[PATCHv2] Handle not explicitly zero terminated strings in merge sections

[Bernd Edlinger]

Hi Jakub,

are your concerns addressed with this answer, or do you have objections
to this patch?


Thanks
Bernd.


On 09/14/18 21:06, Bernd Edlinger wrote:
> On 09/14/18 21:01, Jakub Jelinek wrote:
>> On Fri, Sep 14, 2018 at 06:39:38PM +0000, Bernd Edlinger wrote:
>>> Hi,
>>>
>>> this is an upate of the string-merge section, it is based on the V2-STRING_CST
>>> semantic patch series, which was finally installed yesterday.
>>> It merges single-byte string constants with or without terminating NUL.
>>> The patch has the same Ada and C test cases that were already in the V1 patch.
>>>
>>> Thus there are no pre-requisite patches necessary at this time.
>>>
>>> Bootstrapped and reg-tested on x86_64-pc-linux-gnu.
>>> Is it OK for trunk?
>>
>> I believe SHF_MERGE | SHF_STRINGS sections should only ever contain
>> zero-terminated strings.
>> See e.g. SCO ELF documentation:
>> "The size of each element is specified in the section header's sh_entsize field. If the SHF_STRINGS flag is also set, the data elements consist of null-terminated character strings."
>> http://www.sco.com/developers/gabi/2003-12-17/ch4.sheader.html
>> https://docs.oracle.com/cd/E23824_01/html/819-0690/ggdlu.html
>>
>> So, please never put non-zero terminated strings into MS sections.
>>
> 
> Yes, that is of course clear.
> 
> The idea is, if the string itself is not zero-terminated
> another zero is added, that is only done in at the assembly level.
> 
> Strings with embedded zero bytes are not put in the merge section.
> > 
> Bernd.