linux-kernel - Re: [GIT pull] timer fixes for 2.6.27

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] [day] [month] [year] [list]
Date:	Sat, 6 Sep 2008 18:20:12 -0400
From:	"Jochen Voß" <jochen.voss@...glemail.com>
To:	"Thomas Gleixner" <tglx@...utronix.de>
Cc:	"Linus Torvalds" <torvalds@...ux-foundation.org>,
	"Andrew Morton" <akpm@...ux-foundation.org>,
	LKML <linux-kernel@...r.kernel.org>,
	"Ingo Molnar" <mingo@...e.hu>,
	"Venkatesch Pallipadi" <venkatesh.pallipadi@...el.com>
Subject: Re: [GIT pull] timer fixes for 2.6.27

Hi,

2008/9/6 Thomas Gleixner <tglx@...utronix.de>:
> Dominik Brodowski (2):
>      [...]
>      clocksource, acpi_pm.c: check for monotonicity
I made some comments about this patch which were never answered:

    http://lkml.org/lkml/2008/8/23/28

Not sure whether this was because they were irrelevant or overlooked.
Just to make sure, I replicate them below:

> diff --git a/drivers/clocksource/acpi_pm.c b/drivers/clocksource/acpi_pm.c
> index 5ca1d80..4eee533 100644
> --- a/drivers/clocksource/acpi_pm.c
> +++ b/drivers/clocksource/acpi_pm.c
> @@ -21,6 +21,7 @@
>  #include <linux/errno.h>
>  #include <linux/init.h>
>  #include <linux/pci.h>
> +#include <linux/delay.h>
>  #include <asm/io.h>
>
>  /*
> @@ -151,13 +152,13 @@ DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_SERVERWORKS, PCI_DEVICE_ID_SERVERWORKS_LE,
>  */
>  static int verify_pmtmr_rate(void)
>  {
> -       u32 value1, value2;
> +       cycle_t value1, value2;
>        unsigned long count, delta;
>
>        mach_prepare_counter();
> -       value1 = read_pmtmr();
> +       value1 = clocksource_acpi_pm.read();
>        mach_countup(&count);
> -       value2 = read_pmtmr();
> +       value2 = clocksource_acpi_pm.read();
>        delta = (value2 - value1) & ACPI_PM_MASK;
>
>        /* Check that the PMTMR delta is within 5% of what we expect */
> @@ -175,10 +176,13 @@ static int verify_pmtmr_rate(void)
>  #define verify_pmtmr_rate() (0)
>  #endif
>
> +/* Number of monotonicity checks to perform during initialization */
> +#define ACPI_PM_MONOTONICITY_CHECKS 10
> +
>  static int __init init_acpi_pm_clocksource(void)
>  {
> -       u32 value1, value2;
> -       unsigned int i;
> +       cycle_t value1, value2;
> +       unsigned int i, j, good = 0;
>
>        if (!pmtmr_ioport)
>                return -ENODEV;
> @@ -187,24 +191,32 @@ static int __init init_acpi_pm_clocksource(void)
>                                                clocksource_acpi_pm.shift);
>
>        /* "verify" this timing source: */
> -       value1 = read_pmtmr();
> -       for (i = 0; i < 10000; i++) {
> -               value2 = read_pmtmr();
> -               if (value2 == value1)
> -                       continue;
> -               if (value2 > value1)
> -                       goto pm_good;
> -               if ((value2 < value1) && ((value2) < 0xFFF))
> -                       goto pm_good;
> -               printk(KERN_INFO "PM-Timer had inconsistent results:"
> -                       " 0x%#x, 0x%#x - aborting.\n", value1, value2);
> -               return -EINVAL;
> +       for (j = 0; j < ACPI_PM_MONOTONICITY_CHECKS; j++) {
> +               value1 = clocksource_acpi_pm.read();
> +               for (i = 0; i < 10000; i++) {
> +                       value2 = clocksource_acpi_pm.read();
> +                       if (value2 == value1)
> +                               continue;
> +                       if (value2 > value1)
> +                               good++;
> +                               break;
> +                       if ((value2 < value1) && ((value2) < 0xFFF))
The brackets arout value2 are not needed and look strange.

> +                               good++;
> +                               break;
> +                       printk(KERN_INFO "PM-Timer had inconsistent results:"
> +                              " 0x%#llx, 0x%#llx - aborting.\n",
> +                              value1, value2);
> +                       return -EINVAL;
> +               }
> +               udelay(300 * i);
300*10000 microseconds seems like a long time to me.  Is this the
intended maximal delay?

> +       }
> +
> +       if (good != ACPI_PM_MONOTONICITY_CHECKS) {
> +               printk(KERN_INFO "PM-Timer failed consistency check "
> +                      " (0x%#llx) - aborting.\n", value1);
> +               return -ENODEV;
If the inner loop runs out once, you alreay know that you will later
abort here.  Maybe move the check directly after the inner loop to
avoid the additional delay (10*10000*300 microseconds = 30 seconds) in
case of failure?

>        }
> -       printk(KERN_INFO "PM-Timer had no reasonable result:"
> -                       " 0x%#x - aborting.\n", value1);
> -       return -ENODEV;
>
> -pm_good:
>        if (verify_pmtmr_rate() != 0)
>                return -ENODEV;

I hope this helps,
Jochen
-- 
http://seehuhn.de/
--
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/