From: kuznet@ms2.inr.ac.ru
To: george@mvista.com (george anzinger)
Cc: ak@suse.de, netdev@oss.sgi.com, linux-net@vger.kernel.org,
davem@redhat.com, ak@muc.de, pekkas@netcore.fi
Subject: Re: System crash in tcp_fragment()
Date: Tue, 21 May 2002 04:41:39 +0400 (MSD) [thread overview]
Message-ID: <acc57g$h3q$2@main.gmane.org> (raw)
In-Reply-To: <3CE9960D.15D41380@mvista.com> from "george anzinger" at May 20, 2 05:34:21 pm
Hello!
> > if [ "$CONFIG_SMP" = "n" ]; then
> > bool 'Preemptible Kernel' CONFIG_PREEMPT
> > fi
> >
> That is not a fix! It is dodging the issue. ;)
It is the only real fix, if you are going to follow this instruction:
+RULE #1: Per-CPU data structures need explicit protection
+
+
+Two similar problems arise. An example code snippet:
+
+ struct this_needs_locking tux[NR_CPUS];
+ tux[smp_processor_id()] = some_value;
+ /* task is preempted here... */
+ something = tux[smp_processor_id()];
+
+First, since the data is per-CPU, it may not have explicit SMP locking, but
+require it otherwise. Second, when a preempted task is finally rescheduled,
+the previous value of smp_processor_id may not equal the current. You must
+protect these situations by disabling preemption around them.
If you are not going to break all the kernel just make sure that
tasks preempted in the kernel do not migrate. That's all, simple & stupid.
Alexey
next prev parent reply other threads:[~2002-05-21 0:41 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <3CE9960D.15D41380@mvista.com>
[not found] ` <200205210041.EAA04407@sex.inr.ac.ru>
2002-05-21 0:34 ` System crash in tcp_fragment() David S. Miller
2002-05-21 0:41 ` kuznet [this message]
[not found] <Pine.LNX.4.33.0205201836160.9301-100000@w-nivedita2.des.beaverton.ibm.com>
[not found] ` <3CE9E466.AC2358EE@mvista.com>
2002-05-21 6:00 ` David S. Miller
[not found] ` <20020520.230021.29510217.davem@redhat.com>
2002-05-21 7:25 ` george anzinger
2002-05-21 9:49 ` Andi Kleen
[not found] ` <3CE9F679.90ACF597@mvista.com>
2002-05-21 7:22 ` David S. Miller
2002-05-21 12:47 ` kuznet
2002-05-21 15:42 ` george anzinger
2002-05-21 12:54 ` Andi Kleen
2002-05-21 6:08 ` george anzinger
[not found] <20020520.173416.105610032.davem@redhat.com>
2002-05-21 1:00 ` kuznet
2002-05-21 1:49 ` Nivedita Singhvi
[not found] <20020521015407.A1296@wotan.suse.de>
2002-05-21 0:11 ` kuznet
2002-05-21 0:20 ` Andi Kleen
2002-05-21 0:26 ` george anzinger
[not found] ` <20020521022007.A6248@wotan.suse.de>
2002-05-21 0:34 ` george anzinger
[not found] ` <3CE99434.20E7479C@mvista.com>
2002-05-21 0:18 ` David S. Miller
2002-05-21 0:39 ` Andi Kleen
[not found] <3CE95190.75C52E2D@mvista.com>
2002-05-20 20:29 ` Andi Kleen
[not found] ` <20020520222937.A1467@averell>
2002-05-20 21:18 ` george anzinger
2002-05-20 21:25 ` kuznet
2002-05-20 22:08 ` David S. Miller
[not found] ` <200205202125.BAA03545@sex.inr.ac.ru>
2002-05-20 23:01 ` george anzinger
2002-05-20 23:54 ` Andi Kleen
2002-05-20 19:42 george anzinger
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='acc57g$h3q$2@main.gmane.org' \
--to=kuznet@ms2.inr.ac.ru \
--cc=ak@muc.de \
--cc=ak@suse.de \
--cc=davem@redhat.com \
--cc=george@mvista.com \
--cc=linux-net@vger.kernel.org \
--cc=netdev@oss.sgi.com \
--cc=pekkas@netcore.fi \
/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).