From: Stephane Eranian <eranian@hpl.hp.com>
To: Paul Mackerras <paulus@samba.org>
Cc: Christoph Hellwig <hch@infradead.org>,
Andrew Morton <akpm@linux-foundation.org>,
Greg KH <gregkh@suse.de>, Philip Mucci <mucci@cs.utk.edu>,
William Cohen <wcohen@redhat.com>,
Robert Richter <robert.richter@amd.com>,
linux-kernel@vger.kernel.org, Andi Kleen <andi@firstfloor.org>
Subject: Re: [perfmon] Re: [perfmon2] perfmon2 merge news
Date: Wed, 14 Nov 2007 05:47:13 -0800 [thread overview]
Message-ID: <20071114134713.GD6557@frankl.hpl.hp.com> (raw)
In-Reply-To: <18234.56940.229831.735533@cargo.ozlabs.ibm.com>
Hello,
On Wed, Nov 14, 2007 at 10:39:24PM +1100, Paul Mackerras wrote:
> Christoph Hellwig writes:
>
> > int pfm_read_pmds(int fd, pfarg_pmd_t *pmds, int n)
> >
> > This is basically a read(2) (or for other syscalls a write) on something
> > else than the file descriptor provided to the system call.
>
> No it's not basically a read(). It's more like a request/reply
> interface, which a read()/write() interface doesn't handle very well.
> The request in this case is "tell me about this particular collection
> of PMDs" and the reply is the values.
>
Exactly. This is not a brute force read()! On input you pass the list
of registers you want to read. Upon return, you get the list of values.
Now, I think the current call could be optimized even more by making
the structure smaller. Today, the structure passed read/write
PMD registers is the same. On write, we pass other information such as
the reset values (sampling periods), randomization parameters and some
flags. They are not needed on read.
> It seems to me that an important part of this is to be able to collect
> values from several PMDs at a single point in time, or at least an
> approximation to a single point in time. So that means that you don't
> want a file per PMD either.
>
Yes, we want to be able to read one or many registers in one call.
The number of PMU counters is not going to shrink, so having a file
descriptor per register looks overkill to me.
> Basically we don't have a good abstraction for a request/reply (or
> command/response) type of interface, and this is a case where we need
> one. Having a syscall that takes a struct containing the request and
> reply is as good a way as any, particularly for something that needs
> to be quick.
>
--
-Stephane
next prev parent reply other threads:[~2007-11-14 13:54 UTC|newest]
Thread overview: 116+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-11-07 0:34 [PATCH] fix up perfmon to build on -mm Greg KH
2007-11-07 10:34 ` Stephane Eranian
2007-11-07 17:07 ` Greg KH
2007-11-07 13:42 ` Stephane Eranian
2007-11-07 17:08 ` Greg KH
2007-11-07 17:33 ` Andrew Morton
2007-11-07 17:41 ` Greg KH
2007-11-07 17:50 ` Stephane Eranian
2007-11-07 17:47 ` Greg KH
2007-11-07 17:57 ` Stephane Eranian
2007-11-07 19:53 ` Greg KH
2007-11-07 20:39 ` Stephane Eranian
2007-11-08 15:27 ` Stephane Eranian
2007-11-09 20:06 ` Andrew Morton
2007-11-09 21:38 ` Greg KH
2007-11-10 20:32 ` Andi Kleen
2007-11-13 15:17 ` perfmon2 merge news Robert Richter
2007-11-13 15:35 ` [perfmon2] " William Cohen
2007-11-13 17:55 ` Stephane Eranian
2007-11-13 18:33 ` [perfmon] " William Cohen
2007-11-13 21:13 ` Stephane Eranian
2007-11-13 21:29 ` Andi Kleen
2007-11-13 21:46 ` Stephane Eranian
2007-11-13 21:50 ` Andi Kleen
2007-11-13 22:22 ` Stephane Eranian
2007-11-13 22:25 ` Andi Kleen
2007-11-13 22:58 ` Stephane Eranian
2007-11-14 2:07 ` Andi Kleen
2007-11-14 13:09 ` Stephane Eranian
2007-11-14 14:24 ` Andi Kleen
2007-11-14 15:44 ` William Cohen
2007-11-14 16:13 ` Stephane Eranian
2007-11-14 18:53 ` Philippe Elie
2007-11-14 19:15 ` Andi Kleen
2007-11-15 0:07 ` Stephane Eranian
2007-11-13 18:47 ` Philip Mucci
2007-11-13 18:59 ` Greg KH
2007-11-13 20:07 ` Andrew Morton
2007-11-13 20:14 ` Greg KH
2007-11-13 20:36 ` Andi Kleen
2007-11-14 0:28 ` Philip Mucci
2007-11-14 1:52 ` Andi Kleen
2007-11-16 9:18 ` Philip Mucci
2007-11-16 15:15 ` Andi Kleen
2007-11-16 16:00 ` Stephane Eranian
2007-11-16 16:28 ` Andi Kleen
2007-11-16 17:13 ` William Cohen
2007-11-16 21:56 ` Stephane Eranian
2007-11-16 17:36 ` Stephane Eranian
2007-11-16 17:51 ` dean gaudet
2007-11-17 0:29 ` David Miller
2007-11-17 1:07 ` Greg KH
2007-11-16 20:16 ` Philip Mucci
2007-11-17 0:15 ` David Miller
[not found] ` <1d7226b10711161713j675341b7wdb4f050c59a8be0a@mail.gmail.com>
2007-11-17 1:25 ` Greg KH
[not found] ` <1d7226b10711161748n39b7f195q796d85282ef66134@mail.gmail.com>
2007-11-17 2:13 ` Greg KH
2007-11-14 7:24 ` [perfmon] Re: [perfmon2] " Paul Mackerras
2007-11-14 7:40 ` Andrew Morton
2007-11-14 10:38 ` Christoph Hellwig
2007-11-14 10:43 ` Paul Mackerras
2007-11-14 11:00 ` Christoph Hellwig
2007-11-14 11:12 ` David Miller
2007-11-14 11:14 ` David Miller
2007-11-14 11:44 ` Paul Mackerras
2007-11-13 23:49 ` Nick Piggin
2007-11-14 11:58 ` David Miller
2007-11-14 0:25 ` Nick Piggin
2007-11-14 21:30 ` Paul Mackerras
2007-11-14 10:17 ` Nick Piggin
2007-11-14 22:56 ` Chuck Ebbert
2007-11-14 11:03 ` Nick Piggin
2007-11-14 11:52 ` David Miller
2007-11-14 12:03 ` Paul Mackerras
2007-11-14 12:07 ` David Miller
2007-11-14 0:28 ` Nick Piggin
2007-11-14 21:50 ` Paul Mackerras
2007-11-14 23:03 ` David Miller
2007-11-14 23:12 ` Paul Mackerras
2007-11-14 23:21 ` David Miller
2007-11-15 1:11 ` Paul Mackerras
2007-11-15 1:27 ` David Miller
2007-11-15 2:34 ` Paul Mackerras
2007-11-15 7:48 ` Herbert Xu
2007-11-15 8:19 ` Andi Kleen
2007-11-19 13:08 ` David Miller
2007-11-19 20:53 ` Stephane Eranian
2007-11-20 0:55 ` David Miller
2007-11-19 21:43 ` Paul Mackerras
2007-11-19 22:48 ` Stephane Eranian
2007-11-20 0:53 ` David Miller
2007-12-13 16:00 ` Stephane Eranian
2007-12-14 19:12 ` Frank Ch. Eigler
2007-12-14 21:07 ` Stephane Eranian
2007-12-15 15:54 ` Frank Ch. Eigler
2007-11-15 8:29 ` [perfmon] " Stephane Eranian
2007-11-14 13:51 ` Stephane Eranian
2007-11-14 11:39 ` Paul Mackerras
2007-11-14 11:52 ` David Miller
2007-11-14 13:47 ` Stephane Eranian [this message]
2007-11-14 12:38 ` Andi Kleen
2007-11-14 14:13 ` Stephane Eranian
2007-11-14 14:26 ` Andi Kleen
2007-11-15 0:23 ` Paul Mackerras
2007-11-14 19:48 ` David Miller
2007-11-15 4:20 ` dean gaudet
2007-11-15 4:47 ` Paul Mackerras
2007-11-15 5:14 ` dean gaudet
2007-11-15 8:53 ` Stephane Eranian
2007-11-15 17:01 ` [perfmon2] [perfmon] " Dan Terpstra
2007-11-13 21:33 ` [perfmon] Re: [perfmon2] " Stephane Eranian
2007-11-13 21:45 ` Greg KH
2007-11-13 22:27 ` Christoph Hellwig
2007-11-13 20:42 ` Andi Kleen
2007-11-13 18:32 ` Stephane Eranian
2007-11-13 22:29 ` Christoph Hellwig
2007-11-16 18:25 ` PMC core internal API design 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=20071114134713.GD6557@frankl.hpl.hp.com \
--to=eranian@hpl.hp.com \
--cc=akpm@linux-foundation.org \
--cc=andi@firstfloor.org \
--cc=gregkh@suse.de \
--cc=hch@infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mucci@cs.utk.edu \
--cc=paulus@samba.org \
--cc=robert.richter@amd.com \
--cc=wcohen@redhat.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).