linux-kernel - Re: [PATCH] x86: cleanups to nmi

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Sun, 08 Jun 2008 18:23:38 +0200
From:	Németh Márton <nm127@...email.hu>
To:	Cyrill Gorcunov <gorcunov@...il.com>
CC:	Thomas Gleixner <tglx@...utronix.de>,
	Ingo Molnar <mingo@...hat.com>,
	LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] x86: cleanups to nmi_watchdog variable

Cyrill Gorcunov wrote:
> | From: Márton Németh <nm127@...email.hu>
> | 
> | Use NMI_NONE instead of 0. As the symbol NMI_DISABLED equals to -1,
> | the nmi_watchdog variable should also be signed.
> | 
> | Signed-off-by: Márton Németh <nm127@...email.hu>
> | ---
> | diff -upr linux-2.6.26-rc5.orig/arch/x86/kernel/io_apic_32.c linux-2.6.26-rc5/arch/x86/kernel/io_apic_32.c
> | --- linux-2.6.26-rc5.orig/arch/x86/kernel/io_apic_32.c	2008-06-07 12:35:24.000000000 +0200
> | +++ linux-2.6.26-rc5/arch/x86/kernel/io_apic_32.c	2008-06-07 12:40:20.000000000 +0200
> | @@ -2217,7 +2217,7 @@ static inline void __init check_timer(vo
> | 
> |  	if (nmi_watchdog == NMI_IO_APIC) {
> |  		printk(KERN_WARNING "timer doesn't work through the IO-APIC - disabling NMI Watchdog!\n");
> | -		nmi_watchdog = 0;
> | +		nmi_watchdog = NMI_NONE;
> |  	}
> | 
> |  	printk(KERN_INFO "...trying to set up timer as Virtual Wire IRQ...");
> | diff -upr linux-2.6.26-rc5.orig/arch/x86/kernel/io_apic_64.c linux-2.6.26-rc5/arch/x86/kernel/io_apic_64.c
> | --- linux-2.6.26-rc5.orig/arch/x86/kernel/io_apic_64.c	2008-06-05 18:39:49.000000000 +0200
> | +++ linux-2.6.26-rc5/arch/x86/kernel/io_apic_64.c	2008-06-07 12:37:01.000000000 +0200
> | @@ -1735,7 +1735,7 @@ static inline void __init check_timer(vo
> | 
> |  	if (nmi_watchdog == NMI_IO_APIC) {
> |  		printk(KERN_WARNING "timer doesn't work through the IO-APIC - disabling NMI Watchdog!\n");
> | -		nmi_watchdog = 0;
> | +		nmi_watchdog = NMI_NONE;
> |  	}
> | 
> |  	apic_printk(APIC_VERBOSE, KERN_INFO "...trying to set up timer as Virtual Wire IRQ...");
> | diff -upr linux-2.6.26-rc5.orig/arch/x86/kernel/nmi_32.c linux-2.6.26-rc5/arch/x86/kernel/nmi_32.c
> | --- linux-2.6.26-rc5.orig/arch/x86/kernel/nmi_32.c	2008-06-05 18:39:49.000000000 +0200
> | +++ linux-2.6.26-rc5/arch/x86/kernel/nmi_32.c	2008-06-07 11:19:55.000000000 +0200
> | @@ -43,7 +43,7 @@ static cpumask_t backtrace_mask = CPU_MA
> |   */
> |  atomic_t nmi_active = ATOMIC_INIT(0);		/* oprofile uses this */
> | 
> | -unsigned int nmi_watchdog = NMI_DEFAULT;
> | +int nmi_watchdog = NMI_DEFAULT;
> |  static unsigned int nmi_hz = HZ;
> | 
> |  static DEFINE_PER_CPU(short, wd_enabled);
> | diff -upr linux-2.6.26-rc5.orig/arch/x86/kernel/nmi_64.c linux-2.6.26-rc5/arch/x86/kernel/nmi_64.c
> | --- linux-2.6.26-rc5.orig/arch/x86/kernel/nmi_64.c	2008-06-05 18:39:49.000000000 +0200
> | +++ linux-2.6.26-rc5/arch/x86/kernel/nmi_64.c	2008-06-07 11:20:00.000000000 +0200
> | @@ -43,7 +43,7 @@ static cpumask_t backtrace_mask = CPU_MA
> |  atomic_t nmi_active = ATOMIC_INIT(0);		/* oprofile uses this */
> |  static int panic_on_timeout;
> | 
> | -unsigned int nmi_watchdog = NMI_DEFAULT;
> | +int nmi_watchdog = NMI_DEFAULT;
> |  static unsigned int nmi_hz = HZ;
> | 
> |  static DEFINE_PER_CPU(short, wd_enabled);
> | diff -upr linux-2.6.26-rc5.orig/include/asm-x86/nmi.h linux-2.6.26-rc5/include/asm-x86/nmi.h
> | --- linux-2.6.26-rc5.orig/include/asm-x86/nmi.h	2008-06-05 18:40:12.000000000 +0200
> | +++ linux-2.6.26-rc5/include/asm-x86/nmi.h	2008-06-07 11:21:20.000000000 +0200
> | @@ -61,7 +61,7 @@ extern void enable_timer_nmi_watchdog(vo
> |  extern int nmi_watchdog_tick(struct pt_regs *regs, unsigned reason);
> | 
> |  extern atomic_t nmi_active;
> | -extern unsigned int nmi_watchdog;
> | +extern int nmi_watchdog;
> |  #define NMI_DISABLED    -1
> |  #define NMI_NONE	0
> |  #define NMI_IO_APIC	1
> | 
> 
> We should leave nmi_watchdog as unsigned. The only snippet which
> could be reviewed as controversy point  - is touch_nmi_watchdog()
> where we just test nmi_watchdog for being greater than 0. Actually
> there NMI_NONE should be used instead. But frankly some additional
> simplification is needed.

Don't forget that the NMI_DEFAULT symbol in include/asm-x86/nmi.h also equals to
NMI_DISABLED, which is -1.

	Márton Németh

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/