public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Mathieu Desnoyers <compudj@krystal.dyndns.org>
To: Andi Kleen <ak@suse.de>
Cc: jeremy@goop.org, patches@x86-64.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] [3/11] x86: Stop MCEs and NMIs during code patching
Date: Fri, 20 Jul 2007 12:03:47 -0400	[thread overview]
Message-ID: <20070720160347.GB1960@Krystal> (raw)
In-Reply-To: <20070720153247.75CAD14E7C@wotan.suse.de>

* Andi Kleen (ak@suse.de) wrote:

> Index: linux/arch/x86_64/kernel/nmi.c
> ===================================================================
> --- linux.orig/arch/x86_64/kernel/nmi.c
> +++ linux/arch/x86_64/kernel/nmi.c
> @@ -384,11 +384,14 @@ int __kprobes nmi_watchdog_tick(struct p
>  	return rc;
>  }
>  
> +static unsigned ignore_nmis;
> +
>  asmlinkage __kprobes void do_nmi(struct pt_regs * regs, long error_code)
>  {
>  	nmi_enter();
>  	add_pda(__nmi_count,1);
> -	default_do_nmi(regs);
> +	if (!ignore_nmis) 
> +		default_do_nmi(regs);
>  	nmi_exit();
>  }
>  
> @@ -401,6 +404,18 @@ int do_nmi_callback(struct pt_regs * reg
>  	return 0;
>  }
>  
> +void stop_nmi(void)
> +{
> +	ignore_nmis++;
> +	acpi_nmi_disable();
> +}

I would do:

acpi_nmi_disable();
ignore_nmis++;

Instead, so we minimize the chances to have to drop a do_nmi.

> +
> +void restart_nmi(void)
> +{
> +	ignore_nmis--;
> +	acpi_nmi_enable();
> +}
> +
>  #ifdef CONFIG_SYSCTL
>  
...
>  #endif /* ASM_NMI_H */
> Index: linux/arch/i386/kernel/traps.c
> ===================================================================
> --- linux.orig/arch/i386/kernel/traps.c
> +++ linux/arch/i386/kernel/traps.c
> @@ -774,6 +774,8 @@ static __kprobes void default_do_nmi(str
>  	reassert_nmi();
>  }
>  
> +static int ignore_nmis;
> +
>  fastcall __kprobes void do_nmi(struct pt_regs * regs, long error_code)
>  {
>  	int cpu;
> @@ -784,11 +786,24 @@ fastcall __kprobes void do_nmi(struct pt
>  
>  	++nmi_count(cpu);
>  
> -	default_do_nmi(regs);
> +	if (!ignore_nmis)
> +		default_do_nmi(regs);
>  
>  	nmi_exit();
>  }
>  
> +void stop_nmi(void)
> +{
> +	ignore_nmis++;
> +	acpi_nmi_disable();
> +}
> +

Same here, I would swap the two.

> +void restart_nmi(void)
> +{
> +	ignore_nmis--;
> +	acpi_nmi_enable();
> +}
> +

Besides that, it looks good.

Mathieu

-- 
Mathieu Desnoyers
Computer Engineering Ph.D. Student, Ecole Polytechnique de Montreal
OpenPGP key fingerprint: 8CD5 52C3 8E3C 4140 715F  BA06 3F25 A8FE 3BAE 9A68

  parent reply	other threads:[~2007-07-20 16:04 UTC|newest]

Thread overview: 25+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-07-20 15:32 [PATCH] [0/11] Some more x86 patches for review Andi Kleen
2007-07-20 15:32 ` [PATCH] [1/11] i386: Reserve the right performance counter for the Intel PerfMon NMI watchdog Andi Kleen
2007-07-20 15:32 ` [PATCH] [2/11] x86: Fix alternatives and kprobes to remap write-protected kernel text Andi Kleen
2007-07-20 15:59   ` Mathieu Desnoyers
2007-07-23 14:46   ` Mathieu Desnoyers
2007-07-23 15:01     ` Andi Kleen
2007-07-23 15:25       ` Mathieu Desnoyers
2007-07-24  6:06   ` Rusty Russell
2007-07-24 13:57     ` Mathieu Desnoyers
2007-07-24 15:17       ` Jeremy Fitzhardinge
2007-07-24 15:20         ` Mathieu Desnoyers
2007-07-20 15:32 ` [PATCH] [3/11] x86: Stop MCEs and NMIs during code patching Andi Kleen
2007-07-20 15:45   ` Jeremy Fitzhardinge
2007-07-20 16:03   ` Mathieu Desnoyers [this message]
2007-07-20 15:32 ` [PATCH] [4/11] x86_64: Set K8 CPUID flag for K8/Fam10h/Fam11h Andi Kleen
2007-07-20 15:32 ` [PATCH] [5/11] x86: Unify alternative nop handling between i386 and x86-64 Andi Kleen
2007-07-20 15:32 ` [PATCH] [6/11] i386: Tune AMD Fam10h/11h like K8 Andi Kleen
2007-07-20 15:32 ` [PATCH] [7/11] i386: Do not include other cpus' interrupt 0 in nmi_watchdog Andi Kleen
2007-07-20 15:32 ` [PATCH] [8/11] x86_64: x86_64 - Use non locked version for local_cmpxchg() Andi Kleen
2007-07-20 15:32 ` [PATCH] [9/11] x86: Replace NSC/Cyrix specific chipset access macros by inlined functions Andi Kleen
2007-07-21 12:00   ` Alexey Dobriyan
2007-07-21 12:04     ` Andi Kleen
2007-07-22 10:45     ` Juergen Beisert
2007-07-20 15:32 ` [PATCH] [10/11] i386: Handle P6s without performance counters in nmi watchdog Andi Kleen
2007-07-20 15:32 ` [PATCH] [11/11] i386: Use patchable lock prefix in set_64bit Andi Kleen

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=20070720160347.GB1960@Krystal \
    --to=compudj@krystal.dyndns.org \
    --cc=ak@suse.de \
    --cc=jeremy@goop.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=patches@x86-64.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