From: Liav Rehana <liavr@mellanox.com>
To: Thomas Gleixner <tglx@linutronix.de>
Cc: "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
"john.stultz@linaro.org" <john.stultz@linaro.org>,
Noam Camus <noamca@mellanox.com>,
Elad Kanfi <eladkan@mellanox.com>
Subject: RE: [PATCH] Fix chance of sign extension to nsec after its msb is set during calculation.
Date: Sun, 4 Sep 2016 10:03:32 +0000 [thread overview]
Message-ID: <VI1PR0501MB2461FDE5EE49332C0B7744FBC5E70@VI1PR0501MB2461.eurprd05.prod.outlook.com> (raw)
In-Reply-To: <alpine.DEB.2.20.1609041057540.5647@nanos>
>> The root of the problem is that in case the multiplication of delta
>> and
>> tkr->mult in the line that I've changed is too big that the MSB of the
>> result is set, then the shift will cause an unwanted sign extension.
> I completely understand that, but as I said before:
> > > This typecast is just a baindaid. What happens if you double the
> > > suspend time? The multiplication will simply overflow. So the
> > > proper fix is to sanity check delta and do multiple conversions if
> > > delta is big enough. Preferrably this happens somewhere at the call
> > > site and not in this hotpath function.
> > That sign extension will be avoided completely if the variable nsec
> > was unsigned (u64 instead of s64), so I think the correct solution for
> > this is to change the type of nsec to u64.
> That's a different story and its not a solution for the general problem of
> delta * mult >= (1 << 31) or delta * mult >= (1 << 32)
The case that delta * mult >= 1 << 31 is not a problem by itself, but it causes
an unwanted sign extension since the type of nsec is signed. That sign
extension is what causes the loop to take too long, and not the overflow.
I understand that the typecast is not a general solution, so as I've said, I
think that changing the type of nsec to u64 instead of s64 will be a good and
general solution, as it will indeed solve the problem of the unwanted sign
extension.
To summarize: a sign extension occurs if the nsec variable is signed, and so
I ask if you think it will be a good solution to change its type to unsigned.
Thanks,
Liav.
next prev parent reply other threads:[~2016-09-04 22:41 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-09-01 11:14 [PATCH] Fix chance of sign extension to nsec after its msb is set during calculation Liav Rehana
2016-09-02 8:49 ` Thomas Gleixner
2016-09-02 18:30 ` Thomas Gleixner
2016-09-02 18:47 ` Thomas Gleixner
2016-09-04 6:40 ` Liav Rehana
2016-09-04 9:01 ` Thomas Gleixner
2016-09-04 10:03 ` Liav Rehana [this message]
2016-09-04 10:39 ` Thomas Gleixner
2016-09-07 3:22 ` John Stultz
2016-09-11 6:31 ` Liav Rehana
2016-09-21 21:12 ` John Stultz
2016-09-21 12:52 ` Liav Rehana
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=VI1PR0501MB2461FDE5EE49332C0B7744FBC5E70@VI1PR0501MB2461.eurprd05.prod.outlook.com \
--to=liavr@mellanox.com \
--cc=eladkan@mellanox.com \
--cc=john.stultz@linaro.org \
--cc=linux-kernel@vger.kernel.org \
--cc=noamca@mellanox.com \
--cc=tglx@linutronix.de \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.