Re: Can we backport bug 12296 fix to 2.21?

Re: Can we backport bug 12296 fix to 2.21?

[Terry Guo]

[-- Attachment #1: Type: text/plain, Size: 1034 bytes --]

Hi Tristan,

Thanks for your quick response. The attachment is the final fix in master
branch for PR12296. It is extracted from checkin at
http://sourceware.org/ml/binutils-cvs/2011-04/msg00112.html. Tested for
arm-none-eabi and no regression compared to Binutils 2.21.

Because I don't have write permission, could you or anybody else help me to
backport this fix to Binutils 2.21? Thanks very much.

Best regards,
Terry

> -----Original Message-----
> From: Tristan Gingold [mailto:gingold@adacore.com]
> Sent: Tuesday, August 23, 2011 11:04 AM
> To: Terry Guo
> Cc: binutils@sourceware.org
> Subject: Re: Can we backport bug 12296 fix to 2.21?
> 
> 
> On Aug 22, 2011, at 10:56 PM, Terry Guo wrote:
> 
> > Hello Tristan,
> >
> > Any hints on this issue? Thanks.
> 
> Sorry, I was traveling.
> 
> Yes, it is OK to backport the fix, but note that there is currently no
> plan to make a 2.21.2 release (but you can still use the CVS release
> branch).
> 
> Tristan.
> 
> >
> > BR,
> > Terry
> >

[-- Attachment #2: binutils-pr12296-final-fix.patch --]
[-- Type: application/octet-stream, Size: 682 bytes --]

diff --git a/gas/config/tc-arm.c b/gas/config/tc-arm.c
index 4addafc..1c22ad1 100644
--- a/gas/config/tc-arm.c
+++ b/gas/config/tc-arm.c
@@ -11609,7 +11609,9 @@ do_t_swi (void)
      to ARM_EXT_V6M.  */
   if (ARM_CPU_HAS_FEATURE (cpu_variant, arm_ext_v6m))
     {
-      if (!ARM_CPU_HAS_FEATURE (cpu_variant, arm_ext_os))
+      if (!ARM_CPU_HAS_FEATURE (cpu_variant, arm_ext_os)
+         /* This only applies to the v6m howver, not later architectures.  */
+         && ! ARM_CPU_HAS_FEATURE (cpu_variant, arm_ext_v7))
 	as_bad (_("SVC is not permitted on this architecture"));
       ARM_MERGE_FEATURE_SETS (thumb_arch_used, thumb_arch_used, arm_ext_os);
     }

Re: Can we backport bug 12296 fix to 2.21?

[Matthew Gretton-Dann]

Terry,

On 24/08/11 10:21, Terry Guo wrote:
> Hi Tristan,
>
> Thanks for your quick response. The attachment is the final fix in master
> branch for PR12296. It is extracted from checkin at
> http://sourceware.org/ml/binutils-cvs/2011-04/msg00112.html. Tested for
> arm-none-eabi and no regression compared to Binutils 2.21.
>
> Because I don't have write permission, could you or anybody else help me to
> backport this fix to Binutils 2.21? Thanks very much.
>
> Best regards,
> Terry
>
>> -----Original Message-----
>> From: Tristan Gingold [mailto:gingold@adacore.com]
>> Sent: Tuesday, August 23, 2011 11:04 AM
>> To: Terry Guo
>> Cc: binutils@sourceware.org
>> Subject: Re: Can we backport bug 12296 fix to 2.21?
>>
>>
>> On Aug 22, 2011, at 10:56 PM, Terry Guo wrote:
>>
>>> Hello Tristan,
>>>
>>> Any hints on this issue? Thanks.
>>
>> Sorry, I was traveling.
>>
>> Yes, it is OK to backport the fix, but note that there is currently no
>> plan to make a 2.21.2 release (but you can still use the CVS release
>> branch).
>>
>> Tristan.
>>
>>>
>>> BR,
>>> Terry
>> >
>>
>> binutils-pr12296-final-fix.patch
>>
>>
>> diff --git a/gas/config/tc-arm.c b/gas/config/tc-arm.c
>> index 4addafc..1c22ad1 100644
>> --- a/gas/config/tc-arm.c
>> +++ b/gas/config/tc-arm.c
>> @@ -11609,7 +11609,9 @@ do_t_swi (void)
>>        to ARM_EXT_V6M.  */
>>     if (ARM_CPU_HAS_FEATURE (cpu_variant, arm_ext_v6m))
>>       {
>> -      if (!ARM_CPU_HAS_FEATURE (cpu_variant, arm_ext_os))
>> +      if (!ARM_CPU_HAS_FEATURE (cpu_variant, arm_ext_os)
>> +         /* This only applies to the v6m howver, not later architectures.  */
>> +&&  ! ARM_CPU_HAS_FEATURE (cpu_variant, arm_ext_v7))
>>   	as_bad (_("SVC is not permitted on this architecture"));
>>         ARM_MERGE_FEATURE_SETS (thumb_arch_used, thumb_arch_used, arm_ext_os);
>>       }


This has now been committed upstream with the following ChangeLog:

2011-08-24  Matthew Gretton-Dann  <matthew.gretton-dann@arm.com>

	Backport from mainline.
	2011-04-18  Julian Brown  <julian@codesourcery.com>
	            Nick Clifton  <nickc@redhat.com>
	* config/tc-arm.c (do_t_swi): Exclude v7 and higher variants
	from arm_ext_os test.

Thanks,

Matt


-- 
Matthew Gretton-Dann
Principal Engineer, PD Software - Tools, ARM Ltd

Re: Can we backport bug 12296 fix to 2.21?

[Tristan Gingold]

On Aug 22, 2011, at 10:56 PM, Terry Guo wrote:

> Hello Tristan,
> 
> Any hints on this issue? Thanks.

Sorry, I was traveling.

Yes, it is OK to backport the fix, but note that there is currently no plan to make a 2.21.2 release (but you can still use the CVS release branch).

Tristan.

> 
> BR,
> Terry
> 
>> -----Original Message-----
>> From: Terry Guo [mailto:terry.guo@arm.com]
>> Sent: Friday, August 19, 2011 5:03 PM
>> To: 'binutils@sourceware.org'
>> Cc: 'gingold@adacore.com'
>> Subject: Can we backport bug 12296 fix to 2.21?
>> 
>> Hello,
>> 
>> We recently setup a GCC branch embedded-4_6-branch to provide fixes and
>> enhancements for GCC 4.6.x when used for ARM embedded processors. For
>> this branch we would like to use the latest Binutils 2.21.
>> Unfortunately when trying to use 2.21 to build the latest Newlib 1.19
>> for armv7-m and armv7e-m, we are stopped by Bug 12296 at
>> http://sourceware.org/bugzilla/show_bug.cgi?id=12296. Since it's been a
>> while till 2.22 release. Is it OK to backport this fix to current 2.21
>> branch?
>> 
>> Thanks and regards,
>> Terry
>> 
> 
> 
> 
> 

RE: Can we backport bug 12296 fix to 2.21?

[Terry Guo]

Hello Tristan,

Any hints on this issue? Thanks.

BR,
Terry

> -----Original Message-----
> From: Terry Guo [mailto:terry.guo@arm.com]
> Sent: Friday, August 19, 2011 5:03 PM
> To: 'binutils@sourceware.org'
> Cc: 'gingold@adacore.com'
> Subject: Can we backport bug 12296 fix to 2.21?
> 
> Hello,
> 
> We recently setup a GCC branch embedded-4_6-branch to provide fixes and
> enhancements for GCC 4.6.x when used for ARM embedded processors. For
> this branch we would like to use the latest Binutils 2.21.
> Unfortunately when trying to use 2.21 to build the latest Newlib 1.19
> for armv7-m and armv7e-m, we are stopped by Bug 12296 at
> http://sourceware.org/bugzilla/show_bug.cgi?id=12296. Since it's been a
> while till 2.22 release. Is it OK to backport this fix to current 2.21
> branch?
> 
> Thanks and regards,
> Terry
> 

Can we backport bug 12296 fix to 2.21?

[Terry Guo]

Hello,

We recently setup a GCC branch embedded-4_6-branch to provide fixes and
enhancements for GCC 4.6.x when used for ARM embedded processors. For this
branch we would like to use the latest Binutils 2.21. Unfortunately when
trying to use 2.21 to build the latest Newlib 1.19 for armv7-m and armv7e-m,
we are stopped by Bug 12296 at
http://sourceware.org/bugzilla/show_bug.cgi?id=12296. Since it's been a
while till 2.22 release. Is it OK to backport this fix to current 2.21
branch?

Thanks and regards,
Terry