From: Thomas Gleixner <tglx@linutronix.de>
To: Andy Lutomirski <luto@amacapital.net>
Cc: x86@kernel.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] x86: Remove hpet vclock support
Date: Tue, 4 Feb 2014 20:31:48 +0100 (CET) [thread overview]
Message-ID: <alpine.DEB.2.02.1402042011400.24986@ionos.tec.linutronix.de> (raw)
In-Reply-To: <d348afdb2f6acc41dde0381396226b69a0358f8b.1391210357.git.luto@amacapital.net>
On Fri, 31 Jan 2014, Andy Lutomirski wrote:
> The HPET is so amazingly slow that this is barely a win. It adds
That's nonsense. It's definitely a win to access HPET directly
especially on older systems with TSC wreckage. Why do you want to
enforce a syscall if we can read out the data straight from user
space. The systems which are forced to use HPET have slower syscalls
than those which have a proper TSC.
> complexity, and it scares me a tiny bit to map a piece of crappy
> hardware where every userspace program can poke at it (even if said
> poking is read-only). Let's just remove it.
What's scary about granting user space access to a read only resource?
> This is probably a tiny speedup on kvmclock systems.
Oh, now you are coming to the real point of your change:
You want to save a few cycles in a guest system.
So you justify that by completely unbacked assumptions and a pointless
blurb about scariness.
We are not removing existing functionality which is useful for parts
of our userbase just to make financial online gambling faster.
> #ifdef CONFIG_PARAVIRT_CLOCK
>
> static notrace const struct pvclock_vsyscall_time_info *get_pvti(int cpu)
> @@ -157,8 +151,6 @@ notrace static inline u64 vgetsns(int *mode)
> cycles_t cycles;
> if (gtod->clock.vclock_mode == VCLOCK_TSC)
> cycles = vread_tsc();
> - else if (gtod->clock.vclock_mode == VCLOCK_HPET)
> - cycles = vread_hpet();
> #ifdef CONFIG_PARAVIRT_CLOCK
> else if (gtod->clock.vclock_mode == VCLOCK_PVCLOCK)
> cycles = vread_pvclock(mode);
I really doubt, that your tiny speedup is even measurable simply
because the branch prediction wont fail and the cache foot print is
not that excitingly smaller. I could be wrong as usual, but its up to
you to provide proper numbers which prove that:
- there is no impact for the actual forced to use HPET systems
- there is a measurable improvement for the PV clock case
and not before you tried to revert the evaluation order of VCLOCK_HPET
and VCLOCK_PVCLOCK.
Thanks,
tglx
next prev parent reply other threads:[~2014-02-04 19:31 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-01-31 23:19 Andy Lutomirski
2014-02-01 15:43 ` Clemens Ladisch
2014-02-01 17:10 ` Andy Lutomirski
2014-02-04 19:31 ` Thomas Gleixner [this message]
2014-02-04 19:41 ` Andy Lutomirski
2014-02-04 20:29 ` Thomas Gleixner
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=alpine.DEB.2.02.1402042011400.24986@ionos.tec.linutronix.de \
--to=tglx@linutronix.de \
--cc=linux-kernel@vger.kernel.org \
--cc=luto@amacapital.net \
--cc=x86@kernel.org \
/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
all inboxes | Powered by JetHome®