DPDK patches and discussions
 help / color / mirror / Atom feed
From: "Morten Brørup" <mb@smartsharesystems.com>
To: "Konstantin Ananyev" <konstantin.ananyev@huawei.com>,
	"Konstantin Ananyev" <konstantin.v.ananyev@yandex.ru>,
	<dev@dpdk.org>, "Bruce Richardson" <bruce.richardson@intel.com>,
	"Honnappa Nagarahalli" <Honnappa.Nagarahalli@arm.com>
Cc: "Jan Viktorin" <viktorin@rehivetech.com>,
	"Ruifeng Wang" <ruifeng.wang@arm.com>,
	"David Christensen" <drc@linux.vnet.ibm.com>,
	"Stanislaw Kardach" <kda@semihalf.com>
Subject: RE: [RFC v2] non-temporal memcpy
Date: Fri, 29 Jul 2022 22:26:56 +0200	[thread overview]
Message-ID: <98CBD80474FA8B44BF855DF32C47DC35D87213@smartserver.smartshare.dk> (raw)
In-Reply-To: <66b6efccde5b4d68bffcc47713cdf983@huawei.com>

+TO: @Honnappa, we need input from ARM

> From: Konstantin Ananyev [mailto:konstantin.ananyev@huawei.com]
> Sent: Friday, 29 July 2022 21.49
> >
> > > From: Konstantin Ananyev [mailto:konstantin.ananyev@huawei.com]
> > > Sent: Friday, 29 July 2022 14.14
> > >
> > >
> > > Sorry, missed that part.
> > >
> > > >
> > > > > Another question - who will do 'sfence' after the copying?
> > > > > Would it be inside memcpy_nt (seems quite costly), or would
> > > > > it be another API function for that: memcpy_nt_flush() or so?
> > > >
> > > > Outside. Only the developer knows when it is required, so it
> wouldn't
> > > make any sense to add the cost inside memcpy_nt().
> > > >
> > > > I don't think we should add a flush function; it would just be
> > > another name for an already existing function. Referring to the
> > > required
> > > > operation in the memcpy_nt() function documentation should
> suffice.
> > > >
> > >
> > > Ok, but again wouldn't it be arch specific?
> > > AFAIK for x86 it needs to boil down to sfence, for other
> architectures
> > > - I don't know.
> > > If you think there already is some generic one (rte_wmb?) that
> would
> > > always produce
> > > correct instructions - sure let's use it.
> > >
> >
> > DPDK has generic functions to wrap architecture specific stuff like
> memory barriers.
> >
> > Because they are non-temporal stores, I suspect that rte_mb() is
> required before reading the data from the location it was copied to.
> > Ensuring that STORE operations are ordered (rte_wmb) might not
> suffice. However, I'm not a CPU expert, so I will seek advice from
> > more qualified people in the community on this.
> 
> I think for IA sfence is enough, see citation below,
> for other architectures - no idea.
> What I am trying to say - it needs to be the *same* function on all
> archs we support.

Now I get it: rte_wmb() might be appropriate on x86, but if any other architecture requires something else, we should add a new common function for flushing, e.g. rte_memcpy_nt_flush().

> 
> IA SW optimization manual:
> 9.4.2 Streaming Store Usage Models
> The two primary usage domains for streaming store are coherent requests
> and non-coherent requests.
> 9.4.2.1 Coherent Requests
> Coherent requests are normal loads and stores to system memory, which
> may also hit cache lines
> present in another processor in a multiprocessor environment. With
> coherent requests, a streaming store
> can be used in the same way as a regular store that has been mapped
> with a WC memory type (PAT or
> MTRR). An SFENCE instruction must be used within a producer-consumer
> usage model in order to ensure
> coherency and visibility of data between processors.
> Within a single-processor system, the CPU can also re-read the same
> memory location and be assured of
> coherence (that is, a single, consistent view of this memory location).
> The same is true for a multiprocessor
> (MP) system, assuming an accepted MP software producer-consumer
> synchronization policy is
> employed.
> 

With this reference, I am convinced that you are right about the SFENCE. This puts a checkmark on this item on my TODO list for the patch. Thank you, Konstantin!

Any ARM CPU experts on the mailing list seeing this, not on vacation? @Honnappa, I'm looking at you. :-)

Summing up, the question is:

After a bunch of *non-temporal* stores (STNP instruction) on ARM architecture, does calling rte_wmb() suffice to ensure the data is visible across the system?


  reply	other threads:[~2022-07-29 20:26 UTC|newest]

Thread overview: 57+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-07-19 15:26 Morten Brørup
2022-07-19 18:00 ` David Christensen
2022-07-19 18:41   ` Morten Brørup
2022-07-19 18:51     ` Stanisław Kardach
2022-07-19 22:15       ` Morten Brørup
2022-07-21 23:19 ` Konstantin Ananyev
2022-07-22 10:44   ` Morten Brørup
2022-07-24 13:35     ` Konstantin Ananyev
2022-07-24 22:18       ` Morten Brørup
2022-07-29 10:00         ` Konstantin Ananyev
2022-07-29 10:46           ` Morten Brørup
2022-07-29 11:50             ` Konstantin Ananyev
2022-07-29 17:17               ` Morten Brørup
2022-07-29 22:00                 ` Konstantin Ananyev
2022-07-30  9:51                   ` Morten Brørup
2022-08-02  9:05                     ` Konstantin Ananyev
2022-07-29 12:13             ` Konstantin Ananyev
2022-07-29 16:05               ` Stephen Hemminger
2022-07-29 17:29                 ` Morten Brørup
2022-08-07 20:40                 ` Mattias Rönnblom
2022-08-09  9:24                   ` Morten Brørup
2022-08-09 11:53                     ` Mattias Rönnblom
2022-10-09 16:16                       ` Morten Brørup
2022-07-29 18:13               ` Morten Brørup
2022-07-29 19:49                 ` Konstantin Ananyev
2022-07-29 20:26                   ` Morten Brørup [this message]
2022-07-29 21:34                     ` Konstantin Ananyev
2022-08-07 20:20                     ` Mattias Rönnblom
2022-08-09  9:34                       ` Morten Brørup
2022-08-09 11:56                         ` Mattias Rönnblom
2022-08-10 21:05                     ` Honnappa Nagarahalli
2022-08-11 11:50                       ` Mattias Rönnblom
2022-08-11 16:26                         ` Honnappa Nagarahalli
2022-07-25  1:17       ` Honnappa Nagarahalli
2022-07-27 10:26         ` Morten Brørup
2022-07-27 17:37           ` Honnappa Nagarahalli
2022-07-27 18:49             ` Morten Brørup
2022-07-27 19:12               ` Stephen Hemminger
2022-07-28  9:00                 ` Morten Brørup
2022-07-27 19:52               ` Honnappa Nagarahalli
2022-07-27 22:02                 ` Stanisław Kardach
2022-07-28 10:51                   ` Morten Brørup
2022-07-29  9:21                     ` Konstantin Ananyev
2022-08-07 20:25 ` Mattias Rönnblom
2022-08-09  9:46   ` Morten Brørup
2022-08-09 12:05     ` Mattias Rönnblom
2022-08-09 15:00       ` Morten Brørup
2022-08-10 11:47         ` Mattias Rönnblom
2022-08-09 15:26     ` Stephen Hemminger
2022-08-09 17:24       ` Morten Brørup
2022-08-10 11:59         ` Mattias Rönnblom
2022-08-10 12:12           ` Morten Brørup
2022-08-10 11:55       ` Mattias Rönnblom
2022-08-10 12:18         ` Morten Brørup
2022-08-10 21:20           ` Honnappa Nagarahalli
2022-08-11 11:53             ` Mattias Rönnblom
2022-08-11 22:24               ` Honnappa Nagarahalli

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=98CBD80474FA8B44BF855DF32C47DC35D87213@smartserver.smartshare.dk \
    --to=mb@smartsharesystems.com \
    --cc=Honnappa.Nagarahalli@arm.com \
    --cc=bruce.richardson@intel.com \
    --cc=dev@dpdk.org \
    --cc=drc@linux.vnet.ibm.com \
    --cc=kda@semihalf.com \
    --cc=konstantin.ananyev@huawei.com \
    --cc=konstantin.v.ananyev@yandex.ru \
    --cc=ruifeng.wang@arm.com \
    --cc=viktorin@rehivetech.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 NNTP newsgroup(s).