From: Jan Beulich <jbeulich@suse.com>
To: Hongyu Wang <wwwhhhyyy333@gmail.com>,
"Jiang, Haochen" <haochen.jiang@intel.com>
Cc: "H.J. Lu" <hjl.tools@gmail.com>,
"binutils@sourceware.org" <binutils@sourceware.org>,
Michael Matz <matz@suse.de>
Subject: Re: [PATCH] x86: Warn .insn instruction with length > 15 bytes
Date: Thu, 8 Feb 2024 09:15:20 +0100 [thread overview]
Message-ID: <fe4cc5e9-fa24-41ed-ba10-f4bddfc929b4@suse.com> (raw)
In-Reply-To: <CA+OydWk2aYb8Ec3th3BbFKt9A1LHO_KSvdFUrhB7VqnthOXwbg@mail.gmail.com>
On 08.02.2024 05:47, Hongyu Wang wrote:
>> as_bad() is used to mark errors that result in what we
>> presume to be a useless object file. Say, we ignored
>> something that might have been vital.
>> ...
>>
>> as_warn() is used when we have an error from which we
>> have a plausible error recovery. eg, masking the top
>> bits of a constant that is longer than will fit in the
>> destination. In this case we will continue to assemble
>> the source, although we may have made a bad assumption,
>> and we will produce an object file and return normal exit
>> status (ie, no error).
>> ...
>>
>> It's obvious to me that just continuing to assemble the over-long
>> instruction is a "plausible error recovery". It's even more plausible
>> than "masking the top bits of a constant". Certainly an object file
>> containing a byte sequence correspending to the overlong instruction is
>> not "useless".
>
> We have encountered such issue when running APX workloads with SDE.
> The over-long instruction became either illegal or wrong instruction
> when executing the program, then it crashes after long time running.
> This definitely increases much effort for runtime diagnostics.
Well, that's what you get for not paying attention to diagnostics.
They aren't emitted just for fun.
Jan
> As the over-long instruction would actually break the program, say the
> object crashes or returns with abnormal exit status, the default
> should be an error for assembler and we can add an option to lower the
> error to warning.
next prev parent reply other threads:[~2024-02-08 8:15 UTC|newest]
Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-02-05 20:00 H.J. Lu
2024-02-06 8:19 ` Jan Beulich
2024-02-06 11:36 ` H.J. Lu
2024-02-06 11:41 ` Jan Beulich
2024-02-06 12:26 ` H.J. Lu
2024-02-06 14:43 ` Michael Matz
2024-02-06 14:49 ` H.J. Lu
2024-02-06 15:04 ` Michael Matz
2024-02-06 15:34 ` H.J. Lu
2024-02-06 15:48 ` Michael Matz
2024-02-06 16:28 ` H.J. Lu
2024-02-06 17:05 ` Jan Beulich
2024-02-06 18:06 ` H.J. Lu
2024-02-07 7:51 ` Jan Beulich
2024-02-07 15:24 ` H.J. Lu
2024-02-07 15:32 ` Jan Beulich
2024-02-07 16:53 ` H.J. Lu
2024-02-07 16:59 ` Jan Beulich
2024-02-07 17:03 ` H.J. Lu
2024-02-08 4:09 ` Jiang, Haochen
2024-02-08 4:47 ` Hongyu Wang
2024-02-08 8:15 ` Jan Beulich [this message]
2024-02-08 8:23 ` Jan Beulich
2024-02-08 11:38 ` H.J. Lu
2024-02-08 6:26 ` Sunil Pandey
2024-02-08 6:41 ` Cui, Lili
2024-02-08 8:18 ` Jan Beulich
2024-02-08 11:31 ` H.J. Lu
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=fe4cc5e9-fa24-41ed-ba10-f4bddfc929b4@suse.com \
--to=jbeulich@suse.com \
--cc=binutils@sourceware.org \
--cc=haochen.jiang@intel.com \
--cc=hjl.tools@gmail.com \
--cc=matz@suse.de \
--cc=wwwhhhyyy333@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).