public inbox for binutils@sourceware.org
 help / color / mirror / Atom feed
From: Tsukasa OI <research_trasio@irq.a4lg.com>
To: Nelson Chu <nelson@rivosinc.com>,
	Kito Cheng <kito.cheng@sifive.com>,
	Palmer Dabbelt <palmer@dabbelt.com>
Cc: binutils@sourceware.org
Subject: Re: [PATCH 0/2] RISC-V: Use only real extension names on testsuite
Date: Tue, 25 Oct 2022 23:54:10 +0900	[thread overview]
Message-ID: <131a0f94-91c3-c850-130b-598ee0afb998@irq.a4lg.com> (raw)
In-Reply-To: <cover.1663919347.git.research_trasio@irq.a4lg.com>

Ping.

On 2022/09/23 16:49, Tsukasa OI wrote:
> Hi,
> 
> Most of the RISC-V GAS testsuite uses real extension names but there are a
> few counterexamples.  I will talk about following tests:
> 
> -   b-ext{,-64} for ("Zb*" except "Zbk*")
> -   k-ext{,-64} for ("Zk*" and "Zbk*")
> 
> They:
> -   provide unnecessary confusion (that major "B" and "K" extensions exist),
>     for example shown in:
>     <https://sourceware.org/bugzilla/show_bug.cgi?id=29341> and
> -   do not test separate extension
>     (they are multiple subextensions).
> 
> To avoid confusion and test individual extensions separately, this patchset
> provides per-extension testcases and removes b-ext{,-64} and k-ext{,-64}.
> 
> There's another effect.  Splitting "B" testcases will clarify that "zext.h"
> and "zext.w" (seemingly similar) instructions belong to different extension:
> 
> -   "zext.h" (RV32/64_Zbb)
> -   "zext.w" (RV64_Zba)  <-- not Zbb!
> 
> [Changes]
> 
> -   Split b-ext and b-ext-64 to:
>     -   zba{,-64}
>     -   zbb{,-64}
>     -   zbc
>     -   zbs{,-64}
> -   Reuse existing "zk*" and "zbk*" testcases because they are fine as is,
> -   Add new testcases for separate extension:
>     -   zkr (along with 'Zicsr', to test "seed" CSR accessibility)
>     -   zkt (just to test extension support)
> -   Remove:
>     -   b-ext{,-64}
>     -   k-ext{,-64}
> 
> For vector-insn*, I'll leave this for now but it might possible to split
> (e.g. to three testcases for two instruction classes and INSN_V_EEW64).
> Still, I think this is not as important as b-ext{,-64} and k-ext{,-64} and
> I don't complain about using vector-insn* names themselves.
> 
> Thanks,
> Tsukasa
> 
> 
> 
> 
> Tsukasa OI (2):
>   RISC-V: Reorganize "K"-related subextension tests
>   RISC-V: Reorganize "B"-related subextension tests
> 
>  gas/testsuite/gas/riscv/b-ext-64.d | 72 ------------------------------
>  gas/testsuite/gas/riscv/b-ext-64.s | 64 --------------------------
>  gas/testsuite/gas/riscv/b-ext.d    | 51 ---------------------
>  gas/testsuite/gas/riscv/b-ext.s    | 43 ------------------
>  gas/testsuite/gas/riscv/k-ext-64.d | 47 -------------------
>  gas/testsuite/gas/riscv/k-ext-64.s | 38 ----------------
>  gas/testsuite/gas/riscv/k-ext.d    | 44 ------------------
>  gas/testsuite/gas/riscv/k-ext.s    | 35 ---------------
>  gas/testsuite/gas/riscv/zba-64.d   | 19 ++++++++
>  gas/testsuite/gas/riscv/zba.d      | 12 +++++
>  gas/testsuite/gas/riscv/zba.s      | 14 ++++++
>  gas/testsuite/gas/riscv/zbb-64.d   | 36 +++++++++++++++
>  gas/testsuite/gas/riscv/zbb.d      | 28 ++++++++++++
>  gas/testsuite/gas/riscv/zbb.s      | 34 ++++++++++++++
>  gas/testsuite/gas/riscv/zbc.d      | 12 +++++
>  gas/testsuite/gas/riscv/zbc.s      |  4 ++
>  gas/testsuite/gas/riscv/zbs-64.d   | 34 ++++++++++++++
>  gas/testsuite/gas/riscv/zbs.d      | 25 +++++++++++
>  gas/testsuite/gas/riscv/zbs.s      | 32 +++++++++++++
>  gas/testsuite/gas/riscv/zkr.d      | 10 +++++
>  gas/testsuite/gas/riscv/zkr.s      |  2 +
>  gas/testsuite/gas/riscv/zkt.d      |  5 +++
>  22 files changed, 267 insertions(+), 394 deletions(-)
>  delete mode 100644 gas/testsuite/gas/riscv/b-ext-64.d
>  delete mode 100644 gas/testsuite/gas/riscv/b-ext-64.s
>  delete mode 100644 gas/testsuite/gas/riscv/b-ext.d
>  delete mode 100644 gas/testsuite/gas/riscv/b-ext.s
>  delete mode 100644 gas/testsuite/gas/riscv/k-ext-64.d
>  delete mode 100644 gas/testsuite/gas/riscv/k-ext-64.s
>  delete mode 100644 gas/testsuite/gas/riscv/k-ext.d
>  delete mode 100644 gas/testsuite/gas/riscv/k-ext.s
>  create mode 100644 gas/testsuite/gas/riscv/zba-64.d
>  create mode 100644 gas/testsuite/gas/riscv/zba.d
>  create mode 100644 gas/testsuite/gas/riscv/zba.s
>  create mode 100644 gas/testsuite/gas/riscv/zbb-64.d
>  create mode 100644 gas/testsuite/gas/riscv/zbb.d
>  create mode 100644 gas/testsuite/gas/riscv/zbb.s
>  create mode 100644 gas/testsuite/gas/riscv/zbc.d
>  create mode 100644 gas/testsuite/gas/riscv/zbc.s
>  create mode 100644 gas/testsuite/gas/riscv/zbs-64.d
>  create mode 100644 gas/testsuite/gas/riscv/zbs.d
>  create mode 100644 gas/testsuite/gas/riscv/zbs.s
>  create mode 100644 gas/testsuite/gas/riscv/zkr.d
>  create mode 100644 gas/testsuite/gas/riscv/zkr.s
>  create mode 100644 gas/testsuite/gas/riscv/zkt.d
> 
> 
> base-commit: 8e037eae6823caf5b9cb5b4feb3de838abb25956

      parent reply	other threads:[~2022-10-25 14:54 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-09-23  7:49 Tsukasa OI
2022-09-23  7:49 ` [PATCH 1/2] RISC-V: Reorganize "K"-related subextension tests Tsukasa OI
2022-09-23  7:49 ` [PATCH 2/2] RISC-V: Reorganize "B"-related " Tsukasa OI
2022-10-25 14:54 ` Tsukasa OI [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=131a0f94-91c3-c850-130b-598ee0afb998@irq.a4lg.com \
    --to=research_trasio@irq.a4lg.com \
    --cc=binutils@sourceware.org \
    --cc=kito.cheng@sifive.com \
    --cc=nelson@rivosinc.com \
    --cc=palmer@dabbelt.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).