From: john stultz <johnstul@us.ibm.com>
To: lkml <linux-kernel@vger.kernel.org>
Cc: Andrea Arcangeli <andrea@suse.de>, Andi Kleen <ak@suse.de>,
Ulrich Drepper <drepper@redhat.com>,
Jamie Lokier <jamie@shareable.org>,
"Martin J. Bligh" <mbligh@aracnet.com>,
Wim Coekaerts <wim.coekaerts@oracle.com>,
Joel Becker <Joel.Becker@oracle.com>,
Chris McDermott <lcm@us.ibm.com>
Subject: [RFC][PATCH] vsyscall-gtod_B3 (0/3)
Date: Wed, 03 Mar 2004 16:11:22 -0800 [thread overview]
Message-ID: <1078359081.10076.191.camel@cog.beaverton.ibm.com> (raw)
All,
This is my port of the x86-64 vsyscall gettimeofday code to i386. This
patch moves gettimeofday into userspace, so it can be calledwithout the
syscall overhead, greatly improving performance. This is important for
any application, like a database, which heavily uses gettimeofday for
time-stamping. It supports both the TSC and IBM x44X cyclone time
source.
Example performance gain (using cyclone timesource):
int80 gettimeofday:
gettimeofday ( 1665576us / 1000000runs ) = 1.665574us
systenter gettimeofday:
gettimeofday ( 1239215us / 1000000runs ) = 1.239214us
vsyscall gettimeofday:
gettimeofday ( 875876us / 1000000runs ) = 0.875875us
I've broken the patch into three logical chuncks for clarity and to make
it easier to cherry pick the desired bits.
o Part 1: Renames variables in timer_cyclone.c and timer_tsc.c to avoid
conflicts in the global namespace.
o Part 2: Core vsyscall-gtod implementation.
o Part 3: vDSO hooks to avoid LD_PRELOADing or needing changes to glibc
Please let me know if you have any comments or suggestions.
thanks
-john
Existing issues:
----------------
o Bad pointers cause segfaults, rather then -EFAULT.
Release History:
----------------
B2 -> B3:
o Broke the patch up into 3 chunks
o Added vsyscall-int80.S hooks (4G disables SEP)
B1 -> B2:
o No LD_PRELOADing or changes to userspace required!
o removed hard-coded address in linker script
B0 -> B1:
o Cleaned up 4/4 split code, so no additional patch is needed.
o Fixed permissions on fixmapped cyclone pageo Improved
alternate_instruction workaround
o Use NTP variables to avoid related time inconsistencieso minor code
cleanups
next reply other threads:[~2004-03-04 0:11 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2004-03-04 0:11 john stultz [this message]
2004-03-04 0:12 ` [RFC][PATCH] linux-2.6.4-pre1_vsyscall-gtod_B3-part1 (1/3) john stultz
2004-03-04 0:13 ` [RFC][PATCH] linux-2.6.4-pre1_vsyscall-gtod_B3-part2 (2/3) john stultz
2004-03-04 0:14 ` [RFC][PATCH] linux-2.6.4-pre1_vsyscall-gtod_B3-part3 (3/3) john stultz
2004-03-04 0:55 ` Andrea Arcangeli
2004-03-04 2:16 ` Ulrich Drepper
2004-03-04 2:43 ` john stultz
2004-03-04 3:14 ` Andrea Arcangeli
2004-03-04 8:09 ` Ulrich Drepper
2004-03-04 19:02 ` john stultz
2004-03-04 2:47 ` Andrea Arcangeli
2004-03-04 2:54 ` john stultz
2004-03-04 3:15 ` Andrea Arcangeli
2004-03-04 8:57 ` Jakub Jelinek
2004-03-04 16:45 ` Andrea Arcangeli
2004-03-04 8:00 ` Jamie Lokier
2004-03-04 8:37 ` Jakub Jelinek
2004-03-04 17:48 ` Andrea Arcangeli
2004-03-04 0:15 ` [RFC] vsyscall-gtod_test_B3.tar.gz john stultz
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=1078359081.10076.191.camel@cog.beaverton.ibm.com \
--to=johnstul@us.ibm.com \
--cc=Joel.Becker@oracle.com \
--cc=ak@suse.de \
--cc=andrea@suse.de \
--cc=drepper@redhat.com \
--cc=jamie@shareable.org \
--cc=lcm@us.ibm.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mbligh@aracnet.com \
--cc=wim.coekaerts@oracle.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
all inboxes | Powered by JetHome®