino <quic_mathbern@quicinc.com> Reviewed-by: Brian Cain <bcain@quicinc.com> Message-Id: <17cb32f34d469f705c3cc066a3583935352ee048.1683214375.git.quic_mathbern@quicinc.com>
2023-05-18Hexagon (gdbstub): fix p3:0 read and write via stubBrian Cain1-0/+16
Signed-off-by: Brian Cain <bcain@quicinc.com> Co-authored-by: Sid Manning <sidneym@quicinc.com> Signed-off-by: Sid Manning <sidneym@quicinc.com> Co-authored-by: Matheus Tavares Bernardino <quic_mathbern@quicinc.com> Signed-off-by: Matheus Tavares Bernardino <quic_mathbern@quicinc.com> Reviewed-by: Taylor Simpson <tsimpson@quicinc.com> Signed-off-by: Taylor Simpson <tsimpson@quicinc.com> Message-Id: <32e7de567cdae184a6781644454bbb19916c955b.1683214375.git.quic_mathbern@quicinc.com>
2023-05-18Hexagon: add core gdbstub xml data for LLDBMatheus Tavares Bernardino1-1/+2
Signed-off-by: Matheus Tavares Bernardino <quic_mathbern@quicinc.com> Reviewed-by: Taylor Simpson <tsimpson@quicinc.com> Signed-off-by: Taylor Simpson <tsimpson@quicinc.com> Message-Id: <d25a3a79334d81f0e1ecfb438b6ee82585d02dc4.1683214375.git.quic_mathbern@quicinc.com>
2023-05-18Hexagon (decode): look for pkts with multiple insns at the same slotMatheus Tavares Bernardino1-3/+27
Each slot in a packet can be assigned to at most one instruction. Although the assembler generally ought to enforce this rule, we better be safe than sorry and also do some check to properly throw an "invalid packet" exception on wrong slot assignments. This should also make it easier to debug possible future errors caused by missing updates to `find_iclass_slots()` rules in target/hexagon/iclass.c. Co-authored-by: Taylor Simpson <tsimpson@quicinc.com> Signed-off-by: Taylor Simpson <tsimpson@quicinc.com> Signed-off-by: Matheus Tavares Bernardino <quic_mathbern@quicinc.com> Reviewed-by: Taylor Simpson <tsimpson@quicinc.com> Tested-by: Taylor Simpson <tsimpson@quicinc.com> Message-Id: <f8b829443523568823d062adf8bf6659bc6d4a3f.1683552984.git.quic_mathbern@quicinc.com>
2023-05-18Hexagon (iclass): update J4_hintjumpr slot constraintsMatheus Tavares Bernardino1-2/+4
The Hexagon PRM says that "The assembler automatically encodes instructions in the packet in the proper order. In the binary encoding of a packet, the instructions must be ordered from Slot 3 down to Slot 0." Prior to the architecture version v73, the slot constraints from instruction "hintjr" only allowed it to be executed at slot 2. With that in mind, consider the packet: { hintjr(r0) nop nop if (!p0) memd(r1+#0) = r1:0 } To satisfy the ordering rule quoted from the PRM, the assembler would, thus, move one of the nops to the first position, so that it can be assigned to slot 3 and the subsequent hintjr to slot 2. However, since v73, hintjr can be executed at either slot 2 or 3. So there is no need to reorder that packet and the assembler will encode it as is. When QEMU tries to execute it, however, we end up hitting a "misaliged store" exception because both the store and the hintjr will be assigned to store 0, and some functions like `slot_is_predicated()` expect the decode machinery to assign only one instruction per slot. In particular, the mentioned function will traverse the packet until it finds the first instruction at the desired slot which, for slot 0, will be hintjr. Since hintjr is not predicated, the result is that we try to execute the store regardless of the predicate. And because the predicate is false, we had not previously loaded hex_store_addr[0] or hex_store_width[0]. As a result, the store will decide de width based on trash memory, causing it to be misaligned. Update the slot constraints for hintjr so that QEMU can properly handle such encodings. Note: to avoid similar-but-not-identical issues in the future, we should look for multiple instructions at the same slot during decoding time and throw an invalid packet exception. That will be done in the subsequent commit. Signed-off-by: Matheus Tavares Bernardino <quic_mathbern@quicinc.com> Reviewed-by: Taylor Simpson <tsimpson@quicinc.com> Signed-off-by: Taylor Simpson <tsimpson@quicinc.com> Message-Id: <0fcd8293642c6324119fbbab44741164bcbd04fb.1673616964.git.quic_mathbern@quicinc.com>
2023-05-18Hexagon: list available CPUs with `-cpu help`Matheus Tavares Bernardino2-0/+23
Currently, qemu-hexagon only models the v67 cpu. Nonetheless if we try to get this information with `-cpu help`, qemu just exists with an error code and no output. Let's correct that. The code is basically a copy from target/alpha/cpu.h, but we strip the "-hexagon-cpu" suffix before printing. This is to avoid confusing situations like the following: $ qemu-hexagon -cpu help Available CPUs: v67-hexagon-cpu $ qemu-hexagon -cpu v67-hexagon-cpu ./prog qemu-hexagon: unable to find CPU model 'v67-hexagon-cpu' Signed-off-by: Matheus Tavares Bernardino <quic_mathbern@quicinc.com> Signed-off-by: Taylor Simpson <tsimpson@quicinc.com> Tested-by: Taylor Simpson <tsimpson@quicinc.com> Reviewed-by: Taylor Simpson <tsimpson@quicinc.com> Message-Id: <b946e17c7e17eed9095700b54c5ead36e5d55dfa.1683225804.git.quic_mathbern@quicinc.com>
2023-05-18Hexagon (target/hexagon/*.py): raise exception on reg parsing errorMatheus Tavares Bernardino6-63/+66
Currently, the python scripts used for the hexagon building will not abort the compilation when there is an error parsing a register. Let's make the compilation properly fail in such cases by rasing an exception instead of just printing a warning message, which might get lost in the output. This patch was generated with: git grep -l "Bad register" *hexagon* | \ xargs sed -i "" -e 's/print("Bad register parse: "[, ]*\([^)]*\))/hex_common.bad_register(\1)/g' Plus the bad_register() helper added to hex_common.py. Signed-off-by: Matheus Tavares Bernardino <quic_mathbern@quicinc.com> Reviewed-by: Anton Johansson <anjo@rev.ng> Tested-by: Taylor Simpson <tsimpson@quicinc.com> Reviewed-by: Taylor Simpson <tsimpson@quicinc.com> Signed-off-by: Taylor Simpson <tsimpson@quicinc.com> Message-Id: <1f5dbd92f68fdd89e2647e4ba527a2c32cf0f070.1683217043.git.quic_mathbern@quicinc.com>
generated by cgit v1.1 at 2025-08-10 00:35:16 +0000