From: "Cui, Lili" <lili.cui@intel.com>
To: "Beulich, Jan" <JBeulich@suse.com>, Binutils <binutils@sourceware.org>
Cc: "H.J. Lu" <hjl.tools@gmail.com>
Subject: RE: [PATCH] x86: add missing APX logic to cpu_flags_match()
Date: Mon, 8 Jan 2024 03:17:20 +0000 [thread overview]
Message-ID: <SJ0PR11MB56008304B9A8C6B0703234E99E6B2@SJ0PR11MB5600.namprd11.prod.outlook.com> (raw)
In-Reply-To: <f3f8baa5-efd8-4f24-8bfc-1193adf312ba@suse.com>
> As already indicated during review, we can't get away without certain
> adjustments here: Without these, respective {evex}-prefixed insns are
> assembled to APX encodings even when APX_F is turned off.
>
> While there also extend the respective comment in the opcode table, to
> explain why this construct is used.
> ---
> Strictly speaking we could go with just cpuid|APX_F in the templates, with the
> assertions dropped and instead an "else" added. But I think we're better off
> this way, for being less prone to introducing mistakes later on.
>
> The resulting diagnostics aren't quite correct (because of not mentioning the
> {evex} prefix, which really is what's the problem there), but improving
> diagnostics is a wider topic anyway.
>
Oh, get it, thanks.
Currently, a lot of ugly special handling has been added to i386 in order to merge these vex/evex insns, which increases the complexity of the code. Personally, I don't think it's worth it. Other issues may arise later.
> --- a/gas/config/tc-i386.c
> +++ b/gas/config/tc-i386.c
> @@ -1940,6 +1940,30 @@ cpu_flags_match (const insn_template *t)
> any.bitfield.cpuavx512vl = 0;
> }
> }
> +
> + /* Dual non-APX/APX templates need massaging from what APX_F() in the
> + opcode table has produced. While the direct transformation of the
> + incoming cpuid&(cpuid|APX_F) would be to cpuid&(cpuid) /
> cpuid&(APX_F)
> + respectively, it's cheaper to move to just cpuid / cpuid&APX_F
> + instead. */
> + if (any.bitfield.cpuapx_f
> + && (any.bitfield.cpubmi || any.bitfield.cpubmi2
> + || any.bitfield.cpuavx512f || any.bitfield.cpuavx512bw
> + || any.bitfield.cpuavx512dq || any.bitfield.cpuamx_tile
> + || any.bitfield.cpucmpccxadd))
> + {
> + /* These checks (verifying that APX_F() was properly used in the
> + opcode table entry) make sure there's no need for an "else" to
> + the "if()" below. */
> + gas_assert (!cpu_flags_all_zero (&all));
> + cpu = cpu_flags_and (all, any);
> + gas_assert (cpu_flags_equal (&cpu, &all));
> +
> + if (need_evex_encoding (t))
> + all = any;
> +
> + memset (&any, 0, sizeof (any));
Wouldn't it make sense to put it in the else branch and clean out APX-F specifically? Just like you did before.
if (need_evex_encoding (t))
all = any;
else
any.bitfield.cpuapx_f = 0;
thanks,
Lili.
> + }
> }
>
> if (flag_code != CODE_64BIT)
> --- a/opcodes/i386-opc.tbl
> +++ b/opcodes/i386-opc.tbl
> @@ -143,7 +143,13 @@
>
> #define DstVVVV VexVVVV=VexVVVV_DST
>
> -// The template supports VEX format for cpuid and EVEX format for cpuid &
> apx_f.
> +// The template supports VEX format for cpuid and EVEX format for cpuid &
> APX_F.
> +// While therefore we really mean cpuid|(cpuid&APX_F) here, this can't
> +be // expressed in the generated templates. It's equivalent to just
> +cpuid|APX_F // anyway, but that is not what we want (as APX_F alone
> +isn't a sufficient // prereq for such insns). Instead the assembler
> +will massage the CPU specifier // to the equivalent of either
> +cpuid&(cpuid) or cpuid&(APX_F) (or something // substantially similar),
> depending on what encoding was requested.
> #define APX_F(cpuid) cpuid&(cpuid|APX_F)
>
> // The EVEX purpose of StaticRounding appears only together with SAE. Re-
> use
> --- a/gas/testsuite/gas/i386/x86-64-apx-egpr-promote-inval.l
> +++ b/gas/testsuite/gas/i386/x86-64-apx-egpr-promote-inval.l
> @@ -13,6 +13,13 @@
> .*:25: Error: `andn' is not supported on `x86_64.nobmi'
> .*:28: Error: `bzhi' is not supported on `x86_64.nobmi2'
> .*:29: Error: `bzhi' is not supported on `x86_64.nobmi2'
> +.*:33: Error: .*`andn'.*
> +.*:34: Error: .*`bzhi'.*
> +.*:35: Error: .*`kmovw'.*
> +.*:36: Error: .*`kmovq'.*
> +.*:37: Error: .*`kmovb'.*
> +.*:38: Error: .*`ldtilecfg'.*
> +.*:39: Error: .*`cmpexadd'.*
> GAS LISTING .*
> #...
> [ ]*1[ ]+\# Check illegal 64bit APX EVEX promoted instructions
> --- a/gas/testsuite/gas/i386/x86-64-apx-egpr-promote-inval.s
> +++ b/gas/testsuite/gas/i386/x86-64-apx-egpr-promote-inval.s
> @@ -27,3 +27,13 @@
> .arch .nobmi2
> bzhi %r16,%r15,%r11
> bzhi %r15,%r15,%r11
> +
> + .arch default
> + .arch .noapx_f
> + {evex} andn %r15, %r15, %r11
> + {evex} bzhi %r15, %r15, %r11
> + {evex} kmovw %k1, %r8d
> + {evex} kmovq %k1, %r8
> + {evex} kmovb %k1, %r8d
> + {evex} ldtilecfg (%r8)
> + {evex} cmpexadd %rax, %rcx, (%r8)
next prev parent reply other threads:[~2024-01-08 3:17 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-01-05 12:15 Jan Beulich
2024-01-08 3:17 ` Cui, Lili [this message]
2024-01-08 7:56 ` Jan Beulich
2024-01-08 8:30 ` Cui, Lili
2024-01-08 8:58 ` Jan Beulich
2024-01-08 10:28 ` Cui, Lili
2024-01-08 10:38 ` Jan Beulich
2024-01-09 5:36 ` Cui, Lili
2024-01-09 8:30 ` Jan Beulich
2024-01-09 11:00 ` Cui, Lili
2024-01-09 11:07 ` Jan Beulich
2024-01-10 1:44 ` Cui, Lili
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=SJ0PR11MB56008304B9A8C6B0703234E99E6B2@SJ0PR11MB5600.namprd11.prod.outlook.com \
--to=lili.cui@intel.com \
--cc=JBeulich@suse.com \
--cc=binutils@sourceware.org \
--cc=hjl.tools@gmail.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).