Re: gcc7 log2 compile issues in kernel/time/timekeeping.c

2017-03-02 Thread Linus Torvalds
I can't take this compiler braindamage any more, so I just decided to remove the ilog2_NaN checks, and return 0 for any constant value < 2. It's commit 474c90156c8d ("give up on gcc ilog2() constant optimizations") in my tree. I'll push it out later after my usual build tests (and I have a few

Re: gcc7 log2 compile issues in kernel/time/timekeeping.c

2017-03-02 Thread Markus Trippelsdorf
On 2017.03.01 at 17:39 +, Ard Biesheuvel wrote: > On 1 March 2017 at 00:00, Laura Abbott wrote: > > On 02/25/2017 03:50 AM, Ard Biesheuvel wrote: > >> > >> > >>> On 25 Feb 2017, at 11:23, Ard Biesheuvel > >>> wrote: > >>> > >>> On 25 February 2017 at 11:09, Markus Trippelsdorf > >>> wrote:

Re: gcc7 log2 compile issues in kernel/time/timekeeping.c

2017-03-02 Thread Ard Biesheuvel
On 2 March 2017 at 10:11, Markus Trippelsdorf wrote: > On 2017.03.01 at 17:39 +, Ard Biesheuvel wrote: >> On 1 March 2017 at 00:00, Laura Abbott wrote: >> > On 02/25/2017 03:50 AM, Ard Biesheuvel wrote: >> >> >> >> >> >>> On 25 Feb 2017, at 11:23, Ard Biesheuvel >> >>> wrote: >> >>> >> >>>

Re: gcc7 log2 compile issues in kernel/time/timekeeping.c

2017-03-01 Thread Ard Biesheuvel
On 1 March 2017 at 00:00, Laura Abbott wrote: > On 02/25/2017 03:50 AM, Ard Biesheuvel wrote: >> >> >>> On 25 Feb 2017, at 11:23, Ard Biesheuvel wrote: >>> >>> On 25 February 2017 at 11:09, Markus Trippelsdorf >>> wrote: On 2017.02.25 at 09:11 +, Ard Biesheuvel wrote: >> On 25 Febru

Re: gcc7 log2 compile issues in kernel/time/timekeeping.c

2017-02-28 Thread Laura Abbott
On 02/25/2017 03:50 AM, Ard Biesheuvel wrote: > > >> On 25 Feb 2017, at 11:23, Ard Biesheuvel wrote: >> >> On 25 February 2017 at 11:09, Markus Trippelsdorf >> wrote: >>> On 2017.02.25 at 09:11 +, Ard Biesheuvel wrote: > On 25 February 2017 at 08:18, Markus Trippelsdorf > wrote: >

Re: gcc7 log2 compile issues in kernel/time/timekeeping.c

2017-02-25 Thread Ard Biesheuvel
> On 25 Feb 2017, at 11:23, Ard Biesheuvel wrote: > > On 25 February 2017 at 11:09, Markus Trippelsdorf > wrote: >> On 2017.02.25 at 09:11 +, Ard Biesheuvel wrote: On 25 February 2017 at 08:18, Markus Trippelsdorf wrote: Why not simply get rid of the ilog2_NaN th

Re: gcc7 log2 compile issues in kernel/time/timekeeping.c

2017-02-25 Thread Ard Biesheuvel
On 25 February 2017 at 11:09, Markus Trippelsdorf wrote: > On 2017.02.25 at 09:11 +, Ard Biesheuvel wrote: >> On 25 February 2017 at 08:18, Markus Trippelsdorf >> wrote: >> > >> > Why not simply get rid of the ilog2_NaN thing altogether? >> > >> >> That would remove the issue, sure. But

Re: gcc7 log2 compile issues in kernel/time/timekeeping.c

2017-02-25 Thread Markus Trippelsdorf
On 2017.02.25 at 09:11 +, Ard Biesheuvel wrote: > On 25 February 2017 at 08:18, Markus Trippelsdorf > wrote: > > > > Why not simply get rid of the ilog2_NaN thing altogether? > > > > That would remove the issue, sure. But we lose an opportunity to spot > incorrect code at compile time.

Re: gcc7 log2 compile issues in kernel/time/timekeeping.c

2017-02-25 Thread Ard Biesheuvel
On 25 February 2017 at 08:18, Markus Trippelsdorf wrote: > On 2017.02.24 at 15:33 -0800, Laura Abbott wrote: >> On 02/24/2017 01:45 PM, Ard Biesheuvel wrote: >> > On 24 February 2017 at 21:25, John Stultz wrote: >> >> On Thu, Feb 23, 2017 at 10:43 AM, Laura Abbott wrote: >> >>> Hi, >> >>> >> >>>

Re: gcc7 log2 compile issues in kernel/time/timekeeping.c

2017-02-25 Thread Markus Trippelsdorf
On 2017.02.24 at 15:33 -0800, Laura Abbott wrote: > On 02/24/2017 01:45 PM, Ard Biesheuvel wrote: > > On 24 February 2017 at 21:25, John Stultz wrote: > >> On Thu, Feb 23, 2017 at 10:43 AM, Laura Abbott wrote: > >>> Hi, > >>> > >>> Fedora was previously carrying a workaround for a gcc7 issue repo

Re: gcc7 log2 compile issues in kernel/time/timekeeping.c

2017-02-24 Thread Laura Abbott
On 02/24/2017 01:45 PM, Ard Biesheuvel wrote: > On 24 February 2017 at 21:25, John Stultz wrote: >> On Thu, Feb 23, 2017 at 10:43 AM, Laura Abbott wrote: >>> Hi, >>> >>> Fedora was previously carrying a workaround for a gcc7 issue reported >>> on arm64 >>> http://lists.infradead.org/pipermail/li

Re: gcc7 log2 compile issues in kernel/time/timekeeping.c

2017-02-24 Thread Ard Biesheuvel
On 24 February 2017 at 21:25, John Stultz wrote: > On Thu, Feb 23, 2017 at 10:43 AM, Laura Abbott wrote: >> Hi, >> >> Fedora was previously carrying a workaround for a gcc7 issue reported >> on arm64 >> http://lists.infradead.org/pipermail/linux-arm-kernel/2016-October/461597.html. >> The workar

Re: gcc7 log2 compile issues in kernel/time/timekeeping.c

2017-02-24 Thread John Stultz
On Thu, Feb 23, 2017 at 10:43 AM, Laura Abbott wrote: > Hi, > > Fedora was previously carrying a workaround for a gcc7 issue reported > on arm64 > http://lists.infradead.org/pipermail/linux-arm-kernel/2016-October/461597.html. > The workaround got rid of __ilog2_NaN. I dropped the patch this morn