This is the mail archive of the
mailing list for the binutils project.
Re: [Patch ARM Gas] - strexh/strexb warn for bad addressing modes.
- From: Luke <hazelnusse at gmail dot com>
- To: James Greenhalgh <james dot greenhalgh at arm dot com>
- Cc: binutils at sourceware dot org
- Date: Thu, 22 Sep 2011 09:07:00 -0700
- Subject: Re: [Patch ARM Gas] - strexh/strexb warn for bad addressing modes.
- References: <4E32CCDE.firstname.lastname@example.org> <email@example.com> <4e7b3128.41c8e30a.6565.ffff9e55SMTPIN_ADDED@mx.google.com>
You are correct, the behavior is undefined and your patch does the
right thing. I'll pass this along to the CMSIS folks and give them a
On Thu, Sep 22, 2011 at 5:59 AM, James Greenhalgh
>> If you read this literally, it only says STREX has the restriction that
>> you've added to gas, and that it doesn't apply to any of the other ones
>> particular STREXH and STREXB). ÂI'm not sure if this is the same across
>> the ARM7/9/11 processors, but it certainly seems that this should be an
>> allowed instruction for Cortex M3, given that they are shipping this
>> instruction with their CMSIS stuff.
> In my copy of the ARMARMv7M  strexb, strexh and strex are all listed
> as UNPREDICTABLE when rd == rt || rd == rn.
> I would read the quoted passage of the User Guide Reference Material
> as saying Â"for Store-Exclusive Instructions, Rd must be different
> from both Rt and Rn". Which better fits with the description in the
> There is potentially an argument for making these UNPREDICTABLE behaviours
> warnings, but this would just mask the real issue, which, IMHO is
> that the library is not sufficiently cautious with the registers it uses
> for these instructions.
> If you want further advice on the CMSIS libraries I would suggest going
> through your normal CMSIS support channels.
> James Greenhalgh
>  -
"Those who would give up essential liberty to purchase a little
temporary safety deserve neither liberty nor safety."
-- Benjamin Franklin, Historical Review of Pennsylvania, 1759