From: "H. Peter Anvin" <hpa@zytor.com>
To: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Cc: akpm@linux-foundation.org, linux-kernel@vger.kernel.org,
Andi Kleen <ak@muc.de>, Chuck Ebbert <cebbert@redhat.com>,
Christoph Hellwig <hch@infradead.org>,
Jeremy Fitzhardinge <jeremy@goop.org>
Subject: Re: [patch 5/8] Immediate Values - x86 Optimization (update)
Date: Tue, 13 Nov 2007 13:26:06 -0800 [thread overview]
Message-ID: <473A166E.3070708@zytor.com> (raw)
In-Reply-To: <20071113204033.GB7450@Krystal>
Mathieu Desnoyers wrote:
> * H. Peter Anvin (hpa@zytor.com) wrote:
>> Mathieu Desnoyers wrote:
>>> * H. Peter Anvin (hpa@zytor.com) wrote:
>>>> Mathieu Desnoyers wrote:
>>>>>> - Use "=g" constraint for char immediate value inline assembly.
>>>>>>
>>>>>> "=g" is the same as "=rmi" which is inherently bogus. In your actual
>>>>>> code you use "=r", the correct constraint is "=q".
>>>>> q
>>>>> Any register accessible as rl. In 32-bit mode, a, b, c, and d; in
>>>>> 64-bit mode, any integer register. I am worried that "=q" might exclude
>>>>> the si and di registers in 32-bit mode.
>>>>> What exactly is wrong with "=r" ?
>>>> For "char" (8-bit) values, sp/bp/si/di are illegal in 32-bit mode.
>>>>
>>>> Hence "=q".
>>>>
>>> Ah! yep, I see, so we say:
>>> 1 byte : "=q"
>>> 2 bytes : "=r"
>>> 4 bytes : "=r"
>>> 8 bytes : "=r"
>>> ? (si and di appear to be legal for 2 and 4 bytes in 32-bit mode)
>> That's right.
>>
>> -hpa
>
Something else to watch out for... in 64-bit mode the lengths most of
these will depend on which register is used, since whether or not a REX
prefix is needed will vary.
As far as I can tell, you're assuming fixed length instructions, which
is wrong unless you manually constrain yourself to only legacy registers.
-hpa
next prev parent reply other threads:[~2007-11-13 21:30 UTC|newest]
Thread overview: 40+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-11-13 18:58 [patch 0/8] Immediate Values (now with merged x86 support) Mathieu Desnoyers
2007-11-13 18:58 ` [patch 1/8] Immediate Values - Architecture Independent Code Mathieu Desnoyers
2007-11-13 18:58 ` [patch 2/8] Immediate Values - Kconfig menu in EMBEDDED Mathieu Desnoyers
2007-11-13 18:58 ` [patch 3/8] Immediate Values - Move Kprobes x86 restore_interrupt to kdebug.h Mathieu Desnoyers
2007-11-13 18:58 ` [patch 4/8] Add asm-compat.h to x86 Mathieu Desnoyers
2007-11-13 19:05 ` H. Peter Anvin
2007-11-13 20:37 ` [patch 4/8] Add asm-compat.h to x86 -> use new asm.h instead Mathieu Desnoyers
2007-11-13 18:58 ` [patch 5/8] Immediate Values - x86 Optimization Mathieu Desnoyers
2007-11-13 19:07 ` H. Peter Anvin
2007-11-13 19:24 ` Mathieu Desnoyers
2007-11-13 19:36 ` H. Peter Anvin
2007-11-13 19:45 ` Mathieu Desnoyers
2007-11-13 19:56 ` H. Peter Anvin
2007-11-13 20:40 ` [patch 5/8] Immediate Values - x86 Optimization (update) Mathieu Desnoyers
2007-11-13 21:26 ` H. Peter Anvin [this message]
2007-11-13 22:02 ` Mathieu Desnoyers
2007-11-13 22:35 ` H. Peter Anvin
2007-11-14 0:34 ` Mathieu Desnoyers
2007-11-14 1:02 ` H. Peter Anvin
2007-11-14 1:44 ` [patch 5/8] Immediate Values - x86 Optimization (update 2) Mathieu Desnoyers
2007-11-14 2:58 ` H. Peter Anvin
2007-11-14 14:16 ` Mathieu Desnoyers
2007-11-14 18:52 ` [PATCH] Immediate Values x86 Optimization Declare Discarded Instruction Mathieu Desnoyers
2007-11-14 19:00 ` H. Peter Anvin
2007-11-14 19:16 ` [PATCH] Add __discard section to x86 Mathieu Desnoyers
2007-11-14 19:33 ` H. Peter Anvin
2007-11-15 3:08 ` [patch 5/8] Immediate Values - x86 Optimization Rusty Russell
2007-11-15 4:06 ` Mathieu Desnoyers
2007-11-15 4:45 ` Rusty Russell
2007-11-15 5:37 ` Mathieu Desnoyers
2007-11-15 11:06 ` Rusty Russell
2007-11-16 14:03 ` [patch 5/8] Immediate Values - x86 Optimization (simplified) Mathieu Desnoyers
2007-11-18 23:11 ` Rusty Russell
2007-11-19 14:28 ` Mathieu Desnoyers
2007-11-19 23:06 ` Rusty Russell
2007-11-20 17:02 ` Mathieu Desnoyers
2007-11-19 19:15 ` Mathieu Desnoyers
2007-11-13 18:58 ` [patch 6/8] Immediate Values - Powerpc Optimization Mathieu Desnoyers
2007-11-13 18:58 ` [patch 7/8] Immediate Values - Documentation Mathieu Desnoyers
2007-11-13 18:58 ` [patch 8/8] Scheduler Profiling - Use Immediate Values Mathieu Desnoyers
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=473A166E.3070708@zytor.com \
--to=hpa@zytor.com \
--cc=ak@muc.de \
--cc=akpm@linux-foundation.org \
--cc=cebbert@redhat.com \
--cc=hch@infradead.org \
--cc=jeremy@goop.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mathieu.desnoyers@polymtl.ca \
/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 NNTP newsgroup(s).