1) As multiple people said, it *was* a regression bug fix. It actually
fixed two regressions. (That it fixed the second was discovered only
after I committed it). I'm sorry that it caused problems for you (even
though it's actually lucky for GCC), but I can't help saying that it
might have been
In addition, all but one of the fixes that H.J. made (and for which I
have to thank him) were unrecognizable insns due to a misunderstanding
of how peephole2 worked
I stand corrected; *all* of the fixes. The patch hadn't had a
correctness problem until your message, only ice-on-valids. This
Toon Moene wrote:
H.J. Lu wrote:
If you can provide a testcase, I can take a look. If it isn't easy to
find
a testcase, please disable the second pattern:
(define_peephole2
[(set (match_operand 0 register_operand )
(match_operand 1 register_operand ))
(set (match_dup 0)
Paolo Bonzini wrote:
Toon Moene wrote:
H.J. Lu wrote:
If you can provide a testcase, I can take a look. If it isn't easy to
find
a testcase, please disable the second pattern:
(define_peephole2
[(set (match_operand 0 register_operand )
(match_operand 1 register_operand ))
(set
Attached you'll find the (preprocessed) source of the routine that
printed the Infinity's (of course, I cannot be completely certain that
it actually resulted in the wrong code, but at least it might be studied
to see if it helps to find the culprit).
No, this function is sane (the peephole
Paolo Bonzini wrote:
Attached you'll find the (preprocessed) source of the routine that
printed the Infinity's (of course, I cannot be completely certain that
it actually resulted in the wrong code, but at least it might be studied
to see if it helps to find the culprit).
No, this function is
On Thu, Mar 12, 2009 at 5:53 AM, Toon Moene mo...@knmi.nl wrote:
(*) it would have helped to know the compilation flags and target, of
course.
Yep, sorry: -g -O3 -ffast-math -mcpu=native -march=native
on a x86-64-unknown-linux-gnu system (native 64-bit).
Please show output of 'gcc -v
On Thu, Mar 12, 2009 at 1:48 AM, Toon Moene mo...@knmi.nl wrote:
Ah, but it did: throwing out the second peephole (in stead of both) fixed
the problem too.
Is a bug opened for this? If not, please open one.
Thanks.
--
H.J.
On Thu, Mar 12, 2009 at 5:53 AM, Toon Moene mo...@knmi.nl wrote:
Yep, sorry: -g -O3 -ffast-math -mcpu=native -march=native
on a x86-64-unknown-linux-gnu system (native 64-bit).
Please also try -fno-vectorize to see if this peephole issue is triggered by
vectorizer.
--
H.J.
Toon Moene wrote:
Paolo Bonzini wrote:
Attached you'll find the (preprocessed) source of the routine that
printed the Infinity's (of course, I cannot be completely certain that
it actually resulted in the wrong code, but at least it might be studied
to see if it helps to find the culprit).
On Thu, Mar 12, 2009 at 7:03 AM, Paolo Bonzini bonz...@gnu.org wrote:
Toon Moene wrote:
Paolo Bonzini wrote:
Attached you'll find the (preprocessed) source of the routine that
printed the Infinity's (of course, I cannot be completely certain that
it actually resulted in the wrong code, but
Will REGNO (operands[0]) == REGNO (operands[1]) work here?
Yes. I wanted to be conservative in case one day subregs or who knows
what are allowed. I'll defer to maintainers or other people (Steven?),
either way is fine by me.
Paolo
On Thu, Mar 12, 2009 at 3:34 PM, Paolo Bonzini bonz...@gnu.org wrote:
Will REGNO (operands[0]) == REGNO (operands[1]) work here?
Yes. I wanted to be conservative in case one day subregs or who knows
what are allowed. I'll defer to maintainers or other people (Steven?),
either way is fine
On Thu, Mar 12, 2009 at 3:56 PM, Steven Bosscher stevenb@gmail.com wrote:
- Show quoted text -
On Thu, Mar 12, 2009 at 3:34 PM, Paolo Bonzini bonz...@gnu.org wrote:
Will REGNO (operands[0]) == REGNO (operands[1]) work here?
Yes. I wanted to be conservative in case one day subregs or who
On Thu, Mar 12, 2009 at 8:00 AM, Richard Guenther
richard.guent...@gmail.com wrote:
On Thu, Mar 12, 2009 at 3:56 PM, Steven Bosscher stevenb@gmail.com
wrote:
- Show quoted text -
On Thu, Mar 12, 2009 at 3:34 PM, Paolo Bonzini bonz...@gnu.org wrote:
Will REGNO (operands[0]) == REGNO
On Thu, Mar 12, 2009 at 8:58 AM, H.J. Lu hjl.to...@gmail.com wrote:
Index: i386.md
===
--- i386.md (revision 144796)
+++ i386.md (working copy)
@@ -20813,7 +20813,7 @@
[(match_dup 0)
H.J. Lu wrote:
On Thu, Mar 12, 2009 at 8:58 AM, H.J. Lu hjl.to...@gmail.com wrote:
Index: i386.md
===
--- i386.md (revision 144796)
+++ i386.md (working copy)
@@ -20813,7 +20813,7 @@
[(match_dup 0)
... and it had to be fixed 3-4 times by H.J.Lu before it went so
obscure to only hit people like me.
I run a weather forecasting system 4 times daily to test it out.
Because GCC would-be-4.4 is in regression fixes only during the last
four months, I thought I could use it with impunity.
On Wed, Mar 11, 2009 at 10:54 PM, Toon Moene t...@moene.org wrote:
Unless a very good reason is in my inbox in the next 48 hours, I'll revert
this change under the obvious rule.
It was a regression fix.
See http://gcc.gnu.org/ml/gcc-patches/2009-02/msg00299.html
This took 2 minutes to find
On Wed, Mar 11, 2009 at 10:54 PM, Toon Moene t...@moene.org wrote:
Unless a very good reason is in my inbox in the next 48 hours, I'll revert
this change under the obvious rule.
It doesn't work this way. You would need to start the 48h clock and have two
maintainers of the area approve the
Steven Bosscher wrote:
On Wed, Mar 11, 2009 at 10:54 PM, Toon Moene t...@moene.org wrote:
Unless a very good reason is in my inbox in the next 48 hours, I'll revert
this change under the obvious rule.
It was a regression fix.
See http://gcc.gnu.org/ml/gcc-patches/2009-02/msg00299.html
Richard Guenther wrote:
On Wed, Mar 11, 2009 at 10:54 PM, Toon Moene t...@moene.org wrote:
Unless a very good reason is in my inbox in the next 48 hours, I'll revert
this change under the obvious rule.
It doesn't work this way. You would need to start the 48h clock and have two
maintainers
On Wed, Mar 11, 2009 at 11:20 PM, Toon Moene t...@moene.org wrote:
Richard Guenther wrote:
On Wed, Mar 11, 2009 at 10:54 PM, Toon Moene t...@moene.org wrote:
Unless a very good reason is in my inbox in the next 48 hours, I'll
revert
this change under the obvious rule.
It doesn't work this
On Wed, Mar 11, 2009 at 11:17 PM, Toon Moene t...@moene.org wrote:
Pffft - it was only a regression on optimization - note how many times HJL
had to fix the fix before it became so obscure that only I could run into
it.
*This is not a regression fix, period.*
Well, the PR it was supposed to
On Wed, Mar 11, 2009 at 11:20 PM, Toon Moene t...@moene.org wrote:
I will abide by the rules - but the rules also say that this is not the sort
of fix that goes in at phase 4.
Which rule where says so? Not intended in an offensive manner -- just
curious. I'm not aware of any rules that specify
On Wed, Mar 11, 2009 at 3:03 PM, Steven Bosscher stevenb@gmail.com wrote:
On Wed, Mar 11, 2009 at 10:54 PM, Toon Moene t...@moene.org wrote:
Unless a very good reason is in my inbox in the next 48 hours, I'll revert
this change under the obvious rule.
It was a regression fix.
See
Steven Bosscher wrote:
On Wed, Mar 11, 2009 at 11:20 PM, Toon Moene t...@moene.org wrote:
I will abide by the rules - but the rules also say that this is not the sort
of fix that goes in at phase 4.
Which rule where says so? Not intended in an offensive manner -- just
curious. I'm not aware
On Wed, Mar 11, 2009 at 4:00 PM, Toon Moene t...@moene.org wrote:
Steven Bosscher wrote:
On Wed, Mar 11, 2009 at 11:20 PM, Toon Moene t...@moene.org wrote:
I will abide by the rules - but the rules also say that this is not the
sort
of fix that goes in at phase 4.
Which rule where says
H.J. Lu wrote:
If you can provide a testcase, I can take a look. If it isn't easy to find
a testcase, please disable the second pattern:
(define_peephole2
[(set (match_operand 0 register_operand )
(match_operand 1 register_operand ))
(set (match_dup 0)
On Wed, Mar 11, 2009 at 4:20 PM, Toon Moene t...@moene.org wrote:
H.J. Lu wrote:
If you can provide a testcase, I can take a look. If it isn't easy to find
a testcase, please disable the second pattern:
(define_peephole2
[(set (match_operand 0 register_operand )
(match_operand 1
On Wed, Mar 11, 2009 at 2:54 PM, Toon Moene t...@moene.org wrote:
... and it had to be fixed 3-4 times by H.J.Lu before it went so obscure
to only hit people like me.
I run a weather forecasting system 4 times daily to test it out.
Because GCC would-be-4.4 is in regression fixes only during
31 matches
Mail list logo