netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "H. Peter Anvin" <hpa@zytor.com>
To: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Suresh Siddha <suresh.b.siddha@intel.com>,
	Ingo Molnar <mingo@elte.hu>,
	Wolfgang Walter <wolfgang.walter@stwm.de>,
	"netdev@vger.kernel.org" <netdev@vger.kernel.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"viro@ZenIV.linux.org.uk" <viro@zeniv.linux.org.uk>,
	"vegard.nossum@gmail.com" <vegard.nossum@gmail.com>,
	Thomas Gleixner <tglx@linutronix.de>
Subject: Re: Kernel oops with 2.6.26, padlock and ipsec: probably problem with fpu state changes
Date: Mon, 11 Aug 2008 17:42:21 -0700	[thread overview]
Message-ID: <48A0DC6D.8040409@zytor.com> (raw)
In-Reply-To: <20080812003938.GC18230@gondor.apana.org.au>

Herbert Xu wrote:
> On Mon, Aug 11, 2008 at 01:19:01PM -0700, Suresh Siddha wrote:
> .
>> we can't unconditionally do clts() in the process context. We have
>> to disable pre-emption to avoid interactions with context switch and
>> lazy restore. So there will be RT latency issues aswell.
> 
> Yes disabling preemption is the real killer.
> 
> This is just a quick band-aid.  Longer term we should add a task
> flag that indicates the task is currently doing kernel FPU which
> will tell the scheduler to clear TS the next time it's run.  That
> way we won't need to disable preemtion or pollute the user task's
> FPU used state.

That's not sufficient, though, because you have to track all the state 
and how it relates to everything.  You now have to track both the 
userspace FPU state and the potential kernel FPU state.  The VIA 
instructions are special (in the short bus to school sense) in that they 
use a mechanism intended to protect specific state to protect -- exactly 
nothing.

	-hpa

  reply	other threads:[~2008-08-12  0:47 UTC|newest]

Thread overview: 70+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-07-17 14:53 Kernel oops with 2.6.26 and ipsec (Was: Re: IPSEC in 2.6.25 causes stalled connections) Wolfgang Walter
2008-07-17 20:42 ` Kernel oops with 2.6.26 and ipsec Wolfgang Walter
     [not found] ` <200807301411.01622.wolfgang.walter@stwm.de>
     [not found]   ` <20080806103354.GA31623@gondor.apana.org.au>
2008-08-06 17:33     ` Kernel oops with 2.6.26, padlock and ipsec: probably problem with fpu state changes Wolfgang Walter
2008-08-06 20:14       ` Suresh Siddha
2008-08-06 21:21         ` Suresh Siddha
2008-08-07  0:38           ` Wolfgang Walter
2008-08-08  8:44           ` Wolfgang Walter
2008-08-08 18:31           ` Vegard Nossum
2008-08-07 16:23         ` Wolfgang Walter
2008-08-08 10:36           ` Wolfgang Walter
2008-08-08 18:53             ` Suresh Siddha
2008-08-08 19:01               ` H. Peter Anvin
2008-08-08 19:19                 ` Wolfgang Walter
2008-08-08 23:11                 ` Suresh Siddha
2008-08-09  0:38                   ` Herbert Xu
2008-08-09  1:23                     ` Herbert Xu
2008-08-09  1:54                     ` Wolfgang Walter
2008-08-09  2:16                       ` Herbert Xu
2008-08-09  3:09                         ` Wolfgang Walter
2008-08-09  3:20                           ` Herbert Xu
2008-08-09 14:29                     ` Herbert Xu
2008-08-09 14:32                       ` Herbert Xu
2008-08-09 17:52                         ` Suresh Siddha
2008-08-10  5:30                           ` Herbert Xu
2008-08-10  5:41                             ` H. Peter Anvin
2008-08-11 22:57                               ` Suresh Siddha
2008-08-09 17:48                     ` Suresh Siddha
2008-08-09  1:28                   ` Wolfgang Walter
2008-08-09 13:31                   ` Herbert Xu
2008-08-09 14:37                   ` Herbert Xu
2008-08-09 15:14                     ` Wolfgang Walter
2008-08-09 15:57                     ` Wolfgang Walter
2008-08-09 16:10                       ` H. Peter Anvin
2008-08-09 17:02                         ` Wolfgang Walter
2008-08-09 18:52                         ` Suresh Siddha
2008-08-09 19:37                           ` Suresh Siddha
2008-08-09 22:59                             ` Wolfgang Walter
2008-08-10  3:05                             ` Herbert Xu
2008-08-11 19:01                               ` Suresh Siddha
2008-08-11 19:22                                 ` Ingo Molnar
2008-08-11 19:24                                   ` H. Peter Anvin
2008-08-11 20:19                                     ` Suresh Siddha
2008-08-12  0:39                                       ` Herbert Xu
2008-08-12  0:42                                         ` H. Peter Anvin [this message]
2008-08-12  0:46                                           ` Herbert Xu
2008-08-12  0:48                                             ` H. Peter Anvin
2008-08-12  0:52                                               ` Herbert Xu
2008-08-12  0:38                                 ` Wolfgang Walter
2008-08-12 11:43                                 ` Wolfgang Walter
2008-08-12 12:02                                   ` Herbert Xu
2008-08-12 18:28                                     ` Suresh Siddha
2008-08-12 23:40                                       ` Herbert Xu
2008-08-09 18:12                       ` Suresh Siddha
2008-08-09 18:54                         ` Suresh Siddha
2008-08-09 16:05                     ` H. Peter Anvin
2008-08-09 18:14                       ` Suresh Siddha
2008-08-10  0:29                       ` Herbert Xu
2008-08-10  1:56                         ` Wolfgang Walter
2008-08-10  1:59                           ` Herbert Xu
2008-08-09 17:59                     ` Suresh Siddha
2008-08-10  1:40                       ` Herbert Xu
2008-08-09  1:49                 ` Herbert Xu
2008-08-09  1:59                   ` H. Peter Anvin
2008-08-09  2:43                   ` Wolfgang Walter
2008-08-09  3:30                     ` H. Peter Anvin
2008-08-09 10:50                       ` Wolfgang Walter
2008-08-08 19:09             ` Wolfgang Walter
2008-08-08 19:32               ` Suresh Siddha
2008-08-08 23:10                 ` Wolfgang Walter
2008-08-08 23:15                   ` Suresh Siddha

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=48A0DC6D.8040409@zytor.com \
    --to=hpa@zytor.com \
    --cc=herbert@gondor.apana.org.au \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=netdev@vger.kernel.org \
    --cc=suresh.b.siddha@intel.com \
    --cc=tglx@linutronix.de \
    --cc=vegard.nossum@gmail.com \
    --cc=viro@zeniv.linux.org.uk \
    --cc=wolfgang.walter@stwm.de \
    /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).