Re: [PATCH] drm/i915: Handle full s64 precision for wait-ioctl

From: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
To: Chris Wilson <chris@chris-wilson.co.uk>, intel-gfx@lists.freedesktop.org
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Subject: Re: [PATCH] drm/i915: Handle full s64 precision for wait-ioctl
Date: Fri, 11 Aug 2017 10:18:04 +0300	[thread overview]
Message-ID: <1502435884.3582.4.camel@linux.intel.com> (raw)
In-Reply-To: <150196245251.19198.17738666772629554772@mail.alporthouse.com>

On la, 2017-08-05 at 20:47 +0100, Chris Wilson wrote:
> Quoting Chris Wilson (2017-08-05 20:19:24)

<SNIP>

> > +++ b/drivers/gpu/drm/i915/i915_drv.h
> > @@ -4144,6 +4144,12 @@ static inline unsigned long msecs_to_jiffies_timeout(const unsigned int m)
> >  
> >  static inline unsigned long nsecs_to_jiffies_timeout(const u64 n)
> >  {
> > +#if NSEC_PER_SEC % HZ
> > +       /* nsecs_to_jiffies64() does not guard against overflow */
> > +       if (n >= (u64)MAX_JIFFY_OFFSET * NSEC_PER_SEC / HZ)
> > +               return MAX_JIFFY_OFFSET;
> 
> This still overflows, we need n / NSEC_PER_SEC >= MAX_JIFFY_OFFSET / HZ
> as MAX_JIFFY_OFFSET is ~LONG_MAX/2
> 
> Hmm, so div_u64(n, NSEC_PER_SEC) >= MAX_JIFFY_OFFSET / HZ ?

This reflects the original test best.

Regards, Joonas
-- 
Joonas Lahtinen
Open Source Technology Center
Intel Corporation
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx