From: Ben Lau <benlau@linux.org.hk>
To: linux-kernel <linux-kernel@vger.kernel.org>
Subject: Re: The disappearing sys_call_table export.
Date: Wed, 07 May 2003 10:14:09 +0800 [thread overview]
Message-ID: <3EB86BF1.5010805@linux.org.hk> (raw)
In-Reply-To: <1052127216.2821.51.camel@pc-16.office.scali.no>
Hi,
I am interested with pt2, how NFS did for their syscall?
Terje Eggestad wrote:
> Unfortunately we live in an insane world.
>
> First of all, in the Changelog where the export was removed for 2.5.41
>
> http://www.kernel.org/pub/linux/kernel/v2.5/ChangeLog-2.5.41
>
> Arjan lists 4 reasons for having the export in the first place, and I'm
> on point 3. Here Arjan pretty much acknowledges that there is a
> legitimate need to have a event/hook system to be informed of a syscall.
> The exact quote is: "Eg the use of the export in this just a bandaid due
> to lack of a proper mechanism".
>
> My argument for *why* there should be a mechanism stops here.
>
>
> Since you're bright inquisitive: The exact problem I'm facing is pretty
> complex:
>
>
> 1. performance is everything.
> 2. We're making a MPI library, and as such we don't have any control
> with the application.
> 3a. The various hardware for cluster interconnect all work with DMA.
> 3b. the performance loss from copying from a receive area to the
> userspace buffer is unacceptable.
> 3c. It's therefore necessary for HW to access user pages.
> 4. In order to to 3, the user pages must be pinned down.
> 5. the way MPI is written, it's not using a special malloc() to allocate
> the send receive buffers. It can't since it would break language binding
> to fortran. Thus ANY writeable user page may be used.
> 6. point 4: pinning is VERY expensive (point 1), so I can't pin the
> buffers every time they're used.
> 7. The only way to cache buffers (to see if they're used before and
> hence pinned) is the user space virtual address. A syscall, thus ioctl
> to a device file is prohibitive expensive under point 1.
> 8a. if the app (glibc in practice, but you never know) use sbrk() with a
> negative arg, and then a positive argument, I can get a a different set
> of user pages with the same address.
> 8b ditto with a set of munmap()/mmap().
> 9. since the number of times. any 'realloc' may happen is << than the
> numbers of times any buffer may be used, it's necessary under point 1 to
> to trace changes to virtual addresses to phys pages, rather than test
> every time an address is being used.
> 10. kernel patches are impractical, I must be able to do this with std
> stock, redhat, AND suse kernels.
>
>
>
>
> On Mon, 2003-05-05 at 10:23, Christoph Hellwig wrote:
>
>>On Mon, May 05, 2003 at 10:19:45AM +0200, Terje Eggestad wrote:
>>
>>>Now that it seem that all are in agreement that the sys_call_table
>>>symbol shall not be exported to modules, are there any work in progress
>>>to allow modules to get an event/notification whenever a specific
>>>syscall is being called?
>>
>>No.
>>
>>
>>>We have a specific need to trace mmap() and sbrk() calls.
>>
>>Well, you get mmap events for your driver and I can't imagine a sane
>>reason for intwercepting sbrk(). Do you have a pointer to the driver
>>source doing such strange things?
next prev parent reply other threads:[~2003-05-07 2:02 UTC|newest]
Thread overview: 207+ messages / expand[flat|nested] mbox.gz Atom feed top
2003-05-05 8:19 The disappearing sys_call_table export Terje Eggestad
2003-05-05 8:23 ` Christoph Hellwig
2003-05-05 9:33 ` Terje Eggestad
2003-05-05 9:38 ` Arjan van de Ven
2003-05-05 10:12 ` Terje Eggestad
2003-05-05 10:25 ` Christoph Hellwig
2003-05-05 11:23 ` Terje Eggestad
2003-05-05 11:27 ` Arjan van de Ven
2003-05-05 11:31 ` Terje Eggestad
2003-05-05 11:33 ` Arjan van de Ven
2003-05-05 15:53 ` Tigran Aivazian
2003-05-05 14:57 ` Christoph Hellwig
2003-05-05 14:59 ` Arjan van de Ven
2003-05-05 12:52 ` Christoph Hellwig
2003-05-05 13:41 ` Terje Eggestad
2003-05-05 13:43 ` Christoph Hellwig
2003-05-05 13:50 ` Terje Eggestad
2003-05-05 13:54 ` Arjan van de Ven
2003-05-05 13:55 ` Christoph Hellwig
2003-05-05 14:28 ` Carl-Daniel Hailfinger
2003-05-05 14:34 ` Christoph Hellwig
2003-05-05 15:25 ` Carl-Daniel Hailfinger
2003-05-06 7:30 ` Eric W. Biederman
2003-05-06 8:14 ` Terje Eggestad
2003-05-06 9:21 ` Eric W. Biederman
2003-05-06 11:21 ` Terje Eggestad
2003-05-06 11:37 ` Eric W. Biederman
2003-05-06 12:08 ` Terje Eggestad
2003-05-05 11:16 ` Alan Cox
2003-05-05 13:23 ` Terje Eggestad
2003-05-08 12:25 ` Terje Malmedal
2003-05-08 12:29 ` Christoph Hellwig
2003-05-08 13:18 ` Terje Malmedal
2003-05-08 14:25 ` Christoph Hellwig
2003-05-08 15:29 ` Terje Malmedal
2003-05-08 18:13 ` Jesse Pollard
2003-05-08 19:17 ` Christoph Hellwig
2003-05-09 9:18 ` Terje Malmedal
2003-05-08 14:58 ` Alan Cox
2003-05-09 8:56 ` Terje Malmedal
2003-05-07 2:14 ` Ben Lau [this message]
2003-05-05 8:27 ` Arjan van de Ven
2003-05-05 9:01 ` Dmitry A. Fedorov
2003-05-05 9:19 ` Christoph Hellwig
2003-05-05 9:32 ` Arjan van de Ven
2003-05-05 13:30 Dmitry A. Fedorov
2003-05-05 13:42 ` Christoph Hellwig
2003-05-05 14:46 ` Dmitry A. Fedorov
2003-05-05 13:45 ` viro
2003-05-05 14:29 ` Dmitry A. Fedorov
[not found] <mailman.1052142720.4060.linux-kernel2news@redhat.com>
2003-05-05 20:50 ` Pete Zaitcev
2003-05-06 2:17 ` Dmitry A. Fedorov
2003-05-05 21:29 Chuck Ebbert
2003-05-05 22:49 ` Terje Eggestad
2003-05-06 2:23 ` Dmitry A. Fedorov
2003-05-06 7:27 ` Terje Eggestad
2003-05-06 8:21 ` Dmitry A. Fedorov
2003-05-06 8:45 Yoav Weiss
2003-05-06 9:15 ` David S. Miller
2003-05-06 19:45 ` David Schwartz
2003-05-06 10:06 ` Dmitry A. Fedorov
2003-05-06 17:01 ` Jerry Cooperstein
2003-05-06 17:45 ` Yoav Weiss
2003-05-06 15:51 Yoav Weiss
2003-05-06 20:48 Chuck Ebbert
2003-05-07 15:34 petter wahlman
2003-05-07 15:48 ` Arjan van de Ven
2003-05-07 16:00 ` Richard B. Johnson
2003-05-07 16:08 ` petter wahlman
2003-05-07 16:45 ` Richard B. Johnson
2003-05-07 16:59 ` Richard B. Johnson
2003-05-07 18:07 ` petter wahlman
2003-05-07 18:33 ` Richard B. Johnson
2003-05-08 8:58 ` petter wahlman
2003-05-08 15:11 ` Richard B. Johnson
2003-05-07 21:27 ` Jesse Pollard
2003-05-07 17:21 ` Jesse Pollard
2003-05-07 16:18 ` Steffen Persvold
2003-05-08 12:23 ` Eric W. Biederman
2003-05-07 19:04 Chuck Ebbert
2003-05-08 9:58 ` Terje Eggestad
2003-05-08 9:59 ` Arjan van de Ven
2003-05-08 10:20 ` viro
2003-05-08 12:54 ` Terje Eggestad
2003-05-08 12:58 ` Christoph Hellwig
2003-05-08 19:10 ` Shachar Shemesh
2003-05-08 19:15 ` Christoph Hellwig
2003-05-08 21:48 ` J.A. Magallon
2003-05-09 7:43 ` Muli Ben-Yehuda
2003-05-09 7:42 ` Muli Ben-Yehuda
2003-05-09 8:08 ` Greg KH
2003-05-09 19:07 ` Muli Ben-Yehuda
2003-05-08 14:08 Chuck Ebbert
2003-05-08 14:36 ` Christoph Hellwig
2003-05-08 14:42 ` Alan Cox
2003-05-08 14:56 ` Jesse Pollard
2003-05-08 15:22 ` Alan Cox
2003-05-08 17:02 ` William Stearns
2003-05-08 18:28 ` Jesse Pollard
2003-05-10 14:38 ` Ahmed Masud
2003-05-10 16:50 ` Arjan van de Ven
2003-05-10 17:51 ` Ahmed Masud
2003-05-10 17:56 ` Arjan van de Ven
2003-05-10 18:03 ` Ahmed Masud
2003-05-10 18:09 ` Ahmed Masud
2003-05-10 18:43 ` Werner Almesberger
2003-05-10 18:26 ` Werner Almesberger
2003-05-11 11:01 ` Terje Malmedal
2003-05-11 11:57 ` Ahmed Masud
2003-05-08 19:43 Chuck Ebbert
2003-05-08 19:48 ` Christoph Hellwig
2003-05-08 21:44 ` Alan Cox
2003-05-08 19:43 Chuck Ebbert
2003-05-08 19:58 ` Christoph Hellwig
2003-05-09 13:53 ` Jesse Pollard
2003-05-09 14:37 ` Ragnar =?unknown-8bit?Q?Kj=F8rstad?=
2003-05-12 14:19 ` Jesse Pollard
2003-05-12 15:56 ` Christoph Hellwig
2003-05-08 19:43 Chuck Ebbert
2003-05-09 7:50 Chuck Ebbert
2003-05-09 7:59 ` Christoph Hellwig
2003-05-09 12:18 ` Alan Cox
2003-05-09 17:07 ` Valdis.Kletnieks
2003-05-10 15:34 ` Alan Cox
2003-05-09 7:50 Chuck Ebbert
2003-05-09 7:57 ` Christoph Hellwig
2003-05-09 9:11 Chuck Ebbert
2003-05-09 10:47 ` Christoph Hellwig
2003-05-09 9:43 Chuck Ebbert
2003-05-09 11:09 Chuck Ebbert
2003-05-09 12:41 Chuck Ebbert
2003-05-09 12:47 ` Christoph Hellwig
2003-05-09 17:07 Chuck Ebbert
2003-05-09 17:07 Chuck Ebbert
2003-05-09 18:27 ` Richard B. Johnson
2003-05-09 19:02 ` Valdis.Kletnieks
2003-05-09 19:18 ` Richard B. Johnson
2003-05-09 19:25 ` Valdis.Kletnieks
2003-05-09 21:22 Chuck Ebbert
2003-05-10 19:18 Yoav Weiss
2003-05-10 19:53 ` Muli Ben-Yehuda
2003-05-10 20:06 ` Yoav Weiss
2003-05-11 3:54 ` Ahmed Masud
2003-05-10 20:48 ` David Wagner
2003-05-10 19:32 Chuck Ebbert
2003-05-10 21:45 Yoav Weiss
2003-05-11 16:32 Chuck Ebbert
2003-05-11 17:20 ` David Wagner
2003-05-11 17:53 ` Yoav Weiss
2003-05-11 20:39 Chuck Ebbert
2003-05-11 22:32 ` Yoav Weiss
2003-05-11 21:46 ` Alan Cox
2003-05-11 22:57 ` David Schwartz
2003-05-14 21:08 ` H. Peter Anvin
2003-05-11 23:22 ` Yoav Weiss
2003-05-11 22:32 ` Ahmed Masud
[not found] <20030511164010$5d34@gated-at.bofh.it>
2003-05-12 0:47 ` Ben Pfaff
2003-05-12 16:32 Chuck Ebbert
2003-05-12 16:46 ` Alan Cox
[not found] <20030512164017$6c09@gated-at.bofh.it>
2003-05-12 17:02 ` Pascal Schmidt
2003-05-12 21:51 Chuck Ebbert
2003-05-12 21:05 ` Alan Cox
2003-05-12 22:12 ` Valdis.Kletnieks
2003-05-12 21:19 ` Alan Cox
2003-05-12 22:29 ` Valdis.Kletnieks
2003-05-13 12:31 ` Ahmed Masud
2003-05-12 22:57 Yoav Weiss
2003-05-12 23:58 ` Bryan Andersen
2003-05-13 12:11 ` Jesse Pollard
2003-05-13 13:44 ` Yoav Weiss
2003-05-13 21:26 ` Jesse Pollard
2003-05-13 22:21 ` Yoav Weiss
2003-05-14 13:05 ` Jesse Pollard
2003-05-13 1:57 Chuck Ebbert
2003-05-13 2:25 ` Yoav Weiss
2003-05-13 1:57 Chuck Ebbert
2003-05-13 12:24 ` Jesse Pollard
2003-05-13 9:52 Chuck Ebbert
2003-05-13 13:32 ` Yoav Weiss
2003-05-14 7:44 ` Mike Touloumtzis
2003-05-14 10:34 ` Ahmed Masud
2003-05-14 20:58 ` Mike Touloumtzis
2003-05-14 21:32 ` Richard B. Johnson
2003-05-14 21:37 ` Yoav Weiss
2003-05-14 21:51 ` Richard B. Johnson
2003-05-15 13:17 ` Jesse Pollard
2003-05-15 15:16 ` Chris Ricker
2003-05-15 15:31 ` Richard B. Johnson
2003-05-15 15:33 ` Chris Ricker
2003-05-15 15:46 ` Richard B. Johnson
2003-05-15 16:21 ` Ahmed Masud
2003-05-15 2:06 ` Ahmed Masud
2003-05-13 13:58 Yoav Weiss
2003-05-13 22:51 ` Ahmed Masud
2003-05-13 23:58 ` Yoav Weiss
2003-06-12 23:20 ` Nigel Cunningham
2003-06-15 22:37 ` Yoav Weiss
2003-05-13 14:45 Chuck Ebbert
2003-05-13 21:32 ` Jesse Pollard
2003-05-13 14:45 Chuck Ebbert
2003-05-13 19:00 ` jjs
2003-05-13 21:44 ` Jesse Pollard
2003-05-14 8:41 Chuck Ebbert
2003-05-14 23:24 Chuck Ebbert
2003-05-15 0:49 ` David Schwartz
2003-05-15 8:16 Chuck Ebbert
2003-05-16 16:15 Chuck Ebbert
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=3EB86BF1.5010805@linux.org.hk \
--to=benlau@linux.org.hk \
--cc=linux-kernel@vger.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;
as well as URLs for NNTP newsgroup(s).