From: Athira Rajeev <atrajeev@linux.vnet.ibm.com>
To: Christophe Leroy <christophe.leroy@csgroup.eu>
Cc: "acme@kernel.org" <acme@kernel.org>,
"jolsa@kernel.org" <jolsa@kernel.org>,
"adrian.hunter@intel.com" <adrian.hunter@intel.com>,
"irogers@google.com" <irogers@google.com>,
"namhyung@kernel.org" <namhyung@kernel.org>,
"segher@kernel.crashing.org" <segher@kernel.crashing.org>,
"linux-perf-users@vger.kernel.org"
<linux-perf-users@vger.kernel.org>,
"linuxppc-dev@lists.ozlabs.org" <linuxppc-dev@lists.ozlabs.org>,
"maddy@linux.ibm.com" <maddy@linux.ibm.com>,
"kjain@linux.ibm.com" <kjain@linux.ibm.com>,
"disgoel@linux.vnet.ibm.com" <disgoel@linux.vnet.ibm.com>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
"akanksha@linux.ibm.com" <akanksha@linux.ibm.com>
Subject: Re: [PATCH V2 5/9] tools/perf: Update parameters for reg extract functions to use raw instruction on powerpc
Date: Wed, 22 May 2024 19:36:38 +0530 [thread overview]
Message-ID: <B0117C35-9599-4C86-AC03-5D6B4CCAC1F6@linux.vnet.ibm.com> (raw)
In-Reply-To: <dd7ea638-8a97-4ad1-9e56-5e444bad9252@csgroup.eu>
> On 7 May 2024, at 3:18 PM, Christophe Leroy <christophe.leroy@csgroup.eu> wrote:
>
>
>
> Le 06/05/2024 à 14:19, Athira Rajeev a écrit :
>> Use the raw instruction code and macros to identify memory instructions,
>> extract register fields and also offset. The implementation addresses
>> the D-form, X-form, DS-form instructions. Two main functions are added.
>> New parse function "load_store__parse" as instruction ops parser for
>> memory instructions. Unlink other parser (like mov__parse), this parser
>> fills in only the "raw" field for source/target and new added "mem_ref"
>> field. This is because, here there is no need to parse the disassembled
>> code and arch specific macros will take care of extracting offset and
>> regs which is easier and will be precise.
>>
>> In powerpc, all instructions with a primary opcode from 32 to 63
>> are memory instructions. Update "ins__find" function to have "opcode"
>> also as a parameter. Don't use the "extract_reg_offset", instead use
>> newly added function "get_arch_regs" which will set these fields: reg1,
>> reg2, offset depending of where it is source or target ops.
>
> Yes all instructions with a primary opcode from 32 to 63 are memory
> instructions, but not all memory instructions have opcode 32 to 63.
>
>>
>> Signed-off-by: Athira Rajeev <atrajeev@linux.vnet.ibm.com>
>> ---
>> tools/perf/arch/powerpc/util/dwarf-regs.c | 33 +++++++++++++
>> tools/perf/util/annotate.c | 22 ++++++++-
>> tools/perf/util/disasm.c | 59 +++++++++++++++++++++--
>> tools/perf/util/disasm.h | 4 +-
>> tools/perf/util/include/dwarf-regs.h | 4 +-
>> 5 files changed, 114 insertions(+), 8 deletions(-)
>>
>> diff --git a/tools/perf/arch/powerpc/util/dwarf-regs.c b/tools/perf/arch/powerpc/util/dwarf-regs.c
>> index e60a71fd846e..3121c70dc0d3 100644
>> --- a/tools/perf/arch/powerpc/util/dwarf-regs.c
>> +++ b/tools/perf/arch/powerpc/util/dwarf-regs.c
>> @@ -102,6 +102,9 @@ int regs_query_register_offset(const char *name)
>> #define PPC_OP(op) (((op) >> 26) & 0x3F)
>> #define PPC_RA(a) (((a) >> 16) & 0x1f)
>> #define PPC_RT(t) (((t) >> 21) & 0x1f)
>> +#define PPC_RB(b) (((b) >> 11) & 0x1f)
>> +#define PPC_D(D) ((D) & 0xfffe)
>> +#define PPC_DS(DS) ((DS) & 0xfffc)
>>
>> int get_opcode_insn(unsigned int raw_insn)
>> {
>> @@ -117,3 +120,33 @@ int get_target_reg(unsigned int raw_insn)
>> {
>> return PPC_RT(raw_insn);
>> }
>> +
>> +int get_offset_opcode(int raw_insn __maybe_unused)
>> +{
>> + int opcode = PPC_OP(raw_insn);
>> +
>> + /* DS- form */
>> + if ((opcode == 58) || (opcode == 62))
>
> Can you re-use macros from arch/powerpc/include/asm/ppc-opcode.h ?
>
> #define OP_LD 58
> #define OP_STD 62
Hi Christophe
Sure, I will make these changes
>
>
>> + return PPC_DS(raw_insn);
>> + else
>> + return PPC_D(raw_insn);
>> +}
>> +
>> +/*
>> + * Fills the required fields for op_loc depending on if it
>> + * is a source of target.
>> + * D form: ins RT,D(RA) -> src_reg1 = RA, offset = D, dst_reg1 = RT
>> + * DS form: ins RT,DS(RA) -> src_reg1 = RA, offset = DS, dst_reg1 = RT
>> + * X form: ins RT,RA,RB -> src_reg1 = RA, src_reg2 = RB, dst_reg1 = RT
>> + */
>> +void get_arch_regs(int raw_insn __maybe_unused, int is_source __maybe_unused, struct annotated_op_loc *op_loc __maybe_unused)
>> +{
>> + if (is_source)
>> + op_loc->reg1 = get_source_reg(raw_insn);
>> + else
>> + op_loc->reg1 = get_target_reg(raw_insn);
>> + if (op_loc->multi_regs)
>> + op_loc->reg2 = PPC_RB(raw_insn);
>> + if (op_loc->mem_ref)
>> + op_loc->offset = get_offset_opcode(raw_insn);
>> +}
>
>> diff --git a/tools/perf/util/disasm.c b/tools/perf/util/disasm.c
>> index 85692f73e78f..f41a0fadeab4 100644
>> --- a/tools/perf/util/disasm.c
>> +++ b/tools/perf/util/disasm.c
>> @@ -760,11 +800,20 @@ static void ins__sort(struct arch *arch)
>> qsort(arch->instructions, nmemb, sizeof(struct ins), ins__cmp);
>> }
>>
>> -static struct ins_ops *__ins__find(struct arch *arch, const char *name)
>> +static struct ins_ops *__ins__find(struct arch *arch, const char *name, int opcode)
>> {
>> struct ins *ins;
>> const int nmemb = arch->nr_instructions;
>>
>> + if (arch__is(arch, "powerpc")) {
>> + /*
>> + * Instructions with a primary opcode from 32 to 63
>> + * are memory instructions in powerpc.
>> + */
>> + if ((opcode >= 31) && (opcode <= 63))
>
> Could just be if ((opcode & 0x20)) I guess.
Ok,,
>
> By the way your test is wrong because opcode 31 is not only memory
> instructions, see example below (not exhaustive):
>
> #define OP_31_XOP_TRAP 4 ==> No
> #define OP_31_XOP_LDX 21 ==> Yes
> #define OP_31_XOP_LWZX 23 ==> Yes
> #define OP_31_XOP_LDUX 53 ==> Yes
> #define OP_31_XOP_DCBST 54 ==> No
> #define OP_31_XOP_LWZUX 55 ==> Yes
> #define OP_31_XOP_TRAP_64 68 ==> No
> #define OP_31_XOP_DCBF 86 ==> No
> #define OP_31_XOP_LBZX 87 ==> Yes
> #define OP_31_XOP_STDX 149 ==> Yes
> #define OP_31_XOP_STWX 151 ==> Yes
Thanks for pointing this. I checked through others in this list in arch/powerpc/include/asm/ppc-opcode.h
I will have these filtered in V3
Thanks
Athira
>
>
>
>
>> + return &load_store_ops;
>> + }
>> +
>> if (!arch->sorted_instructions) {
>> ins__sort(arch);
>> arch->sorted_instructions = true;
next prev parent reply other threads:[~2024-05-22 14:07 UTC|newest]
Thread overview: 29+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-05-06 12:18 [PATCH V2 0/9] Add data type profiling support for powerpc Athira Rajeev
2024-05-06 12:18 ` [PATCH V2 1/9] tools/perf: Move the data structures related to register type to header file Athira Rajeev
2024-05-06 12:18 ` [PATCH V2 2/9] tools/perf: Add "update_insn_state" callback function to handle arch specific instruction tracking Athira Rajeev
2024-05-06 12:19 ` [PATCH V2 3/9] tools/perf: Fix a comment about multi_regs in extract_reg_offset function Athira Rajeev
2024-05-07 4:40 ` Namhyung Kim
2024-05-07 14:52 ` Arnaldo Carvalho de Melo
2024-05-06 12:19 ` [PATCH V2 4/9] tools/perf: Add support to capture and parse raw instruction in objdump Athira Rajeev
2024-05-07 4:57 ` Namhyung Kim
2024-05-07 9:35 ` Christophe Leroy
2024-05-09 17:26 ` Athira Rajeev
2024-05-09 20:55 ` Namhyung Kim
2024-05-10 14:26 ` Arnaldo Carvalho de Melo
2024-05-22 13:58 ` Athira Rajeev
2024-05-06 12:19 ` [PATCH V2 5/9] tools/perf: Update parameters for reg extract functions to use raw instruction on powerpc Athira Rajeev
2024-05-07 9:48 ` Christophe Leroy
2024-05-22 14:06 ` Athira Rajeev [this message]
2024-05-06 12:19 ` [PATCH V2 6/9] tools/perf: Update instruction tracking for powerpc Athira Rajeev
2024-05-07 9:52 ` Christophe Leroy
2024-05-23 13:58 ` Athira Rajeev
2024-05-06 12:19 ` [PATCH V2 7/9] tools/perf: Update instruction tracking with add instruction Athira Rajeev
2024-05-07 9:58 ` Christophe Leroy
2024-05-23 13:55 ` Athira Rajeev
2024-05-06 12:19 ` [PATCH V2 8/9] tools/perf: Add support to find global register variables using find_data_type_global_reg Athira Rajeev
2024-05-07 10:03 ` Christophe Leroy
2024-05-24 12:17 ` Athira Rajeev
2024-05-24 12:47 ` Christophe Leroy
2024-05-06 12:19 ` [PATCH V2 9/9] tools/perf: Add support for global_die to capture name of variable in case of register defined variable Athira Rajeev
2024-05-07 4:39 ` [PATCH V2 0/9] Add data type profiling support for powerpc Namhyung Kim
2024-05-13 7:32 ` Athira Rajeev
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=B0117C35-9599-4C86-AC03-5D6B4CCAC1F6@linux.vnet.ibm.com \
--to=atrajeev@linux.vnet.ibm.com \
--cc=acme@kernel.org \
--cc=adrian.hunter@intel.com \
--cc=akanksha@linux.ibm.com \
--cc=christophe.leroy@csgroup.eu \
--cc=disgoel@linux.vnet.ibm.com \
--cc=irogers@google.com \
--cc=jolsa@kernel.org \
--cc=kjain@linux.ibm.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-perf-users@vger.kernel.org \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=maddy@linux.ibm.com \
--cc=namhyung@kernel.org \
--cc=segher@kernel.crashing.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®