public inbox for binutils@sourceware.org
 help / color / mirror / Atom feed
From: Hau Hsu <hau.hsu@sifive.com>
To: Tsukasa OI <research_trasio@irq.a4lg.com>
Cc: Binutils <binutils@sourceware.org>
Subject: Re: [PATCH v2] RISC-V: Add --march=help
Date: Tue, 30 Jan 2024 20:32:17 +0800	[thread overview]
Message-ID: <F515EF28-DCED-430F-83B2-78662411B887@sifive.com> (raw)
In-Reply-To: <cd0fe966-c47e-45a6-9396-748cbdfc2e57@irq.a4lg.com>

[-- Attachment #1: Type: text/plain, Size: 6487 bytes --]

Hi Tsukasa,

Indeed this idea is from Clang.
Let me see if I can have another implementation based on your suggestions.
Thanks for your quick reply :).

Best,

Hau Hsu
Software Engineer
hau.hsu@sifive.com




> Tsukasa OI <research_trasio@irq.a4lg.com> 於 2024年1月30日 下午4:43 寫道:
> 
> Hi Hau,
> 
> I'm not yet ready to come back to Binutils development (because I
> haven't finished writing my research paper) but I'd like to comment.
> 
> While your idea seems great at first glance, I would never like to see
> "help" handling in the riscv_parse_subset function because many other
> functions call it (some calls this function with an argument retrieved
> from an assembly / object file and your design seems vulnerable from
> adversarial files e.g. when disassembling files with unknown origin).
> 
> Besides that, I think that's not a bad idea (quite rare to see in the
> GNU toolchain, though).
> 
> IMHO, the right place to put "help" handling (at least, its entry point)
> is the md_parse_option function in gas/config/tc-riscv.c.  Exiting
> successfully from here is rare but at least cris and kvx does so for
> help-like options.
> 
> Thanks,
> Tsukasa
> 
> On 2024/01/30 15:36, Hau Hsu wrote:
>> Use --march=help to print all supported extensions and versions.
>> 
>> This patch assumes that the supported extensions with the same versions
>> are listed together.
>> 
>> For example
>> static struct riscv_supported_ext riscv_supported_std_ext[] =
>> {
>>  ...
>>  {"i",         ISA_SPEC_CLASS_20191213,        2, 1, 0 },
>>  {"i",         ISA_SPEC_CLASS_20190608,        2, 1, 0 },
>>  {"i",         ISA_SPEC_CLASS_2P2,             2, 0, 0 },
>>  ...
>> };
>> 
>> For "i" extension, 2.1.0 with different spec class are listed together.
>> This patch records the previous printed extension and version.  If the
>> current extension and version are the same as the previous one, skip
>> printing.
>> 
>> Here is part of the output of `as -march=help`:
>> 
>> All available -march extensions for RISC-V:
>>        e                                       1.9
>>        i                                       2.1, 2.0
>>        m                                       2.0
>>        a                                       2.1, 2.0
>>        f                                       2.2, 2.0
>>        d                                       2.2, 2.0
>>        q                                       2.2, 2.0
>>        c                                       2.0
>>        v                                       1.0
>>        h                                       1.0
>>        zicbom                                  1.0
>>        zicbop                                  1.0
>> 
>> Signed-off-by: Hau Hsu <hau.hsu@sifive.com>
>> ---
>> bfd/ChangeLog     |  9 ++++++
>> bfd/elfxx-riscv.c | 72 +++++++++++++++++++++++++++++++++++++++++++++++
>> 2 files changed, 81 insertions(+)
>> 
>> diff --git a/bfd/ChangeLog b/bfd/ChangeLog
>> index 97d0c585a56..0807a2e0bcb 100644
>> --- a/bfd/ChangeLog
>> +++ b/bfd/ChangeLog
>> @@ -1,3 +1,12 @@
>> +2024-01-17 Hau Hsu  <hau.hsu@sifive.com>
>> +
>> +	* elfxx-riscv.c (riscv_parse_subset): Parse 'help' keyword to print
>> +	  available extension and versions.
>> +	(riscv_print_extensions): New function.
>> +	(riscv_same_extension_version): New function.
>> +	(riscv_same_extension_diff_version): New function.
>> +	(riscv_valid_ext): New function.
>> +
>> 2024-01-15  Nick Clifton  <nickc@redhat.com>
>> 
>> 	* 2.42 branch point.
>> diff --git a/bfd/elfxx-riscv.c b/bfd/elfxx-riscv.c
>> index 9a121b47121..bb394a589b1 100644
>> --- a/bfd/elfxx-riscv.c
>> +++ b/bfd/elfxx-riscv.c
>> @@ -2051,6 +2051,71 @@ riscv_set_default_arch (riscv_parse_subset_t *rps)
>>     }
>> }
>> 
>> +static
>> +bool riscv_same_extension_version(
>> +  const struct riscv_supported_ext* ext1,
>> +  const struct riscv_supported_ext* ext2)
>> +{
>> +  return (strcmp(ext1->name, ext2->name) == 0
>> +          && ext1->major_version == ext2->major_version
>> +          && ext1->minor_version == ext2->minor_version);
>> +}
>> +
>> +static
>> +bool riscv_same_extension_diff_version(
>> +  const struct riscv_supported_ext* ext1,
>> +  const struct riscv_supported_ext* ext2)
>> +{
>> +  return (strcmp(ext1->name, ext2->name) == 0
>> +          && !(ext1->major_version == ext2->major_version
>> +               && ext1->minor_version == ext2->minor_version));
>> +}
>> +
>> +static
>> +bool riscv_valid_ext(const struct riscv_supported_ext *ext)
>> +{
>> +  return (ext->isa_spec_class != ISA_SPEC_CLASS_NONE
>> +          && ext->major_version != RISCV_UNKNOWN_VERSION
>> +          && ext->minor_version != RISCV_UNKNOWN_VERSION);
>> +}
>> +
>> +static
>> +void riscv_print_extensions(void)
>> +{
>> +  /* Record the previous pritned extension.
>> +     Print the current one if they are not the same.  */
>> +  const struct riscv_supported_ext *cur = NULL, *prev = NULL;
>> +
>> +  int i, j;
>> +  printf ("All available -march extensions for RISC-V:");
>> +  for (i = 0; riscv_all_supported_ext[i] != NULL; i++)
>> +    {
>> +      const struct riscv_supported_ext *exts = riscv_all_supported_ext[i];
>> +      prev = NULL;
>> +      for (j = 0; exts[j].name != NULL; j++)
>> +        {
>> +          cur = &exts[j];
>> +          if (!riscv_valid_ext (cur))
>> +            continue;
>> +
>> +          if (prev && riscv_same_extension_version (prev, cur))
>> +            continue;
>> +
>> +          if (!prev || !riscv_same_extension_diff_version (prev, cur))
>> +            {
>> +              printf("\n\t%-40s%d.%d", cur->name, cur->major_version, cur->minor_version);
>> +              prev = &exts[j];
>> +            }
>> +          else
>> +            {
>> +              printf(", %d.%d", cur->major_version, cur->minor_version);
>> +              prev = &exts[j];
>> +            }
>> +        }
>> +    }
>> +  printf ("\n");
>> +}
>> +
>> /* Function for parsing ISA string.
>> 
>>    Return Value:
>> @@ -2089,6 +2154,13 @@ riscv_parse_subset (riscv_parse_subset_t *rps,
>>     }
>> 
>>   p = arch;
>> +  /* List all avaiable archs. */
>> +  if (strcmp (p, "help") == 0)
>> +    {
>> +      riscv_print_extensions();
>> +      exit (EXIT_SUCCESS);
>> +    }
>> +
>>   if (startswith (p, "rv32"))
>>     {
>>       *rps->xlen = 32;


      reply	other threads:[~2024-01-30 12:32 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-01-30  6:36 Hau Hsu
2024-01-30  8:43 ` Tsukasa OI
2024-01-30 12:32   ` Hau Hsu [this message]

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=F515EF28-DCED-430F-83B2-78662411B887@sifive.com \
    --to=hau.hsu@sifive.com \
    --cc=binutils@sourceware.org \
    --cc=research_trasio@irq.a4lg.com \
    /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;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).