* patch to fix PR70703
@ 2017-04-05 15:11 Vladimir Makarov
2017-04-05 15:25 ` Jakub Jelinek
0 siblings, 1 reply; 6+ messages in thread
From: Vladimir Makarov @ 2017-04-05 15:11 UTC (permalink / raw)
To: gcc-patches
[-- Attachment #1: Type: text/plain, Size: 172 bytes --]
The following patch fixes
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=70703
The patch was successfully bootstrapped and tested on x86-64.
Committed as rev. 246707
[-- Attachment #2: pr70703.patch --]
[-- Type: text/x-patch, Size: 2739 bytes --]
Index: ChangeLog
===================================================================
--- ChangeLog (revision 246706)
+++ ChangeLog (working copy)
@@ -1,3 +1,10 @@
+2017-04-05 Vladimir Makarov <vmakarov@redhat.com>
+
+ PR rtl-optimization/70703
+ * ira-color.c (update_costs_from_allocno): Use the smallest mode.
+ (update_conflict_hard_regno_costs): Use long instead of unsigned
+ arithmetic for cost calculation.
+
2017-04-05 Jakub Jelinek <jakub@redhat.com>
Bernd Edlinger <bernd.edlinger@hotmail.de>
Index: testsuite/ChangeLog
===================================================================
--- testsuite/ChangeLog (revision 246706)
+++ testsuite/ChangeLog (working copy)
@@ -1,3 +1,8 @@
+2017-04-05 Vladimir Makarov <vmakarov@redhat.com>
+
+ PR rtl-optimization/70703
+ * gcc.target/i386/pr70703.c: New.
+
2017-04-05 Jakub Jelinek <jakub@redhat.com>
PR sanitizer/80308
Index: ira-color.c
===================================================================
--- ira-color.c (revision 246536)
+++ ira-color.c (working copy)
@@ -1367,6 +1367,16 @@ update_costs_from_allocno (ira_allocno_t
|| ALLOCNO_ASSIGNED_P (another_allocno))
continue;
+ if (GET_MODE_SIZE (ALLOCNO_MODE (cp->second)) < GET_MODE_SIZE (mode))
+ /* If we have different modes use the smallest one. It is
+ a sub-register move. It is hard to predict what LRA
+ will reload (the pseudo or its sub-register) but LRA
+ will try to minimize the data movement. Also for some
+ register classes bigger modes might be invalid,
+ e.g. DImode for AREG on x86. For such cases the
+ register move cost will be maximal. */
+ mode = ALLOCNO_MODE (cp->second);
+
cost = (cp->second == allocno
? ira_register_move_cost[mode][rclass][aclass]
: ira_register_move_cost[mode][aclass][rclass]);
@@ -1512,7 +1522,7 @@ update_conflict_hard_regno_costs (int *c
index = ira_class_hard_reg_index[aclass][hard_regno];
if (index < 0)
continue;
- cost = (int) ((unsigned) conflict_costs [i] * mult) / div;
+ cost = (int) (((long) conflict_costs [i] * mult) / div);
if (cost == 0)
continue;
cont_p = true;
Index: testsuite/gcc.target/i386/pr70703.c
===================================================================
--- testsuite/gcc.target/i386/pr70703.c (nonexistent)
+++ testsuite/gcc.target/i386/pr70703.c (working copy)
@@ -0,0 +1,9 @@
+/* { dg-do compile } */
+/* { dg-options "-O2 -fomit-frame-pointer" } */
+/* { dg-require-effective-target ia32 } */
+/* { dg-final { scan-assembler "movl\t\\\$6700417, %eax" } } */
+
+unsigned ud_x_641_mul(unsigned x) {
+ /* optimized version of x / 641 */
+ return ((unsigned long long)x * 0x663d81) >> 32;
+}
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: patch to fix PR70703
2017-04-05 15:11 patch to fix PR70703 Vladimir Makarov
@ 2017-04-05 15:25 ` Jakub Jelinek
2017-04-05 16:07 ` Vladimir Makarov
0 siblings, 1 reply; 6+ messages in thread
From: Jakub Jelinek @ 2017-04-05 15:25 UTC (permalink / raw)
To: Vladimir Makarov; +Cc: gcc-patches
On Wed, Apr 05, 2017 at 11:11:54AM -0400, Vladimir Makarov wrote:
> --- ira-color.c (revision 246536)
> +++ ira-color.c (working copy)
> @@ -1367,6 +1367,16 @@ update_costs_from_allocno (ira_allocno_t
> || ALLOCNO_ASSIGNED_P (another_allocno))
> continue;
>
> + if (GET_MODE_SIZE (ALLOCNO_MODE (cp->second)) < GET_MODE_SIZE (mode))
> + /* If we have different modes use the smallest one. It is
> + a sub-register move. It is hard to predict what LRA
> + will reload (the pseudo or its sub-register) but LRA
> + will try to minimize the data movement. Also for some
> + register classes bigger modes might be invalid,
> + e.g. DImode for AREG on x86. For such cases the
> + register move cost will be maximal. */
> + mode = ALLOCNO_MODE (cp->second);
> +
> cost = (cp->second == allocno
> ? ira_register_move_cost[mode][rclass][aclass]
> : ira_register_move_cost[mode][aclass][rclass]);
> @@ -1512,7 +1522,7 @@ update_conflict_hard_regno_costs (int *c
> index = ira_class_hard_reg_index[aclass][hard_regno];
> if (index < 0)
> continue;
> - cost = (int) ((unsigned) conflict_costs [i] * mult) / div;
> + cost = (int) (((long) conflict_costs [i] * mult) / div);
If you want something wider than unsigned, wouldn't it be better to
use HOST_WIDE_INT then? Otherwise it will work differently between
32-bit and 64-bit hosts.
Can any of those 3 values be negative? If not, perhaps
unsigned HOST_WIDE_INT?
Jakub
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: patch to fix PR70703
2017-04-05 15:25 ` Jakub Jelinek
@ 2017-04-05 16:07 ` Vladimir Makarov
2017-04-05 16:18 ` Vladimir Makarov
0 siblings, 1 reply; 6+ messages in thread
From: Vladimir Makarov @ 2017-04-05 16:07 UTC (permalink / raw)
To: gcc-patches
On 04/05/2017 11:25 AM, Jakub Jelinek wrote:
> On Wed, Apr 05, 2017 at 11:11:54AM -0400, Vladimir Makarov wrote:
>> --- ira-color.c (revision 246536)
>> +++ ira-color.c (working copy)
>> @@ -1367,6 +1367,16 @@ update_costs_from_allocno (ira_allocno_t
>> || ALLOCNO_ASSIGNED_P (another_allocno))
>> continue;
>>
>> + if (GET_MODE_SIZE (ALLOCNO_MODE (cp->second)) < GET_MODE_SIZE (mode))
>> + /* If we have different modes use the smallest one. It is
>> + a sub-register move. It is hard to predict what LRA
>> + will reload (the pseudo or its sub-register) but LRA
>> + will try to minimize the data movement. Also for some
>> + register classes bigger modes might be invalid,
>> + e.g. DImode for AREG on x86. For such cases the
>> + register move cost will be maximal. */
>> + mode = ALLOCNO_MODE (cp->second);
>> +
>> cost = (cp->second == allocno
>> ? ira_register_move_cost[mode][rclass][aclass]
>> : ira_register_move_cost[mode][aclass][rclass]);
>> @@ -1512,7 +1522,7 @@ update_conflict_hard_regno_costs (int *c
>> index = ira_class_hard_reg_index[aclass][hard_regno];
>> if (index < 0)
>> continue;
>> - cost = (int) ((unsigned) conflict_costs [i] * mult) / div;
>> + cost = (int) (((long) conflict_costs [i] * mult) / div);
> If you want something wider than unsigned, wouldn't it be better to
> use HOST_WIDE_INT then? Otherwise it will work differently between
> 32-bit and 64-bit hosts.
> Can any of those 3 values be negative? If not, perhaps
> unsigned HOST_WIDE_INT?
>
Thank you for finding this, Jakub. The overflow might happen only when
65K is used for ira_register_move_cost which means an impossible move in
most cases. Still if it happens in rare cases, the behavior might be
different for the native and a cross.
I'll correct the patch.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: patch to fix PR70703
2017-04-05 16:07 ` Vladimir Makarov
@ 2017-04-05 16:18 ` Vladimir Makarov
2017-04-06 10:24 ` Richard Biener
0 siblings, 1 reply; 6+ messages in thread
From: Vladimir Makarov @ 2017-04-05 16:18 UTC (permalink / raw)
To: gcc-patches
On 04/05/2017 12:07 PM, Vladimir Makarov wrote:
>
>
> I'll correct the patch.
>
Here is the patch I've committed.
2017-04-05 Vladimir Makarov <vmakarov@redhat.com>
PR rtl-optimization/70703
* ira-color.c (update_conflict_hard_regno_costs): Use
HOST_WIDE_INT instead of long.
Index: ira-color.c
===================================================================
--- ira-color.c (revision 246707)
+++ ira-color.c (working copy)
@@ -1522,7 +1522,7 @@
index = ira_class_hard_reg_index[aclass][hard_regno];
if (index < 0)
continue;
- cost = (int) (((long) conflict_costs [i] * mult) / div);
+ cost = (int) (((HOST_WIDE_INT) conflict_costs [i] *
mult) / div);
if (cost == 0)
continue;
cont_p = true;
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: patch to fix PR70703
2017-04-05 16:18 ` Vladimir Makarov
@ 2017-04-06 10:24 ` Richard Biener
2017-04-07 16:07 ` Vladimir Makarov
0 siblings, 1 reply; 6+ messages in thread
From: Richard Biener @ 2017-04-06 10:24 UTC (permalink / raw)
To: Vladimir Makarov; +Cc: GCC Patches
On Wed, Apr 5, 2017 at 6:18 PM, Vladimir Makarov <vmakarov@redhat.com> wrote:
>
>
> On 04/05/2017 12:07 PM, Vladimir Makarov wrote:
>>
>>
>>
>> I'll correct the patch.
>>
> Here is the patch I've committed.
Note that in such contexts it's better to just use [u]int64_t.
Richard.
> 2017-04-05 Vladimir Makarov <vmakarov@redhat.com>
>
> PR rtl-optimization/70703
> * ira-color.c (update_conflict_hard_regno_costs): Use
> HOST_WIDE_INT instead of long.
>
> Index: ira-color.c
> ===================================================================
> --- ira-color.c (revision 246707)
> +++ ira-color.c (working copy)
> @@ -1522,7 +1522,7 @@
> index = ira_class_hard_reg_index[aclass][hard_regno];
> if (index < 0)
> continue;
> - cost = (int) (((long) conflict_costs [i] * mult) / div);
> + cost = (int) (((HOST_WIDE_INT) conflict_costs [i] * mult) /
> div);
> if (cost == 0)
> continue;
> cont_p = true;
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: patch to fix PR70703
2017-04-06 10:24 ` Richard Biener
@ 2017-04-07 16:07 ` Vladimir Makarov
0 siblings, 0 replies; 6+ messages in thread
From: Vladimir Makarov @ 2017-04-07 16:07 UTC (permalink / raw)
To: Richard Biener; +Cc: GCC Patches
On 04/06/2017 06:24 AM, Richard Biener wrote:
> On Wed, Apr 5, 2017 at 6:18 PM, Vladimir Makarov <vmakarov@redhat.com> wrote:
>>
>> On 04/05/2017 12:07 PM, Vladimir Makarov wrote:
>>>
>>>
>>> I'll correct the patch.
>>>
>> Here is the patch I've committed.
> Note that in such contexts it's better to just use [u]int64_t.
>
>
You are right, Richard. I've committed the following patch:
Index: ChangeLog
===================================================================
--- ChangeLog (revision 246764)
+++ ChangeLog (working copy)
@@ -1,5 +1,11 @@
2017-04-07 Vladimir Makarov <vmakarov@redhat.com>
+ PR rtl-optimization/70703
+ * ira-color.c (update_conflict_hard_regno_costs): Use
+ int64_t instead of HOST_WIDE_INT.
+
+2017-04-07 Vladimir Makarov <vmakarov@redhat.com>
+
PR rtl-optimization/70478
* lra-constraints.c (process_alt_operands): Disfavor alternative
insn memory operands.
Index: ira-color.c
===================================================================
--- ira-color.c (revision 246763)
+++ ira-color.c (working copy)
@@ -1522,7 +1522,7 @@ update_conflict_hard_regno_costs (int *c
index = ira_class_hard_reg_index[aclass][hard_regno];
if (index < 0)
continue;
- cost = (int) (((HOST_WIDE_INT) conflict_costs [i] *
mult) / div);
+ cost = (int) (((int64_t) conflict_costs [i] * mult) / div);
if (cost == 0)
continue;
cont_p = true;
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2017-04-07 16:07 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-04-05 15:11 patch to fix PR70703 Vladimir Makarov
2017-04-05 15:25 ` Jakub Jelinek
2017-04-05 16:07 ` Vladimir Makarov
2017-04-05 16:18 ` Vladimir Makarov
2017-04-06 10:24 ` Richard Biener
2017-04-07 16:07 ` Vladimir Makarov
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).