Re: [PATCH] Support for known unknown alignment

2012-05-13 Thread H.J. Lu
On Fri, Apr 27, 2012 at 5:02 AM, Martin Jambor wrote: > Hi, > > On Tue, Apr 24, 2012 at 12:31:38PM +0200, Martin Jambor wrote: >> Hi, >> >> On Mon, Apr 23, 2012 at 03:30:19PM +0200, Richard Guenther wrote: >> > On Mon, 23 Apr 2012, Martin Jambor wrote: >> > >> > > Hi, >> > > >> > > On Mon, Apr 23,

Re: [PATCH] Support for known unknown alignment

2012-04-27 Thread Martin Jambor
Hi, On Tue, Apr 24, 2012 at 12:31:38PM +0200, Martin Jambor wrote: > Hi, > > On Mon, Apr 23, 2012 at 03:30:19PM +0200, Richard Guenther wrote: > > On Mon, 23 Apr 2012, Martin Jambor wrote: > > > > > Hi, > > > > > > On Mon, Apr 23, 2012 at 12:50:51PM +0200, Richard Guenther wrote: > > > > On Fri

Re: [PATCH] Support for known unknown alignment

2012-04-24 Thread Martin Jambor
Hi, On Mon, Apr 23, 2012 at 03:30:19PM +0200, Richard Guenther wrote: > On Mon, 23 Apr 2012, Martin Jambor wrote: > > > Hi, > > > > On Mon, Apr 23, 2012 at 12:50:51PM +0200, Richard Guenther wrote: > > > On Fri, 20 Apr 2012, Martin Jambor wrote: > > > > > > > Hi, > > > > > > > > two days ago I

Re: [PATCH] Support for known unknown alignment

2012-04-23 Thread Richard Guenther
On Mon, 23 Apr 2012, Jay Foad wrote: > On 23 April 2012 14:30, Richard Guenther wrote: > > Well, CCP simply tracks known-bits and derives the alignment > > value from that.  If tem & -tem computes as zero that means > > val->mask.low is all zeros. > > Doesn't that mean that all bits are known? S

Re: [PATCH] Support for known unknown alignment

2012-04-23 Thread Jay Foad
On 23 April 2012 14:30, Richard Guenther wrote: > Well, CCP simply tracks known-bits and derives the alignment > value from that.  If tem & -tem computes as zero that means > val->mask.low is all zeros. Doesn't that mean that all bits are known? So you could set: pi->align = 1 << 32; // or som

Re: [PATCH] Support for known unknown alignment

2012-04-23 Thread Richard Guenther
On Mon, 23 Apr 2012, Martin Jambor wrote: > Hi, > > On Mon, Apr 23, 2012 at 12:50:51PM +0200, Richard Guenther wrote: > > On Fri, 20 Apr 2012, Martin Jambor wrote: > > > > > Hi, > > > > > > two days ago I talked to Richi on IRC about the functions to determine > > > the expected alignment of ob

Re: [PATCH] Support for known unknown alignment

2012-04-23 Thread Martin Jambor
Hi, On Mon, Apr 23, 2012 at 12:50:51PM +0200, Richard Guenther wrote: > On Fri, 20 Apr 2012, Martin Jambor wrote: > > > Hi, > > > > two days ago I talked to Richi on IRC about the functions to determine > > the expected alignment of objects and pointers we have and he > > suggested that get_obje

Re: [PATCH] Support for known unknown alignment

2012-04-23 Thread Richard Guenther
On Fri, Apr 20, 2012 at 8:41 PM, Jay Foad wrote: > On 20 April 2012 16:54, Martin Jambor wrote: >> two days ago I talked to Richi on IRC about the functions to determine >> the expected alignment of objects and pointers we have and he >> suggested that get_object_alignment_1 and get_pointer_align

Re: [PATCH] Support for known unknown alignment

2012-04-23 Thread Richard Guenther
On Fri, 20 Apr 2012, Martin Jambor wrote: > Hi, > > two days ago I talked to Richi on IRC about the functions to determine > the expected alignment of objects and pointers we have and he > suggested that get_object_alignment_1 and get_pointer_alignment_1 > should return whether the alignment is a

Re: [PATCH] Support for known unknown alignment

2012-04-20 Thread Jay Foad
On 20 April 2012 16:54, Martin Jambor wrote: > two days ago I talked to Richi on IRC about the functions to determine > the expected alignment of objects and pointers we have and he > suggested that get_object_alignment_1 and get_pointer_alignment_1 > should return whether the alignment is actuall

[PATCH] Support for known unknown alignment

2012-04-20 Thread Martin Jambor
Hi, two days ago I talked to Richi on IRC about the functions to determine the expected alignment of objects and pointers we have and he suggested that get_object_alignment_1 and get_pointer_alignment_1 should return whether the alignment is actually known and return the actual alignment in a refe